ASD Technical Guide 3rd Ed.

Section 0-1

Section 0
Analytical Spectral Devices, Inc. (ASD)
Technical Guide
4th Ed.

Managing Editor: David C. Hatchell

Copyright©1999
Analytical Spectral Devices, Inc.
5335 Sterling Drive
Suite A
Boulder, CO 80301-2344
USA

Phone: 303-444-6522
Fax: 303-444-6825
web: www.asdi.com
email: [email protected]

FieldSpec®, LabSpec®, QualitySpec®, SeaSpec®, Spectrode®, RxSpec®, ViewSpec™, TerraSpec™, and

SpecSupport™ are registered and unregistered trademarks, and intellectual property of Analytical Spectral
Devices, Inc., Boulder, Colorado, USA. Other brand and product names are trademarks of their respective
holders. Unauthorized use of these trademarks is illegal and punishable by law.

The information and specifications contained in this guide are subject to change without notice. Analytical
Spectral Devices, Inc. shall not be held liable for technical or editorial omissions or errors made herein; nor for
incidental or consequential damages resulting from furnishing, performance or use of this material.

This document contains proprietary information protected under US Copyright Laws.

Contact ASD for formal quotations.

 ASD Technical Guide 3rd Ed. Section 0-2

Table of Contents

Welcome to the ASD Technical Guide....................................................................................................................0-4

Company Profile......................................................................................................................................................0-5
FieldSpec® FR........................................................................................................................................................1-1
FR Post Dispersive Spectrometer...........................................................................................................................1-1
FR Direct Path Fiberoptic Input...............................................................................................................................1-1
FR Speed and Sensitivity........................................................................................................................................1-1
FR Longer Fiberoptic Cable Comparison ...............................................................................................................1-3
FR Sampling Interval and Spectral Resolution .......................................................................................................2-1
FR Repeatability......................................................................................................................................................3-1
FR Large Field-of-View ...........................................................................................................................................3-2
FR Stray Light Rejection .........................................................................................................................................4-1
Stray Light Look Alikes............................................................................................................................................4-7
FR Driftlock..............................................................................................................................................................4-7
FR Suggested Set-up..............................................................................................................................................5-1
Artificial Illumination and Mineral Reflectance ........................................................................................................5-5
FR Foreoptics..........................................................................................................................................................6-1
FR Fiberoptic Cable ................................................................................................................................................7-1
FieldSpec® FR 1 deg Field-of-View Lens Foreoptic...............................................................................................7-1
FR 1 Deg Field-of-View Mirror Foreoptic ................................................................................................................7-2
FR 5 deg Field-of-View Lens Foreoptic ..................................................................................................................7-3
FR 8 deg Field-of-View Lens Foreoptic ..................................................................................................................7-3
FR 18 deg Field-of-View Tube ................................................................................................................................7-4
FR PistolGrip ...........................................................................................................................................................7-5
FieldSpec® JR (350 -2500 nm) ..............................................................................................................................8-1
JR versus FR Plots .................................................................................................................................................8-1
JR Unique Highlights...............................................................................................................................................8-4
FieldSpec® NIR JR (1000 - 2500 nm) ....................................................................................................................8-5
FieldSpec® UV/VNIR (350 -1050 nm) ....................................................................................................................9-1
VNIR Post Dispersive Spectrometer.......................................................................................................................9-1
VNIR Direct Path Fiberoptic Input...........................................................................................................................9-1
VNIR Fiberoptic Cable ............................................................................................................................................9-2
VNIR Speed and Sensitivity ....................................................................................................................................9-2
VNIR Sampling Interval and Spectral Resolution ...................................................................................................9-3
VNIR Ground Truthing Hyperspectral Imagery .......................................................................................................9-4
VNIR Stray Light Rejection .....................................................................................................................................9-5
VNIR Driftlock..........................................................................................................................................................9-7
VNIR Large Field-of-View .......................................................................................................................................9-7
VNIR Light Weight, Small, and Wear-able..............................................................................................................9-7
VNIR Upgrade to Dual Spectrometer System.........................................................................................................9-7
VNIR Upgrade to a FieldSpec® FR ........................................................................................................................9-7
FieldSpec® Dual UV/VNIR (350 -1050 nm)............................................................................................................9-8
FieldSpec® UV/VNIR/CCD .....................................................................................................................................9-8
FieldSpec® HandHeld...........................................................................................................................................10-1
HandHeld Post Dispersive Spectrometer .............................................................................................................10-1
HandHeld Speed and Sensitivity...........................................................................................................................10-1
HandHeld Sampling Interval and Spectral Resolution..........................................................................................10-2
HandHeld Small Size and Weight.........................................................................................................................10-4
HandHeld Large Field-of-View ..............................................................................................................................10-4
HandHeld Stray Light Rejection ............................................................................................................................10-5
HandHeld Driftlock ................................................................................................................................................10-8
HandHeld Foreoptics ............................................................................................................................................10-9
LabSpec® Pro Portable Spectrophotometer.........................................................................................................11-1
LabSpec® Pro Spectral Ranges ...........................................................................................................................11-1
LabSpec® Pro Features & Advantages: ...............................................................................................................11-1
Fundamentals of Spectroradiometry.....................................................................................................................13-1
Illumination Geometry ...........................................................................................................................................14-1
Wavenumber and Wavelength..............................................................................................................................15-1
 ASD Technical Guide 3rd Ed. Section 0-3

Spectral Regions...................................................................................................................................................16-1
Reflectance ...........................................................................................................................................................17-1
Wavelength and Radiometric Calibration Methods...............................................................................................18-1
Blackbody Radiance and Spectroradiometric Calibration.....................................................................................19-1
The Benifits of Portable NIR Analysis ...................................................................................................................20-1
Making an NIR Analyzer Work For You ................................................................................................................21-1
Derivation of Bouguer-Lambert-Beer (BLB) Law ..................................................................................................22-1
Approximating Spot Size.......................................................................................................................................23-1
Foliar Spectral Features........................................................................................................................................24-1
Vegetation Index Formulas ...................................................................................................................................24-2
Water Spectral Features .......................................................................................................................................24-3
Glossary ................................................................................................................................................................25-1
 ASD Technical Guide 3rd Ed. Section 0-4

Welcome to the ASD Technical Guide

In this information age we are constantly bombarded by copious quantities of information, both wanted and
unwanted, good and bad. Those of us in the business of creating information using the latest in technology are
constantly answering requests for information, some simple, others complex and involved. We often end up
answering the same questions again and again. In order to increase the efficiency of our work force and provide
you with concise information, all wanted and good, we have put together a technical guide to ASD’s
spectrophotometer/spectroradiometer product line along with a glossary of terms that are used in the light
measurement field.

We welcome comments, criticisms and suggestions for the next issue, and we hope that our products will fill your
application needs. The management and employees of ASD take pride in our customer service reputation. Our
aim is to make sure that each customer understands how our instruments work and how to derive the most
information from the data acquired. Please contact us by phone, fax, or email:

Phone: 303-444-6522
Fax: 303-444-6825
email: [email protected]
web: www.asdi.com
 ASD Technical Guide 3rd Ed. Section 0-5

Company Profile

Analytical Spectral Devices, Inc. (ASD) was founded in February, 1990. ASD manufactures and sells the
FieldSpec®®, and LabSpec®® lines of optical spectrometers (all trademarked by ASD). ASD's facilities occupy
11,000 square feet and include laboratories and equipment for the development, testing, calibration and
manufacturing of high spectral resolution spectroradiometers, spectrophotometers, and associated run-time and
data analysis software. This equipment includes: analog and digital oscilloscopes, function generators, power
supplies, digital logic analyzers, optical benches and fixtures, fiber optic illuminators, monochromators, baffled
tanks with optical windows for underwater calibrations, spectral line sources, and NIST traceable calibrated
irradiance and radiance sources, computers, CAD software, and embedded software development tools. ASD
currently employees a variety of Ph.D. Scientists, Engineers, and Bachelors Degreed technicians.
 ASD Technical Guide 3rd Ed. Section 1-1

Section 1
FieldSpec® FR

Several important features of the FieldSpec® FR make it the best and unique choice for a field portable
spectroradiometer:

The FieldSpec® FR is a post dispersive spectrometer.

The FieldSpec® FR spectroradiometer uses a direct path fiber optic input.
The FieldSpec® FR spectroradiometer has an unbeatable combination of speed and sensitivity.
The FieldSpec® FR spectroradiometer has the spectral resolution and sampling interval required for ground
truthing hyperspectral imagery.
The FieldSpec® FR spectroradiometer has excellent repeatability.
The FieldSpec® FR has the largest standard field-of-view of any portable spectroradiometer.
The FieldSpec® FR spectroradiometer has excellent stray light rejection.
The FieldSpec® FR spectroradiometer features the unique Driftlock™ dark-drift correction system.

FR Post Dispersive Spectrometer

All ASD spectrometers are “post-dispersive”. Because portable spectrometers are typically used outside the
controlled laboratory environment, they are exposed to much higher levels of ambient light. In almost all cases,
some of this ambient light will stray into the sample being measured. The errors produced by this ambient stray
light are much greater for a pre-dispersive spectrometer than they are for a spectrometer that is post-dispersive.

In a pre-dispersive spectrometer, the sample is illuminated with monochromatic light. Light scattered off or
transmitted through the sample is then collected and delivered to the instrument’s detector. Ambient light that
strays into the sample being measured is also collected. Thus, both the monochromatic illumination from the
instrument and all wavelengths of the ambient stray light are delivered to the detector. Because the stray ambient
light signal can represent a large fraction of the total light signal measured by the detector, it is a major source of
error. While this source or error can be minimized by completely shielding the sample form all sources of ambient
illumination, this often precludes the use of most reflectance and transmittance fiber optic probes.

In a post-dispersive spectrometer, the sample is illuminated with white light. Light scattered off or transmitted
through the sample is then dispersed and delivered to the entrance slit of the instrument’s spectrometer. As with
the pre-dispersive spectrometer, ambient light that strays into the sample being measured is also collected. The
difference is that in the post-dispersive instrument, only ambient stray light of the same wavelength as that being
measured by the detector is added to the signal resulting from the instrument’s illumination of the sample. Thus,
the stray ambient light signal represents a small fraction of the total light signal measured by the detector.

FR Direct Path Fiberoptic Input

The FieldSpec® FR spectroradiometer uses a 1 meter fiber optic input that feeds directly into the spectrometer.
There are two advantages to this arrangement.

First, the fiber optic input allows the user to quickly move and aim the very lightweight fiber optic probe from point
to point without having to move the entire spectroradiometer.

Secondly, since the fiber optic is connected directly into the spectroradiometer there is none of the signal losses
otherwise associated with detachable couplings (detachable couplings typically result in as high as 50% signal
loss).

FR Speed and Sensitivity

The FieldSpec® FR spectroradiometer can record a complete 350 - 2500 nm spectrum in 0.1 seconds. This
amazing speed allows the convenience of collecting more data in less time, as well as minimizing errors
associated with clouds and wind under solar illumination.

Another thing to consider is the number of scans that can be collected in a specified time period. The FieldSpec®
FR also the only portable spectroradiometer designed with a unique type of high speed bi-directional parallel
interface to the controlling computer to allow the averaging of continuous sequences of spectra. A serial interface
 ASD Technical Guide 3rd Ed. Section 1-2

(used in some competitors' instruments) would severely limit the number of scans per second that can be
averaged.

But, speed is not the only critical factor. Fast speed in combination with extremely low Noise-equivalent-Radiance
(NeDL) are what make the optimal spectroradiometer. The following typical NeDL and Signal-to-Noise plots
shows the superb sensitivity of the FieldSpec® FR spectroradiometer (Note: every spectrometer has a unique
NeDL).

-6
10
Noise Equivalent delta Radiance
NEdL, W/cm /nm/sr

Unit FSFR 6122/1, 9/29/98

-7
10
2

-8
10

-9
10
-9 -9 -9
1.3824 x 10 1.2012 x 10 3.2362 x 10
-10 @ 703 nm @ 1508 nm @ 2208 nm
10
500 1000 1500 2000 2500
Wavelength (nm)
SIGNAL-TO-NOISE RATIO

SOLAR ILLUMINATION
100% reflectance target
10000 60° solar zenith angle
atmospheric water absorption

Mid-latitude summer
atmospheric conditions
atmospheric water absorption
atmospheric water absorption

1000

100 ANALYTICAL
SPECTRAL
DEVICES, INC.
FieldSpec FR
10
500 1000 1500 2000 2500
Wavelength (nm)
 ASD Technical Guide 3rd Ed. Section 1-3

It is important to note that under solar illumination the signal-to-noise is low in the regions around 1400 nm and
1900 nm due to the very low or zero solar energy that reaches the Earth's surface due to atmospheric water
absorption in these regions.

Detachable fiberoptic 'jumper' cables, will result in about a 50 percent loss in signal. Also, longer 'permanent'
fiberoptic cables will result in signal attenuation, which can bee seen when comparing the NeDl of an FR that has
a standard 1 meter 'permanent' fiberoptic cable with an FR that has a 10 meter 'permanent' fiberoptic cable (see
plots below).

FR Longer Fiberoptic Cable Comparison

-6
10
Noise Equivalent delta Radiance
NEdL, W/cm /nm/sr

Unit FSFR 675/1, Lamp S802, 6/19/97

10
-7 With 1 meter 'permanent' fiberoptic cable
2

-8
10

-9
10
-9 -9 -9
1.3054 x 10 1.3704 x 10 3.8671 x 10
-10
@ 700 nm @ 1500 nm @ 2200 nm
10
500 1000 1500 2000 2500
Wavelength (nm)

-6
10
Noise Equivalent delta Radiance
NEdL, W/cm /nm/sr

Unit FSFR 665/1, Lamp S802, 2/27/97

10
-7 With 10 meter 'permanent' fiberoptic cable
2

-8
10

-9
10
-9 -9 -9
1.7598 x 10 2.0888 x 10 5.3368 x 10
-10
@ 700 nm @ 1500 nm @ 2200 nm
10
500 1000 1500 2000 2500
Wavelength (nm)
ASD Technical Guide 3rd Ed. Section 1-4
 ASD Technical Guide 3rd Ed. Section 2-1

Section 2
FR Sampling Interval and Spectral Resolution

The FieldSpec®FR spectroradiometer has a sampling interval of 2 nm and a spectral resolution of 10 nm, which
meet the nominal sampling and resolution requirements for hyperspectral remote sensing applications (Curtiss B.
and Goetz, A. F. H., Field Spectrometry: Techniques and Instrumentation, Presented at the International
Symposium on Spectral Sensing Research, July 1994 ).

Spectral sampling interval is the spacing between sample points in the spectrum. Sampling is independent of
resolution and in ASD spectroradiometers is between 2 and 5 times per FWHM. The sampling interval for the
FieldSpec®FR is 1.4 nm for the region 350 - 1000 nm and 2 nm for the region 1000 - 2500 nm.

Spectral resolution is defined as the full-width-half-maximum (FWHM) of the instrument response to a

monochromatic source. This is in fact the definition ASD uses when stating spectral resolution specifications.
Careful! When reading manufacturers specifications, do not make the mistake of interpreting sampling interval as
resolution. The FWHM spectral resolution of the FieldSpec®FR spectroradiometer is 3 nm for the region 350 -
1000 nm and 10 nm for the region 1000 - 2500 nm. These spectral resolution values have been measured by
calculating the FWHM of a near monochromatic peak in a spectrum acquired when viewing the output of a
monochromator with the FieldSpec®FR spectroradiometer.

Sampling Interval
Spectrometer offset corrected raw DN

Spectral
Resolution
50% of peak height

Wavelength

IMPORTANT! Please be careful not to make the mistake often made when reading our competitor’s
specifications. If you look at the number of channels they quote and calculate the minimum possible sampling
interval, you come up with a number that looks like what some competitors present as ‘bandwidth’. Many readers
of these ‘bandwidth’ numbers mistakenly interpret them as FWHM spectral resolution, thereby greatly
overestimating the instrument’s capability.
 ASD Technical Guide 3rd Ed. Section 2-2

The FieldSpec® FR is comprised of three separate spectrometers all in the same enclosure. We call these three
spectrometers: the VNIR spectrometer (VNIR = Visible & Near Infrared), the SWIR1 spectrometer, and the
SWIR2 spectrometer (SWIR = Short Wave Infrared).

Diagram of VNIR spectrometer

The VNIR spectrometer uses a fixed concave holographic reflective grating that disperses the light onto a fixed
photodiode array that has 512 individual detection points or ‘elements’, in a line. Associated with each of these
elements is a distinct signal whose magnitude is determined by the total integrated amount of light energy falling
on that element. For now, we’ll ignore the magnitude of the signal and first concentrate how an element position
corresponds to a narrow band or ‘wavelength channel’. We assign each element a position, X1, X2, X3, . . . ,
X512

Diagram of SWIR1 spectrometer

The SWIR1 spectrometer uses a concave holographic reflective grating that rotates up and then down on its axis
thereby scanning the dispersed light across a fixed single InGaAs graded index TE cooled detector. As the
dispersed light is scanned across the detector in the up scan direction, a distinct signal is recorded for each of
530 positions of an encoder in-line with the rotation axis of the grating. We assign each SWIR1 upward encoder
position XA1, XA2, XA3, . . . , XA530. The upper limit can actually be as high as A535 depending upon the
mechanical limits of the scanning device, which are unique for each spectrometer. And we assign each SWIR1
downward encoder position XB1, XB2, XB3, . . . , XB530.

Therefore, there are at least 1060 distinct encoder positions each with its associated energy signal, for each full
cycle scan of SWIR1.
 ASD Technical Guide 3rd Ed. Section 2-3

The SWIR2 is constructed and operates exactly the same way as the SWIR1 spectrometer except; the grating
and detector are manufactured for the longer SWIR2 wavelength region. The positions are assigned similarly. The
number of encoder positions is similar for SWIR2.

After the element and encoder positions are assigned, ASD chooses the positions where the transitions should
occur between the VNIR, SWIR1, and SWIR2 spectrometers. These ‘splice positions’ are chosen so that the
energy signal-to-noise ratio is optimized at these critical points. The splice between the VNIR spectrometer and
the SWIR1 spectrometer turns out to occur at around 1000 nm, where the response of the SWIR1 spectrometer is
superior to that of the VNIR photodiode array. The splice between SWIR1 and SWIR2 turns out to occur at
around 1800 nm. The exact splice positions are rarely the same for every instrument because the optimum
response positions of the detectors and gratings are slightly different for each machine.

This splicing optimization results in a few VNIR elements being ignored in exchange for the superior signal-to-
noise ratio of the SWIR1 at those positions. A few SWIR1 positions might also be ignored in exchange for the
superior signal-to-noise ratio of the SWIR2 at those positions.

The FieldSpec®FR is set to view NIST traceable wavelength references such as emission sources, reflectance
standards, and the output of a triple monochromator. ASD chooses at least 40 of these references that produce
signals with center-peaks at “known” wavelengths, Wk1, Wk2, Wk3, . . . , Wk40 sufficiently distributed throughout
the region 350 – 2500 nm. Each known Wk is paired with the known element-or-encoder position Xk where the
strongest energy signal is read while viewing that Wk reference. The results are data points with known element-
or-encoder-position and wavelength-channel coordinates, (Xk, Wk).
Wavelength Channel (W)

Wavelength Channel (W)

Group A Group B

Interpolated Interpolated

Known
Known

VNIR Array Element Position (X) SWIR1 Encoder position (A) or (B)

An appropriate cubic spline interpolation function is used to pair up the remaining element-or-encoder positions
with interpolated wavelength-channels, to 1 nm intervals. The results are a complete set of 2,625 or so ordered
pairs (assuming 7 ignored VNIR elements, 1056 SWIR1 and 1056 SWIR2 positions, splices at 1000 and 1800
nm, and group C and D for SWIR2):

(X1,350), (X2, 351), (X3, 352), . . . , (X505, 999),

(XA1, 1000), (XA2, 1001), (XA3, 1002), . . . , (XA528, 1799),
(XB1, 1000), (XB2, 1001), (XB3, 1002), . . . , (XB528, 1799),
(XC1, 1800), (XC2, 1801), (XC3, 1802), . . . , (XC528, 2500),
(XD1, 1800), (XD2, 1801), (XD3, 1802), . . . , (XD528, 2500)

The interpolated wavelength-channels are then verified using the output of the triple monochromator.

So far, we’ve only talked about pairing up the element-or-encoder-positions with wavelength-channels. For clarity,
we have been ignoring the third coordinate in the total data system. That third coordinate is the detector ‘energy’
signal E, which results in the set of triplets:

(X1,350, E1), (X2, 351, E2), (X3, 352, E3), . . . , (X505, 999, E555),
(XA1, 1000, EA1), (XA2, 1001, EA2), (XA3, 1002, EA3), . . . , (XA528, 1799, EA528),
(XB1, 1000, EB1), (XB2, 1001, EB2), (XB3, 1002, EB3), . . . , (XB528, 1799, EB528),
(XC1, 1800, EC1), (XC2, 1801, EC2), (XC3, 1802, EC3), . . . , (XC528, 2500, EC528),
 ASD Technical Guide 3rd Ed. Section 2-4

(XD1, 1800, ED1), (XD2, 1801, ED2), (XD3, 1802, ED3), . . . , (XD528, 2500, ED528)

The above complete set of triplets shows, that interpolation results in two independent energy measurements for
each SWIR1 and SWIR2 wavelength-channel, i.e., two independent data points for each SWIR1 and SWIR2
wavelength-channel, one data point for the up scan, and one for the down scan. Each of these up-down pairs is
too closely spaced spectrally to plot them separately within the 1 nm interpolation resolution. Therefore, the only
practical thing to do is to co-add the paired energy signals:

EAB1 = (EA1 + EB1)/2, EAB2 = (EA2 + EB2)/2, and so forth,

forming the new set of ordered pairs:

(350, E1), (351, E2), (352, E3), . . . , (999, E505),

(1000, EAB1), (1001, EAB2), (1002, EAB3), . . . , (1799, EAB528),
(1800, ECD1), (1801, ECD2), (1802, ECD3), . . . , (2500, ECD528)

Hence, after interpolation and up-down co-adding we have:

649 1 nm data points for the region 350 – 999 nm
799 1 nm data points for the region 1000 – 1799 nm
700 1 nm data points for the region 1800 – 2500 nm

If the customer requests it, the FieldSpec® software can also contain, the VNIR-array-only-mode software, which
allows the user to switch off the FR software and use all available VNIR array elements for the region 350 - 1050
nm only. This VNIR-array-only-mode also allows the user to manually select integration times of 2n x 17
milliseconds for n = 0, 1, ... , 15.

If you intend to measure radiance or irradiance in the "VNIR-array-only-mode", you will also need to purchase the
appropriate "VNIR-array-only-mode" radiometric calibrations, which are in addition to, and completely separate
from, any "FR-mode" radiometric calibrations you may have also purchased. For example, if you want to measure
radiance in the "VNIR-array-only-mode" using item no. 11925 (1 Deg NIR/FR FOV Lens Foreoptic), then you will
also need item no. 70051 (Radiometric Calibration of 1 Deg UV/VNIR FOV Lens Foreoptic). Of course, if you also
want to measure radiance in the "FR-mode" using item no. 11925, then you will also need item no. 70151
(Radiometric Calibration of 1 Deg NIR/FR FOV Lens Foreoptic).
 ASD Technical Guide 3rd Ed. Section 3-1

Section 3
FR Repeatability

The FieldSpec®FR spectroradiometer has excellent repeatability. The following plots are self-explanatory.

0.48
0 minutes
0.47 5 minutes
10 minutes
15 minutes
0.46 20 minutes
25 minutes
0.45 30 minutes
Reflectance

0.44

0.43
Repeatability test
using Kaolinite reflectance
0.42 automatically collected
every 5 minutes using the
same white reference
0.41 and dark current

0.40
No wavelength difference.
0.39
± 0.2% for reflectance.
0.38
2150 2160 2170 2180 2190 2200 2210 2220
Wavelength (nm)
1 1

0.9 0.9

0.8 0.8
Reflecta nce

Reflectance

0.7 0.7

0.6 0.6

0.5 0.5

0.4 Kaolinite Reflectance Spectra, March 16th, 1996 0.4 Kaolinite Reflectance Spectra, March 16th, 1996
ASD FieldSpec FR, instrument #644, 5 sec avg. ASD FieldSpec FR, instrument #644, 5 sec avg.
50W Tungsten Halogen Source, 30 cm from sample 50W Tungsten Halogen Source, 30 cm from sample
0.3 0.3
500 1000 1500 2000 2500 2000 2044 2089 2133 2178 2222 2267 2311 2356 2400

Wavelength, nm Wavelength, nm
 ASD Technical Guide 3rd Ed. Section 3-2

FR Fiberoptic Cable Bending Effects

1.5%

Fiberoptic cable
Difference due to 'Curling' FieldSpec FR Fiberoptic Cable

1% Curled to 1.75 inch diameter

Effect is less than 1%
Curled = complete circle of 1.75 inch radius
[100 percent * (Straight - Curled)/Straight]

0.5%

0%

-0.5%

-1%
350 780 1210 1640 2070 2500
Wavelength (nm)

FR Large Field-of-View

Many in-situ field applications require a broad field-of-view, for example, for a spot size that closely matches the
hyperspectral image pixels. Other examples include collecting data of large area backgrounds as well as other
large area targets. The standard built-in fiberoptic input of the FieldSpec® FR has the largest field-of-view of any
portable spectroradiometer (25 degrees full conical angle). Optional narrower field-of-view attachable foreoptics
are also available at additional cost. For more details see section 7.
 ASD Technical Guide 3rd Ed. Section 4-1

Section 4
FR Stray Light Rejection

Stray light is one form of constant systematic noise. There are two types of stray light that are considered
significant in terms of spectrometer performance: Diffraction that is not first order and internal light scattering that
is not diffraction related. It also important to recognize anomalies that look like stray light but really are not and
we'll discuss those here as well.

In general, recognizing stray light problems is not an easy task when comparing instruments. However, in the
case of field applications, there are several clues to look for in spectra that can reveal stray light problems as a
possibility.

Stray light results in computed reflectance values that are different from the actual values. The appearance of
reflectance signal in spectral regions of zero illumination energy and the appearance of biased signals in spectral
regions of low illumination energy are indicators of stray light problems.

The case of untrue signal in regions of zero illumination energy can be discovered in the near infrared regions by
observing the 'water bands' under solar illumination. In the figure below, solar energy reaching the earth’s surface
is essentially zero in the bands around 1400 nm and 1900 nm due to the atmospheric water vapor absorbing light
in these regions over the long path length of the atmosphere. There are some occasions when total column water
vapor is low, such as very dry cloud free atmospheric conditions over a desert, where it is sometimes possible to
measure some true signal in the 1400 nm band, but signal is always too low in the 1900 nm band.

0.30
Fraunhofer lines
Chappius O
3
absorption band
0.25
H O doublet
2
absorption
Solar Radiance W/m /steradian/nm

0.20
H O absorption
2
2

0.15

0.10 O H O absorption
2 2
absorption

0.05 O absorption CO
3 2
absorption

0.00
400 600 800 1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000
Wavelength (nm)

In order to understand what the computed reflectance should look like in these zero energy bands it is useful to
review the meaning of measured signal and computed reflectance:
Measured signal = true signal + dark current + stray light + random noise
As we'll discuss later, dark current can be easily recorded and subtracted so that it is a negligible contributor.
Therefore, assuming that the dark current is pre-subtracted we can re-write the above formula as follows:
Dark corrected measured signal = true signal + stray light + random noise
Using dark corrected measured signal, computed reflectance is written as follows:
 ASD Technical Guide 3rd Ed. Section 4-2

Computed reflectance
= (dark corrected measured signal from target) / (dark corrected measured signal from reference)
= (true target signal + stray light + random noise) / (true reference signal + stray light + random noise)

If stray light is negligible, then for regions of near zero illumination energy, i.e., signal less than or equal to the
instrument random noise at the time the signals are recorded, computed reflectance is written as:
Computed reflectance
= (random noise at time of target measurement) / (random noise at time of reference measurement)

Even though the random noise signals are extremely small, for example 5 DN out of 65,536 DN the ratio of one
random noise signal to another can range between near zero and greater than one. For example, suppose at the
time the reference measurement was take the random noise signal was 3 DN at 1900 nm wavelength channel,
and suppose at the time the target measurement was taken the random noise signal was 6 DN at the same 1900
nm. Then the computed reflectance at that 1900 nm channel would be 200 percent. Graphically, this would be a
vertical line that shoots upward from the last wavelength channel with a non-zero measured signal. Likewise let's
say that at the 1901 nm wavelength channel the random noise levels are such that the ration is near zero.
Graphically, this would be another vertical line next to the last one that shoots straight down to zero. And so forth.
This effect is what should correctly be observed in the water bands as shown in the figure below.

Water band noise

1.00
(This is right!)
FieldSpec FR Reflectannce of Vegetation

0.80

ASD
0.60 FieldSpec FR

Vegetation
0.40 Reflectance

Solar
Illumination
0.20

Normal
Humidity
0.00
400 900 1400 1900 2400
Wavelength (nm)

As previously mentioned, under very dry conditions, it is sometimes possible to observe some true signal in the
1400 nm band, but even under these conditions random noise is still dominate in the 1900 nm water band as
shown in the figure below.

