Geospatial Applications in

F
A
R
M
I
N
G
Precision
Farming

with

TNTmips
TNTedit
TNTview, TNTatlas
Precision Farming

Before Getting Started

TNTmips provides a variety of tools to assist you with the spatial asset allocation
and management tasks referred to as precision farming. An integrated system that
handles both raster and vector objects is a necessity for these types of tasks. The
purpose of this booklet is to acquaint you with the concepts and tools used in
precision farming applications. All of the features required for a robust precision
farming system are available in TNTmips, which is used to demonstrate the
concepts presented.
Prerequisite Skills This booklet assumes you have completed the exercises in
Getting Started: Displaying Geospatial Data and Getting Started: Navigating.
Those exercises introduce essential skills and basic techniques that are not
covered again here. Please consult these booklets for any review you need.
Sample Data This booklet does not use exercises with specific sample data to
develop the topics presented. You can, however, use the sample data distributed
with the TNT products to explore the ideas discussed on these pages. If you do
not have access to a TNT products CD, you can download data from MicroImages
web site. Make a read-write copy on your hard drive of data sets you want to use
so changes can be saved.
TNTmips and TNTlite TNTmips comes in two versions: the professional version
and the free TNTlite version. This booklet refers to both versions as TNTmips.
If you did not purchase the professional version (which requires a software license
key), TNTmips operates in TNTlite mode, which limits the size of your project
materials. All of the processes described in this booklet, with the exception of
creating your own atlases for distribution on CD-ROM with TNTatlas, can be
performed in TNTlite.

Merri P. Skrdla, Ph.D., 2 January 2004

MicroImages 20012004

It may be difficult to identify the important points in some illustrations without

a color copy of this booklet. You can print or read this booklet in color from
MicroImages web site. The web site is also your source of the newest Getting
Started booklets on other topics. You can download an installation guide,
sample data, and the latest version of TNTlite.
http://www.microimages.com

page 2
 Precision Farming

Introducing Precision Farming

Precision farming involves the use of spatial asset
allocation and management to distribute available
time and money where it is most needed and will
provide the best return. Data useful for precision
farming may come from a variety of sources includ-
ing accounting packages, spreadsheets, databases,
Geographic Information Systems (GIS), calendar
schedules, and a variety of web sites. A Geographic
Management Information System (GMIS) allows
you to pull together all the different pieces of data
you generate and publish them as map-based
information for all parties with an interest in a
particular property.
This booklet is not intended to tell you all
about precision farming. It is merely intended
to introduce you to one approach to some of the
tasks on the digital side of precision farming for
crop production using TNTmips and acquaint
you with the fundamentals of GMIS. There are other
approaches that can be taken in TNTmips and in
other software packages. The approach taken in this
booklet is fairly simple. More complex approaches,
such as using principal components to determine
which of your input layers contribute the most to
variability or automatic classification to di-
vide fields into management zones, are also
possible with TNTmips.
The approach described here divides the data
necessary to develop a spatially based decision
support system into four types: background,
activity, current, and management. Background
layers, such as Digital Elevation Models (DEMs)
and soil maps, can often be acquired by download
from the Internet. Activity layers are created over the
background layers. Current layers include recent
satellite images and airphotos of your land.
Management layers become work orders for day
to day operations.

page 3
Precision Farming

Background Data: DOQQs

Background data becomes the base layers for your
spatially based decision support system. Typical
background layers include Digital Ortho Quarter
a portion of a DOQQ
Quads (DOQQs), DEMs, soil maps, topographic
maps (Digital Raster Graphics [DRG]
quad sheet maps at 1:24000 scale),
and public land survey data. You can
build the base layers using inexpen-
sive or free public data. Your base
layers will come in many different
data formats, but that is not a problem
for TNTmips, which imports nearly
90 different raster formats and over 25
vector formats, as well as CAD, TIN,
and database, including all the ones
discussed here.
DOQQs (also known as DOQs or as
COQs or COQQs when compressed)
are rectified digital images of aerial
photographs processed to remove dis-
tortion and displacement resulting
from such factors as camera tilt and
terrain relief. These images cover
one-quarter of a standard 7.5' map quadrangle pro-
duced by the US Geological Survey (USGS) in
Universal Transverse Mercator (UTM) projection
with a 1-meter cell size. Uncompressed DOQQs are
USGS EROS Data Center
http://edcwww.cr.usgs.gov/ available from the USGS EROS DataCenter on a
dsprod/prod.html variety of media for a fee that includes a base charge
and a per file fee. They also offer CD-ROM sets for
COQQs for the state of entire counties. Your state, if in the USA, may offer
Nebraska are available for these same files for sale and also offer the files in
free download from its De- compressed format for free download.
partment of Natural Re-
sources (DNR) web site at The resolution of DOQQs is sufficient for you to
http://www.nrc.state.ne.us/ clearly identify many features on your land. You can
databank/doqqs/doq.html.
use this layer as a base over which to digitally draw
Check the Department of
Natural Resources in your your fence lines and pasture or field boundaries,
state or a similar agency in which then automatically acquire georeference in-
your country. formation from the DOQQs.