0.5
ASD FieldSpec FR

0.4
Vegetation
Reflectance

Solar Illumination
Reflectance

0.3
Very Dry Conditions

0.2

0.1

0
400 900 1400 1900 2400
Wavelength (nm)
 ASD Technical Guide 3rd Ed. Section 4-3

Now, going back to the formula for computed reflectance:

computed reflectance
= (true target signal + stray light + random noise) / (true reference signal + stray light + random noise)

Again, we consider the water bands where illumination energy is zero, but this time we consider the case of
significant stray light. That is, stray light significantly greater than the lowest level random noise. Computed
reflectance is then written as follows:
computed reflectance
= (stray light + random noise) / (stray light + random noise)

In this case, the computed reflectance is no longer a simple ratio of random noise signals but the ratio of two
significant signals. This anomaly results in what appears to be real reflectance signals where there should only be
random noise (see figure below).

Significant stray light signal in water bands (problem).

1.00
Vegetation reflectance with stray light problem.

0.80

Vegetation
0.60
Reflectance
w/Stray Light
Problem
0.40
Solar
Illumination
0.20
Normal
Humidity

0.00
400 900 1400 1900 2400
Wavelength (nm)

The question arises, so what if the spectrometer has such a low random noise level that these anomalies really
are signals? It is physically impossible for any scientific instrument to have a zero random noise component.
Therefore, if the source energy is zero, and stray light is truly negligible, a ratio of dark corrected signals should
only be a ratio of random noise signals.

We modeled these effects with two reflected solar radiance spectra collected with the FieldSpec®
spectroradiometer. The first reflected solar radiance spectrum is for a 100% reflectance target, the second is for a
50% reflectance target. The reflectance spectrum is computed by dividing the 50% reflected solar radiance
spectrum by the spectrum for the 100% target as shown in the figure below. If these measurements were made
using artificial illumination (i.e. a light source without deep absorption bands), the computed reflectance spectrum
would be a straight line at 50%. The noise seen at 1400 nm and 1900 nm confirms the excellent stray light
rejection of the FieldSpec® spectroradiometer. This excellent stray light rejection ensures that collected spectra
accurately represent the true reflectance of materials observed in the field.
 ASD Technical Guide 3rd Ed. Section 4-4

0.8

0.7

0.6

Reflectance
0.5

0.4

0.3

0.2

0.1

0
1000 1200 1400 1600 1800 2000 2200 2400
Wavelength (nm)

Figure above is reflectance spectrum for a 50% reflectance target computed by dividing the reflected solar
radiance spectra for a 50% reflectance target by that for a 100% reflectance reference panel.

We can model the effects of scattered light inside the spectrometer by adding a 0.5% stray light component to
these same spectra. Because this results in a signal in the deep water vapor absorption bands, the computed
reflectance spectrum plot below, does not show the noise seen in the 1400 and 1900 nm region in figure below.

0.8

0.7

0.6
Reflectance

0.5

0.4

0.3

0.2

0.1

0
1000 1200 1400 1600 1800 2000 2200 2400
Wavelength (nm)

The figure above is reflectance spectrum for a 50% reflectance target with the addition of a 0.5% stray light
component. This reflectance spectrum is computed by dividing the reflected solar radiance spectra for a 50%
reflectance target by that for a 100% reflectance reference panel after adding a offset term to each radiance
spectrum. The stray light offset terms were computed by multiplying the average value for each radiance
spectrum by 0.5%.

In order to further demonstrate the errors produced by stray light, we collected reflected solar radiance spectra for
both a polyester fabric and a 100% reflectance reference panel. The computed reflectance spectrum (see Figure
below) shows the effects of the deep atmosphere water vapor absorption band centered near 1900 nm.
 ASD Technical Guide 3rd Ed. Section 4-5

0.8
0.7
0.6

Reflectance
0.5
0.4
0.3
0.2
0.1
0
1600 1800 2000 2200 2400
Wavelength (nm)

When the same material is measured in the laboratory using a tungsten filament illumination source (see Figure
below), the reflectance spectrum shows absorption features in the 1900 nm region that were obscured in the
spectrum collected using solar illumination.

0.8
0.7
0.6
Reflectance

0.5
0.4
0.3
0.2
0.1
0
1600 1800 2000 2200 2400
Wavelength (nm)

The figure above is the reflectance spectrum for a polyester fabric with the addition of a 0.5% stray light
component. This reflectance spectrum is computed by dividing the reflected solar radiance spectra for a 50%
reflectance target by that for a 100% reflectance reference panel after adding a offset term to each radiance
spectrum. Note that the features present in the laboratory spectrum near 1900 nm are not present.

Stray light clues can also be observed at the shorter wavelengths (see figure below). Because the illumination
energy is low in this region, stray light produces a positive bias error. Because solar irradiance is dropping as
wavelength decreases in the ultraviolet and blue wavelengths, the stray light bias error is large and increases, i.e.,
smiles upward at shorter wavelengths.

0.4

0.3
Reflectance

0.2

0.1

0
350 400 450 500 550
Wavelength (nm)
 ASD Technical Guide 3rd Ed. Section 4-6

The figure above is the short wavelength end of the reflectance spectrum for a polyester fabric. The solid line is
the correct spectrum; the dotted line is the spectrum with the addition of a 1% stray light component. Because of
the low intensity of solar radiation in the UV and blue wavelengths, stray light inside the spectrometer produces
the largest positive bias error at shorter wavelengths. This effect is also observed in the UV region for the
vegetation example below.

Significant stray light signal in water bands (problem).

1.00

Significant stray light problem in UV.

Vegetation reflectance with stray light problem.

0.80

Vegetation
0.60
Reflectance
w/Stray Light
Problem
0.40
Solar
Illumination
0.20
Normal
Humidity

0.00
400 900 1400 1900 2400
Wavelength (nm)

When comparing spectra from different instruments, it is important to note the illumination conditions as they
relate to stray light. For example, solar illumination provides much greater energy in the UV then a typical close-
up DC powered tungsten filament lamp. Therefore, stray light problems in the UV tend to reveal themselves more
so with artificial light sources then with solar illumination. On the other hand, close-up tungsten filament lamps
provide plenty of energy throughout the near-infrared spectrum since there are no long path atmospheric water
absorption columns. Therefore, stray light problems in these areas are revealed under solar illumination.

Of course, under artificial illumination, a good spectrometer should show no stray light problems as in the figure
below.

FR Reflectance of Vegetation

®
FieldSpec FR
Spectroradiometer
0.8 Reflectance of
Vegetation

ARTIFICIAL
0.6 ILLUMINATION
Reflectance

0.4

0.2

0
500 1000 1500 2000
Wavelength (nm)
 ASD Technical Guide 3rd Ed. Section 4-7

The stray light component for a given spectrometer can be approximated as a fraction of the raw signal integrated
over the wavelength range of the spectrometer. Typically, this fraction is in the range of 0.02% to 0.1% for a well-
designed single monochromator system such as that used in ASD’s FieldSpec® spectroradiometers.

In summary, poor stray light performance results in spectra that, while 'looking' better in the regions around the
deep atmospheric water vapor bands, have significant errors. Typical errors include large biases in both the short
and long wavelength ends of spectra measured in the field (due to the low solar irradiance in these regions).

Stray Light Look Alikes

As mentioned earlier, there are some anomalies that look like stray light effects but really are not. The most
significant of these look-alike effects is what is know as 'dark-current-drift', or simply 'dark-drift'. Dark-current is
systematic noise from the instrument electronics and detectors.

Dark-current can be measured by either viewing a black, near zero reflectance target or by closing a shutter on
the spectrometer input so that zero illumination energy strike the detectors. The dark-current signal can then be
stored and subtracted from all subsequent measurements. The FieldSpec® spectroradiometer includes
mechanically controlled shutters and software for recording and automatically subtracting dark-current.

Within short time periods, dark-current is relatively constant. However, during the initial start-up the spectrometer
goes through a period where the internal components and external ambient temperature attempt to reach thermal
equilibrium. During this period, the dark-current changes slowly as the change in temperature effects the
efficiency of the internal components. These effects are most significant at the outer boundaries of the detectors'
quantum efficiencies, i.e. at the low and high spectral regions of a given detector. Even after thermal equilibrium is
reached, less significant dark-drift can occur with fluctuations in external ambient temperature.

Two approaches can be used to minimize the effects of dark-drift. First, the detectors can be cooled to a very low
and very stable temperature. The FieldSpec® FR spectroradiometer includes thermo-electric (TE) cooling as
standard for the region 1000 - 2500 nm, and optional cooling is also available at additional cost for the region 350
- 1000 nm.

The second approach is to update the dark-current frequently during the first 15 minutes of operation, and then
less frequently after the system has reached thermal equilibrium. The frequency for dark-current updating after
thermal equilibrium should be based on the stability of the ambient conditions and the priority of the regions most
effected by dark-drift.

As previously mentioned, dark-drift is most noticeable in the regions of least quantum efficiency. For example, in
the lower UV regions of a silicon detector. Such drift looks very similar to the upward smiling effect that can also
be caused by stray light. The difference is that a dark-drift smile forms gradually over time while stray light effects
are constant. Another example of dark-drift can be recognized by a smile-down at the upper most NIR regions of
the silicon detector. If stray light is negligible, these dark-drift effects can be easily dealt with by simply updating
dark-current more frequently. Even with cooled detectors, however, subtle changes in thermal equilibrium over a
long time period necessitates dark-current updating.

FR Driftlock

In addition to these methods of dealing with dark-drift, the more recent FieldSpec®FR's incorporate a unique
combination of hardware and software known as Driftlock ('Driftlock' is a trademark of ASD). Driftlock
automatically updates dark-current for every measurement from masked pixels in the silicon array.
ASD Technical Guide 3rd Ed. Section 4-8
 ASD Technical Guide 3rd Ed. Section 5-1

Section 5
FR Suggested Set-up

1. In the control configuration window, set the sample, white reference** and dark current averages to 10 scans
(type: "cntl-C" to open this window).
2. Then, optimize the instrument (type “cntl-O” or click on the OPT button to initiate optimize).
3. Reopen the control configuration window, and set the white reference and dark current averages to 60 for
indoor illumination*, or 150 for outdoor solar illumination.
4. Take white reference (type F4 key or click on the WR button).
5. Reopen the control configuration window, and set the spectrum average to 30 for indoor illumination*, or 100
for outdoor solar illumination.
6. Observe samples.

* Stable DC power supply with quartz halogen lamp and metal reflector operating at a color temperature of at
least 3000 Deg K. For example, Lowel Pro-Light, P1-10 & 50 Watt Lowel Bulb LOL12V50WPL #JC145V50WC
12-14V/50W 3200K LAMP (B&H Photo-Video, 119 West 17th Street, New York, NY 10011, PH: 212-444-6601,
FX: 212-444-5001) and Tamura SWA1506W 15VDC Table Top Power Supply (Tamura Corp. of America,
Temecula California). Lowel Pro-Light and Tamura power supply are also available from ASD (See Price List).

** For example, White Spectralon (Labsphere, Inc., Shaker St., PO Box 70, North Sutton, NH 03260-0070, PH:
603-927-4266, FX: 603-927-4694, E-mail: [email protected], Website: www.labsphere.com). Spectralon is
also available from ASD (See Price List).

FACTORS TO CONSIDER UNDER SOLAR ILLUMINATION

Clouds Sensor
Solar
Angle
Path
Radiance Reflected
Light
From Wind
Direct Target
Sun Light
Sky Light
Scatter From
Surroundings
Shadows
Target (maintain same angle to sun)

FACTORS TO CONSIDER UNDER ARTIFICIAL ILLUMINATION

Diffuse Reflectance FieldSpec
Tungsten Quartz Halogen Lamp Fiberoptic Input
Metal Reflector
Either fill the field-of-view
Stable
with sample or use back-
DC
ground metal plate painted
Power
Sample with ultra flat black paint,
Supply
or white or black Spectralon.

Maintain same distance & angle of lamp to sample.

Transmission Through Liquids Collimation Lenses

Tungsten Quartz Halogen Lamp Cuvette
FieldSpec
for Liquid
Metal Reflector Fiberoptic
Input
Stable
DC
Power
Supply
Maintain same distance & angle of lamp to lenses.
 ASD Technical Guide 3rd Ed. Section 5-2

Shown below is aTypical Vegetation Spectrum Under Solar Illumination.

Water band noise

1.00
(This is right!)
Fie ldSpec FR Re flectannce of Vege tation

0.80

ASD
0.60 FieldSpec FR

Vegetation
0.40 Reflectance

Solar
Illumination
0.20

Normal
Humidity
0.00
400 900 1400 1900 2400
Wavelength (nm)

Shown below is a Typical Solar Energy Spectrum at Earth's surface.

0.30
Fraunhofer lines
Chappius O
3
absorption band
0.25
H2O doublet
absorption
Solar Radiance W/m /steradian/nm

0.20
H O absorption
2
2

0.15

0.10 O2 H O absorption
2
absorption

0.05 O absorption CO
3 2
absorption

0.00
400 600 800 1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000
Wavelength (nm)
 ASD Technical Guide 3rd Ed. Section 5-3

Shown below is a Typical Vegetation Spectrum under DC powered Tungsten Quartz Halogen Artificial
Illumination.

®
FieldSpec FR
Spectroradiometer
0.8 Reflectance of
Vegetation

ARTIFICIAL
0.6 ILLUMINATION
Reflectance

0.4

0.2

0
500 1000 1500 2000
Wavelength (nm)

Shown below is a typical energy spectrum of a DC powered tungsten quartz halogen lamp. Tungsten is the
filament, quartz is what the glass envelope is made of, and halogen is the gas inside the bulb. Filament emitters
with the appropriate transparent envelope containing a non-absorbing gas, give nice smooth energy curves.
Because this energy is measured 'close-up' in the lab, there are no long-path atmospheric absorption features.
Also, there are no solar atmosphere absorption features; i.e., no Fraunhofer lines.

0.3

Tungsten-Quartz-Halogen
0.25
Lamp ('Gray Body')
Irradiance, W/m^2/nm/sr

0.2

0.15

0.1

0.05

0
500 1000 1500 2000 2500
Wavelength (nm)
 ASD Technical Guide 3rd Ed. Section 5-4

In contrast to filament emitters, low pressure, gas emitters give energy spectra with sharp 'emission lines' (high-
pressure gases produce a smooth spectrum). For example, Xenon solar simulators give a rough approximation at
best with several sharp discrepancies from a true solar spectrum (Figure below).
6.0

Irradiance (W/m^2/nm) 5.0

Typical Xenon
4.0 Solar Simulator
3.0

2.0

1.0

0.0
300 800 1300 1800 2300
Wavelength (nm)

Shown below is a typical energy spectrum of room fluorescent lights powered by 60 cycle AC. Notice that their is
little or no energy long of 800 nm.
Energy

Spectrum of Typical
Fluorescent Room Lights

350 400 450 500 550 600 650 700 750 800 850 900 950
Wavelength (nanometers)

Typical emission spectrum of Mercury Argon tubes (shown below).

Mercury Argon
Emission Spectrum
Energy

763.0
546.4 852.0

300 400 500 600 700 800 900 1000 1100

Wavelength (nm)
 ASD Technical Guide 3rd Ed. Section 5-5

Artificial Illumination and Mineral Reflectance

Sometimes solar illumination is impractical for collecting field spectra, such as in a cave. Also, sometimes all that
is needed is in-situ sample measurements that are not necessarily needed for ground truthing hyperspectral
imagery. In those cases, a portable, battery powered, artificial light source that interfaces with the FR fiberoptic
input would be just the ticket. ASD offers such an optional device at addition cost, and it is the High Intensity
Reflectance Probe, a.k.a, 'potato masher' (see picture and drawing below).

In the picture above right, the High Intensity Reflectance Probe is also equipped with the optional FOV limiter plat
for smaller spot sizes (sold separately). The FieldSpec®FR fiberoptic cable is inserted through the gray strain
relief spring, into the illumination cavity to view the sample. Under artificial illumination the following FieldSpec®
FR reflectance spectra are of Calcite, Kaolinite, Gypsum, Talc, Calcium Phosphate Tribasic, Oxtriphylline,
Diphenehydramine HCL, Sugar, Magnesium Carbonate, Procainamide HCL, Quinapril, and Oil Sand Bitumen.

1.0

0.8
ASD FieldSpec FR
Reflectance of Calcite
0.6
Artificial Illumination

0.4

0.2

0.0
500 900 1300 1700 2100 2500
Wavelength (nm)
 ASD Technical Guide 3rd Ed. Section 5-6

0.9

0.8
Reflectance

0.7

0.6

0.5

Kaolinite Reflectance Spectra, March 16th, 1996

0.4
ASD FieldSpec FR, instrument #644, 5 sec avg.

0.3
500 1000 1500 2000 2500

Wavelength (nm)

1.0

0.8

0.6
ASD FieldSpec FR
Reflectance of Gypsum
0.4 Artificial Illumination

0.2

0.0
500 900 1300 1700 2100 2500
Wavelength (nm)
 ASD Technical Guide 3rd Ed. Section 5-7

1.0

0.8
ASD FieldSpec FR
Reflectance of Talc
0.6
Artificial Illumination

0.4

0.2

0.0
500 900 1300 1700 2100 2500
Wavelength (nm)

FieldSpecFR
Reflectance (10% intervals). Spectra are offset for clarity.

Calcium
Phosphate
Tribasic
Oxtriphylline
Diphenehydra-
mine HCl
Sugar
Magnesium Carbonate
Procainamide HCl
Quinapril
1000 1200 1400 1600 1800 2000 2200
Wavelength (nm)
 ASD Technical Guide 3rd Ed. Section 5-8

0.70

Oil Sand 5.1% Bitumen

Oil Sand 10.4% Bitumen
0.60 Oil Sand 14.5% Bitumen

0.50 FieldSpecFR
Reflectance

0.40

0.30

0.20

0.10

0.00
500 1000 1500 2000 2500
Wavelength (nm)
 ASD Technical Guide 3rd Ed. Section 6-1

Section 6
FR Foreoptics

The FieldSpec®'s fiber optic cable provides the flexibility to adapt the instrument to a wide range of applications.
While many people choose to perform field measurements while handholding the FieldSpec®'s fore optics, the
FieldSpec®'s pistol grip and other fore optic accessories are equally suited to tripod mounting. Shown on the
below, the fore optic pistol grip is oriented perpendicular to the ground surface while user is oriented such that the
user's shadow is 180° from the instrument's field-of-view.

Shown below, the fore optic pistol grip is mounted on a tripod.

 ASD Technical Guide 3rd Ed. Section 6-2

Typically, reflected radiance and surface reflectance measurements are made using the hand-held configuration.
When required for more precise orientation, the FieldSpec®'s pistol grip is available with both a sighting scope
and leveling device as shown below.

These accessories allow the user to view the exact spot where the fore optic is pointed, and allow the fore optic to
be oriented in precise, nadir-viewing, geometry. Because of the need to orient the irradiance collector in precise
geometry, the majority of irradiance measurements are performed using the fore-optics mounted on a tripod as
shown below.

The small size of the FieldSpec®'s fore optics greatly reduce errors associated with instrument self-shadowing.
Even when the area viewed by the fore optic is outside the direct shadow of the instrument, the instrument still
blocks some of the illumination (either diffuse skylight or light scattered off surrounding objects) that would
normally be striking the surface under observation. Thus, the instrument (as well as other objects -- including the
 ASD Technical Guide 3rd Ed. Section 6-3

user) should be placed as far as possible from the surface under observation. This orientation requirement also
applies white reference measurements as shown below.

When the fore optics are built into the instrument, it is difficult to minimize this source of error. The small size of
the FieldSpec®'s fore optics also allows the user to more easily position the fore optics at a greater distance from
the surface under observation. In many cases, the purpose of the field observations is the collection of ground-
truth spectra. Because the pixel size of most imaging sensor systems is several meters or more, a field
spectroradiometer with a larger field-of-view means that fewer measurements are needed to approximate the
spatial resolution of the imaging sensor. While many field portable spectrometers have fore optics that are integral
to the instrument, ASD's FieldSpec® line of spectroradiometers is available with a wide selection of fore optics.
These interchangeable fore optics provide the user the flexibility required to adapt the FieldSpec®
spectroradiometer to a wide range of applications.

25 deg (59.9 cm diam)

18 deg (42.8 cm diam)

8 deg (18.8 cm diam)

1 deg (2.4 cm diam)

5 deg (11.8 cm diam)

3 deg (7 cm diam)

Drawing Scale = 1:7

 ASD Technical Guide 3rd Ed. Section 6-4

The figure above shows the available fields-of-view (FOV) for the FieldSpec® FR with an instrument fore optic
height of 135 cm. The dashed circle represents the FOV of a non-ASD instrument with a fixed 3° FOV. The solid
circles are for ASD's FieldSpec® FR. The largest circle is the FOV of the FieldSpec®'s standard built-in fiberoptic
cable, with optional foreoptics providing 1°, 5°, 8°, or 18°. Fore optics covering approximately the same range of
angular FOVs are available for the other FieldSpec® instruments.

Irradiance Observations

ASD has several types of fore optic for irradiance measurements. These include:
ASD's in-air cosine corrected receptors for measurement of total irradiance
ASD's accessories for measurement of direct irradiance
ASD's under-water cosine corrected receptors for measurement of in-water up- and down-welling irradiance

Shown below left is a FieldSpec® FR spectroradiometer (Item no. A109000) with the FieldSpec® Full Sky
Irradiance Remote Cosine Receptor (Item no. A124500), mounted on the Bogan tripod (Item no. 128560). The
tripod tray is not a standard item.

Shown above right top is a close-up of the FieldSpec® Full Sky Irradiance Remote Cosine Receptor (Item no.
A124500), mounted on the Bogan tripod (Item no. 128560). The base of the irradiance receptor has built-in
bubble levels for horizontal alignment of the receptor. The FieldSpec®'s fiber optic cable enters the receptor
through the gray strain relief fitting.

Shown above right bottom is the Reflective Cosine Receptor (iten no. A124700) for measuring full-sky-irradiance
with optimum S/N and optimum cosine accuracy throughtout the entire 350 - 2500 nm spectral region. This
version of the reflective cosine receptor includes a protective dome and is designed for long-term installation.
Radiometric Calibration of RCR / S701550, which is sold separately is also required for irradiance (W/m^2/nm)
calculation.
 ASD Technical Guide 3rd Ed. Section 6-5

The Direct Irradiance Attachment (Item no. A119720), shown in the below left, fits over the Full Sky Irradiance
Remote Cosine Receptor (Item no. A124700) that is available separately. It acts to limit the angular field-of-view
(FOV) of the irradiance receptor. Three interchangeable FOV modules are included for 2°, 1.5°, and 0.5°. FOV
aiming sight allows for proper alignment to the solar disk. This accessory is used for measuring the direct
component of solar irradiance. Because it uses the same optics as are used for the RCR, relative radiometric
errors between total and direct irradiance measurements are minimized.

The Fore Optic Equatorial Mount (Item no. 119730), shown above right, is used to maintain the orientation of the
Direct Irradiance Attachment (Item no. A119720-- available separately) and the Full Sky Irradiance RCR (Item no.
A124700-- available separately) relative to the sun. Once the mount is manually aligned with the Earth's rotational
axis and proper solar alignment is obtained using the directional keypad, this mount scans in synchronization with
the motion of the sun through the sky. The mount is powered by a rechargeable 12 volt gel cell battery pack
included with the mount. The heavy duty tripod on which the mount is installed is also included. The
FieldSpec®FR is mounted on a Bogan tripod (Item no. 128560-- available separately). The tripod tray is not a
standard item.
ASD Technical Guide 3rd Ed. Section 6-6
 ASD Technical Guide 3rd Ed. Section 7-1

Section 7
FR Fiberoptic Cable

PROBE PVC
END Monocoil
Sheathing
FR FO CABLE No Splices
DIMENSIONS ARE O.D. ø0.190
IN INCHES UNLESS CM 0.020 X 45°
OTHERWISE NOTED
CM 0.028 X 45°

ø0.250

1.000
ø0.125 0.500
ø0

EXPLODED VIEW OF PROBE END

FIBERS ARE RANDOM MIX OF:
.1

38 ø.220 MICRONS
25

19 ø.110 MICRONS

.2 50
ø0

ø0.067

FieldSpec® FR 1 deg Field-of-View Lens Foreoptic

ø

1/2-20 UNF Thread

0
0.

97
15

Thread relief as needed

1.
6

0.380
ø 2.000

0.500 FR 1 DEG LENS ø1.900 I.D.

DIMENSIONS ARE
4.623 IN INCHES
5.173
 ASD Technical Guide 3rd Ed. Section 7-2

FR 1 Deg Field-of-View Mirror Foreoptic

FR 1 DEG MIRROR
DIMENSIONS ARE
IN INCHES

2.75
Rifle Scope

10.00[25.4]

Concave Mirror

Tripod Mountable Grip

12.25[31.1]
M2

Fiber Optic Cable

1° Mirror FOV
45° Mirror
 ASD Technical Guide 3rd Ed. Section 7-3

FR 5 deg Field-of-View Lens Foreoptic

ad

ø 0.810
re
Th
ø
FR 5 DEG F
0.

DIMENSIONS ARE N
U
15

IN INCHES 0
-2 5
1/
2 71
6

0.
ø

0.380

ø
0.
0.500

62
1.454

5
2.000

FR 8 deg Field-of-View Lens Foreoptic

a d
ø 0.810

re
Th
ø

FR 8 DEG F
0.

DIMENSIONS ARE N
IN INCHES U
15

20 15
2-
1/
6

7
0.
ø

0.380
ø
0.

5.00
62
5

1.500
 ASD Technical Guide 3rd Ed. Section 7-4

FR 18 deg Field-of-View Tube

ø 0

d
ea

10
FR 18 DEG r
DIMENSIONS ARE Th
.15

ø 0.8

5
IN INCHES F

71
N
U
6

20

0.
/2-

ø
1

0.380
0.500 ø
0.
62
2.130
2.930 5
 ASD Technical Guide 3rd Ed. Section 7-5

FR PistolGrip

0.7700"
C.B. ø0.50 x 0.300 Dp.
ru 1/2-20 UNF x 0.450 Dp.
0.3850" T
h 2.610
0
6 0.1650"
.2
0
ø

0.5000"
1.00
1.00

1.50
0.40

4.30
5°
15

0.3250"
PISTOL GRIP °
DIMENSIONS ARE
IN INCHES
0.7700"

+0.003 0.6443"
-0.000 0.3222"

"
50
C.B. ø0.375 12
x 0.450 Dp.
0.
R

details
End View of
0 threaded
0.3850"

5
2
.1 insert
0
R

0.750
0.4750
0.3750
0.135 4-40 UNC Thru
To Other Side
ø 0.1

0.1000
260

ø0.450 Thread Relief

0.1000 ø0.500

ø 0.2570"

1/2-20 UNF Thread

 ASD Technical Guide 3rd Ed. Section 7-6

Approximating Spot Size

FO
V

β α
h
x

y
w z

arctan(y/x) = α = (FOV full angle)/2

y/x = tan α
y = x tan α

z + w = h tan (β + α)
w = [h tan (β + α)] - z
and
w = h tan β
Therefore:
z = h [tan(β)+ tan(β + α)]

Also w = square root of (x2 + h2 - 2xh cosβ)

EXAMPLE:
1 degree FOV Tube at x = 2 meters from perpendicular target (small ellipse):
y = (2 meters) tan (0.5 deg) = 0.0175 meters = 1.75 cm
So, for a perpendicular target the spot has a diameter of 3.5 cm

EXAMPLE:
Suppose we were limited to a 12.7 x 12.7 cm (.127 x .127 m) oblique target. So, to be on the safe side we will
want z = 0.635m. For h = 1 meter, β = 45 deg, 1 degree foreoptic:

w = [h tan (β + α)] - z = [tan(45.5 deg)] - 0.635]meters = 0.38 m

Also, in this case x2 = h2 + w2 = 1.144

 ASD Technical Guide 3rd Ed. Section 8-1

Section 8
FieldSpec® JR (350 -2500 nm)

ASD has developed the FieldSpec® JR spectroradiometer, a lower cost alternative to our popular FieldSpec® FR
spectroradiometer. The use of lower performance detectors and optics allow ASD to price the FieldSpec® JR
significantly lower that ASD’s existing FieldSpec® FR instrument. The FieldSpec® JR spectroradiometer is fully
upgradeable to the FieldSpec® FR spectroradiometer (at additional cost). The FieldSpec® JR spectroradiometer
has the same look and feel as the FieldSpec® FR but has poorer spectral resolution and noise equivalent
radiance. The following table and figures document the performance differences.

JR versus FR Spectral Resolution

@ 700 nm: 3 nm versus 3 nm
@ 1400 nm: 30 nm versus 10 nm
@ 2100 nm: 30 nm versus 10 nm
JR versus FR Noise Equivalent delta Radiance (NeDL)
@ 700 nm: 9.6 x 10-9 versus 4.8 x 10-9
@ 1400 nm: 2.4 x 10-9 versus 2.4 x 10-9
@ 2100 nm: 1.1 x 10-8 versus 8.8 x 10-9

JR versus FR Plots

Spectra in all plots represent a one second measurement (an average of ten 0.1 second scans). Samples were
illuminated using a 50 watt tungsten filament light source powered by a stable DC power source (ASD part
number A128930). The first plot is transmission spectrum of Mylar film; entire spectrum (shown below).
Reflectance (20% intervals)

Mylar, JR
Mylar, FR
Spectra have been offset for improved viewing

500 1000 1500 2000 2500

Wavelength, nm
 ASD Technical Guide 3rd Ed. Section 8-2

Shown below is the transmission spectrum of Mylar film; detail of the region of 2150 nm triplet.

Mylar, JR
Mylar, FR
Reflectance (10% intervals)

Spectra have been offset for improved viewing

2050 2100 2150 2200 2250

Wavelength, nm

Shown below is the transmission spectrum of Mylar; 1600-2000 nm region detail.

Reflectance (10% intervals)

Mylar, JR
Mylar, FR
Spectra have been offset for improved viewing

1600 1650 1700 1750 1800 1850 1900 1950 2000

Wavelength, nm
 ASD Technical Guide 3rd Ed. Section 8-3

Shown below is the reflectance spectrum of the mineral Kaolinite; entire spectrum.

1.0

0.9
Reflectance (10% intervals)

0.8

0.7

0.6

0.5
Kaolinite, JR
0.4 Kaolinite, FR
Spectra have been offset for improved viewing
0.3
500 1000 1500 2000 2500
Wavelength, nm

Shown below is the reflectance spectrum of the mineral kaolinite; detail showing the region of the 2200 nm
doublet.

0.8 Kaolinite, JR
Kaolinite, FR
Reflectance (10% intervals)

Spectra have been offset for improved viewing

0.7

0.6

0.5

0.4

2050 2100 2150 2200 2250 2300 2350 2400

Wavelength, nm
 ASD Technical Guide 3rd Ed. Section 8-4

Shown below is the reflectance spectrum of the mineral Kaolinite, detail showing the region of the 1400 nm
doublet.

0.9
Kaolinite, JR
Kaolinite, FR
Reflectance (10% intervals)

Spectra have been offset for improved viewing

0.8

0.7

0.6

0.5
1320 1340 1360 1380 1400 1420 1440 1460 1480
Wavelength, nm

JR Unique Highlights

1. A small sampling interval of 2 to 3 nm provides the necessary oversampling for ground truthing current
hyperspectral sensors. This size sampling interval is also necessary for calculating derivative spectra of nitrogen
and lignin. The FieldSpec® JR has sufficiently small sampling intervals of 1.4 nm for the region 350 - 1050 nm
and 2 nm for the region 1000 - 2500 nm. No other portable spectroradiometer meets these sampling interval
requirements.

2. A 'full' spectral range of 350 - 2500 nm is essential if the data is to be useful for analyzing the most up-to-date
hyperspectral imagery. No other portable spectroradiometer covers this spectral region while also meeting
requirement 1.

3. A very low Noise Equivalent Delta Radiance (NeDL) is required to measure many important yet subtle spectral
features. The FieldSpec® JR has an NeDL that is only a little higher than the NeDL of the FieldSpec® FR.

4. A very fast scan is required to minimize the effects of wind. Shadowing effects caused by wind can introduce
large errors. The FieldSpec® JR has a sufficiently fast scan time of 100 milliseconds for a complete 350 - 2500
nm spectrum.

Another thing to consider is the number of scans that can be collected in a specified time period. The FieldSpec®
JR is designed with a unique type of high-speed parallel interface to the controlling computer to allow the
averaging of continuous sequences of spectra. A serial interface (used in some competitors' instruments) would
severely limit the number of scans per second that can be averaged.

5. A fiber optic input is required for negligible self-shadowing effects. For some spectroradiometers, the input is
via lens optics mounted on the spectrometer enclosure. It is impractical if not impossible to make good reflectance
measurements of small spots with a spectrometer using this arrangement, since the spectrometer shadows the
sample. The FieldSpec® JR comes standard with a 1 meter, small diameter fiber optic input that carries signal
 ASD Technical Guide 3rd Ed. Section 8-5

directly into the spectrometer (signal carried directly into the spectrometer avoids large signal losses associated
with detachable / coupled fiber optic options that other manufactures may offer).

6. Many in-situ field applications require a broad field-of-view, for example, for a spot size that closely matches
the hyperspectral image pixels. Other examples include collecting data of large area backgrounds as well as other
large area targets. The standard built-in fiberoptic input of the FieldSpec® JR has the largest field-of-view of any
portable spectroradiometer (25 degrees full conical angle). Optional narrower field-of-view attachable foreoptics
are also available at additional cost.

7. When designing the FieldSpec® JR spectrometer, we paid particular attention to the potential problem of
scattered light. Careful baffling of the incoming light and special light traps for the zero-order light from the grating
have been implemented. All interior surfaces are painted with ultra-low reflectance coatings. The detectors are
covered with order separation filters. These filters reject not only second-order-diffracted light but also cut out
other stray scattered light reaching the detector elements. The holographic gratings and extensive baffling used in
the FieldSpec® JR spectrometers also significantly reduce scattered light noise. The stray light level within the
FieldSpec® JR spectrometers is less than 0.02%.

8. The FieldSpec® JR incorporates a unique combination of hardware and software known as Driftlock ('Driftlock'
is a trademark of ASD). Driftlock automatically updates dark-current for every measurement from masked pixels
in the silicon array.

9. Many in-situ field measurements require a portable instrument that is lightweight and carried in a way that
allows fast measurements while moving from target to target. The FieldSpec® can actually be worn around the
waist for easy walk-along, spot-to-spot measurements.

10. The FieldSpec® JR is fully upgradable to a FieldSpec® FR at additional cost.

FieldSpec® NIR JR (1000 - 2500 nm)

ASD has developed a JR, a lower cost alternative to our NIR version of the FieldSpec®. The use of lower
performance detectors and optics allow ASD to price the FieldSpec® NIR JR significantly lower that ASD’s
existing FieldSpec® NIR instrument. The FieldSpec® NIR JR spectroradiometer is fully upgradeable to the
FieldSpec® NIR spectroradiometer (at additional cost). The FieldSpec® NIR JR spectroradiometer has the same
look and feel as the FieldSpec® NIR but has poorer spectral resolution and noise equivalent radiance. Please
review the 1000 - 2500 nm region of the FieldSpec® JR specifications above for equivalent NIR JR specifications.
ASD Technical Guide 3rd Ed. Section 8-6
 ASD Technical Guide 3rd Ed. Section 9-1

Section 9
FieldSpec® UV/VNIR (350 -1050 nm)

The FieldSpec® UV/VNIR looks the same as a FieldSpec® FR. The primary differences are in spectral range total
weight and user interface settings. The UV/VNIR (or as we call it here at ASD,'VNIR') has user selectable
integration times, rather than the fixed scan time and auto optimization software of the FR.

Several important features of the FieldSpec® UV/VNIR make it the best and unique choice for a field portable
spectroradiometer:

The FieldSpec® UV/VNIR is a post dispersive spectrometer.

The FieldSpec® UV/VNIR spectroradiometer uses a direct path fiber optic input.
The FieldSpec® UV/VNIR spectroradiometer has an unbeatable combination of speed and sensitivity.
The FieldSpec® UV/VNIR spectroradiometer has the nominal spectral range of 350 - 1050 nm for analyzing many
current hyperspectral images, while maintaining optimal: spectral resolution, signal-to-noise.
The FieldSpec® UV/VNIR has excellent stray light rejection.
The FieldSpec® UV/VNIR features the unique Driftlock™ dark-drift correction system.
The FieldSpec® UV/VNIR has the largest standard field-of-view of any portable spectroradiometer.
The FieldSpec® UV/VNIR spectroradiometer is very lightweight, small and 'wear-able'.
The FieldSpec® UV/VNIR spectroradiometer has the ability to be upgraded to a dual spectrometer system.
The FieldSpec® UV/VNIR spectroradiometer has the ability to be upgraded to a FieldSpec® FR.

VNIR Post Dispersive Spectrometer

All ASD spectrometers are “post-dispersive”. Because portable spectrometers are typically used outside the
controlled laboratory environment, they are exposed to much higher levels of ambient light. In almost all cases,
some of this ambient light will stray into the sample being measured. The errors produced by this ambient stray
light are much greater for a pre-dispersive spectrometer than they are for a spectrometer that is post-dispersive.

In a pre-dispersive spectrometer, the sample is illuminated with monochromatic light. Light scattered off or
transmitted through the sample is then collected and delivered to the instrument’s detector. Ambient light that
strays into the sample being measured is also collected. Thus, both the monochromatic illumination from the
instrument and all wavelengths of the ambient stray light are delivered to the detector. Because the stray ambient
light signal can represent a large fraction of the total light signal measured by the detector, it is a major source of
error. While this source or error can be minimized by completely shielding the sample form all sources of ambient
illumination, this often precludes the use of most reflectance and transmittance fiber optic probes.

In a post-dispersive spectrometer, the sample is illuminated with white light. Light scattered off or transmitted
through the sample is then dispersed and delivered to the entrance slit of the instrument’s spectrometer. As with
the pre-dispersive spectrometer, ambient light that strays into the sample being measured is also collected. The
difference is that in the post-dispersive instrument, only ambient stray light of the same wavelength as that being
measured by the detector is added to the signal resulting from the instrument’s illumination of the sample. Thus,
the stray ambient light signal represents a small fraction of the total light signal measured by the detector.

VNIR Direct Path Fiberoptic Input

The FieldSpec®UV/VNIR spectroradiometer uses a 1.5 meter fiber optic input that feeds directly into the
spectrometer. There are two advantages to this arrangement.

First, the fiber optic input allows the user to quickly move and aim the very lightweight fiber optic probe from point
to point without having to move the entire spectroradiometer.

Secondly, since the fiber optic is connected directly into the spectroradiometer there is none of the signal losses
otherwise associated with detachable couplings (detachable couplings typically result in as high as 50% signal
loss).
 ASD Technical Guide 3rd Ed. Section 9-2

VNIR Fiberoptic Cable

PVC Monocoil PROBE END

PROBE
Sheathing

50
END Random Mix

ø0
No Splices

.2
O.D. ø0.207 19 ø.110

.1

ø0
FIBERS

25
CM
0.020
X 45°
CM 0.028 X 45°
PROBE END
ø0.250
All dimensions are in inches.

1.0000"
ø0.026
0.5000"
ø0.125

VNIR Speed and Sensitivity

The FieldSpec®UV/VNIR spectroradiometer can record a complete 350 - 1050 nm spectrum in as little as 17
milliseconds (user selectable integration times = 2n x 17 milliseconds for n = 0, 1, ... , 15). This amazing speed
allows the convenience of collecting more data in less time, as well as minimizing errors associated with clouds
and wind under solar illumination.

Another thing to consider is the number of scans that can be collected in a specified time period. The FieldSpec®
UV/VNIR also the only portable spectroradiometer designed with a unique type of high speed bi-directional
parallel interface to the controlling computer to allow the averaging of continuous sequences of spectra. A serial
interface (used in some competitors' instruments) would severely limit the number of scans per second that can
be averaged.

But, speed is not the only critical factor. Fast speed in combination with extremely low Noise-equivalent-Radiance
(NeDL) are what make the optimal spectroradiometer. The following typical NeDL and Signal-to-Noise plots
shows the superb sensitivity of the FieldSpec®UV/VNIR spectroradiometer (Note: every spectrometer has a
unique NeDL).

-7
10

Noise Equivalent Radiance

(NEdL), Unit FS VNIR 729/1,
-8
10 1-Second Measurements,
W/cm^2/nm/sr

Taken 2/4/98

-9
10

-10
10 400 500 600 700 800 900 1000
Wavelength (nm)
 ASD Technical Guide 3rd Ed. Section 9-3

Detachable fiberoptic 'jumper' cables, will result in about a 50 percent loss in signal. Also, longer 'permanent'
fiberoptic cables will result in some very minor signal attenuation for this spectral region.

VNIR Sampling Interval and Spectral Resolution

Spectral sampling interval is the spacing between sample points in the spectrum. Sampling is independent of
resolution and in ASD spectroradiometers is between 2 and 5 times per FWHM. The sampling interval for the
FieldSpec® UV/VNIR is 1.4 nm for the region 350 - 1050 nm.

Spectral resolution is defined as the full-width-half-maximum (FWHM) of the instrument response to a

monochromatic source. This is in fact the definition ASD uses when stating spectral resolution specifications.
Careful! When reading manufacturers specifications, do not make the mistake of interpreting sampling interval
as resolution. The FWHM spectral resolution of the FieldSpec® UV/VNIR spectroradiometer is 3 nm for the region
350 - 1050 nm. These spectral resolution values have been measured by calculating the FWHM of a near
monochromatic peak in a spectrum acquired when viewing the output of a monochromator with the FieldSpec®
UV/VNIR spectroradiometer.

Sampling Interval
Spectrometer offset corrected raw DN

Spectral
Resolution
50% of peak height

Wavelength

IMPORTANT! Please be careful not to make the mistake often made when reading our competitor’s
specifications. If you look at the number of channels they quote and calculate the minimum possible sampling
interval, you come up with a number that looks like what some competitors present as ‘bandwidth’. Many readers
of these ‘bandwidth’ numbers mistakenly interpret them as FWHM spectral resolution, thereby greatly
overestimating the instrument’s capability.

The FieldSpec® UV/VNIR spectrometer uses a fixed concave holographic reflective grating that disperses the
light onto a fixed photodiode array that has 512 individual detection points or ‘elements’, in a line. Associated with
each of these elements is a distinct signal whose magnitude is determined by the total integrated amount of light
energy falling on that element. For now, we’ll ignore the magnitude of the signal and first concentrate how an
element position corresponds to a narrow band or ‘wavelength channel’. We assign each element a position, X1,
X2, X3, . . . , X512 as shown in the diagram below.
 ASD Technical Guide 3rd Ed. Section 9-4

The FieldSpec® UV/VNIR is set to view NIST traceable wavelength references such as emission sources,
reflectance standards, and the output of a triple monochromator. ASD chooses about 14 of these references that
produce signals with center-peaks at “known” wavelengths, Wk1, Wk2, Wk3, . . . , Wk12 sufficiently distributed
throughout the region 350 – 1050 nm. Each known Wk is paired with the known element-position Xk where the
strongest energy signal is read while viewing that Wk reference. The results are data points with known element-
position and wavelength-channel coordinates, (Xk, Wk).
Wavelength Channel (W)

Interpolated

Known

VNIR Array Element Position (X)

An appropriate cubic spline interpolation function is used to pair up the remaining element-positions with
interpolated wavelength-channels, to 1 nm intervals. The results are a complete set of 700 ordered pairs:

(X1,350), (X2, 351), (X3, 352), . . . , (X512, 1050)

The interpolated wavelength-channels are then verified using the output of the triple monochromator.

So far, we’ve only talked about pairing up the element-positions with wavelength-channels. For clarity, we have
been ignoring the third coordinate in the total data system. That third coordinate is the detector ‘energy’ signal E,
which results in the set of triplets:

(X1,350, E1), (X2, 351, E2), (X3, 352, E3), . . . , (X512, 1050, E700).

VNIR Ground Truthing Hyperspectral Imagery

For analyzing many current hyperspectral images, a spectral range of at least 350 - 1050 nm is essential while
maintaining optimal spectral resolution and signal-to-noise. For example, strong correlations have been found
between: total wet and dry biomass and its reflectance in the region 350 - 450 nm, and leaf water content and the
region 450 - 850 nm [Tucker, Remote Sensing of Environment, 6, 11-26, 1977]. Other examples are strong
correlations between: chlorphyll and other accessory pigments and the region 350 - 800 nm [Goetz, CSES
Hyperspectral Imaging & Data Analysis Course, July 1997]; cellular scattering in the region 700 - 1050 nm
[Goetz, CSES Hyperspectral Imaging & Data Analysis Course, July 1997]; nitrogen status in corn and the region
 ASD Technical Guide 3rd Ed. Section 9-5

550 - 710 nm and nitrogen stress and ratios of 550-600/800 - 900 nm [Blackmere, Agronomy Journal, vol. 88, no.
1, 1-5, 1996]; and foliar protein at 910 nm and 1020nm , foliar oil at 930 nm and 1040 nm, foliar water at 970 nm,
foliar starch at 990 nm [Remote Sensing of Foliar Chemistry]. The Normalized difference Vegetation Index (NDVI)
uses the absorption at 680 nm and 900 nm [Field Analytical Chemistry and Technology, 1(2): 67-76, 1996].
Important absolute energy measurements include: Photosynthetic Flux Density (PPFD) in the region 400 - 700
nm, and photometric illuminance in the region 380 - 770 nm.

While there are also other applications that require data outside this spectral region, this particular FieldSpec®
model FieldSpec® UV/VNIR (350 - 1050 nm) is designed to provide the highest possible spectral resolution for
the fore mentioned in-situ foliar measurements necessary for the most current high spectral resolution imaging
sensors, while also maintaining optimal signal-to-noise.

VNIR Stray Light Rejection

Stray light is one form of constant systematic noise. There are two types of stray light that are considered
significant in terms of spectrometer performance: Diffraction that is not first order and internal light scattering that
is not diffraction related. It also important to recognize anomalies that look like stray light but really are not and
we'll discuss those here as well.

In general, recognizing stray light problems is not an easy task when comparing instruments. However, in the
case of field applications, there are several clues to look for in spectra that can reveal stray light problems as a
possibility.

Stray light results in computed reflectance values that are different from the actual values. The appearance of
reflectance signal in spectral regions of zero illumination energy and the appearance of biased signals in spectral
regions of low illumination energy are indicators of stray light problems.
In order to recognize stray light problems it is useful to review the meaning of measured signal and computed
reflectance:
Measured signal = true signal + dark current + stray light + random noise
As we'll discuss later, dark current can be easily recorded and subtracted so that it is a negligible contributor.
Therefore, assuming that the dark current is pre-subtracted we can re-write the above formula as follows:
Dark corrected measured signal = true signal + stray light + random noise
Using dark corrected measured signal, computed reflectance is written as follows:
Computed reflectance
= (dark corrected measured signal from target) / (dark corrected measured signal from reference)
= (true target signal + stray light + random noise) / (true reference signal + stray light + random noise)

If stray light is negligible, then for regions of near zero illumination energy, i.e., signal less than or equal to the
instrument random noise at the time the signals are recorded, computed reflectance is written as:
Computed reflectance
= (random noise at time of target measurement) / (random noise at time of reference measurement)

Even though the random noise signals are extremely small, for example 5 DN out of 65,536 DN the ratio of one
random noise signal to another can range between near zero and greater than one. For example, suppose at the
time the reference measurement was take the random noise signal was 3 DN at 1000 nm wavelength channel,
and suppose at the time the target measurement was taken the random noise signal was 6 DN at the same 1000
nm. Then the computed reflectance at that 1000 nm channel would be 200 percent. Graphically, this would be a
vertical line that shoots upward from the last wavelength channel with a non-zero measured signal. Likewise let's
say that at the 1001 nm wavelength channel the random noise levels are such that the ration is near zero.
Graphically, this would be another vertical line next to the last one that shoots straight down to zero. And so forth.

Now, going back to the formula for computed reflectance:

computed reflectance
= (true target signal + stray light + random noise) / (true reference signal + stray light + random noise)
 ASD Technical Guide 3rd Ed. Section 9-6

Again, we consider the water bands where illumination energy is zero, but this time we consider the case of
significant stray light. That is, stray light significantly greater than the lowest level random noise. Computed
reflectance is then written as follows:
computed reflectance
= (stray light + random noise) / (stray light + random noise)

In this case, the computed reflectance is no longer a simple ratio of random noise signals but the ratio of two
significant signals. This anomaly results in what appears to be real reflectance signals where there should only be
random noise.

Stray light clues can also be observed at the shorter wavelengths (see figure below). Because the illumination
energy is low in this region, stray light produces a positive bias error. Because solar irradiance is dropping as
wavelength decreases in the ultraviolet and blue wavelengths, the stray light bias error is large and increases, i.e.,
smiles upward at shorter wavelengths.

0.4

0.3
Reflectance

0.2

0.1

0
350 400 450 500 550
Wavelength (nm)

The figure above is the short wavelength end of the reflectance spectrum for a polyester fabric. The solid line is
the correct spectrum; the dotted line is the spectrum with the addition of a 1% stray light component. Because of
the low intensity of solar radiation in the UV and blue wavelengths, stray light inside the spectrometer produces
the largest positive bias error at shorter wavelengths.

When comparing spectra from different instruments, it is important to note the illumination conditions as they
relate to stray light. For example, solar illumination provides much greater energy in the UV then a typical close-
up DC powered tungsten filament lamp. Therefore, stray light problems in the UV tend to reveal themselves more
so with artificial light sources then with solar illumination. On the other hand, close-up tungsten filament lamps
provide plenty of energy throughout the near-infrared spectrum since there are no long path atmospheric water
absorption columns. Therefore, stray light problems in these areas are revealed under solar illumination.

The stray light component for a given spectrometer can be approximated as a fraction of the raw signal integrated
over the wavelength range of the spectrometer. Typically, this fraction is in the range of 0.02% to 0.1% for a well-
designed single monochromator system such as that used in ASD’s FieldSpec® spectroradiometers.

In summary, poor stray light performance results in spectra that, while 'looking' better in low true signal regions
have significant errors. Typical errors include large biases in both the short and long wavelength ends of spectra
measured in the field.

Stray Light Look Alikes

As mentioned earlier, there are some anomalies that look like stray light effects but really are not. The most
significant of these look-alike effects is what is know as 'dark-current-drift', or simply 'dark-drift'. Dark-current is
systematic noise from the instrument electronics and detectors.

Dark-current can be measured by either viewing a black, near zero reflectance target or by closing a shutter on
the spectrometer input so that zero illumination energy strike the detectors. The dark-current signal can then be
 ASD Technical Guide 3rd Ed. Section 9-7

stored and subtracted from all subsequent measurements. The FieldSpec® spectroradiometer includes
mechanically controlled shutters and software for recording and automatically subtracting dark-current.

Within short time periods, dark-current is relatively constant. However, during the initial start-up the spectrometer
goes through a period where the internal components and external ambient temperature attempt to reach thermal
equilibrium. During this period, the dark-current changes slowly as the change in temperature effects the
efficiency of the internal components. These effects are most significant at the outer boundaries of the detectors'
quantum efficiencies, i.e. at the low and high spectral regions of a given detector. Even after thermal equilibrium is
reached, less significant dark-drift can occur with fluctuations in external ambient temperature.

Dark-drift can be minimized by updatng the dark-current frequently during the first 15 minutes of operation, and
then less frequently after the system has reached thermal equilibrium. The frequency for dark-current updating
after thermal equilibrium should be based on the stability of the ambient conditions and the priority of the regions
most effected by dark-drift.

As previously mentioned, dark-drift is most noticeable in the regions of least quantum efficiency. For example, in
the lower UV regions of a silicon detector. If dark current is not updated, such drift looks very similar to the upward
smiling effect that can also be caused by stray light. The difference is that a dark-drift smile forms gradually over
time when dark current is not updated, while stray light effects are constant. Another example of dark-drift can be
recognized by a smile-down at the upper most NIR regions of the silicon detector. If stray light is negligible, these
dark-drift effects can be easily dealt with by simply updating dark-current more frequently.

VNIR Driftlock

In addition to these methods of dealing with dark-drift, the more the FieldSpec® VNIR incorporate a unique
combination of hardware and software known as Driftlock ('Driftlock' is a trademark of ASD). Driftlock
automatically updates dark-current for every measurement from masked pixels in the silicon array.

VNIR Large Field-of-View

Many in-situ field applications require a broad field-of-view, for example, for a spot size that closely matches the
hyperspectral image pixels. Other examples include collecting data of large area backgrounds as well as other
large area targets. The standard built-in fiberoptic input of the FieldSpec® has the largest field-of-view of any
portable spectroradiometer (25 degrees full conical angle). Optional narrower field-of-view attachable foreoptics
are also available at additional cost. For more details see page 7-3.

VNIR Light Weight, Small, and Wear-able

In-situ measurements or measurements made from a light payload elevation device, such as a balloon, require a
portable instrument that is light-weight and carried in a way that allows fast measurements while moving from spot
to spot or elevation to elevation. The FieldSpec® UV/VNIR can actually be worn around the waist for easy walk-
along, spot-to-spot measurements and has the lightest payload weight for a spectrometer of its performance
specifications.

VNIR Upgrade to Dual Spectrometer System

The FieldSpec® UV/VNIR spectroradiometer has the ability to be upgraded to a dual spectrometer system (see
price list for upgrade cost). For many foliar and oceanographic applications, a dual spectrometer system is greatly
prefered to a single spectrometer system. Particularly under conditions of rapidly changing illumination, e.g., on a
day with partly cloudy skies, it is essential to simultaneously measure downwelling irradiance in conjunction with
the upwelling radiance reflected from the target.

VNIR Upgrade to a FieldSpec® FR

There are many applications that require measurements beyond the spectral range of the FieldSpec® UV/VNIR
(350 - 1050 nm). Accordingly, ASD offers an option at additional cost for upgrading the FieldSpec® UV/VNIR to a
FieldSpec® FR (350 - 2500 nm).
 ASD Technical Guide 3rd Ed. Section 9-8

FieldSpec® Dual UV/VNIR (350 -1050 nm)

The FieldSpec® Dual UV/VNIR looks the same as the standard 'single-beam' FieldSpec® UV/VNIR. The primary
differences are in total weight and user interface settings.

In addition to all the unique benefits of the 'single-beam' FieldSpec® UV/VNIR the FieldSpec® Dual UV/VNIR is
the only portable spectroradiometer that can measure two separate light components simultaneously. For many
applications, a 'dual-beam' system is necessary instead of a single 'single-beam' system. Particularly under
conditions of changing illumination (e.g. on a day with partly cloudy skies), it is essential to simultaneously
measure downwelling irradiance in conjunction with the measurement of upwelling reflected radiance.

FieldSpec® UV/VNIR/CCD

The FieldSpec® UV/VNIR/CCD looks the same as the standard FieldSpec® UV/VNIR. The primary differences
are the detectors and the illumination conditions under which each are considered optimal. The FieldSpec®
UV/VNIR/CCD uses Charge-Coupled Device (CCD) array while the standard FieldSpec® UV/VNIR uses a Photo-
diode array (PDA).

A CCD is a two dimensional silicon array of detectors manufactured into a single chip and gated for sequential
transfer and reading of pixels. A Photodiode Array (PDA) is a one demensional silicon array of detectors
manufactued into a single chip. For a CCD, a whole column of pixels at a specific wavelength are co-added or
"binned" in a summing register and read as a single pixel. The sensitivity of a CCD is superior to a photo diode
array (PDA), but its dynamic range is lower. This type of detector is ideally suited for low light level applications.
These include: in situ measurements of apparent optical properties of lake and sea water, fluorescence, and
bioluminescence.

In general, the S/N of a CCD is superior to a PDA only at VERY LOW LIGHT LEVELS. However, for normal to
high light levels the S/N of a CCD is not as good as that of a PDA. The CCD is best used for things like
fluorescence, bioluminescence, and solar energy measurements under DEEP, dark natural waters. The PDA is
best used for surface measurements under solar illumination and solar energy measurements under shallow,
clear natural waters.

While the S/N of a PDA is not as good as a CCD at VERY low light levels, the PDA CAN measure very low light
levels by setting a long integration time. So, if your mission is to purchase an instrument for a variety of
applications, the PDA is probably your best bet. The CCD should be considered only if you wish to specialize in
VERY low light level measurements.

The plot below shows and example NeDL for the standard 64 element high CCD model.
 ASD Technical Guide 3rd Ed. Section 9-9

-7
10
CCD NER @ 1.4 sec.
64 Element High x 1024 channel CCD

CCD NER @ 704 ms

-8 FieldSpec UV/VNIR/CCD (350-1050 nm)

10
NeDL W/cm /nm

Analytical Spectral Devices, Inc. (ASD)

2

-9
10

-10
10
350 450 550 650 750 850 950 1050
Wavelength (nm)
ASD Technical Guide 3rd Ed. Section 9-10
 ASD Technical Guide 3rd Ed. Section 10-1

Section 10
FieldSpec® HandHeld

The FieldSpec® HandHeld is the smallest and lightest of all the FieldSpec® models. It has a different look and
feel from the other FieldSpec® models. The FieldSpec® HandHeld also has operational features that are different
from the other FieldSpec® models.

Several important features of the FieldSpec® HandHeld make it the best and unique choice for a field portable
spectroradiometer:

The FieldSpec® HandHeld is a post dispersive spectrometer.

The FieldSpec® HandHeld spectroradiometer has an unbeatable combination of speed and sensitivity.
The FieldSpec® HandHeld spectroradiometer has the spectral resolution and sampling interval required for many
applications including ground truthing hyperspectral imagery.
The FieldSpec® HandHeld spectroradiometer is very lightweight, small and can be hand-held.
The FieldSpec® HandHeld spectroradiometer has a large field-of-view in its standard configuration.
The FieldSpec® HandHeld spectroradiometer has excellent stray light rejection.
The FieldSpec® HandHeld spectroradiometer features the unique Driftlock™ dark-drift correction system.

HandHeld Post Dispersive Spectrometer

All ASD spectrometers are “post-dispersive”. Because portable spectrometers are typically used outside the
controlled laboratory environment, they are exposed to much higher levels of ambient light. In almost all cases,
some of this ambient light will stray into the sample being measured. The errors produced by this ambient stray
light are much greater for a pre-dispersive spectrometer than they are for a spectrometer that is post-dispersive.

In a pre-dispersive spectrometer, the sample is illuminated with monochromatic light. Light scattered off or
transmitted through the sample is then collected and delivered to the instrument’s detector. Ambient light that
strays into the sample being measured is also collected. Thus, both the monochromatic illumination from the
instrument and all wavelengths of the ambient stray light are delivered to the detector. Because the stray ambient
light signal can represent a large fraction of the total light signal measured by the detector, it is a major source of
error. While this source or error can be minimized by completely shielding the sample form all sources of ambient
illumination, this often precludes the use of most reflectance and transmittance fiber optic probes.

In a post-dispersive spectrometer, the sample is illuminated with white light. Light scattered off or transmitted
through the sample is then dispersed and delivered to the entrance slit of the instrument’s spectrometer. As with
the pre-dispersive spectrometer, ambient light that strays into the sample being measured is also collected. The
difference is that in the post-dispersive instrument, only ambient stray light of the same wavelength as that being
measured by the detector is added to the signal resulting from the instrument’s illumination of the sample. Thus,
the stray ambient light signal represents a small fraction of the total light signal measured by the detector.

HandHeld Speed and Sensitivity

The FieldSpec® HandHeld spectroradiometer can record a complete 325 - 1075 nm spectrum in as little as 17
milliseconds (user selectable integration times = 2n x 17 milliseconds for n = 0, 1, ... , 15). This amazing speed
allows the convenience of collecting more data in less time, as well as minimizing errors associated with clouds
and wind under solar illumination.

But, speed is not the only critical factor. Fast speed in combination with extremely low Noise-equivalent-Radiance
(NeDL) are what make the optimal spectroradiometer. The following typical NeDL shows the superb sensitivity of
the FieldSpec® HandHeld spectroradiometer (Note: every spectrometer has a unique NeDL).
 ASD Technical Guide 3rd Ed. Section 10-2

10-6

Noise Equivalent delta Radiance

Unit FSHH 1107/1, 5/26/99
0.544-second Scans (30)
NEdL, W/cm^2/nm/sr

-7 Light Adjusted to Fill

10
System Dynamic Range

10-8

10-9
400 500 600 700 800 900 1000
Wavelength (nm)

NeDL = 5 x 10-8 W/cm2/nm/str @ 700 nm

HandHeld Sampling Interval and Spectral Resolution

Spectral sampling interval is the spacing between sample points in the spectrum. Sampling is independent of
resolution and in ASD spectroradiometers is between 2 and 5 times per FWHM. The sampling interval for the
FieldSpec® HandHeld is 1.5 nm for the region 325 - 1075 nm.

Spectral resolution is defined as the full-width-half-maximum (FWHM) of the instrument response to a

monochromatic source. This is in fact the definition ASD uses when stating spectral resolution specifications.
Careful! When reading manufacturers specifications, do not make the mistake of interpreting sampling interval
as resolution. The FWHM spectral resolution of the FieldSpec® HandHeld spectroradiometer is 3.5 nm for the
region 325 - 1075 nm. These spectral resolution values have been measured by calculating the FWHM of a near
monochromatic peak in a spectrum acquired when viewing the output of a monochromator with the FieldSpec®
HandHeld spectroradiometer.

Sampling Interval
Spectrometer offset corrected raw DN

Spectral
Resolution
50% of peak height

Wavelength
 ASD Technical Guide 3rd Ed. Section 10-3

IMPORTANT! Please be careful not to make the mistake often made when reading our competitor’s
specifications. If you look at the number of channels they quote and calculate the minimum possible sampling
interval, you come up with a number that looks like what some competitors present as ‘bandwidth’. Many readers
of these so called ‘bandwidth’ numbers mistakenly interpret them as FWHM spectral resolution, thereby greatly
overestimating the instrument’s capability.

The FieldSpec® HandHeld spectrometer uses a fixed concave holographic reflective grating that disperses the
light onto a fixed photodiode array that has 512 individual detection points or ‘elements’, in a line. Associated with
each of these elements is a distinct signal whose magnitude is determined by the total integrated amount of light
energy falling on that element. For now, we’ll ignore the magnitude of the signal and first concentrate how an
element position corresponds to a narrow band or ‘wavelength channel’. We assign each element a position, X1,
X2, X3, . . . , X512 as shown in the diagram below.

The FieldSpec® HandHeld is set to view NIST traceable wavelength references such as emission sources,
reflectance standards, and the output of a triple monochromator. ASD chooses about 14 of these references that
produce signals with center-peaks at “known” wavelengths, Wk1, Wk2, Wk3, . . . , Wk14 sufficiently distributed
throughout the region 325 – 1175 nm. Each known Wk is paired with the known element-position Xk where the
strongest energy signal is read while viewing that Wk reference. The results are data points with known element-
position and wavelength-channel coordinates, (Xk, Wk).
Wavelength Channel (W)

Interpolated

Known

VNIR Array Element Position (X)

An appropriate cubic spline interpolation function is used to pair up the remaining element-positions with
interpolated wavelength-channels, to 1 nm intervals. The results are a complete set of 700 ordered pairs:

(X1,325), (X2, 326), (X3, 327), . . . , (X512, 1075)

The interpolated wavelength-channels are then verified using the output of the triple monochromator.
 ASD Technical Guide 3rd Ed. Section 10-4

So far, we’ve only talked about pairing up the element-positions with wavelength-channels. For clarity, we have
been ignoring the third coordinate in the total data system. That third coordinate is the detector ‘energy’ signal E,
which results in the set of triplets:

(X1,325, E1), (X2, 326, E2), (X3, 327, E3), . . . , (X512, 1075, E700).

NOTE: DriftLock™ requires masking of several channels which results in 'dark' or zero true signal energy for
those channels.

HandHeld Small Size and Weight

At 2.7 pounds (1.2 kg) including battery, and 22 x 15 x 18 cm, the HandHeld is the lightest self contained
spectrometer system available.

HandHeld Large Field-of-View

Many in-situ field applications require a broad field-of-view, for example, for a spot size that closely matches the
hyperspectral image pixels. Other examples include collecting data of large area backgrounds as well as other
large area targets. The standard built-in input of the FieldSpec® HandHeld has the largest field-of-view of any
portable spectroradiometer (25 degrees full conical angle). Optional narrower field-of-view attachable foreoptics
are also available at additional cost. Shown below is the front of the FieldSpec® HandHeld showing the standard
field-of-view and the threaded fixture that receives other field-of-view foroptics.
 ASD Technical Guide 3rd Ed. Section 10-5

Shown below is the back of the FieldSpec® HandHeld with the operating buttons and ports.

A – Quick-save Buttons: Save next spectrum delivered from instrument;

B – Low Battery Indicator; C – RS-232 Serial-to-PC Interface Socket;
D – AC Power Jack; E – Power On/Off

Important notes: System batteries must be charged with separately supplied charger. 12V AC wall adaptor is used for both
powering the unit and charging the batteries (using charger). Unit only runs attached to PC with DOS or Win 95/98.

Collecting and Saving Reflectance Data:

1. Turn on instrument and controlling computer.
2. Be sure PC and HH are connected via Serial Cable.
3. Run HH executable (c:\hh\hh.exe [unit num] [cal num], or double-click on shortcut).
4. Point instrument at fully illuminated White Reference standard.
5. Set integration time for maximum signal (F1 and F2 keys).
6. Set spectrum averaging and spectrum save preferences.
7. Pointing unit at White Reference standard, collect a White Reference (F4 key or WR button).
8. After a uniform line at 1.00 appears (this is the Reflectance Factor of the reference panel), point the instrument at the
target of interest. The target must be illuminated at about the same distance and intensity as the reference panel.
9. When incoming data stabilizes (after one complete average is delivered) use save command (spacebar or quicksave
buttons) to save next spectrum delivered. To save and current spectrum, use the FR button or the F6 key to freeze the
spectrum on the screen. The save command will now save the current spectrum to file. Don’t forget to unfreeze the
screen (F6 again).
10. Postprocess data into ASCII with STABLE.EXE (arrays) or PORTSPEC.EXE (individual files with headers).

HandHeld Stray Light Rejection

Stray light is one form of constant systematic noise. There are two types of stray light that are considered
significant in terms of spectrometer performance: Diffraction that is not first order and internal light scattering that
is not diffraction related. It also important to recognize anomalies that look like stray light but really are not and
we'll discuss those here as well.

In general, recognizing stray light problems is not an easy task when comparing instruments. However, in the
case of field applications, there are several clues to look for in spectra that can reveal stray light problems as a
possibility.
 ASD Technical Guide 3rd Ed. Section 10-6

Stray light results in computed reflectance values that are different from the actual values. The appearance of
reflectance signal in spectral regions of zero illumination energy and the appearance of biased signals in spectral
regions of low illumination energy are indicators of stray light problems.
In order to recognize stray light problems it is useful to review the meaning of measured signal and computed
reflectance:
Measured signal = true signal + dark current + stray light + random noise
As we'll discuss later, dark current can be easily recorded and subtracted so that it is a negligible contributor.
Therefore, assuming that the dark current is pre-subtracted we can re-write the above formula as follows:
Dark corrected measured signal = true signal + stray light + random noise
Using dark corrected measured signal, computed reflectance is written as follows:
Computed reflectance
= (dark corrected measured signal from target) / (dark corrected measured signal from reference)
= (true target signal + stray light + random noise) / (true reference signal + stray light + random noise)

If stray light is negligible, then for regions of near zero illumination energy, i.e., signal less than or equal to the
instrument random noise at the time the signals are recorded, computed reflectance is written as:
Computed reflectance
= (random noise at time of target measurement) / (random noise at time of reference measurement)

Even though the random noise signals are extremely small, for example 5 DN out of 65,536 DN the ratio of one
random noise signal to another can range between near zero and greater than one. For example, suppose at the
time the reference measurement was take the random noise signal was 3 DN at 1000 nm wavelength channel,
and suppose at the time the target measurement was taken the random noise signal was 6 DN at the same 1000
nm. Then the computed reflectance at that 1000 nm channel would be 200 percent. Graphically, this would be a
vertical line that shoots upward from the last wavelength channel with a non-zero measured signal. Likewise let's
say that at the 1001 nm wavelength channel the random noise levels are such that the ration is near zero.
Graphically, this would be another vertical line next to the last one that shoots straight down to zero. And so forth.

Now, going back to the formula for computed reflectance:

computed reflectance
= (true target signal + stray light + random noise) / (true reference signal + stray light + random noise)

Again, we consider the water bands where illumination energy is zero, but this time we consider the case of
significant stray light. That is, stray light significantly greater than the lowest level random noise. Computed
reflectance is then written as follows:
computed reflectance
= (stray light + random noise) / (stray light + random noise)

In this case, the computed reflectance is no longer a simple ratio of random noise signals but the ratio of two
significant signals. This anomaly results in what appears to be real reflectance signals where there should only be
random noise.

Stray light clues can also be observed at the shorter wavelengths (see figure below). Because the illumination
energy is low in this region, stray light produces a positive bias error. Because solar irradiance is dropping as
wavelength decreases in the ultraviolet and blue wavelengths, the stray light bias error is large and increases, i.e.,
smiles upward at shorter wavelengths.
 ASD Technical Guide 3rd Ed. Section 10-7

0.4

0.3

Reflectance
0.2

0.1

0
350 400 450 500 550
Wavelength (nm)

The figure above is the short wavelength end of the reflectance spectrum for a polyester fabric. The solid line is
the correct spectrum; the dotted line is the spectrum with the addition of a 1% stray light component. Because of
the low intensity of solar radiation in the UV and blue wavelengths, stray light inside the spectrometer produces
the largest positive bias error at shorter wavelengths.

When comparing spectra from different instruments, it is important to note the illumination conditions as they
relate to stray light. For example, solar illumination provides much greater energy in the UV then a typical close-
up DC powered tungsten filament lamp. Therefore, stray light problems in the UV tend to reveal themselves more
so with artificial light sources then with solar illumination. On the other hand, close-up tungsten filament lamps
provide plenty of energy throughout the near-infrared spectrum since there are no long path atmospheric water
absorption columns. Therefore, stray light problems in these areas are revealed under solar illumination.

The stray light component for a given spectrometer can be approximated as a fraction of the raw signal integrated
over the wavelength range of the spectrometer. Typically, this fraction is in the range of 0.02% to 0.1% for a well-
designed single monochromator system such as that used in ASD’s FieldSpec® spectroradiometers.

In summary, poor stray light performance results in spectra that, while 'looking' better in low true signal regions
have significant errors. Typical errors include large biases in both the short and long wavelength ends of spectra
measured in the field.

Stray Light Look Alikes

As mentioned earlier, there are some anomalies that look like stray light effects but really are not. The most
significant of these look-alike effects is what is know as 'dark-current-drift', or simply 'dark-drift'. Dark-current is
systematic noise from the instrument electronics and detectors.

Dark-current can be measured by either viewing a black, near zero reflectance target or by closing a shutter on
the spectrometer input so that zero illumination energy strike the detectors. The dark-current signal can then be
stored and subtracted from all subsequent measurements. The FieldSpec® spectroradiometer includes
mechanically controlled shutters and software for recording and automatically subtracting dark-current.

Within short time periods, dark-current is relatively constant. However, during the initial start-up the spectrometer
goes through a period where the internal components and external ambient temperature attempt to reach thermal
equilibrium. During this period, the dark-current changes slowly as the change in temperature effects the
efficiency of the internal components. These effects are most significant at the outer boundaries of the detectors'
quantum efficiencies, i.e. at the low and high spectral regions of a given detector. Even after thermal equilibrium is
reached, less significant dark-drift can occur with fluctuations in external ambient temperature.

Dark-drift can be minimized by updatng the dark-current frequently during the first 15 minutes of operation, and
then less frequently after the system has reached thermal equilibrium. The frequency for dark-current updating
after thermal equilibrium should be based on the stability of the ambient conditions and the priority of the regions
most effected by dark-drift.

As previously mentioned, dark-drift is most noticeable in the regions of least quantum efficiency. For example, in
the lower UV regions of a silicon detector. If dark current is not updated, such drift looks very similar to the upward
 ASD Technical Guide 3rd Ed. Section 10-8

smiling effect that can also be caused by stray light. The difference is that a dark-drift smile forms gradually over
time when dark current is not updated, while stray light effects are constant. Another example of dark-drift can be
recognized by a smile-down at the upper most NIR regions of the silicon detector. If stray light is negligible, these
dark-drift effects can be easily dealt with by simply updating dark-current more frequently.

HandHeld Driftlock

In addition to these methods of dealing with dark-drift, the more the FieldSpec® HandHeld incorporate a unique
combination of hardware and software known as Driftlock ('Driftlock' is a trademark of ASD). Driftlock
automatically updates dark-current for every measurement from masked pixels in the silicon array.
The following HandHeld Spectra Reflectance of Maple Leaf and Reflected Solar Radiance.

1.0

FieldSpec HandHeld
0.8 Reflectance of
Maple Leaf
Solar Illumination
Reflectance

0.6

0.4

0.2

0.0
400 500 600 700 800 900 1000
Wavelength (nm)

0.40

0.35

0.30
Radiance, W/m^2/nm/sr

0.25

0.20

Solar Radiance
0.15 Reflected from
Spectralon
0.10 Standard Surface
FieldSpec HandHeld
0.05
Unit No. 1107/1, 5/26/99

0.00
400 500 600 700 800 900 1000
Wavelength (nm)
 ASD Technical Guide 3rd Ed. Section 10-9

HandHeld Foreoptics
ASD Technical Guide 3rd Ed. Section 10-10
 ASD Technical Guide 3rd Ed. Section 11-1

Section 11
LabSpec® Pro Portable Spectrophotometer

LabSpec® Pro is the first, truly portable, near infrared analyzer that offers unprecedented ruggedness, durability
and reliability in a laboratory instrument. Equipped with a built in handle and internal batteries, LabSpec® Pro
offers the flexibility to get out of the lab, collect samples, make demonstrations, prove concepts or trouble shoot
problems. Its ultra tough case, made of the same material as industrial power tools, and chemically resistant
aluminum front panel makes LabSpec® Pro ideal for use in virtually any environment.

LabSpec® Pro has laboratory accuracy to meet your QA/QC measurement needs. A host of probes and sampling
accessories are available for liquids, powders, and slurries. With a wavelength range of 350 nm – 2500 nm that
can penetrate many plastic bags, bottles or glass vials, LabSpec® Pro offers the convenience of testing samples
in their original containers. With a super fast scanning time of 0.1 seconds and a sampling interval of 2 nm,
LabSpec® Pro is ideal for non-destructive quality control applications.

Industries with NIR Applications:

Chemical, Pharmaceutical, Petrochemical, Food & Beverage, Pulp & Paper, Textile, Paint, & Tobacco.

LabSpec® Pro Spectral Ranges

Model Number
1000-1800 nm A108280
1000-2500 nm A108300
350-2500 nm A108310

LabSpec® Pro Features & Advantages:

Portable
Make measurements at-line for model building, verification, proof of concept, or trouble shooting.
Ideal for raw materials inspection, stability testing, or blending/batching verification.
Built in handle for easy transportation.
No power cord required.
Internal NiMH batteries provide up to 3 hours of portable testing.

Accurate & Precise

High signal-to-noise ratio for superior repeatability of results.
Better discrimination of materials.

Rugged
Designed from the start to be a portable instrument operating in harsh environments.
Not affected by vibration or changes in temperature or humidity.
Use it with confidence in virtually any environment.

Fast 10 spectra per second data collection for the entire 350-2500 nm range.
0.1 second scanning time provides an accurate average every second.
Allows the user to quickly scan several areas when analyzing bulk samples.
Rapid analysis allows screening all or most samples vs. spot-checking.

Small Foot Print

12.5 x 4.5 x 16 (LxWxH) inches. (Base is only 8 inches wide)
Upright design maximizes bench space in crowded laboratories.

The Unscrambler Software by Camo is Standard

The Unscrambler by Camo is a superb multivariate-modeling program used for chemometrics.
Simplifies model building, calibration, prediction and classification.
Incorporates many practical features that make daily use faster and easier.

Optional Carrying Case

Provides additional compartments for a laptop and accessories.
Easy to transport from the lab.
 ASD Technical Guide 3rd Ed. Section 11-2

Shown below is the LabSpec® Pro with optional carrying case (each sold seperately).

Near Infrared Technology

Ideal for testing samples in their original containers.
NIR energy penetrates several millimeters into many organic substances including solids.
NIR measures many properties simultaneously.
Quick, accurate non-destructive testing.

Fiber Optic Input

Flexible cable allows the probe to go to the sample for easier product sampling.
Allows remote monitoring in challenging environments.
Easily interfaced to a wide variety of sampling devices.
 ASD Technical Guide 3rd Ed. Section 12-1

Section 12Field Spectrometry: Techniques and Instrumentation

FIELD SPECTROMETRY: TECHNIQUES AND INSTRUMENTATION
(Presented at the International Symposium on Spectral Sensing Research, July, 1994)

Brian Curtiss
Analytical Spectral Devices, Inc.
4760 Walnut Street Suite 105
Boulder, CO 80301-2561, USA
303-444-6522 FAX: 303-444-6825

Alexander F. H. Goetz
University of Colorado
Center for the Study of Earth from Space/
Cooperative Institute for Research in the Environmental Sciences
Boulder, Colorado 80309-0449, USA

ABSTRACT

Applications such as aircraft and satellite sensor calibration, development of remote sensing data exploitation methods,
remote sensing feasibility studies, and geologic mapping greatly benefit from the use of field spectrometry. The collection of
accurate spectra in the field requires an awareness of the influences of the various sources of illumination, atmospheric
characteristics and stability, winds, instrument field of view, target viewing and illumination geometry, instrument scanning
time, and the spatial and temporal variability of the target characteristics. The first step in the development of a field
experiment is the definition of the overall experimental design. Unfortunately, the formulation of an appropriate experimental
design is not always obvious. Issues such as the timing of the data collection, spatial scale of the field measurement, target
viewing and illumination geometry, and the collection of ancillary data sets must be considered in light of the objectives of
the study. The lack of the appropriate ancillary data sets that often makes previously collected data sets unusable for a new
application. Frequently, the experimental design must be modified to account for the characteristics of the available
instrumentation. Instrument characteristics, such as signal to noise ratio, radiometric calibration, spectral resolution, spectrum
acquisition time, and angular field of view, all place limitations on the types of spectral measurements that can be made in the
field. For example, vegetation canopy spectra collected using a slow scanning instrument will sometimes have small wind-
induced "absorption" features in those portions of the spectra when the instrument was viewing more shadow.

1. INTRODUCTION

Field spectrometry is the quantitative measurement of radiance, irradiance, reflectance or transmission in the field. Portable,
battery powered spectroradiometers are typically used to make these measurements. In this paper, the discussion will be
limited to: 1) visible to near infrared wavelengths, 300 to 2500 nm; 2) examination of geological, man-made, and vegetative
materials; and, 3) instrumentation acquiring a continuous spectrum.

There are many reasons why it is desirable to perform spectral measurement in the field, not all related to remote sensing.
Field spectra of ground targets that are homogeneous at the scale of the imaging sensor and collected using ambient solar
illumination can be used to convert radiance images to reflectance (Conel et al., 1987a & 1987b). Often, field spectra of
target materials are collected to allow for more precise image analysis and interpretation (Goetz and Srivastava, 1985). Field
spectroscopy is also used as a tool to perform feasibility studies to understand if and how a process or material of interest can
be detected using remote sensing. Field spectra of both the material(s) of interest and spectra of other materials present in the
environment can be used to address such questions as: 1) What spectral resolution is required for detection?; 2) What spatial
resolution is required for detection?; 3) What is the best time of year/day for detection?; and 4) What signal-to-noise ratio is
required for detection?. Aside from remote sensing applications, field spectrometers are used to make direct material
identifications in the field rather than collecting samples for later laboratory analysis.
 ASD Technical Guide 3rd Ed. Section 12-2

2. ILLUMINATION

2.1 ASSUMPTIONS

In order to determine the reflectance or transmittance of a material, two measurements are required: the spectral response of a
reference sample and that of the target material. The reflectance or transmittance spectrum is then computed by dividing the
spectral response of the target material by that of a reference sample. Using this method, all parameters which are
multiplicative in nature and present in both the spectral response of a reference sample and the target material, are ratioed out.
These parameters include the spectral irradiance of the illumination source and the optical throughput of the field
spectrometer. Thus, when determining the reflectance or transmittance of a material in the field, an inherent assumption is the
characteristics of the illumination are the same for the reference and target materials. Variability of the illumination
characteristics between the time the reference and target materials are measured will result in errors in the resultant spectra.

2.2 CHARACTERISTICS OF NATURAL ILLUMINATION

Spectral measurements are typically made in the field using ambient solar illumination. In the field, the target material is
illuminated by three or more sources (see Fig. 1), each with its own spectral characteristics (Curtiss and Ustin, 1988).
Unless the target is in a shadow, the direct solar illumination is the dominant source of illumination. Parameters such as solar
elevation angle and atmospheric conditions will effect the overall intensity and spectral characteristics of direct solar
illumination. Diffuse skylight illumination can contribute as much as 5-10% of the total illumination reaching a surface. At
shorter wavelengths, diffuse skylight can contribute as much as 20-25% of the total. The spectral characteristics of the
illumination scattered off of surrounding objects is determined by their reflectance characteristics. In the case of a forest
clearing, as much as 20% of the illumination in the 750 - 1200 nm wavelength range can be attributed to sunlight scattered
off the surrounding forest canopy (Curtiss and Ustin, 1988). One important source of surroundings scattered light is the
person and the instrumentation making the measurement. Objects in the surroundings also effect the overall illumination of
the target surface by obscuring a portion of the diffuse skylight and, possibly, shading the target from direct solar
illumination. The magnitude of both the diffuse skylight and scattered from surrounding illumination components is
determined by the solid angle subtended by these sources when viewed from the reference frame of the target surface. The
surface texture of the material being measured also effect the relative proportion of the various sources of illumination. When
compared to a smooth surface, a surface with a rough texture will tend to have a higher proportion of illumination from the
diffuse and scattered-from -surroundings sources relative to the direct solar illumination (Curtiss and Ustin, 1988).

diffuse
direct illumination
illumination

diffuse
illumination scatter from
surroundings

Figure 1. The major sources of illumination. Note that it is possible to have several sources of light scattered off of
surrounding objects, each with its own unique spectral distribution.
 ASD Technical Guide 3rd Ed. Section 12-3

2.3 CHARACTERISTICS OF ARTIFICIAL ILLUMINATION

While in most cases it is desirable to use ambient solar illumination to maintain equivalence between field spectral
measurements and remotely sensed images, there are some cases where the use of artificial illumination is desirable. The use
of artificial illumination allows: 1) more control over illumination and viewing geometry; 2) more control over sample
geometry; 3) measurements during non-optimal conditions (e.g. cloud cover or at night); and 4) measurement of reflectance
and transmittance in the deep atmospheric absorption bands. Several problems with using artificial illumination include: 1)
difficulty in maintaining a constant distance between the sample or reference and the light source when measuring samples
with irregular geometry; and 2) lights can 'cook' vegetation samples (water loss, chlorophyll degradation). A typical lamp
configuration for indoor use is shown in Figure 2. Alternatively, the light source can be either incorporated into the field
spectrometer (often precluding the use of solar illumination) or can be provided in the form of an optional accessory that
mounts to the light collecting optics of the instrument.

Figure 2. A typical lamp configuration for artificial illumination; the sample is viewed with the collecting optics of
the spectrometer nadir to the sample. Use 1 or 2 200 to 500 Watt quartz-halogen cycle tungsten filament
lamps (~3400°K color temperature) in housings with aluminum reflectors about 1 meter above the surface
being measured.

3. ATMOSPHERIC CHARACTERISTICS

3.1 ATMOSPHERIC TRANSMISSION

Absorbing molecules in the atmosphere strongly modify the incoming solar irradiance (Goetz, 1992). All of absorption
features described in this section will increase in intensity as the atmospheric path length of the incoming solar radiation
increases (e.g. with changing solar elevation angle). By far, water vapor is the strongest modifier of the incoming solar
spectrum (Gao and Goetz, 1990). Water vapor has absorption features spanning the solar reflected region of the spectrum
(see Fig. 3), and varies both spatially and temporally. Carbon dioxide has strong features in the 2000-2200 nm range (see Fig.
4), a region of major interest for the identification of layered silicate minerals (Goetz, 1992). Carbon dioxide is a well mixed
gas, thus the intensity of the absorption features associated with carbon dioxide are not as variable as those of water vapor,
but they do decrease with increasing altitude. Other major atmospheric components that influence the atmospheric
transmission spectrum are shown in Figure 5.
 ASD Technical Guide 3rd Ed. Section 12-4

1.2
Water Vapor (H O)
2
1.0
% Transmitance

0.8

0.6

0.4

0.2

0.0
0.5 1 1.5 2 2.5
W avelength, µm
Figure 3. Transmission spectrum of water vapor for typical atmospheric conditions.

1.2
Carbon Dioxide (CO )
2
% Transmittance

1.0

0.8

0.6

0.4

0.2

0.0
0.5 1 1.5 2 2.5
Wavelength, µm
Figure 4. Transmission spectrum of carbon dioxide for typical atmospheric conditions.
 ASD Technical Guide 3rd Ed. Section 12-5

Ozone (O )
3

Transmittance (5% Intervals)

Carbon Monoxide (CO)
Nitrous Oxide (N O)
2
Methane (CH )
4

Oxygen (O )
2

0.6
0.4 0.7 1 1.3 1.6 1.9 2.2 2.5
Wavelength, µm

Figure 5. Transmission spectrum of various gases for a typical atmospheric conditions.

3.2 CLOUDS

The presence of partial cloud cover is indicative of highly spatially and temporally variable atmospheric water vapor (Gao
and Goetz, 1992). Because of the large influence of water vapor on the atmospheric transmission (see Fig. 3), variability of
atmospheric water vapor between the time when the reference and target measurements are made will result in errors in the
resultant spectrum. This error can be reduced by minimizing the length of time between the measurement of the reference
sample and the target.

While they are difficult to see and often appear inconsequential, the presence of cirrus clouds tends to produce significant
variability in atmospheric water vapor (Gao and Goetz, 1992). The field spectrometer itself can be used to measure the
magnitude of the effect. Simply standardize the instrument on the reference panel, then continue to view the reference panel
with the instrument. If the atmospheric conditions are stable, the computed reflectance of the panel will be a flat spectrum
with near 100% reflectance. If atmospheric conditions are unstable, the computed reflectance of the panel will vary over time
and will show absorption minima or maxima (depending on whether atmospheric water vapor is increasing or decreasing) at
the wavelengths corresponding to the water vapor absorption features. In this way, it can be determined whether spectral data
with sufficient accuracy can be acquired.

In addition to the errors produced by time varying atmospheric water vapor, partial cloud cover also greatly increases the
intensity of diffuse skylight illumination (Curtiss and Ustin, 1988). This tends to "fill in" shadows and reduce the contrast
between surfaces with dissimilar surface textures. If the goal is to collect field spectra for image calibration or interpretation,
spectra should be collected under illumination conditions similar to those at the time the image was collected.

3.3 WIND

Wind can be a source of error if the material being measured moves during the time the spectrum is acquired. If a spectrum is
slowly scanned, changes in the amount of shadow in the instrument field-of-view will result in erroneous "features" in the
 ASD Technical Guide 3rd Ed. Section 12-6

spectrum. Vegetation canopies, with their large proportion of shadow, are especially susceptible to wind induced errors.
Instruments using an array detector or that scan the spectrum rapidly are not significantly affected by wind.

4. CHARACTERISTICS OF TARGET MATERIALS

4.1 VEGETATION

Because of the complex three dimensional geometry of a plant canopy, light returned from the canopy is a complex mixture
of multiply reflected and/or transmitted components (Curtiss, 1990; Curtiss and Maecher, 1991; Curtiss and Ustin, 1989).
The canopy level optical signal is dependence upon illumination and viewing geometry, canopy structure, leaf optical
properties, and the optical properties of other vegetative and non-vegetative components within and below the canopy. The
strong dependent on illumination and viewing geometry can be seen in Figure 6. Both the overall brightness of the canopy
and the shape of the spectral signature (e.g. the red to infrared ratio for the canopies in Figure 6) are dependent on the
illumination and viewing geometry. Thus, it is only by controlling the viewing and illumination geometry, that changes in
canopy reflectance attributable to the canopy itself can be detected.

0.16 1.00
Reflectance

0.12 Reflectance 0.80

0.60
0.08
0.40
0.04 0.20
0.00 0.00

Figure 6. Bi-directional reflectance functions (BRDF) of a conifer, Mugo Pine, (circles) and a broadleaf shrub,
Viburnum, (squares). For reference, the BRDF of a lambertian reflector is plotted as a dashed line. The left
plot is for a red band (650 nm) and the right plot is for a near infrared band (850 nm). Both canopies were
illuminated from the left at an elevation of about 30°.

True differences in canopy reflectance between several canopies may be due to either differences in canopy structure or in
leaf/needle optical properties (Curtiss and Maecher, 1991). Using the data presented in Figure 6 as an example, for almost all
viewing geometries, a conifer canopy will appear darker in the infrared than the broadleaf canopy even though the reflectance
of the individual needles and leaves may be almost identical. Observed differences in canopy level reflectance may be
attributable to differences in either leaf level optical properties or other, larger scale, structural properties of the canopy.
Important structural properties include leaf/needle size, leaf/needle density at the branch level, number of years of needles
retained (conifers), and branching angles. When differences are observed between canopies of the same species, it is almost
always due to differences in canopy or branch level structure.

The absorption features seen in vegetation spectra are all related to organic compounds common to the majority of plant
species (Peterson et al., 1988; Gao and Goetz, 1992). Thus, the information about a plant canopy is contained in the relative
intensity of the various absorption features rather than in the presence or absence of a specific absorption feature. The major
spectral absorption features can be attributed to plant pigments (chlorophylls, zanthophyll, and carotenoids) and water. Other,
minor, absorption features are attributable to other chemical components; these include cellulose, lignin, proteins, starches,
and sugars. Non-photosynthetic components of the canopy have spectra which are dominated by absorption features
attributed to lignin and cellulose.
 ASD Technical Guide 3rd Ed. Section 12-7

4.2 ROCKS, SOILS, AND MAN-MADE MATERIALS

Unlike vegetation, the shape of the spectral signature of rocks and soil tend to be invariant with varying viewing geometry.
Due to changes in the amount of shadow in the field-of-view of the spectrometer, the overall brightness of the observed
spectrum does change with illumination and viewing geometry. Absorption features in the spectra of rocks and minerals (See
Figure 7) are due to the presence of specific molecular groups and are often diagnostic of the minerals present in the sample
(Abrams et al., 1877; Hunt, 1980).
Reflectance (10% intervals)

500 1000 1500 2000

Wavelength, nm
Figure 7. Reflectance spectra of alunite (triangles), kaolinite (squares), illite (circles), and buddingtonite (diamonds)
measured with a FieldSpec®™ FR field portable spectroradiometer.

The dependence of the optical properties of man-made objects on the viewing and illumination geometry will fall somewhere
between that for vegetation and rocks/soils. This will depend on the amount of light transmission through elements of the
target and on the surface texture of the object. The reflectance spectra of objects with translucent elements and highly
textured surfaces have an angular dependence more like vegetation, while those with opaque elements have an angular
dependence more like spectra of rocks and soils.

5. INSTRUMENTATION

5.1 SPECTRAL RESOLUTION VS. SPECTRAL SAMPLING INTERVAL

While the terms "spectral resolution" and "spectral sampling interval" are often used interchangeably, they refer to very
different characteristics of a spectrometer. Spectral resolution is a measure of the narrowest spectral feature that can be
resolved by a spectrometer. It is also defined as the full width at half maximum (FWHM) response to a spectral line source.
The spectral sampling interval of a spectrometer is the interval, in wavelength units, between data points in the measured
spectrum.

For hyperspectral remote sensing applications, a spectral resolution of about 10 nm and a spectral sampling interval of about
2 to 3 nm is required (Goetz and Calvin, 1987, Goetz, 1992b). The requirement for 10 nm spectral resolution is driven by the
spectral resolution of the hyperspectral sensor (about 10 nm). A spectral sampling interval of about 2 to 3 nm provides 3-4
data points in the field spectral data. This oversampling of the spectrum results in less degradation of spectral resolution when
resampling the field spectral data to match the wavelengths of the hyperspectral sensor channels. Also, analysis methods
utilizing derivative spectra greatly benefit from the oversampling of the spectrum.
 ASD Technical Guide 3rd Ed. Section 12-8

5.2 SPECTROMETER DESIGN

Spectrometers used in currently available field spectrometers are either based upon a fixed grating and an array detector, or a
single element detector and a scanning grating. One of the drawbacks of an array based spectrometer is the signal-to-noise
ratio (SNR) is tied to the sampling interval as well as the spectral resolution (Smith, 1992), while in a scanning spectrometer
the SNR is independent of spectral sampling interval. Array detectors in the 350 to 1000 nm region of the spectrum (VNIR)
have sufficient sensitivity to allow spectra to be collected with spectral resolutions well above the 10 nm resolution required
for hyperspectral remote sensing studies. Thus, in the VNIR, fixed grating, array detector, based spectrometers provide the
best performance for most remote sensing applications. Typical VNIR spectrometer designs provide 3 to 5 nm spectral
resolution with a spectral sampling interval of better than 2 nm.

Due to the rapid fall-off in energy of the solar spectrum in the 1000 to 2500 nm region of the spectrum (SWIR), spectrometer
design is strongly driven by the need to maintain an acceptable SNR. In general, a loss of SNR is not acceptable in exchange
for decreased spectral sampling interval. Thus, the use of an array detector in the SWIR spectral region generally precludes
the oversampling of the spectrum. While scanning spectrometers do not have this limitation, they have the drawback of not
measuring all wavelengths simultaneously. This can result in errors due to changes in the target during the measurement of
the spectrum (see Section 3.3). This limitation is overcome by rapidly scanning the spectrum.

Traditionally, scanning SWIR spectrometers have used lead sulfide (PbS) detectors. Due to the lack of sufficient frequency
response, the use of PbS detectors precludes rapid scanning. Indium gallium arsinide (InGaAs) detectors have the high
frequency response required for rapid scanning (less than 100 milliseconds to scan from 1000 to 2500 nm). InGaAs detectors
have the added advantage of having a higher sensitivity than PbS and, also, do not require an optical chopper as does PbS.
These increases in performance over PbS allow the design of a rapid scanning spectrometer with equivalent or better SNR
performance than a PbS array detector based spectrometer. A scanning design has the added advantage of being having a
spectral sampling interval that is less than one fifth the spectral resolution without a reduction in SNR.

While it is possible to include the illumination source within the spectrometer, this limits the applications that can be
addressed and the types of targets that can be measured. Because of the need to hold illumination and viewing geometry
constant between the field and image spectral data, field spectra collected for either hyperspectral sensor calibration or for
direct comparison with hyperspectral image data are best collected using ambient solar illumination. The use of a built-in
light source also precludes the measurement of radiance or irradiance. Additionally, targets such as vegetation with a
complex three dimensional structure can only be measured in the field using solar illumination.

5.3 SPECTROMETER FIELD OF VIEW

Existing field spectrometers use one of two basic approaches to collect light energy and deliver it to the spectrometer. In
some spectrometer designs, foreoptics are used to form an image of the target upon the entrance slit of the spectrometer. This
approach results in a ground field of view (GFOV) that has the same shape as the spectrometer entrance slit (often a rectangle
with a height to width ratio of more than 10:1). If more than one spectrometer is used in an instrument (e.g. a VNIR and
SWIR spectrometer), it is often difficult to ensure both spectrometers are viewing the same GFOV. The use of optical fibers
to deliver the light to the spectrometer results in a circular GFOV. The use of more than one spectrometer is accommodated
by splitting the optical fiber bundle within the instrument to deliver light to the various spectrometers. The size of the GFOV
is determined by the angular field of view of the instrument and the distance to the target. While optics can be added to
modify the angular field of view of a field instrument, this is often practical only for those instruments utilizing optical fibers
for light collection.

6. EXPERIMENTAL DESIGN

6.1 TIME OF DAY / YEAR

The timing of the field data collection is important if field spectra are to be used to calibrate or interpret a hyperspectral
image. Reflectance spectra used to interpret a hyperspectral image should be collected under illumination conditions similar
to those when the image was acquired. Time of day and date are the major controlling factors of direct illumination geometry,
while atmospheric conditions relating to scattering and clouds are the major factors controlling the geometry and relative
intensity of indirect illumination (Curtiss and Ustin, 1988). If field spectra are to be used to convert a hyperspectral sensor
image to reflectance, spectra should be acquired simultaneously with image acquisition. If this is not possible, a water vapor
correction should be made to the image prior to its conversion to reflectance (Gao and Goetz, 1990).
 ASD Technical Guide 3rd Ed. Section 12-9

6.2 SAMPLING STRATEGY

The selection of a particular sampling strategy should be based upon the objectives of the study. If the objective is to
determine the detectability of either a target material or of a process effecting the spectral signature of a target material, the
sampling strategy must encompass examples of the target material under all expected conditions as well as all other
background materials. Consideration must be given to all the processes modifying target and background spectral signatures;
these include illumination, slope and aspect, and target surface architecture. Adequate collection of vegetation is even more
problematic due to the wide range of processes that effect their spectral signature. These include soil chemical properties, soil
reflectance, water availability, time of day, time of year, cloud cover, temperature, and relative humidity.

If the purpose of a study is to understand how a process effects the spectral signature of a target without consideration of how
background materials and other sources of variability effect the spectral signature, variance from sources other than the one
of interest should be minimized in the selected targets. This is achieved by carefully controlling viewing and illumination
geometry, often in a laboratory. This type of study is often performed as a feasibility study prior to a full study evolving
collection of spectra of background material.

6.3 VIEWING GEOMETRY

The selection of an appropriate viewing geometry depends upon whether the spectra are used for image analysis or for a
feasibility study. For image analysis, a viewing geometry similar to the airborne sensor is required. For a feasibility study it is
possible to eliminate much of the confounding variance typically present in an image data set by fixing the viewing and
illumination geometry.

For vegetation, spectra can be acquired at the spatial scale of the leaf/needle, branch, or canopy. Canopy level spectra are
most suitable for image analysis and interpretation. Rarely can leaf/needle or branch level spectra be used to directly to
interpret an image. Leaf/needle and branch level spectra are useful to understand how the canopy level spectra are influenced
by changes in leaf/needle and branch spectral changes as opposed to changes in the canopy structure.

6.4 ANCILLARY DATA SETS

Collection of appropriate ancillary data sets is as important as the collection of field spectra. The measurement of processes
and material properties directly influencing, or correlated with, the collected spectra are essential to the understanding of the
variance observed in a hyperspectral image data set. It is often necessary to collect a set of spectra with ancillary data in order
to validate predictions made using the main field spectra plus image data set.

REFERENCES

Abrams, M.J., R. Ashley, L.C. Rowan, A.F.H. Goetz, and A.B. Kahle (1977) Mapping of hydrothermal alteration in the
Cuprite Mining District, Nevada, using aircraft scanner imagery for the 0.46 - 2.36 µm spectral region, Geol., 5, 713-
718.

Conel J.E., C.J. Bruegge and B. Curtiss (1987a) Correcting Airborne Imaging Spectrometer measurements for the
atmosphere: a comparison of methods. Proc. 31st. S.P.I.E. International Technical Symposium on Optical and
Optoelectronic Applied Science and Engineering.

Conel J.E., R.O. Green, G. Vane, C.J. Bruegge, R.E. Alley and B. Curtiss (1987b) AIS-2 Radiometry and a comparison of
methods for the recovery of ground reflectance. Proc. of the Third Airborne Imaging Spectrometer Data Analysis
Workshop..

Curtiss B. (1990) From leaf to landscape: The parameterization of vegetation canopy radiative transfer models from remote
sensing data. Proc. Ann. Meeting of the Ecol. Society of Amer..

Curtiss B. and A.G. Maecher (1991) Changes in forest canopy reflectance associated with chronic exposure to high
concentrations of soil trace metals. Proc. of the 8th Thematic Conference on Geologic Remote Sensing, Denver,
Colorado.

Curtiss B. and S.L. Ustin (1988) Spectral changes in Ponderosa Pine associated with natural ozone exposure. Proc. U.S.
Forest Service Forest Response Program Annual Meeting, Corpus Christi, Texas, February 23 - 26.
 ASD Technical Guide 3rd Ed. Section 12-10

Curtiss B. and S.L. Ustin (1989a) The Remote Detection of Early Stages of Air Pollution Injury in Coniferous Forests using
imaging Spectrometry. Proc. European Joint Research Center Remote Sensing of Forests Workshop. Ispara, Italy.

Gao, B.-C. and A.F.H. Goetz (1990) Column atmospheric water vapor retrievals from airborne imaging spectrometer data,
J. Geophys. Res.-Atmospheres, 95, 3549-3564.

Gao, B.-C., and A.F.H. Goetz (1992) A Linear Spectral Matching Technique for Retrieving Equivalent Water Thickness and
Biochemical Constituents of Green Vegetation, Proc. of the Third Airborne Annual JPL Geoscience Workshop
(AVIRIS, TIMS, and AIRSAR), Jet Propulsion Laboratory, Pasadena, CA.

Goetz, A.F.H. (1992a) Imaging Spectrometry for Earth Remote Sensing, in Imaging Spectroscopy: Fundamentals and
Prospective Applications, F. Toselli and J. Bodechtel, Eds., pp. 1-19, Brussels and Luxembourg.

Goetz, A.F.H. (1992b) Principles of Narrow Band Spectrometry, in The Visible and IR: Instruments and Data Analysis in
Imaging Spectroscopy: Fundamentals and Prospective Applications, F. Toselli and J. Bodechtel, Eds., pp. 21-32,
Brussels and Luxembourg.

Goetz, A.F.H. and W.M. Calvin (1987) Imaging spectrometry: Spectral resolution and analytical identification of spectral
features, Imaging Spectros. II, G.Vane, Ed., Proc. SPIE, 834, 158-165.

Goetz, A.F.H. and V. Srivastava (1985) Mineralogical mapping in the Cuprite mining district, Nevada, Proceedings of the
Airborne Imaging Spectrometer Data Analysis Workshop, April 1985, JPL Publication , no. 85-41.

Hunt, G.R. (1980) Electromagnetic radiation: The communications link in remote sensing. In Remote Sensing in Geology,
B.S. Siegal and A.R. Gillespie, eds, Wiley, NY, pp 5-45.

Peterson, D.L., J.D. Aber, P.A. Matson, D.H. Card, N. Swanberg, C.A. Wessman and M. Spanner (1988) Remote sensing of
forest canopy and leaf biochemical contnets, Rem. Sens. Environ., 24, 85-108.

Smith, M.W. (1992) Design and initial performance evaluation of a portable short wave infrared spectrometer, Proc. SPIE,
Vol. 1762-14., San Diego, CA.
 ASD Technical Guide 3rd Ed. Section 13-1

Section 13
Fundamentals of Spectroradiometry
Fundamentals of Spectroradiometry
David Hatchell, Analytical Spectral Devces, Inc.

Radiant flux F . The amount of light energy that crosses a defined surface in a unit time. The units typically are:
Joules
Second = Watts. Also known as "radiant power". Flux is pictorially indicated by arrows indicating the direction
of travel of the energy.

Defined Surface

Photons

Intensity I : For an infinitesimal solid angle element dΩ (discussed in detail later), and infinitesimal flux
dF
element dF the intensity I is defined as .
dΩ

Solid angle: "When a portion of space is separated from the rest by three or more planes which meet in but one
point, the planes are said to form, or to include, a polyhedral angle. A polyhedral angle is also called a solid
angle." 1

Examples of Solid Angles

The 'curve'
(heavy line)

"A solid angle is similarly formed at a point, also called the vertex, by a conical surface or cone. The cone, in turn,
is the surface that contains all possible straight lines (i.e., is the locus of those lines) that extend from the vertex
point to a point on some closed, simply-connected curve in space that does not pass through the vertex. Such a
'curve', using the term in its broadest sense, may include straight-line segments and discontinuous changes of
direction or angles. It is only required that, starting from any point on the 'curve' and traveling along it far enough
in either direction, you return to the starting point after passing once, and only once, through every other point on
the 'curve'. In particular, when the 'curve' is made up entirely of straight-line segments that form a polygon, the
'cone' is a pyramid. On the other hand, when a pencil of rays converges on the axis of a cylindrical optical system,
with circular optical components, the solid angle formed at the focus is bounded by a right circular cone and the
solid angle is often so depicted. In fact, speaking loosely, we often say that the solid angle is a right circular cone.
However, there is really no such limitation on the concept of a solid angle which, as we have just seen, can be
formed at the vertex of a pyramid, or of the conical surface formed by the straight lines joining the vertex to a
closed curve of almost any shape."2

1Wooster Woodruff Beman and David Eugen Smith, New Plane and Solid Geometry, (New York: Ginn and
Company, 1900), p274.
2Fred E. Nicodemus and Henry J. Kostkowski, Self-Study Manual on Optical Radiation Measurements, Part 1-
Concepts Chapters 1 to 3 (Washington, D.C.: U.S. Department of Commerce / National Bureau of Standards,
1976), p.65.
 ASD Technical Guide 3rd Ed. Section 13-2

Projected Area: The projected area (shaded portion) along the radius arm D:

Projected Area ∆ A proj

a d
e D
θ
c
∆A

where line segments a,b,c,d, and D are normal to the plane ∆Aproj and e is normal to the plane ∆A . Trigonometry
shows that ∆Aproj = ∆A cosθ.

Radiance L : The radiant flux per unit projected area per unit solid angle. The units of radiance are
Watts/meter2/steradian. In this quantity the projected area may be that of a detector surface, a source surface, or an
imaginary surface in space. The flux may be that falling on a detector or being emitted by a source. In other
words, the vector along the radius arm D may point into or away from the surface.

Referring to the figure below, A general model for radiance may be explained as follows: Letting ∆A be a part of
a surface located at the center of a sphere of radius D.

Condition I: Let ∆A become so small that it is essentially a point. With this condition, a solid angle ∆Ω can be
defined extending from the point to the sphere where an area ∆Aproj is defined. Using formal terminology we
say, "the solid angle ∆Ω subtended at ∆A by ∆Aproj".

Condition II: Now, ∆Aproj isn't truly the projected area of ∆A unless the rays defining the solid angle are normal
to ∆Aproj (see definition of projected area above). The equivalent of this conditions is to let ∆Ω be so narrow that
is essentially a ‘line’, or 'beam' of rays.

Allowing conditions I and II and following the definition of radiance L :

∆F ∆F
L = = , under conditions I and II
∆ Aproj ∆Ω ∆ A cosθ ∆Ω
 ASD Technical Guide 3rd Ed. Section 13-3

4π
Steradians ∆A proj
Sphere

D ∆Ω

zθ
Surface
∆A
y
x φ

z
D
θ ∆Ω

y
∆A
φ
x
Using the mathematical equivalent of conditions I and II gives:
 ASD Technical Guide 3rd Ed. Section 13-4

lim
∆F → 0 ∆F
L =
∆ A → 0 ∆A cosθ ∆Ω
∆Ω → 0
This expression is equivalent to the second partial derivative:
1 dF 2
L =
cosθ dAdΩ
which can be re-written in integral form as:
dF
∫L cos Ω dΩ = dA
Ω
dF
where the ratio dA is known as the irradiance .

Irradiance E : The amount of flux per unit area of a defined surface is the irradiance . The units of irradiance
are Watts/meter2.

⇒ ∫ L cosθ dΩ = E
Ω
It is clear that for a particular ‘beam’ of radiant flux (radiance L ) within solid angle dΩ at a particular angle
from the surface normal, irradiance E is equal to the radiance L multiplied times cosine θ. Since cosine θ is
always less than or equal to 1 and decreases as θ increases, irradiance decreases as θ increases. A simple example
of this relationship is shown in the following illustration.

Beam of radiant
flux lines projects
small gray circle.
Surface Normal

Same beam of
radiant flux lines,
but at greater
angle to surface
normal.

Same number of
flux lines, but
spread over larger
area (large circle).
Unit Area
Ie., lower energy
(small box)
per unit area. Ie.,
Surface lower irradiance.
 ASD Technical Guide 3rd Ed. Section 13-5

Cosine Receptor

If a radiometer is to measure radiant energy in irradiance units, then the sensor surface must emulate a response
similar to the example above, and it must integrate all such ‘beams’ of radiant flux incident on the surface as in
the integral shown above. In order that the sensor surface emulates this ‘cosine’ response, various designs of
optical accessories have been created, which are located between the incident radiant energy and the sensor
surface. Such accessories are typically called ‘cosine receptors’, or ‘cosine collectors’. Most cosine collectors
come in one of two forms, the ‘integrating sphere’ (IS) or ‘diffusion-disc collector’ (DDC). Both forms have there
advantages and disadvantages.

An IS consists of a hollow sphere that is coated on the inside with a diffuse nearly 100 percent reflectance
material. An entrance port for the incident illumination is constructed into the IS so that it is at a 90 degree angle
to the exit port in the IS, which is viewed by the radiometer sensor. A very basic design of IS is shown below.

Diffuse nearly 100%

reflectance coating

Entrance Port

Baffle
Exit Port
A DDC is constructed of a tube with one end covered by a diffusion-disc. Both the tube and disc are designed
with a geometry and material that optimizes the cosine response. The radiometer sensor views the radiant energy
that transmits through the diffusion-disc. Many designs have been tested over the years and the most current
designs compete well with IS’s in cosine response. A basic mechanical drawing of the DDC manufactured by
Analytical Spectral Devices, Inc. (ASD) is shown below and it is called the ‘Remote Cosine Receptor’ (RCR).

Diffuser Fiberoptic Input

to Spectrometer

Guard Ring

This ASD RCR is designed to interface with the FieldSpec® Portable Spectroradiometer fiberoptic input such that
the instrument has a cosine response for measuring irradiance within the geometric hemisphere above the diffuser.

Spectral Irradiance
 ASD Technical Guide 3rd Ed. Section 13-6

For simplicity the above discussions on irradiance have neglected the concept of wavelength interval or ‘band’. A
‘broad-band’ radiometer measures the sum total of integrated irradiances over all wavelengths within the band.
Such a radiometer does not discriminate the integrated radiances for discrete wavelength channels or ‘narrow-
bands’.

A ‘spectro-radiometer’ such as the ASD FieldSpec® Portable Spectroradiometer measures the integrated
irradiance for many contiguous discrete wavelength channels, all within a single scan. This kind of irradiance
measurement is called ‘spectral irradiance’ since the discrete values are stored and displayed as a spectral plot of
irradiance versus wavelength. Spectral irradiance has units Watts/meter2/nanometer, which indicates that a
discrete irradiance integration is measured for each discrete narrow-band interval of one nanometer.

Its is also possible to emulate the broad-band radiometer using a spectro-radiometer, assuming the total spectral
range is compatable. This can be done by integrating under the spectral curve of the spectro-radiometer in post
processing.

Spectral Radiance

A spectro-radiometer such as the ASD FieldSpec® Portable Spectroradiometer also measures ‘spectral radiance’
for contiguous discrete narrow band intervals of one nanometer and the units of spectral radiance are,
Watts/meter2/steradian/nanometer. Either the bare fiberoptic cable or a variety of optional Field-of-View Lens
Foreoptics can be used with the ASD FieldSpec® Portable Spectroradiometer for measuring radiance.

Radiometric Calibrations

One other area of detector response not covered in this paper, is the absolute radiometric response. That is, when
the spectro-radiometer sensor says that it is measuring one Watt/ meter2/nanometer, it is in fact measuring such a
value to within some known error. This area involves ‘radiometric calibration’, which relates the spectro-
radiometer sensor real response to a know, traceable calibration standard. ASD uses NIST traceable irradiance
standards for creating its radiance and irradiance calibration files for the ASD FieldSpec® Portable
Spectroradiometer. The procedures and other information regarding these radiometric calibration files are
discussed in great detail in the ASD FieldSpec® Portable Spectroradiometer User’s Guide.
 ASD Technical Guide 3rd Ed. Section 13-7

Case A: Simplification of Irradiance Integral for a Common Hemispherical Case.

Conditions: Illumination falling from all directions in a hemisphere onto a ‘point-unit-area’ of a surface through
an isotropic medium.
dA D proj
2π Steradians
Hemisphere

'Beam' of Radiant Flux

D dΩ

zθ

y
x

Surface
z
D
dΩ
dA 0proj θ

y
dA 0
φ
x

∫ L cosθ dΩ = E
Ω
 ASD Technical Guide 3rd Ed. Section 13-8

Let dAproj = dAD proj .

⎧ D2sin θ dθdφ ⎫
⇒ ⎨ dΩ = = sin θ dθdφ ⎬ 3
⎩ D 2
⎭

⇒ ∫ ∫ L cosθ sin θ dθ dφ =
φ θ
E

is true for any ‘point-unit-area’ on this surface so long as location of the ‘point-unit-area’ is not an influence on
the radiant energy. Also, since the medium that the radiant energy is passing through is isotropic, L is
independent of θ and φ:

⇒ L ∫ ∫ cosθ sin θ dθ dφ =
φ θ
E

Since the integration is over a hemisphere, φ varies from 0 to 2π,

⇒ 2πL ∫ cosθ
θ
sinθ dθ = E

θ2

⇒ 2πL ∫ cosθ
θ1
sinθ dθ = E

Let cosθ = u. Then, sinθ dθ = – du. And: u1 = cosθ1 , u2 = cosθ2 by chain rule for anti derivatives also known
as "anti differentiation by substitution" (see standard calculus texts such as James Hurley, Calculus, Wadsworth
Publishing, pp 211-216).
u2

⇒ 2πL ∫ u du
u1
= E

(u 22 - u 12 ) = E
1
⇒ 2πL
2

⇒ πL (cos θ 2
2 - cos 2 θ1 ) = E
θ varies from 0 to π.
π
⇒ θ2 = , θ =0
2 1
⇒ πL (cos θ 2 - cos θ1 ) = E
2 2

⇒ E =πL
⇒ The amount of total integrated irradiance E falling on the ‘point-unit-area’ is equal to π times the
radianceL passing through the hemisphere to the ‘point-unit-area’.

3see any good calculus book.

 ASD Technical Guide 3rd Ed. Section 14-1

Section 14
Illumination Geometry
David C. Hatchell, Analytical Spectral Devices, Inc.

Measurement consistency and signal-to-noise are dependent on Illumination Geometry. A thorough

understanding illumination geometry is essential for successful analysis and model building. The primary areas of
concern are 'Point Sources', 'Collimated Sources', Source distance and angle, and the Diffuse and Specular
characteristics of the target.

Point Sources

As shown in the illustration to the right, a ideal Point Source is just that,
a point that radiates light equally in all directions. The radially diverging
lines represent energy 'Flux'.

Of course, there is no such thing as a perfect 'real' Point Source.

However, an illumination source may be considered a Point Source if
its dimensions are small relative to the target distance.

The most important thing to note regarding a Point Source is that the
Flux density for a unit area of a target drops off as that area moves Unit
farther away from the Point Source, assuming angular orientation of area
the area stays the same. In other words, there are fewer Flux lines
impinging on the area the farther away it is from the Point Source.

As long as a point source is set at a constant distance and angle to the target, the Flux density remains the same.
Not a problem for flat targets that can be fixed in position. But what about targets with large textures, or targets
that move? Notice in the illustration below that the Flux density at surface A is greater than at surface B when the
object is close to the Point Source (for clarity, only the impinging Flux lines are shown).

However, increasing the distance of the object from the Point Source, results in a more equitable Flux density for
both surfaces. Of course, there is always some degree of difference, i.e., error. The goal is to set the distance so
that this error is negligible. For outdoor environmental measurements, using the sun as the Point Source, any
differences are certainly insignificant by virtue of the 93 million mile distance to the sun.

Fiber Optic Illuminators Fiber Optic

Bundle Illumination
All fiber optics have a numerical aperture, i.e., a conical Flux
shaped optical field. When used to carry Illumination to the
target, the Illumination output Flux for each fiber in the bundle
is radially divergent, so all the concerns of a Point Source
apply as shown on the right.
 ASD Technical Guide 3rd Ed. Section 14-2

Collimated Sources

According to Webster, the word 'collimate' means to make parallel. A

Collimated Source has all of its Flux lines parallel. As shown on the right, the
preferred way to Collimate light is with a lens. The most common
configuration uses a point source backed with a curved reflector
with a convex lens on the front.

A lens Collimator can be thought of as a reverse telescope. That is, if you

replace the telescope eyepiece with a Point Source, the light that emerges
from the large lens end will be Collimated.

The most important thing to note regarding a Collimated Source is that the
Flux density for a unit area of a target stays constant as that area moves
farther away from the Collimated Source, assuming angular orientation of the
area stays the same.

If you must use close Illumination for targets with large textures or that are moving, then a Collimated Source is
the easy choice. Collimation is also the best choice if you want to Illuminate a limited area or 'spot' such as with a
reflectance type integration sphere.

As shown on the right, another way to Collimate light is with a series of black
absorbing baffles.

Here, the baffles absorb all but the most nearly parallel Flux lines.

Still another way to Collimate light is with a parabolic reflector as shown below.

While this reflector approach Collimates most of the light (solid lines), much of the light also radiates divergently
out of the reflector opening (dashed lines). These radially divergent components can be dealt with by using a
combination of lens and reflector or by inserting a small blocking baffle just in front of the point source. Good
examples of lamps that use all three are modern automobile headlights.

Angular Dependence

In the previous examples, the angle of the source to the target is kept constant. However, even with small areas
Flux density depends upon the angle of Illumination. One of the most fundamental examples is Illumination on an
isotropic flat surface, i.e., a surface that physically interacts with light exactly the same way at every point on the
surface regardless of angle or position.
 ASD Technical Guide 3rd Ed. Section 14-3

To demonstrate Angular Dependence only, distance effects are eliminated by using a Collimated Source. The
illustration below shows a beam of Collimated light impinging on the isotopic Surface, forming a small spot (solid
line circle). Notice that as the angle α to the surface normal increases, the Flux lines (dashed) at the Surface
become more spread out forming a larger spot (dashed circle), i.e., the Flux density decreases.

Normal
to
Surface
a

Spectroradiometer
FOV Input

Surface

In geometric terms this Angular Dependency can be described as follows:

Flux Density at angle α = Flux Density at Normal * cos α

which is known as Lambert's Cosine Law or simply the Cosine Effect.

This effect can be measured using a spectroradiometer and an isotropic surface that reflects the Illumination
uniformly for the spectral range, such as a Diffuse White Reference Panel. Fix the position and the field of view
(FOV) of the spectroradiometer input so that it is limited to the smaller solid line circle. In other words, set the FOV
so that the smaller circle is the unit area. The spectroradiometer will measure the response to the light reflecting
off the surface only within the unit area. Move the Collimated Source to an off normal angle α and measure the
response of the spectroradiometer. Then, move the Collimated Source to progressive larger angles and measure
the response at each angle. Analysis of the response data will show that they decrease as the cosine of the
angle.

Remember that it is the Illumination Flux density that is really decreasing with angle. Therefore, the light energy
that is available to reflect within the unit area must therefore also decrease. It is essential to understand this
distinction. That is, that the Cosine Effect is about Illumination Angle, not sensor (spectroradiometer input) angle,
assuming sensor FOV and position are fixed. Also, the Cosine Effect is not about some mysterious physical
interaction between light and matter - it is strictly about geometry!

As mentioned earlier, a Collimated Source was used to study Angular Dependence separate from distance
effects. However, the Cosine Effect is present in light from a Point Source as well.
 ASD Technical Guide 3rd Ed. Section 14-4

Diffuse and Specular Characteristics of Targets

All matter Transmits, Reflects, and Absorbs optical radiant energy. The difference we see is due to the matter
having greater Reflecting or Transmitting or Absorbing characteristics. For example, even the most solid reflector
Transmits optical radiant energy, though these Transmitted components may be below the level of any known
detector. Generally, we call matter a Transmitter, Reflector, or Absorber depending upon its dominate
characteristics.

Diffuse Characteristics are associated with Reflectors and Transmitters, and are those Illumination components
that are Reflected or Transmitted with equal energy in all directions.

According to Webster, the word 'Specular' means, of, relating to, or having the qualities of a mirror (note the
difference between the words 'Spectral' and 'Specular'). Specular Characteristics are associated only with
Reflection. A Specular component is a Flux line that Reflects off the target at the same angle to the surface
normal as an Illumination Flux line. Specular components have greater Flux density than Diffuse components.

All Reflectors produce both Diffuse and Specular Reflected components. Generally, we call a Reflector Diffuse if it
mostly produces Diffuse components and only limited Specular components. A Specular Reflector is one that
mostly produces Specular components, like a shinny mirror.
About 90°

Lamp Specular
Spectrometer Reflected
Normal Components
Fiber optic
to
Input
Surface
Face

Diffuse
Reflected
Components
Target
Reflection spectroscopists are usually interested in measuring only Diffuse Reflected components since they are
abundant in most natural materials, easiest to find, and cause fewer measurement problems. As shown in the
illustration above, the easiest way to avoid Specular components (solid lines) and measure only Diffuse
components (dashed lines) is to set the spectrometer sensor input far off the Specular plane. For most set ups a
right angle to the Source axis is convenient, while maintaining good signal-to-noise. For example, as shown
above and below, the face of the input sensor is parallel to the side of the Lamp and away from the front of the
Lamp.
Spectrometer
Fiber optic
Top View Input

Specular
Reflected
Lamp Components

Target
 ASD Technical Guide 3rd Ed. Section 14-5

Shadowing

Another important Illumination Geometry issue that is rather

straightforward but often overlooked is shadowing. As shown in the
illustration to the right, shadowing can leave some areas of the target with
less illumination than others (dark area). Always try to position the
illumination source to minimize shadowing. Rotating the object and
averaging many spectra for different object positions is one approach to
minimizing the effects of shadowing. Of course, when using the rotation
approach, be sure that the source distance is sufficient to minimize
distance difference effects. In some cases multiple light sources might be
necessary to achieve more uniform illumination. When using multiple sources, be careful to configure them so
that the output, distance, angle, and flux density at the target are equivalent. When using lamps, always change
the bulbs at the same time so that the output from both is consistent for the set up. It is also a good idea to update
the White Reference frequently to accommodate any on going differences between multiple Sources.

Side Note: Source Physical Properties

Even though a study of the Physical Properties of Sources is off the subject of Geometry, a review is order to
avoid confusing these properties with the Geometry.

Sources can be any of a variety of generators of radiant energy, such as those included in the following list:

Filament emitters excited by electrical power, such as an incandescent lamp

Gaseous emitters excited by electrical power, such as a fluorescent lamp
Burning gaseous emitters
Burning solid emitters
Heated gaseous emitters
Heated solid emitters
Phosphorescent emitters, which continue to emit light after limited exposure to an excitation light source
Fluorescent emitters, which emit light while under constant excitation by shorter wavelength light
Bioluminescent emitters, which generate light internally by chemical reaction

Reflected light can also be considered a Source if that is in fact what is being used to Illuminate the target. For
example, moonlight.

Illumination output by any of these Sources can vary according to the generating mechanism. For example a
filament emitter varies according to the supplied electrical power. Ideally one wants an electrical power supply
that operates at a constant current setting such as a steady, linear Direct Current (DC) power supply. Alternating
Current (AC) generates undesirable sinusoidal fluctuations in output.

Illumination output can also vary by virtue of the design of the emitter. For example, some lamps have both a
Reflector and a protective front glass or lens; either of which can cause constructive/destructive interference
fringes in certain exact positions. Some lamps also include an adjustable position on the Reflector or front lens,
which can increase or decrease the Flux density output. Ideally one wants, a lamp with negligible levels of these
effects and to set /fix the adjustable features to optimal levels.

While electrically powered filament emitters and heated solid emitters generate nice Gaussian distributions of
energy, low pressure gaseous emitters such as most off the shelf fluorescent tubes generate energy with sharp
narrow features (high pressure gaseous emitters produce a smooth Gaussian distribution). For most applications
one wants the smooth Gaussian distribution of a solid emitter. However, if the application involves only very
limited, narrow spectral regions then perhaps a gaseous emitter with strong features in those regions is
appropriate.

Most filament emitters are designed with an internal gas and glass envelope, either of which can absorb radiant
energy from the filament thereby corrupting a smooth energy output. Ideally one wants these factors eliminated
from the design. For the region, 350 -2500 nm, a bulb with a tungsten filament, non-absorbing gas for the region
such as halogen, and a non absorbing envelope such as quartz, is the easy choice (tungsten-quartz-halogen).
 ASD Technical Guide 3rd Ed. Section 14-6

In outdoor environmental applications one is often limited to the Sun as the Source. Beyond the Earth's
atmosphere the Sun can be considered a heated high pressure gaseous emitter, which produces a smooth
Gaussian energy curve except at the Absorption features of its own outer low pressure atmosphere (Fraunhofer
Absorption Lines). However, here at the Earth's surface one also must deal with Absorption features of Earth's
low pressure atmosphere, the strongest being water vapor. Whenever possible, one wants to make outdoor
measurements on days with low atmospheric humidity and no clouds. Also, whenever possible, if the desired
feature falls within one of these low energy regions, one can also search for correlated features in more energetic
regions.

Another Illumination Property to consider is the medium in which the Source energy must travel through. As in the
atmospheric example above, if the Source energy must travel through a gas, liquid, or solid, then the Absorption
characteristics of those materials may fall in spectral regions that reduce energy where you most want it. If
altering the gaseous mixture, liquid solution, or solid matrix is out of the question, then the only practical
alternative is to reduce the path length, with the hope that the desired feature stays stronger than the undesired
Absorbing feature. This is in fact the case for most applications. For example when measuring Transmission of
analytes in water solutions in the NIR region, use a cuvette or immersion probe with the smallest practical path
length.

One final Source Property to consider is the energy output relative to the dynamic range of the spectrometer. It is
not practical to design any single measuring device to cover all ranges in the universe. Instead these devices are
designed to suit the ranges of most applications. If it is not possible to optimize the sensitivity settings of the
spectrometer to an Illumination Source that is too bright then Source attenuation may be necessary. Attenuation
can be easily accomplished by moving the Source away from the target (Point Source), turning down the Source
power, or inserting spatial attenuators or filters or apertures in the Illumination path. Turning down the source
power is usually not a viable option for tungsten-quartz-halogen bulbs, which require current within a limited range
to maintain output lifetime.

As best as possible, always configure your Illumination Source to produce constant energy output, and realize the
difference between the Source Physical Properties and Illumination Geometry.

Measurement Approaches to Suite the Illumination Geometry

The point of understanding Illumination Geometry is not necessarily to avoid one Geometry for the other. Instead,
approach your measurement to accommodate the Geometry available, which in many cases cannot be changed.
For example, in outdoor environmental applications using the Sun as the Source, one has little choice about the
Illumination Geometry. Even for indoor laboratory applications, sample size and handling, signal-to-noise,
application variety, and budgetary constraints might render the easy choice as the less practical one.

The good news is that almost all limitations of a particular Geometry are possible to accommodate by using the
correct measurement approach. In, general if the Illumination field changes re-do your measurement under the
new Illumination field.

If the Illumination field changes while making a Reflectance or Transmittance measurement, update the White
Reference and then re-do the target measurement so the changes drop out in the ratio calculation. If the
Illumination field changes significantly, it may also be necessary to update the Optimization (sensitivity) prior to
updating the White Reference and re-doing the measurement.

If the Illumination field changes while making absolute energy measurements such as Radiance or Irradiance,
then be sure to also update your baseline Reference (usually downwelling Irradiance) if you intend to ratio the
measurements in post processing. Again, re-Optimization may be necessary prior to the updates.

If the target has large textures or is moving, use a Collimated Source or distance the Point Source far enough
away to minimize delta-distance effects.

If the target texture is large enough or moving enough to create significant shadowing effects, then average
multiple spectra for different target positions, or use a spatially broader Source or multiple Sources. In the case of
multiple sources be careful to configure them so that the output, distance, angle, and flux density at the target are
equivalent. When using lamps, always change the bulbs at the same time so that the output from both is
 ASD Technical Guide 3rd Ed. Section 14-7

consistent for the set up. It is also a good idea to update the White Reference frequently to accommodate any on
going differences between multiple Sources.

If the target texture is large enough to create significant angular effects, then average multiple spectra for different
target positions. If the Illumination Angle has changed significantly then take a new Optimization and White
Reference, and re-do the target measurement.

Transmitting Diffuser filters in the Illumination path can also accommodate many Geometry effects. Choose a
Diffuser with negligible Absorption characteristics for the spectral region being used. Of course, such Diffusers will
reduce the overall Illumination energy reaching the target.

Integrating spheres can also be used to produce a more Diffuse Illumination field. Also, a contact Diffuse
Reflectance type integrating sphere can be used to combine all Specular and Diffuse Reflected components into
a uniform average. Of course, any Absorption features inside the sphere will reduce Reflectance signal by virtue
of the many bounces the Reflected Flux must make before reaching the spectrometer input.

In conclusion, the overall goal for Illumination Geometry is to make it consistent. If Illumination consistency is not
possible, then update White Reference and target measurement frequently.
ASD Technical Guide 3rd Ed. Section 14-8
 ASD Technical Guide 3rd Ed. Section 15-1

Section 15
Wavenumber and Wavelength

by David Hatchell, ASD

Y = 107 / X
where
Y = the number of nanometers (nm)
X = the number of wavenumbers (cm-1)

Y = 107 / 28571 = 350

Resolution in cm-1, (Rcm-1) is dependent upon wavelength position.

So, Resolution in nanometers, (Rnm) is calculated as follows:

Rnm = +/- [Y - Y']

= +/- {[107 / X] - [107 / X']}

= +/- 107 * {[1 / X] - [1 / X']}

= +/- 107 * {[1 / X] - [1 / (X - Rcm-1)]}

Example:

Rcm-1= +/- 8 cm-1 @ 4,000 cm-1

Therefore,

Rnm = +/- 107 * abs[[1 / 4,000] - [1 / (4,000 - 8)]}

= +/- 107 * abs{[0.00025] - [1 / (3992)]}

= +/- 107 * abs{[0.00025] - [0.0002505]}

= +/- 107 * 0.0000005

= +/- 5 nm

= 10 nm FWHM.

TABLE FOR ABOVE EXAMPLE:

cm-1: 3,992 <------ 8 ------> 4,000 <------ 8 ------> 4,008

nm: 2,495<------ 5 ------>2,500<------ 5 ------>2,505

ASD Technical Guide 3rd Ed. Section 15-2
 ASD Technical Guide 3rd Ed. Section 16-1

Section 16
Spectral Regions

Spectral region refers to a specific range of electromagnetic radiation. This range is characterized by wavelength
in nanometers (nm) or micrometers (µm) a.k.a. microns, or millimeters (mm), wave number
(cm-1), or frequency (Hz).

Electromagnetic Radiation

Region Names wavelength

nanometers (nm)
or microns (µ)
Cosmic
Ray
Gamma
Ray
X Ray
Ultra- Far Ultra-Violet 1 - 200 nm
Violet Ultra Violet C (UVC) 200 - 280 nm
(UV) Ultra Violet B (UVB) 280 - 315 nm
Ultra Violet A (UVA) 315 - 400 nm
Visible Photosynthetically Blue Light 400 - 525 nm
(VIS) Active Green Light 525 - 605 nm
Radiation Yellow Light 605 - 655 nm
(PAR) Red Light 655 - 725 nm
Far Red 725 - 750 nm
Infrared (IR) Near-Infrared Short Wave Near Infrared (SW-NIR) 750 - 1100 nm
(Near-IR)
*Optical Typical 1st NIR region detector 1000 - 1800 nm
(NIR1) or (SWIR1)
Typical 2nd NIR region detector 1800 - 2500 nm
(NIR2) or (SWIR2)
'Conventional' 1000 - 2500 nm
Near Infrared (NIR) or 1.0 - 2.5 µ
Mid Infrared 2.5 - 50 µ
(Mid-IR) Thermal (emitted) 8 - 15 µ
Far Infrared (Far-IR) 50 - 1000 µ
Micro
wave
&
Radar
UHF TV
VHF TV
&
FM
Radio
AM
Radio

*Optical Electromagnetic Radiation: 'Optical' means controllable by lenses, mirrors, prisms, and fiber optics.
ASD Technical Guide 3rd Ed. Section 16-2
 ASD Technical Guide 3rd Ed. Section 17-1

Section 17
Reflectance
by David Hatchell

By definition, reflectance, R, is a calculation, such that for a given wavelength channel:

R = Energy reflected from target / Energy incident on target

In the case of optical radiant energy, the construction known as radiance, L, is used to quantify the optical radiant
energy. Radiance has the units, Watts per square meter per steradian per nanometer (W/m2/str/nm). So, in this
case, let Lt be radiance reflected from the target, and Li be radiance incident on the target. Then, for a given
wavelength channel:

R = Lt / Li

You can see that in this calculation of R, the units (W/m2/str/nm) cancel out. So, as long as our instrument is
configured to properly view Lt and Li, it really doesn't matter that it be calibrated to absolute units. All we really
need is the response of the instrument for Lt in some unit of measure and the response of the instrument for Li in
the same unit of measure. In the case of an analog photo detector, PD, the current produced by the PD is linear in
response to radiance (voltage is approximately logarithmic). This analog current signal can be converted to digital
numbers, DN, as in the FieldSpec®, giving the following for a given wavelength channel:

R = DNt / DNi

Reference panel method for measuring reflectance:

Now as far as the forementioned "properly view Lt and Li" is concerned, one can easily measure Lt by pointing the
fiberoptic input of the FieldSpec® at the target. But where does one point the fiberoptic to measure Li, especially
in the case of outdoor measurements using solar illumination? Do you point it at the sun? Do you point it at the
sky? Do you point it somewhere in the middle? In other words, how, with the limited field-of-view, FOV, of any
optical input do you view all the incident energy of the sun and sky per unit steradian? The answer is to aim the
fiberoptic input at a diffuse white reflectance panel (sometimes called a calibration panel) that provides a diffuse
homogeneous mix of all the full sky and sun radiance reflected at nearly 100 percent up to the fiberoptic input. By
the way, chemists like to call this white reference measurement the "baseline".

(Ir)radiance method for measuring reflectance:

A second method for measuring R involves the construction known as irradiance, E, which has units Watts per
square meter per nanometer (W/m2/nm). Again, E is another way of quantifying optical radiant energy. E was also
constructed to have a relationship with L. There are many good references discussing the concepts and
derivations associated with E and L, such as those listed at the end of this essay. I would like to point out, that
these concepts are often introduced by relating L to point emitters and E to flat receivers. While such a
presentation offers one example of applying these concepts, this example is not the whole story. Without
digressing to much, in general, L and E are geometric constrictions for quantifying the optical radiant energy
within a space as it impinges on a geometric surface, regardless of where it came from or where it is going.

Well, getting back to E, there is a relation between E and L such that in the case of solar, full-sky incident energy
Ei, for a given wavelength channel:

Lt = R Ei / π

Therefore:

R = π Lt / Ei

I'm assuming the distance from the target to the fiberoptic input is close enough that atmospheric attenuation of
transmission, T, is negligible. Such atmospheric effects become noticeable for great distances between target and
 ASD Technical Guide 3rd Ed. Section 17-2

spectrometer, such as from a flying aircraft to the ground. I'm also neglecting stray light radiance, S, back-
scattered from the atmosphere or reflected from buildings, trees or other objects too near the target. A more
thorough equation for such cases would be of the form:

Lt = S + REi T / π
Assuming conditions allow the simpler equation (R = π Lt / Ei) to hold, one can measure R for a given wavelength
channel by measuring Lt and then attach a cosine collection device to the fiberoptic input to measure Ei. Of
course, measuring R this way, does require absolute radiometric calibrations, one for radiance using the fiberoptic
input, (and/or fiberoptic input with FOV) and one for irradiance using the cosine collector attached to the fiberoptic
input.

This second approach also offers the ability for near continuous monitoring by using two spectroradiometers. One
spectroradiometer continuously monitors Ei while the other is used for Lt measurements.

NOTE: Calibrated White Spectralon is a near 100% diffuse (Lambertian) reference reflectance panel made from a
sintered poly - tetra - fluorethylene based material. You may also want to consider material such as Sintered
Halon or Barium Sulfate (BaSO4) powder. We prefer Spectralon, which you may purchase through ASD or
directly from the vender: Labsphere, Inc., Shaker St., PO Box 70, North Sutton, NH 03260-0070, PH: 603-927-
4266, FX: 603-927-4694, E-mail: [email protected], Website: www.labsphere.com.

References:

Remote Sensing Optics and Optical Systems, by Phillip N. Slater, 1980, Addison-Wesly Publishing Company,
Reading, MA

Remote Sensing and Image Interpretation, by Thomas M. Lillesand and Ralph W. Kiefer, 1994, John Wiley &
Sons, Inc., New York

Introduction to the Physics and Techniques of Remote Sensing, by Charles Elachi, 1987, John Wiley & Sons,
Inc., New York

Fundamentals of Radiometry: Calculation, Measurement, and Calibration, by James M. Palmer, 1995, SPIE,
Bellingham, WA

"Ground Reflectance Measurement Techniques: A Comparison", by M. J. Duggin and T. Cunia, in Applied Optics,
December 1993, vol 22, No. 23

"Radiance", by Fred E. Nicodemus, in American Journal of Physics, Vol. 31(5), 1963

"Distribution of Optical Radiation with Respect to Position and Direction -- Radiance", by Fred E. Nicodemus and
Henry J. Kostkowski, in National Bureau of Standards Technical Notes, Vol. 910-1: 10-44, 58-59, 61, 1976
 ASD Technical Guide 3rd Ed. Section 18-1

Section 18
Wavelength and Radiometric Calibration Methods
WAVELENGTH AND RADIOMETRIC CALIBRATION METHODS
by Dave Beal, Analytical Spectral Devices, Inc.

Introduction:

Reliable field spectrometry data collection depends, to perhaps the greatest extent, upon accurate calibration of the
instruments used. The Analytical Spectral Devices FieldSpec®™ line of spectroradiometers is commonly used in the analysis
and cross-referencing of Reflectance, Transmittance, and Absorption characteristics of materials and land/sea surfaces. For
this reason, accurate wavelength calibration is a necessity and a standard feature included in the acquisition of any
FieldSpec®™ instrument package. Periodic examination of the absorption features in the spectra of materials with known
characteristics is highly recommended for the SWIR (short wave infrared; 1000 - 2500 nm) detectors. A mercury discharge
tube or other known discrete emission light source works well for verifying calibration in the VNIR (visible and near
infrared; 350 - 1000 nm) portion of the spectrum. After the instruments are calibrated at the ASD factory with a 0.1 nm
bandpass monochromator, their responses to several known elements and emission sources are checked, before shipment, as
a verification of method and equipment.

Due to the consistency of the detector performance in the FieldSpec®™ instruments, they are also commonly used,
as their name implies, in radiometry applications. This, naturally, requires accurate and reliable radiometric calibrations for
the instrument and accessories used in such applications, as well as software for manipulation of calibration and data files.
Radiometric calibration is an optional feature in the purchase of any FieldSpec®™ instrument and, in most cases, must be
performed at Analytical Spectral Devices’ factory, in Boulder, CO, USA. This paper is intended as a guide for current and
potential ASD customers who would like to know more about our methodology in calibration so that they can make a better
informed decision about how they would like to proceed with their instruments; and perhaps even in how they would prefer
to collect and analyze their data.

Reflectance or Radiance?

Many papers written on the subject refer to the use of spectroradiometers in field reflectance factor measurements.
This is because with some instruments, it is best to use the ratio of the reflected radiance of the observed sample to the
reflected radiance of a known reference panel (taken at the same time), in order to determine the reflectance factor of the
sample. Of course, when this is done, all units fall out in the result, which is typically a number between zero and one for
each channel measured. However, instruments such as the FieldSpec®™ spectroradiometer, with a well-characterized linear
response, can be used for the same purpose without going to the trouble and expense of radiometric calibration. Raw signal
inputs from the sample and the reference panel can be ratioed to achieve the same result.

This does not mean that FieldSpec®™ instruments must be used only as spectrometers, or spectrophotometers, as
the case may be. On the contrary, at least one-fourth of ASD’s customers have a true need for spectroradiometric data.
Spectroradiometric needs might include: Ground-truthing of reflected radiance data collected from high altitudes or from
different instruments, radiance of molten solids such as lava, steel and glass, solar irradiance at various levels of the Earth’s
atmosphere and points on the Earth’s surface, studies in photosynthesis, flame studies, and synthetic light studies, among
many others. But for most of our customers, who use the instruments for remote sensing ground truthing exercises, material
absorptance/reflectance characteristics, water/snow/ice studies and mineral identification, an instrument with a good linear
response to intensity across the spectral region of interest is all that is needed, in addition to an accurate wavelength
calibration, of course.

Wavelength Calibration:

Wavelength calibration is a standard feature of all ASD instruments. Typically, one calibration will be accurate to
within one nanometer and is guaranteed for one year. ASD will always recalibrate a FieldSpec®™ instrument’s wavelength
values once, within that time, at no extra charge.

The UV/VNIR detector array and housing combine to provide a simple linear relationship between wavelength and
channel number. For this reason, all that is needed for calibration is a few well-characterized emission lines, spread
throughout the region from 350 - 1000 nm wavelength, and a finely focused instrument. At ASD, the emission lines come
from a separate monochromator, set to emit at 50 nm intervals, which are plotted against the responding channel numbers,
 ASD Technical Guide 3rd Ed. Section 18-2

and the first channel number’s wavelength is extrapolated from a linear regression fit of the data. The final equation is a
simple linear formula, in the form:

wavelength = lamstart + (lamstep)(channelnumber)

The constants, lamstart and lamstep, are calculated as above, at the ASD factory and installed into the controlling
computer’s asd.ini file for access by the controlling software. The channelnumber domain is 0 to 511. Wavelengths of
certain emission lines from a Mercury-Argon source lamp are then measured as a cross-calibration of the monochromator
values. Some of the HgAr lines may be used for the calibration itself. But the presence of doublets means that not all of the
lines should be used. Contact ASD for the purchase price of your own HgAr portable emission source. The monochromator is
also frequently cross-calibrated using up to four orders of Helium-Neon laser diffraction.

Mercury-Argon Emission Spectrum

4
7 10

4
6 10
852.1

4
5 10
763.5
866.8

4
4 10
Raw DN

912.3

4
3 10
546.1
706.7
4
2 10
965.8
435.8 696.5
4
1 10

400 500 600 700 800 900 1000

Wavelength

The SWIR scanning spectrometers, covering the range from 1000 - 2500 nm, use much the same calibration principles, with
two major differences: 1. We use a monochromator as our only emission source, and 2. Two third order polynomials are
calculated for each SWIR detector, to account for both forward and backward scans of the gratings. This amounts to the
calculation of eight constants for each detector, listed in the ini file, as noted above. We then check the SWIR wavelength
calibrations with well-defined absorption features in a material such as Mylar or Polystyrene.
 ASD Technical Guide 3rd Ed. Section 18-3

Mylar Transmission, Unit FSFR 650

9/25/96
1

0.8

1128.6
Transmission

0.6

1659.5
0.4
2130

0.2

2443

500 1000 1500 2000 2500

Wavelength in nm

If you would like any further information on wavelength calibration of the FieldSpec®™ instruments, please contact Sales or
Technical Support at ASD.

Radiometric Calibration:

Radiometric readings with the FieldSpec®™ come in two basic flavors:

1. Irradiance (E), which is the radiant flux (Φ) per unit of area, or E = dΦ/dA, given in W/m2. This term is only definable
at a given distance from a given radiant energy source, or through a given surface in space, without regard to sources.
2. Radiance (L) is the radiant flux emitted from a source per unit of solid angle (ω) per unit area. It is useful at this time to
define Radiant Intensity as I = dΦ/dω, such that L = dI/dA cosθ, given in W/sr-m2 (Pedrotti, F. L. & L. S. (1993)
Introduction to Optics, 10 - 13).

Thus, the irradiance (E) at a given surface may be converted to the radiance (L) emitted from that surface by
dividing E by π. It is important to note, though, that the conversion requires a couple of important properties inherent in the
surface, in order to be even a close approximation. First, that the surface be as near to 100% reflective (at all wavelengths) as
possible; and second, the surface must be perfectly lambertian (or at least nearly so). That is, it must be perfectly reflective,
and it must radiate uniformly in all directions.

As stated earlier, FieldSpec®™ instruments used for radiometric readings must first be calibrated to do so. Even so,
by definition, the quantity actually measured by a spectroradiometer that measures irradiance is E per ∆λ, while radiance
measurements are actually L per ∆λ (better known as spectral irradiance and radiance), as the data is collected across a given
spectrum. The collected curves must be integrated with respect to wavelength to arrive at the definitive terms. All
FieldSpec®™ instruments currently must be set to collect raw data, later to be converted to radiometric data in
postprocessing. The following is a rough compilation of the procedures and equipment used in calibrating a unit at the ASD
factory.
 ASD Technical Guide 3rd Ed. Section 18-4

The first requirement is, of course, a well-focused and wavelength-calibrated instrument. An irradiance source is
powered by a stable DC current-regulated supply as the calibration standard. We use a lamp and power supply from Optronic
Laboratories Inc., which is traceable by NIST (the National Institute of Standards and Technology), in the USA. The 3200
Kelvin lamp irradiance data, at a given distance from the bulb and a given current, is supplied by the manufacturer and
guaranteed to be accurate within 1-2% in total irradiance, relative to NIST uncertainty (Optronic Laboratories, Inc. (1994)
Condensed Catalog, 7). The system is installed inside a flat black baffled box, to insure that the only energy projected upon
the screen is that coming directly from the bulb, itself – Not reflected from equipment, walls, etc. The bulb is calibrated at a
healthy 50 cm distance, so uncertainty due to the intersection of the spherical irradiance front with the planar Spectralon®
panel can be considered negligible, when viewing a small portion of the surface in the center of normal energy incidence.

The screen surface is a 12 X 12 inch panel of calibrated Spectralon® made by Labsphere. This panel was calibrated
from at least one of a number of NIST-calibrated reference standards, which were in turn calibrated “using the highly
accurate NIST reference reflectometer” (Labsphere, (1995) Calibration Certificate, 8° / Hemispherical Spectral Reflectance
Factor, Report No.: 14440-D). The random error in the calibration file is estimated to be within 0.5% over the spectral region
from 300 - 2200 nm, and within 2% from 2200 - 2500 nm.

An increase in a Full Range instrument’s internal ambient temperature leads to an increase in VNIR detector
sensitivity from about 700 - 1000 nm, so each unit is allowed to warm up for about 1.5 hours before calibration data is
collected. To learn more about this VNIR temperature/sensitivity effect, contact technical support at ASD. ASD monitors the
usage of the calibration lamp to maintain the guaranteed irradiance, and periodically checks its performance against at least
two backup lamps of the same style and wattage.

Providing a buffer against ground vibrations (caused by construction equipment, trucks, Great American Bison
herds, etc.) is a set of six pneumatic high-frequency shock absorbers, placed beneath the optical benchtop upon which the
entire apparatus resides. This gives great stability to the actual lamp filament, and increases the lifetime of the bulb as well as
its guaranteed accuracy, when compared with our backup.

The power to the bulb is ramped upward over several minutes, to a final current value of 8.0000 Amperes. The
calibration room is completely darkened, except for the lamp, and the instrument’s fiber optic probe is placed in a stationary
position at a 45 degree angle to the illuminated Spectralon® panel. The controlling software is entered, the instrument is
optimized, and spectra are saved, as raw DN files, with each foreoptic attachment requested, with a spectrum average of 50
scans per spectrum, including 50 dark current scans. If a unit is being calibrated for a Remote Cosine Receptor, the
Spectralon® panel is replaced with the fiber optic, fitted with the RCR, in the same irradiance calibration plane.

At this time, it is usually convenient to collect data for the NE∆L (Noise Equivalent change in Radiance) properties
of the instrument. We do this with no foreoptic attached, in the configuration for calibration with bare fiber, at 10 spectrum
averaging. 30 spectra are saved and the standard deviations from the means at all channels are calculated and converted to
radiance data. This data will tell us here at ASD, as well as the customer, what the true performance of their unit is like. And
it assures everybody that our units are meeting noise specifications.
 ASD Technical Guide 3rd Ed. Section 18-5

Noise Equivalent Change in Radiance,

Instrument #645, October 14, 1996
-6
10

-7
10
NEL, W/m^2/nm/sr

-8
10

-9
10

-10
10

500 1000 1500 2000 2500

Wavelength

The ASD radiometry software, RCALC.EXE (postprocessing) uses the following formula in calculating Radiance:

(lampfile)(calpanelfile)(inputfile)(calITG)
L = ⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯
(calibrationfile)(inputITG)(π)

where L is the radiance to be calculated (on a channel by channel basis)

lampfile is the calibrated irradiance file for the lamp
calpanelfile is the calibrated Spectralon® reflectance file
inputfile is the unknown, dark current corrected input file
calITG is the integration time and/or gain of the calibration file
calibrationfile is the dark current corrected raw data collected at ASD
inputITG is the integration time and/or gain of the input file

The divisor of π is automatically left out in the calculation of an Irradiance (E) measurement. The software “knows” which
formula to use by looking at the foreoptic specified in the header of the data file.

Conclusion:

The expense involved in the calibration process is not limited to calibration equipment and its maintenance, though
this is a major portion of it. There is also a considerable outlay of time and research in software development and quantitative
analysis of our own results and repeatability. This is not to say that our users would not wish to conduct their own
calibrations. Many researchers need the control over and accountability for uncertainties that an outside entity simply cannot
provide.

However, when conducting studies concerning the viability of performing their own calibrations (radiometric, in
particular) FieldSpec®™ users should consider the necessity of building postprocessing software from what may amount to a
whole new platform, from the ground up. On one hand, this allows great freedom and flexibility in their options; on the other,
it also requires great resources and patience.
ASD Technical Guide 3rd Ed. Section 18-6
 ASD Technical Guide 3rd Ed. Section 19-1

Section 19
Blackbody Radiance and Spectroradiometric Calibration
David C. Hatchell
Analytical Spectral Devices, Inc.

Blackbody

In 1860, G. R. Kirchhoff showed that an object that efficiently emits radiant energy when heated, also efficiently
absorbs radiant energy when the radiant energy source is external. He concluded that a “perfect” emitter is also a
“perfect” absorber and such an object was given the name “black body”.

12
1 10

3000 deg K
11
8 10
Radiance W/m /sr/m

Planck
11
Blackbody
2

6 10
Radiance

11
4 10

11
2 10
2000 deg K

1500 deg K
0
0 -6 -6 -6 -6 -6
0 10 1 10 2 10 3 10 4 10 5 10
Wavelength m

Planck Radiation Law

In late 1900, Max Planck used quantum mechanical theory to derive a mathematical expression that nicely fits
the measured radiance curves of many blackbodies. This expression is know as Planck’s Radiation law:

Lλ,T = 2•h•n2•c2 • λ–5 • {–1 + exp(h • c / λ • κ • T)}–1 [ Watts • meter–2 • steradian–1 • micron–1 ]

Lλ,T = Spectral radiance of a Blackbody at wavelength λ and true temperature T

h = Planck’s constant = 6.626176(36) • 10–34 [ Joules • seconds]
n = the index of refraction of the medium through which the radiant energy is traveling (eg. nair = 1)
c = velocity of light = 2.997992458(1.2) • 108 [ meters • seconds–1 ]
κ = the Boltzmann constant = 1.380662(44) • 10–23 [Joules • °K–1 ]

Imperfect Blackbody

If an object is not a perfect emitter then the radiance, LO of the object is found from the following expression:

LO,λ,T = εO,λ,T Lλ,T which can be rewritten as: εO,λ,T = LO,λ,T / Lλ,T

where εO,λ,T is the 'emissivity', a.k.a., 'emittance' of the imperfect emitter at wavelength λ and true temperature T.

In other words, emissivity is the ratio that relates non-blackbody radiance to blackbody radiance.
 ASD Technical Guide 3rd Ed. Section 19-2

If εO,λ,T is constant for all wavelengths, the object is called a 'grey body'. A good example of a grey body is a
tungsten filament, as shown in the plot below.

0.3

Tungsten-Quartz-Halogen
0.25
Irradiance, W/m^2/nm/sr Lamp ('Gray Body')

0.2

0.15

0.1

0.05

0
500 1000 1500 2000 2500
Wavelength (nm)

National Institute of Standards & Technology (NIST) “Gold-Point” Blackbody Primary Standard

Somewhere near Paris is the International Bureau of Weight and Standards (IBWS). In a environmentally control
vault at IBWS is a particular platinum iridium cylinder. The mass associated with this cylinder is what the whole
world has agreed to call 1 kilogram. Since a perfect measure of mass is only a theoretical concept, we just have
to pick some physical object and call it our primary standard of mass. In other words, all mass measurements
made relate back to the mass in the vault. Like the measure of mass, a perfect blackbody is a theoretical concept,
and we must simply agree to a physical representation. The physical representation that most of the world has
agreed to in this case, is the Gold-Point Blackbody at the National Institute of Standards & Technology, in
Gathersburg, Maryland, USA.
Molten State
Temperature of Gold

Freeze
Point
Plateau

Time
 ASD Technical Guide 3rd Ed. Section 19-3

A critical factor in the decision to use a Gold-Point Blackbody was temperature predictability and stability over a
long period of time. This long period of time is required so many data points can be recorded over the entire
spectral region, and so that many data points can be averaged to increase signal-to-noise for the calibration
instrument. When very pure gold metal is heated to the molten state and then allowed to cool, a very stable
“freezing temperature plateau” occurs with the change of state from molten , i.e, liquid, to solid. This plateau
typically lasts about an hour.

Gold-Point Blackbody Construction and Operation

The Gold Point Primary Standard is constructed with heater elements that can evenly heat the gold core well
above the melting temperature. An appropriate thermocouple is placed to measure the temperature of the gold
core.

According to Kirchhoff, absorptance = emissivity, and for a blackbody both of these values are 1. Therefore, a
cavity is constructed of a shape and material such that the absorptance is nearly 1. From conservation of energy
we know that nearly 100% absorptance is equivalent to nearly zero reflectance (interesting note: the converse is
also true, i.e., a perfect reflector will not emit any radiant energy when heated). So, the cavity shape is chosen to
allow many internal reflections and a cavity material with uniformly low reflectance such as graphite is used.
Realistically, the best achievable cavity absorptance values are around 0.99 and these absorptance values
become the emissivities for the Gold-Point Primary Standard, i.e.:

LG,λ,T= εG,λ,T Lλ,T

Where, the difference between εG,λ,T and 1 are known 'traceable' uncertainties for this standard.

When heating elements are activated the gold turns to molten metal. The heating elements are then turned off
and the gold begins to cool. When the freezing-point thermocouple begins to show no change in temperature, the
radiant energy exiting through the aperture is the primary standard of blackbody radiance at the freezing-point of
gold which is, 1337.33 ± 0.34 °K.

Pure Gold
Metal

Aperture Heating
elements
Cavity made
of material Ceramic
with refractory
emissivity material
value very
close to 1
at all
wavelengths Freezing-point
Thermocouple

During the Freeze Point, a well-characterized Primary Spectroradiometer is used to record the signal associated
with viewing the Primary Standard output at various wavelengths.
 ASD Technical Guide 3rd Ed. Section 19-4

Other Primary Standards

Other Primary Freeze Point standards are also used to add more calibration points to the Primary
Spectroradiometer. These include:

Silver, freeze point 1234.93 °K

Copper, freeze point 1357.77 °K
Zinc, freeze point 419.58 °C
Tin, freeze point 231.9681 °C
Lead, freeze point 327.502 °C

Appropriate interpolation functions are then used to complete the radiance response characterization of the
Primary Spectroradiometer. This characterization model is recalled to characterize other 'non-Primary' standards.

Lamp Transfer Standards

Using the Primary Spectroradiometer, its radiance response characterization model, and an appropriate
integrating sphere, the radiance of secondary 'Lamp Transfer Standards' are measured at a specific current
settings. The lamp, radiance data file, current and distance settings, and traceable uncertainty numbers are sold
as a package so that other similarly constructed, third level, 'Working Lamp Standards' can be calibrated.

Calibration of Commercial Spectroradiometer

The Working Lamp Standards, and their calibration files are used to calibrate commercial spectroradiometers.
Since the radiance output of the lamps is know for the specified current and distance, the correction factors for the
spectroradiometer signal output can be determined, to within the traceable uncertainties.

References

Klaus D. Mielenz, Robert D. Saunders, Albert C. Parr, and Jack J. Hsia, The 1990 NIST Scales of Thermal
Radiometry, Journal of Research of the National Institute of Standards and Technology, Volume 95, Number 6,
November-December 1990.

T. P. Merritt and F. F. Hall, Jr., Blackbody Radiation, Selected Papers on Radiometry, SPIE Milestone Series,
Volum MS14, Page598-604, Reprinted from Proceedings of the IRE, Vol. 47(2), 1435-1441 (1959).

Arthur J. Cussen, Overview of blackbody radiation sources, SPIE Milestone Series, Volum MS14, Page 647-660,
Reprinted from Infrared Sensor Technology, Proc, SPIE 344, 2-15, (1982).

J.C. Richmond and D. P. DeWitt, Applications of Radiation Thermometry, A symposium sponsored by ASTM
Committee E-20 on Temperature Measurement in cooperation with the National Bureau of Standards, pages 4-7,
(1984).
 ASD Technical Guide 3rd Ed. Section 20-1

Section 20
The Benifits of Portable NIR Analysis

Portable NIR analysis is a new concept that has been made possible through advances in solid state and optical
materials as well as computer miniaturization. Because portable NIR analysis is a new concept, its time and labor
saving potential are limited only by the imagination of the user.

But bench-top systems require that the sample be carried to the lab and carefully prepared for examination. The
lab instrument performs its own sequence of steps that can take several minutes. Finally, the lab results must be
communicated to the appropriate personnel. Such a procedure requires the time and effort of several persons,
and may also require undesirable handling and disposal of materials.

In contrast, a Portable NIR analyzer can be transported to the site of interest and measurements taken very
quickly.

NIR SPECTRUM

The near infrared region is immediately adjacent to the visible range and falls between 750 and 3,000 nanometers
in wavelength. Most organic materials have excellent reflectance or transmittance properties at these
wavelengths.

The biggest advantage of NIR over Mid-IR and Far-IR is sample preparation. Little or no sample preparation is
required for NIR.

However, significant preparation required for Mid-IR and Far-IR. For example, due to the extremely high
absorption coefficients of materials in the Mid-IR and Far-IR, transmission measurements require that samples be
extremely thin, including liquid samples. The samples must be placed between cell window materials that are
transparent in the IR, e.g., KBr, CsBr, AgCl, and NaCl, which must be cleaved and polished. These materials are
very hygroscopic and must be stored in a desiccator. 'Mulling agents' such as mineral oil must also be used in
preparing solid samples. The solid sample, as a fine powder, is suspended in the mineral oil and then smeared
between two cell plates. Extreme care must be taken in handling and cleaning the cells. Any oils on one's hands
could contaminate the cells.

Another advantage of NIR over Mid-IR and Far-IR is 'thermal' noise. All internal electronic components are a
source of thermal noise in the Mid-IR and Far-IR. However, internal sources of IR are either insignificant to NIR
detectors or can be made insignificant by minor shielding.

The disadvantage of NIR analysis is that most of the useful features in a spectrum consist of 'overtones' or
combinations of overtones, which are much subtler than the distinct fundamentals found in Mid-IR and Far-IR
spectra. However, recent developments in off-the-shelf chemometrics software and powerful PC's have made NIR
analysis the practical choice for most applications.

SPECTRAL FEATURES

According to the principles of quantum physics, molecules may only assume discrete energy levels. Similar to the
vibrating string of a musical instrument, the vibration of a molecule has a fundamental frequency, i.e., wavelength
as well as a series of overtones. For molecules, the fundamental vibrations involve no change in center of gravity
of the molecule. The spectrum shape for any material is the result of these characteristic fundamentals and
overtones. There are many fundamentals in the mid and far infrared regions. Near-infrared spectra are mainly the
result of overtones.

Since the molecular structure of most compounds is very complex, the resulting spectra are actually the result of
many overlapping peaks and valleys. Generally speaking, persons performing NIR analysis must then identify and
characterize specific features in the spectra by means of statistical methods. Chemometrics software is designed
to accomplish this task.

CHEMOMETRICS MODELS
 ASD Technical Guide 3rd Ed. Section 20-2

In order for any NIR analyzer to make quantitative measurements or qualitative discriminations, the controlling
computer must have access to one or more chemometrics models, which represent the type of material being
tested. The model is a mathematical construct that had been developed using samples of the same product or
class of products. The controlling computer applies the model(s) to the target spectrum and returns a model
result.

A chemometrics model is developed by collecting spectral readings from a group of samples that display (A) the
maximum variability of the characteristic of interest, and (B) non-correlating or random variability in all other
characteristics. The same samples are submitted for independent testing to measure the characteristic of interest
by a standard analytical method.

The spectral data and independent test data are then analyzed using commercially available chemometrics
software. The statistical processes used in quantitative spectral analysis include multiple linear regression,
classical least squares, inverse least squares and principal component regression. The statistical processes used
in qualitative spectral analysis include K-nearest neighbors, SIMCA and others.

When a sufficient number of samples has been collected and properly analyzed, a mathematical model is
constructed that describes the relationship between specific spectral features and the sample characteristic of
interest. Thereafter, a chemist or technician may quickly measure that same characteristic in a new target sample
by applying the chemometrics model to the spectrum of the target sample.

An NIR analyzer is also able to identify an unknown target sample if its spectral signature matches any of the
models already developed.

ASD LABSPEC® PRO

The LabSpec® Pro is a highly portable, general-purpose spectrophotometer that has many applications for
product analysis. The instrument measures the unique pattern that light reflects off or is transmitted through a
product sample.

The LabSpec® Pro uses a fiber-optic sampling probe to deliver illumination to the physical sample and to collect
reflected or transmitted light from the sample. A tungsten-halogen bulb provides the source light, which is carried
by one optical fiber or bundle of fibers to the tip of the probe. A separate optical fiber carries the reflected or
transmitted light back to the instrument where it passes through the entrance slit of the spectrometer.

The heart of the LabSpec® Pro is its proprietary spectrometer, which depending upon the model measures
incoming light in the Visible (VIS), and/or near-infrared (NIR) region of the electromagnetic spectrum (see price
list for different models according to spectral coverage).

The instrument functions well with many available reflectance, transmission or trans-reflectance fiber-optic
probes. Generally, reflectance probes are used for solids, trans-reflectance probes are used for clear liquids and
transmittance probes are used for liquids containing particulates.

ASD consultants can assist in the selection of appropriate fiber-optic probes for specific applications.

BUILT-IN LIGHT SOURCE

NIR analysis requires an adequate level of incident light on the sample. The LabSpec® Pro provides a controlled
source of incident light having a color temperature of 3,000°K. Even though the built-in light source is designed for
stability and accuracy, characteristics of the light source are eliminated when the baseline spectrum is
automatically applied to each raw sample spectrum. The internal light source reaches maximum stability after
approximately 5 minutes of warm-up time. If it is necessary to acquire spectra during this warm-up period, take a
new baseline spectrum immediately before taking each sample spectrum.
 ASD Technical Guide 3rd Ed. Section 21-1

Section 21
Making an NIR Analyzer Work For You
MAKING AN NIR ANALYZER WORK FOR YOU

Zhao Lu, Analytical Chemist

Analytical Spectral Devices, Inc.
Boulder, CO 80301 USA

ABSTRACT

In recent years, NIR analysis has steadily grown in popularity because of its ability to quickly provide qualitative
and quantitative information on many products, especially raw materials. To determine if NIR spectroscopy is a
reasonable alternative to more traditional methods, many factors must be considered. These factors include
sample characteristics, experiment configuration, and data analysis.

SAMPLE CONSIDERATION

The chemical constituents and physical phenomena of interest should have direct or indirect absorbance in the
NIR region. Virtually all organic compounds do, particularly those with functional groups like hydroxyl,
carboxyl, amine and carbon-hydrogen. A good reference for researching near infrared spectra is The Atlas of Near
Infrared Spectra, Bio-Rad Sadtler Division, Philadelphia, Pennsylvania. For calibration samples, the amount of
analyte in the sample set should be above the detection limit and have sufficient variability. Some analytes are
reportedly detectable at the ppm level. For most analytes, the nominal detection limit will be 1% or above.

The analytical chemist must have an accurate independent method for measurement of the properties and must
know the level of error in the reference methods. Errors in NIR prediction most often arise from errors in the
reference methods, instability of the NIR spectrometer, and inappropriate choice of the calibration model.

The samples used in the development of calibration sets must be representative. All the variations in the future
unknown samples should be covered in the "training" calibration sets, for example, sample composition and
particle size, homogeneity, and temperature variation at the working environment. As a rule of thumb, the more
samples you have for the training set, the more reliable the calibration model.

EXPERIMENT CONFIGURATION

When using an NIR analyzer, instrument characteristics such as sensitivity, resolution, and signal-to-noise ratio
parameters need to be evaluated. The quality of these values is a function of the light source stability, optics
throughput, dispersion/filter element accuracy, and detector sensitivity in the instrument.

The choice of accessories is application dependent. For liquid samples, transmission and transflectance modes are
commonly employed using fiber optic probes or cuvettes. The path length is sample dependent, usually ranging
from 0.1 to 1 cm. The advantage of using a fiber-optic probe is that sample preparation is significantly reduced
and noninvasive or nondestructive measurements become possible. For solid samples, diffuse-reflectance spectra
collected by a reflectance probe will provide information for analytes. Diffuse reflectance should be measured
without interference from specular reflectance. The setup configuration, such as the angle of incident light and the
distance of light illumination / collection ought to be consistent through all the measurements, including those
taken in developing the calibration set and for predicting the future unknowns.

For solid samples, the sample should be rotated and measurements done on different spots of the sample to
average out the sampling error. A group of spectra may be averaged to increase the signal-to-noise ratio. Random
noise is reduced by the factor, square root of the number of spectra averaged. For ASD’s portable NIR
 ASD Technical Guide 3rd Ed. Section 21-2

spectrometer, it takes 0.1 seconds to acquire one spectrum. Therefore, a 10 second measurement reduces the
random noise by a factor of 10.

DATA ANALYSIS

NIR spectroscopy is an extremely rapid method of measurement, capable of performing an analysis within a
minute. The time-consuming part of NIR work is the data analysis phase, where chemists try to find the
correlation between near-infrared spectral characteristics and the property, or properties, of interest as measured
by more traditional methods. There are several commercially available software packages for doing this task. Data
analysis involves the following steps.

Data preprocessing:
When the spectral data plots are presented, first determine if there is any baseline drift or slope in the spectra,
which often occurs in diffuse-reflectance measurements. Baseline subtraction, first derivative and second
derivative transformations may be performed to reduce these effects if necessary. There is a trade-off though, each
successive degree of derivative that is taken introduces more noise into the spectral data.

Outlier detection:
An outlier is a data point that falls well outside of the main population. Outliers result from lab measurement
errors, samples from different categories, and instrument error. It is important to check for and remove outliers in
both the training set and the set of unknowns that the calibration is to be tested on (see "validation" and
"prediction").

Building a good calibration model:

This is one of the most important steps in NIR analysis. Developing a calibration model involves calculating the
regression equation based on the NIR spectra and the known analyte information. The model is then used to
predict the future unknowns. Multiple Linear Regression (MLR), Principal Component Regression (PCR) and
Partial Least Squares (PLS) are commonly used linear calibration methods, along with Locally Weighted
Regression (LWR) for nonlinear models. In developing a calibration model, several parameters are evaluated:
factors, loadings, and scores. When choosing the number of factors, one should try to avoid under-fitting, i.e. too
few factors, and over-fitting, i.e. too many factors. If an insufficient number of factors are chosen, the prediction
is not reliable because useful information has been left out. If too many factors are chosen, however, more
uncertainty is included in the calibration set which will result in errors in prediction. Scores are used to check the
sample homogeneity and possible clusters, while loadings are used to interpret how the variables are weighted in
principal component space.

Validation:
The validity of the model must be tested. Usually the way this is done is to split the whole sample set into two
sets; one set for calibration and the other for validation. If there are not enough samples, “leave-one-out” cross
validation can be performed. This means leaving one sample out, using the rest of the samples to build a
calibration model and then using the model to predict the one that was left out. The advantage of doing cross
validation is that, unlike calibration with a full data set, the sample being predicted is not included in the
calibration model. Thus, the model can be tested independently.

Prediction:
Finally, the calibration can be used to predict future unknowns, assuming that the unknowns are in the same
sample population as those used in the calibration set. Whether the unknown is an outlier needs to be tested.
SUMMARY

Applying an NIR analyzer to a particular application requires the development of a reliable calibration model. The
most important steps involve a thorough consideration of experimental design and multivariate calibration. Once
this is established, one can enjoy the advantages of the NIR analysis. The speed of the analysis will save time and
avoid many mistakes instantaneously. The speed advantage is so valuable to engineers involved with on-line
process monitoring that instruments are routinely installed in the process line with feedback loops. With a
 ASD Technical Guide 3rd Ed. Section 21-3

portable NIR analyzer such as the LabSpec®® Pro spectrometer, samples can be noninvasively analyzed on-the-
spot, dramatically reducing costly and time consuming laboratory analysis. The low absorptivity in the NIR
region allows measurements to be taken on raw materials without elaborate sample preparation. In the food,
agriculture, pharmaceutical, polymer, cosmetics, environmental, textile, and medical fields, portable NIR analysis
finds a wide range of applications. There are still many unknown applications waiting to be discovered and with
the maturity of this technique, more and more people will use portable NIR analyzers for convenience and
flexibility.

REFERENCES

Burns D. A. and E. W. Ciurczak (Eds.), Handbook of Near-Infrared Analysis, (Volume 13 in Practical

Spectroscopy Series), Marcel Dekker, Inc., New York, 1992

Hildrum K. I., T. Isaksson, T. Naes and A. Tandberg (Eds.), Near Infra-red Spectroscopy, (Ellis Horwood Series
in Analytical Chemistry), Ellis Horwood, Ltd., England, 1992

Murray I. and I. A. Cowe (Eds.), Making Light Work: Advances in Near Infrared Spectroscopy, 4th International
Conference on Near Infrared Spectroscopy, Aberdeen, Scotland, August 19-23, 1991, Weinheim, New York,
Basel, Cambridge, VCH, 1992
ASD Technical Guide 3rd Ed. Section 21-4
 ASD Technical Guide 3rd Ed. Section 22-1

Section 22
Derivation of Bouguer-Lambert-Beer (BLB) Law
Derivation of Bouguer-Lambert-Beer (BLB) Law
David Hatchell, ASD

Bouguer-Lambert-Beer Law (BLB): The most useful concept in spectrophotometry is the Bouguer-Lambert-Beer Law
(BLB). When reflection, fluorescence, and Raman scattering are negligible BLB works very well to determine the
concentration of absorbing species in dilute solutions. Another important condition is that the source be monochromatic
(same wavelength) or all measurements must be made as a function of wavelength4. Omitting the wavelength nomenclature
for clarity, BLB is derived as follows:

Let N equal the total number of photons for one second striking a sample of thickness x and a defined surface area of one
square centimeter. For the infinitesimal thickness element dx there is an associated decrease –dN in the number of photons
allowed to pass through to the next infinitesimal thickness element. This decrease is caused by absorption and it must be
proportional to the total number N for thickness x . Therefore:

dN N
- dx ∝ x

Now since x is partially determined by the number of molecules m in the sample, x must be proportional to the total number
1 1
of molecules. That is x ∝ m . Letting the constant of proportionality = k gives, x = kcm Letting k also absorb the constant
dN N
of proportionality for - dx ∝ x gives:

dN
- dx = kmN

Since the intensity I of a flux component is in part determined by the number of photons, I must be proportional to N and
dI must be proportional to dN . Therefore, Letting k also absorb this constant of proportionality gives:

dI
- dx = km I

This equation can then be rewritten in integral form as follows:

Iι
ι
- ⌠ I = km ⌠
dI
⌡ ⌡ dx
Io 0
where ι is the total length of the path length traversed by the transmitted flux and Io is the intensity of the source flux, and
Iι is the intensity of the transmitted flux after passing through the matter.

If the incident flux Fo is collimated, that is, all incident flux lines are parallel, and if it is normal to the sample surface, the
path length is simply a straight line, giving:

Iι Iο
- ln I = kmι , which can be rewritten as: ln I = kmι
o ι

km
Letting K = ln10 , gives (see note5):

4Gordon Marc Loudon, "The Necessity of Using Monochromatic Radiation in Spectrometry", Journal of Chemical
Education, v. 41 (July 1964), p. 391.
1
ln10 converts "ln" to "log10" hereafter "log"
5 "natural log" = log = ln, the factor
e
 ASD Technical Guide 3rd Ed. Section 22-2

Iο
log I = Kι which is the Bouguer-Lambert part of BLB.
ι

Iο
Absorbance (Abs): The quantity, log I is known as the 'absorbance'. Experiment also shows that absorbance is
ι
proportional to the concentration c of the absorbing species. Letting the proportionality constant be α gives the experimental
result of:

Abs = αc which is Beer's contribution to the Bouguer-Lambert-Beer law.

Beer's and Bouguer-Lambert can be combined via the total differential to form the combined BLB law (see Lohman for the
combined derivation)6:

Abs = αλιc

where αλ is called the molar absorption coefficient and has units of liters per mole per centimeter. The λ subscript indicates
the wavelength dependence of molar absorption coefficient.

When referring to the sphectrophotometric Bouguer-Lambert type measurement the units are usually indicated by AU for
absorbance units (eg. absorbance = 6 AU). Abs is used when referring to Beer's expression.

6FredH. Lohman, "The Mathematical Combination of Lambert's Lw and Beer's Law", Journal of Chemical
Education, v. 32 (March 1955), p. 155.
 ASD Technical Guide 3rd Ed. Section 23-1

Section 23

Approximating Spot Size

FO
V

α
β
h

y
w z

arctan(y/x) = α = (FOV full angle)/2

y/x = tan α
y = x tan α

z + w = h tan (β + α)
w = [h tan (β + α)] - z
and
w = h tan β
Therefore:
z = h [tan(β)+ tan(β + α)]

Also w = square root of (x2 + h2 - 2xh cosβ)

EXAMPLE:
1 degree FOV Tube at x = 2 meters from perpendicular target (small ellipse):
y = (2 meters) tan (0.5 deg) = 0.0175 meters = 1.75 cm
So, for a perpendicular target the spot has a diameter of 3.5 cm

EXAMPLE:
Suppose we were limited to a 12.7 x 12.7 cm (.127 x .127 m) oblique target. So, to be on the safe side we will want z =
0.635m. For h = 1 meter, β = 45 deg, 1 degree foreoptic:

w = [h tan (β + α)] - z = [tan(45.5 deg)] - 0.635]meters = 0.38 m

Also, in this case x2 = h2 + w2 = 1.144

ASD Technical Guide 3rd Ed. Section 23-2
 ASD Technical Guide 3rd Ed. Section 24-1

Section 24
Foliar Spectral Features
Paul J. Curran, (1989), Remote Sensing of Foliar Chemistry, ‘Remote Sensing of Environment’ 30:271-278
(1989)

Wavelength(nm)__Mechanism__Chemical(s)
0430_Electron transition_Chlorophyll a
0460_Electron transition_Chlorophyll b
0640_Electron transition_Chlorophyll b
0660_Electron transition_Chlorophyll a
0910_C-H stretch, 3rd overtone_Protein
0930_C-H stretch, 3rd overtone_Oil
0970_O-H bend, 1st overtone_Water, starch
0990_O-H stretch, 2nd overtone_Starch
1020_N-H stretch_Protein
1040_C-H stretch, C-H deformation_Oil
1120_C-H stretch, 2nd overtone_Lignin
1200_O-H bend, 1st overtone_Water, cellulose, starch, lignin
1400_O-H bend, 1st overtone_Water
1420_C-H stretch, C-H deformation_Lignin
1450_O-H stretch, 1st overtone,C-H stretch, C-H deformation_Starch, sugar, lignin, water
1490_O-H stretch, 1st overtone_Cellulose, sugar
1510_N-H stretch, 1st overtone_Protein, nitrogen
1530_O-H stretch, 1st overtone_Starch
1540_O-H stretch, 1st overtone_Starch, cellulose
1580_O-H stretch, 1st overtone_Starch, sugar
1690_C-H stretch, 1st overtone_Lignin, starch, protein, nitogen
1780_C-H stretch, 1st overtone/O-H stretch/H-O-H deformation_Cellulose, sugar, starch
1820_O-H stretch/C-O stretch, 2nd overtone_Cellulose
1900_O-H stretch, C-O stretch_Starch
1940_O-H stretch, O-H deformation_Water, lignin, protein, nitrogen, starch, cellulose
1960_O-H stretch/O-H bend_Sugar, starch
1980_N-H asymmetry_Protein
2000_O-H deformation, C-O deformation_Starch
2060_N=H bend, 2nd overtone/N=H bend/N-H stretch_Protein, Nitrogen
2080_O-H stretch/O-H deformation_Sugar, starch
2100_O=H bend/C-O stretch/C-O-C stretch, 3rd overtone_Starch, cellulose
2130_N-H stretch, Protein
2180_N-H bend, 2nd overtone/C-H stretch/C-O stretch/C=O stretch/ C-N stretch_Protein, nitrogen
2240_C-H stretch, Protein
2250_O-H stretch, O-H deformation_Starch
2270_C-H stretch/O-H stretch, CH2 bend/CH2 stretch_Cellulose, sugar, starch
2280_C-H stretch/CH2 deformation_Starch, cellulose
2300_N-H stretch, C=O stretch, C-H bend, 2nd overtone_Protein, nitrogen
2310_C-H bend, 2nd overtone_Oil
2320_C-H stretch/CH2 deformation_Starch
2340_C-H stretch/O-H deformation/C-H deformation/O-H stretch_Cellulose
2350_CH2 bend, 2nd overtone, C-H deformation, 2nd overtone_Cellulose, protein, nitrogen

Near infrared absorption of cellulose

(CU, Boulder, CSES Course on Hyperspectral Data Analysis, 1997, page 87)
Wavelength(nm)__Mechanism
1480_OH stretching overtone: cellulose and adsorbed water
1930_Adsorbed water
2100_OH and CH deformation modes+OH stretching
2280_CO stretching+OH stretching CH2 bending+CH2 stretching
2340_OH and CH deformation modes + OH stretching
2480_CO stretching+CH and CH2 stretching
 ASD Technical Guide 3rd Ed. Section 24-2

Vegetation Index Formulas

(Chistopher D. Elvidge and Zhikang Chen, Comparison of Broad-band and Narrow-band Red Versus Near
Infrared Vegetation Indices, Desert Research Institute, June 29, 1994)

Ratio Vegetation Index (RVI) = NIR/Red

Normalized Difference Vegetation Index (NDVI) = (NIR - red) / (NIR + Red)
Soil Adjusted Vegetation Index (SAVI) = (1 + L)(NIR - red) / (NIR + red + L)
Perpendicular Vegetation Index (PVI) = -cos α (red) + sin α (NIR), where α = angle between the soil line and the
NIR axis.
Difference Vegetation Index (DVI) = NIR - red
 ASD Technical Guide 3rd Ed. Section 24-3

Water Spectral Features

(CU, Boulder, CSES Course on Hyperspectral Data Analysis, 1997, pages 35 & 36)

Water Vapor Absorption Fundamentals:

ν1 = 3657.05 cm-1 = 2.734 µm symmetric stretch

ν2 = 1594.75 cm-1 = 6.271 µm bend
ν3 = 3755.93 cm-1 = 2.662 µm asymmetric stretch
Important Water Vapor Absorptions:

ν2 + ν3 = 1.865 µm
ν1 + ν3 = 1.379 µm
ν1 + ν2 + ν3 = 1.135 µm
2ν1 + ν3 = 0.942 µm

Liquid Water Absorption Fundamentals:

ν1 = 3219.57 cm-1 = 3.106 µm

ν2 = 1644.74 cm-1 = 6.08 µm
ν3 = 3444.71 cm-1 = 2.903 µm
ASD Technical Guide 3rd Ed. Section 24-4
 ASD Technical Guide 3rd Ed. Section 25-1

Section 25
Glossary
GLOSSARY
Absorbance (Abs): (see Bouguer-Lambert-Beer (BLB) law and see Absorbance Units)

Absorbance Units: When referring to Absorbance measurements the units are usually indicated by AU for
absorbance units (eg. Absorbance = 6 AU).

Absorbance = log(1/Transmittanceλ)
= -log(Transmittanceλ)
Example:
Transmittance = 10–6
⇒ Absorbance = –log (10–6)
⇒ –Absorbance = log (10–6)
⇒ 10–Absorbance = 10–6
⇒ Absorbance = 6 AU

Absorbed Energy: A portion of incident illumination on a sample is converted to charge carrier energy or into
mechcanical energy. Both of these conversions involve the pushing and pulling, i.e., accelerating and
decelerating of net charges by the oscillating electric and magnetic fields of the incident illumination. These
conversions can involve individual charge carriers such as electrons, or the dipole moment of entire molecules.
Sometimes, if the incident illumination frequency matches the resonant frequency of the charge carrier or
molecule, these conversions are one-way, in the process known as 'Absorption'. For example, electrons excited
into the conduction band to propagate electric current, or the mechanical vibration of a molecule being converted
to internal heat. Sometimes, these conversions are two-way. For example, electrons are accelerated and
decelerated, which results in a nearly instantaneous release or 'emission' of energy, e.g., conversion back to light
energy. Some of these two-way conversion energies continue to propagate or 'scatter' within the sample to be
absorbed again, some simply continue through the sample and exit out the other side, and some simply return
back out of the sample.

Absorption: (see Absorbed Energy).

Absorptance (Absorptivity, Absorption Coefficient, Absorption Power): (also see Conservation of Energy)
Absorptance is the ratio of the radiant energy absorbed by a sample to the radiant energy incident on the sample.

Blackbody (Planckian Radiator, Full Radiator): A theoretical model of a perfect radiator and absorber of
radiant energy. Radiant energy of a blackbody is computed using 'Planck's Formula'.

Bouguer-Lambert-Beer Law (BLB Law, 'Beer's Law): (also see Absorbance Units) For dilute analyte solutions,
collimated incident illumination, negligible luminescence, and negligible scattering:

log(1/Transmittanceλ) = αλlc

where αλ is the molar absorption coefficient, l is the path length, and c is the analyte concentration. This equation
is called the BLB law and the quantity log(1/Transmittanceλ) is called 'absorbance'. Absorbance is a unitless
quantity, however, the term absorbance units (AU) is often used to indicate this type of measurement. BLB is
valid only for transmittance measurements and much has been written on the mathematics and physics of this
law. There is no rigorous derivation of a similar law that relates reflectance to analyte concentration (see
Log(1/Reflectance)). Absorbance cannot be measured directly since it there's no way to directly count the number
of photons as they disappear one-by-one. So, what is actually measured is transmittance.

Calibration Model: Also known as the 'Prediction Model'. A polynomial or set of polynomials that relate spectral
measurements to physical or chemical properties of a sample. Such a model is typically developed using Multiple
Linear Regression Methods. This model development is typically referred to as 'Chemometrics'.

Calibration for 'Real-Time' Reflectance: Real-time calulation of ratios such as reflectance require a diffuse
 ASD Technical Guide 3rd Ed. Section 25-2

measurement of incident illumination, which is also known as the 'white reference' or 'baseline'. Typically
reflectance is used for solids, although it can be used with liquids under the right viewing geometry. The white
reference can be measured with the instrument viewing the incident illumination, that passes through a nearly
transparent lambertian difussor, or is collected by an integrating sphere, or that is reflected off of a nearly 100%
reflective lambertian panel. The most convenient and most common approach is the reflective panel. Aim the fiber
optic probe the panel then click on the WR (white reference) button in the display. During this baseline
measurement, the ASD software is used to adjust instrument gain for optimal performance, and dark-offsets are
also determined and stored for automatic subtraction from reflectance calculations. A good white reference
material is Spectralon® which is a registered trademark of Labsphere, Inc.

Calibration for 'Real-Time' Trasmittance: Real-time calculation of ratios such as transmittance require a
measurement of incident illumination, which is also known as the 'baseline'. Typically transmittance is used for
liquids, although it can be useful with clear or translucent solids as well. For liquids, either an immersion probe or
a cuvette holder is used. With the immersion probe in air, or dipped into a non-absorbing liquid, click on the
baseline button on the computer screen. The baseline measurement is stored for automatic real-time calculation
of target transmittance measurements. The procedure is the same except for the cuvette in the holder either
empty or full of a non-absorbing liquid. During this baseline measurement, the ASD software is used to adjust
instrument gain for optimal performance, and dark-offsets are also determined and stored for automatic
subtraction from transmittance calculations.

Calibration of Wavelength: Wavelength calibrations are performed in our ASD factory laboratory with NIST
Traceable sources and a triple monochromator. Under typical use these calibrations should remain stable for
approximately one year.

Calibration for Radiance and Irradiance Measurements: The FieldSpec® spectroradiometer software comes
complete with the necessary structure and files for creating spectral radiance and irradiance calibrations. These
calibrations correlate the instrument response to the radiant energy of a NIST-traceable standards. A calibration
file must be generated for the bare fiberoptic/foreoptic configuration to be used. Irradiance measurements require
a Remote Cosine Collector (RCR) foreoptic. There are two options for generating the calibrations:

1. ASD can generate the calibrations here at our factory laboratory in Boulder, Colorado, USA (Product & Price
Lists). Under typical use, these calibrations should remain stable for about a year (new calibrations are covered
under warranty if the unit is returned within the 1 year warranty period). The radiometric calibration accuracy of
the FieldSpec® spectroradiometer is ±5% from 400 to 900 nm (UV/VNIR and OpenSkies units), rising to ±8% at
2200 nm over a 10 to 30°C temperature range (NIR and FR units). These values have been determined by
viewing a stable, NIST traceable, radiance source with the FieldSpec® spectroradiometer and from the
radiometric accuracy of the calibration source quoted by its manufacturer, Optronics Laboratory.

2. If the user has the appropriate calibration facilities, they can install their own calibration files.

Charge-Coupled Device (CCD): (also see Photodiode Array) A CCD is a two dimensional silicon array of
detectors manufactured into a single chip and gated for sequential transfer and reading of pixels. A Photodiode
Array (PDA) is a one demensional silicon array of detectors manufactued into a single chip. For a CCD, a whole
column of pixels at a specific wavelength are co-added or "binned" in a summing register and read as a single
pixel. The sensitivity of a CCD is superior to a photo diode array (PDA), but its dynamic range is lower. This type
of detector is ideally suited for low light level applications. These include: in situ measurements of apparent optical
properties of lake and sea water, fluorescence, and bioluminescence.

In general, the S/N of a CCD is superior to a PDA only at VERY LOW LIGHT LEVELS. However, for normal to
high light levels the S/N of a CCD is not as good as that of a PDA. The CCD is best used for things like
fluorescence, bioluminescence, and solar energy measurements under DEEP, dark natural waters. The PDA is
best used for surface measurements under solar illumination and solar energy measurements under shallow,
clear natural waters.

While the S/N of a PDA is not as good as a CCD at VERY low light levels, the PDA CAN measure very low light
levels by setting a long integration time. So, if your mission is to purchase an instrument for a variety of
applications, the PDA is probably your best bet. The CCD should be considered only if you wish to specialize in
VERY low light level measurements.
 ASD Technical Guide 3rd Ed. Section 25-3

Chromophore: Some materials absorb energy by electron transition and are called 'chromophores.'
Chromophores typically involve covalently bonded unsaturated groups such as, phenyl.

Collimated light: Light for which all flux lines are parallel.

Conservation of energy: By the law of Conservation of Energy, the total incident energy Ei interacting with
matter is:

Ei = EReflection + ETransmission + EAbsorption + ERamon+ELuminescence

Dividing through by Ei gives:

100% = (EReflection / Ei) + (ETransmission / Ei) + (EAbsorption / Ei) + (ERamon/ Ei) + (ELuminescence / Ei)

Some of the quotients above are known as follows:

EReflection / Ei= Reflectance, ETransmission / Ei = Transmittance, EAbsorption / Ei = Absorptance

Cosine Collector: (see Measurements and see Remote Cosine Receptor)

Data Acquisition: (also see Measurements) Data are stored on the standard built-in sub notebook computer
internal hard drive and displayed on the sub notebook computer screen. The ASD operating / data collection
software uses 1 MB of hard disk space.

A single 350 - 2500 nm spectrum including typical headers requires 2.5 MB of hard disk space.

A single 350 - 1050 nm spectrum including typical headers requires 2.5 KB of hard disk space.

Derivative Spectrum: A spectrum that is the result of applying a derivative transform to the data of the original
spectrum. Derivatives of spectra are very useful for two reasons:

1. First, and second derivatives may swing with greater amplitude than the primary spectra. For example, a
spectrum suddenly changes from a positive slope to a negative slope, such as at the peak of a narrow feature
(see the figure below). The more distinguishable derivatives are especially useful for separating out peaks of
overlapping bands.

2. In some cases derivative spectra can be a good noise filter since changes in base line have negligible effect on
derivatives. For example, scattering increases with wavelength for some biologically active macromolecules
causing an increasing slope of the absorbance baseline.

A commonly used approximation of the first derivative is: dα/dλ = [α (λ + ∆λ) - α (λ - ∆λ)] / 2∆λ.

A more accurate approximation of the first and higher order derivatives is presented in thorough explanations by
Whitaker7 and Morrey8. Still other methods involve a best fit match to the curve on the features of interest and
performing higher order derivatives with numerical analysis.

7Stephen Whitaker and R. L. Pigford, "Numerical Differentiation of Experimental Data", Industrial and Engineering
Chemistry, vol. 52, no. 2 February 1960, pp.185 - 187.
8J. R. Morrey, "On Determining Spectral Peak Positions from Composit Spectra with a Digital Computer",
Analytical Chemistry, vol. 40, no. 6, May 1968, pp. 905 - 914.
 ASD Technical Guide 3rd Ed. Section 25-4

Derivative spectra yield good signal-to-noise ratios only if the difference of noise levels at the endpoints of the
interval is small enough to yield a noise equivalent ∆dα/dλ calculation much smaller than the absorbance.

Sudden change in the curve

Absorbance spectrum

First derivative

Second derivative

Difference Spectrum: A spectrum that is the result of subtracting all the signal channels of one spectrum from
another.

Digital numbers (DN): Values between 0 and 65,535 except CCD units which use values between 0 and 4096.

Digitization: All ASD spectrometers except CCD units uses 16 bit digitization. CCD units use 12 bit.

Emissivity (Emittance): The ratio of non-Blackbody radiance to Blackbody radiance.

Field-of-view (FOV): ASD offers optional foreoptics for all models of the FieldSpec® spectrometer that attach
to the Pistol Grip and reduce the FOV to as little as 1° (see Product and Price list). The bare fiber optic cable has
a field-of-view angle of 25°. These Foreoptics are also available with pistol grip telescopic sights (see Product and
Price list).

Fluorescence: (see Luminescence)

Flux (F): (also see Intensity) Radiant flux, is the amount of light energy that crosses a defined surface in a
period of time. The units are typically Joules per Second (note: 1 Watt = 1 Joule per Second)). Also known as
'radiant power', Flux is pictorially indicated by arrows, i.e., 'vectors' showing the flux magnitude and direction of
travel.

Immersion probes: ASD offers lightweight and easy to use fiber optic probes for transmittance measurements.
These probes feature SMA connectors that interface directly with ASD LabSpec® Pro.

Intensity: Intensity is flux density. That is, for an infinitesimal solid angle element dΩ, and infinitesimal flux
element dF, the intensity is dF/ dΩ.
 ASD Technical Guide 3rd Ed. Section 25-5

Isotropic Medium: Material in which light energy may impinge upon or travel through without any energy
changes due to position or direction.

Log(1/Reflectance): A commonly used math pretreatment, useful for linearizing reflectance data. This expression
is often abbreviated as log(1/R). In most cases it is possible to find a linear correlation of log(1/R) data to
concentration of an analyte in the target matrix. However, a general derivation relating reflectance to
concentration cannot be rigorously derived, such as, the Bouguer-Lambert-Beer law for transmittance.

Luminescence: Sometimes, a photon incident on a sample is absorbed, causing an electron to jump from its
-8
ground state to a higher, excited, energy state. Around 10 seconds later, the electron returns to the ground state
resulting in an emission of a photon of a different wavelength than the incident photon. This return to the ground
state also results in an excited vibrational state which reaches equilibrium in about
-12
10 seconds. Altogether, these events are called luminescence. There are two types of luminescence that are
distinguished by their ground and excited states: fluorescence and phosphorescence. The incident photons
associated with luminescence are called the "excitation source". Luminescence and inelastic scattering are
different and separate events even though it is difficult to differentiate the resulting photons. Except for highly
viscous liquids, fluorescence of solutions is always unpolarized even if the excitation source is highly polarized.
The wavelength of the fluorescing photons is always longer than the excitation source. The ratio of the energies of
fluorescing photons to excitation source is typically less than 0.01.

Maintenance: All ASD manufactured products come with a full 1 year limited warranty. Optional service contracts
that take effect at warranty expiration are also available at additional cost (see Product & Price lists).

Monochromator: Optical device for isolating a very narrow wavelength region of a radiant energy source. The
isolation element may be any or combinations of the following: filters, gratings, prisms, and slits.

Noise-equivalent-Delta-Radiance (Ne∆L or NeDL or NER): Defined as the radiance at a signal-to-noise ratio

of unity. These values have been determined by viewing a stable, NIST traceable, radiance source with the
FieldSpec® spectroradiometer, and from the radiometric accuracy of the calibration source quoted by its
manufacturer, Optronics Laboratory. NOTE: Standard built-in spectrum averaging function allows up to 31,800
spectra to be averaged thereby reducing the associated NeDL by a factor of the square root of the number of
scans averaged. (also see Signal-to-noise ratio).

Offset: Systematic noise including 'Dark-Current'.

Photon: There are two models for radiant energy, the particle model and the wave model. In the particle model,
radiant energy is comprised of discrete packets called photons. A photon's electromagnetic energy is defined as:

Ephoton = hf, where h is Planck's constant and f is the frequency of the electromagnetic energy

The energy amplitude measured at a specific wavelength of frequency f is determined by the number of photons
at that wavelength, i.e., amplitude for the wave model is equivalent to numbers of photons for the particle model.

Photodiode Array (PDA): (also see Charge Coupled Device) A PDA is a line of contiguous detectors
manufactued into a single chip.

Post-dispersive Spectrophotometers: All ASD spectrometers are “post-dispersive”. Because portable

spectrometers are typically used outside the controlled laboratory environment, they are exposed to much higher
levels of ambient light. In almost all cases, some of this ambient light will stray into the sample being measured.
The errors produced by this ambient stray light are much greater for a pre-dispersive spectrometer than they are
for a spectrometer that is post-dispersive.

In a pre-dispersive spectrometer, the sample is illuminated with monochromatic light. Light scattered off or
transmitted through the sample is then collected and delivered to the instrument’s detector. Ambient light that
strays into the sample being measured is also collected. Thus, both the monochromatic illumination from the
instrument and all wavelengths of the ambient stray light are delivered to the detector. Because the stray ambient
light signal can represent a large fraction of the total light signal measured by the detector, it is a major source of
error. While this source or error can be minimized by completely shielding the sample form all sources of ambient
 ASD Technical Guide 3rd Ed. Section 25-6

illumination, this often precludes the use of most reflectance and transmittance fiber optic probes.

In a post-dispersive spectrometer, the sample is illuminated with white light. Light scattered off or transmitted
through the sample is then dispersed and delivered to the instrument’s detector. As with the pre-dispersive
spectrometer, ambient light that strays into the sample being measured is also collected. The difference is that in
the post-dispersive instrument, only ambient stray light of the same wavelength as that being measured by the
detector is added to the signal resulting from the instrument’s illumination of the sample. Thus, the stray ambient
light signal represents a small fraction of the total light signal measured by the detector.

Power Ratings:
Instrument Voltage (volts DC) Power (watts)
FieldSpec®UV/VNIR 12 18
FieldSpec®FR 12 20
LabSpec® Pro 12 20

Radiance: (see Calibrations for Radiance Measurements)

Radiometer: (also see Spectroradiometer) A radiometer is instrument that produces an electrical signal, which
corresponds to the total integrated radiant flux falling on its detector for a given spectral region.

Raman Scattering: One form of inelastic scattering. If the incident illumination is strongly polarized, the resulting
Raman energy will also be strongly polarized. The sum of the energies of Raman photons is usually thousands of
times less than the sum of the energies of fluorescing photons (also see scattering and also see Luminescence).

Reflectance: (also see Conservation of Energy) Reflectance is the ratio of the radiant energy reflected from a
surface to the radiant energy incident on the surface.

Reflectance Standards: Lambertian diffuse reflectance panels for which the reflectance is nearly the same at all
wavelengths. These panels are available in a variety of reflectance levels and sizes either calibrated or standard.
Spectralon® is a registered trademark of Labsphere, Inc.

Reflection: The process by which incident illumination reacts with the sample and is converted to radiant energy

θi θr
Incident Specular reflection
illumination

Diffuse reflection
Sample

that subsequently travels back away from the sample surface (also see Absorbed Energy). All real reflection
involves varying degrees of specular reflection (θr = θi) and diffuse reflection (θr ≠ θi).

Remote Cosine Receptor (RCR): (also see Calibrations for Radiance and Irradiance) Attaches to the
FieldSpec® spectroradiometer fiber optic cable for measuring full-sky irradiance. Resolution: (see Spectral
sampling interval and resolution)

Sampling interval: (see Spectral sampling interval and resolution)

Scattering: Sometimes a photon incident on a sample is absorbed by an electron in the sample (see Absorbed
Energy). Shortly thereafter, this electron releases a photon. This electron-photon process is called scattering. (see
also Raman scattering).

Signal-to-noise ratio (S/N or SNR): Ratio of the radiance measured to the root-mean-square (RMS) noise
created by the detector and associated electronics. Parameters such as reflectance, sun angle and others
affecting the signal must be defined to make the value meaningful (also see Noise-equivalent-radiance). Standard
 ASD Technical Guide 3rd Ed. Section 25-7

built-in spectrum averaging function allows up to 31,800 spectra to be averaged thereby reducing the noise and
increasing signal-to-noise.

SMA Adapter: FR to SMA adapter for attaching SMA accessories or jumper cables to the FieldSpec® FR.

Spectral Resolution and Sampling interval: Spectral resolution is defined as the full-width-half-maximum
(FWHM) of the instrument response to a monochromatic source. Spectral sampling Interval is the spacing
between sample points in the spectrum. Sampling is independent of resolution and in ASD spectrometers is
between 2 and 5 times per FWHM. These values are based on wavelength calibrations performed by calculating
the peak position when viewing the output of a monochromator with the FieldSpec® spectrometer.

Sampling Interval
Spectrometer offset corrected raw DN

Spectral
Resolution
50% of peak height

Wavelength

Spectrograph: Originally, the word 'spectrograph' meant a device that would disperse light into a spectrum and
record that spectrum onto photographic film. Early on, such devices were mostly used for identifying absorption
lines, in solar spectra and the spectra of burning gasses. Recently the word has also been used to describe
devices that use photo detectors instead of film.

Spectrometer (optical): From the Latin word 'specere' meaning image, and the Greek word 'metron' meaning to
measure. A spectrometer is a spectroscope that uses some sort of mechanical or electrical detection device
instead of the human eye. The words, 'spectrometer' and 'spectroscope' are often used interchangeably.

Spectrophotometer: From the word 'spectrometer' and the Greek 'phot' meaning light. A spectrophotometer is
an optical spectrometer that is designed for computation of the ratio of the radiant energy of two different light
beams as a function of wavelength. One of the beams is typically called the ‘baseline’, 'reference', or ‘white
reference’ and is often the equivalent of the ‘incident’ radiation. The output is typically ‘transmittance’,
‘reflectance’, or transformations of either.

Spectroradiometer: (also see Radiometer) A spectroradiometer is an optical spectrometer that produces

electrical signals, which correspond to radiant flux energy falling on its detectors for a series of discrete
wavelength intervals. Spectroradiometers use single photoelectric detectors over which a spectrum is scanned
and/or an array of detectors aligned in the spectrum path. Photoelectric detector current responds linearly to
radiant flux energy while voltage responds logarithmically. Early spectroradiometers produced analogue electrical
signals, while more recent spectroradiometers typically have built-in electronics for converting the analogue signal
to digital. The signal output of a spectroradiometer may be calibrated to NIST traceable standards to produce
measurements with Radiance or Irradiance units (ASD radiance and irradiance calibration files cost extra).
 ASD Technical Guide 3rd Ed. Section 26-8

Spectroscope: From the Latin word 'specere' meaning image, and the Greek word 'scopos' meaning to observe.
A spectroscope was an early instrument for visually studying dispersed light. These devices included either a
prism or diffraction grating for dispersing the light, telescopes for viewing narrow regions of the dispersed light,
and protractors or other calibrated scales for measuring the position of a narrow region.

Spectroscopy: Originally, the word spectroscopy meant the study of spectra using a spectroscope. Recently,
the word has also been used to describe any study of spectra.

Spectrum (optical): A spectrum is an image of radiant energy dispersed into its wavelength constituents or a
two dimensional plot of radiant energy or radiant energy ratio versus wavelength.

Spectrum Data Size: (see Data Acquisition)

Transmission: A portion of the incident illumination energy passes directly through the sample without any
interactions with electrons or molecules in the sample. Another portion of energy is absorbed and re-emmited
over and over again in the same direction as the incident illumination. Transmission is the process of these
events. In either case, the photons are "transmitted" through the sample.

Transmittance Reference Standard: (see Calibrations for Real-Time Transmittance) ASD recommends air or
carbon tetrachloride as a transmittance reference.

Transmittance: (also see Conservation of Energy) Transmittance is the ratio of the radiant energy transmitted
through a sample to the radiant energy incident on the surface of the sample.

Wavelength: The distance, λ traversed by one period of an electromagnetic wave. λ = c/f, where c is the
velocity of light and f is the frequency.

λ
Electric field
amplitude

Distance

Magnetic field
amplitude

ction 26