page 4
 Precision Farming

Background Data: DEMs

DEMs are sampled arrays of elevations at regu- auto-normalized contrast
larly spaced ground intervals produced in ras-
ter format by the USGS as part of the National
Mapping Program. DEMs are produced at a
variety of scales. The largest scale available,
which is the scale most useful for your back-
ground layers, corresponds to a 7.5' map quad-
rangle at 1:24,000. The newer level 2 DEMs
provide a much smoother surface than the
original level 1 DEMs. DEMs have tradition-
ally had a 30-meter cell size, but 10-meter
DEMs are now available for some areas.
The 30-meter level 2 DEMs are available for
free FTP download from the USGS Geographic
Data Download page (http://edcwww.cr.usgs.gov/ shaded relief
doc/edchome/ndcdb/ndcdb.html). You can
also check your states DNR site for the avail-
ability of 10-meter DEMs.
DEMs are useful for calculating drainage pat-
terns and runoff areas on your land. They also
let you view your DOQQs and other imagery
in 3D perspective and create fly-bys. To find
out more about these topics consult the Get-
ting Started: Modeling Watersheds and Land
Surfaces, 3D Perspective Visualization, and
Operating the 3D Simulator booklets.

partial 3D per-
spective view of
DOQQ over
DEM

page 5
Precision Farming

Background Data: DRGs, PLS

The Digital Raster Graphics A digital raster graphic (DRG) is a scanned image of
home page (http://
a USGS standard series topographic map, including
topomaps.usgs.gov/drg)
provides information on how all map collar information. DRGs are available at
to order DRGs for $50 plus 1:24,000 and 1:100,000 scales. Again, the 1:24,000
$1 for each DRG desired. scale, or 7.5' map, is most useful for your background
layers. These digital maps are in Universal
Transverse Mercator (UTM) projection.
DRGs for the majority of the United States,
its trusts, and territories are available from
USGS, but most of California and part of
Tennessee and some of its neighboring
states are available only from USGS Data
Partners. DRGs may be available for free
download in your area. In Nebraska, they
are available from the Conservation and
Survey Division of the University of Ne-
braskaLincoln, which is a designated
USGS Earth Science Information Center. A
metadata file is also available for down-
load. This file provides such information as
the map extents, the date of the paper map,
DRGs of Nebraska are the date the DRG was created, and general informa-
available for free download tion about DRGs.
from http://csd.unl.edu/csd-
esic/index.html. Check the You can import DRGs into TNTmips from the
designated Earth Science
GEOTIFF format in which they are distributed. If
Information Center in your
state for free DRG availabil- your land happens to fall over more than one map
ity. quad, TNTmips has a process that automatically
trims off the collar area so you can mosaic your
maps.
Public Land Survey (PLS) data (section/town-
ship/range) can also be obtained by free download
to add to the completeness of your background
layers. In Nebraska this download is for the entire
state (77,625 polygons), so you may want to limit
the extents of your import of this Arc/Info Export
A unique record for each format (E00) file to the bounding coordinates pro-
polygon with the information
shown above is attached to vided by your DRG(s).
each section polygon during
import.

page 6
 Precision Farming

Background Data: SSURGO

Soils information can also be useful in the decision Obtain MUIR data from
http://soils.usda.gov and
making process. For example, when planning to
SSURGO data from http://
spray invasive weed species, you can evaluate your ortho.ftw.nrcs.usda.gov/
potential return by the suitability of the soil for muir.
growing your intended replacement.
SSURGO (Soil Survey Geographic Database)
is the most detailed level of digital soil geo-
graphic data developed by the National Co-
operative Soil Survey. SSURGO data is avail-
able for download in modified Digital Line
Graph (DLG-3) optional or E00 format, which
are both imported by TNTmips. Additional
attributes for the soil types in an individual
survey are available from the Map Unit Inter-
pretation (MUIR) Database. MUIR data con-
tains about 88 estimated soil physical and chemical
properties, interpretations, and performance data
such as available water capacity, soil erodibility
factors, and yields for common
crops. TNTmips has a special
import for MUIR data that re-
quires the ELEMENTS table, which
provides information about all
the fields in all the other tables.
This table is not included with MUIR import is
found with the
MUIR data from some down- import wizard in
load sites. The web address listed the Database
at the top of the page includes Editor. All the
this table in its download, but be connections
between MUIR
sure to click on Select All Tables tables are made
and yes for Include header on import, you
rows in tables before you click just need to
on the Process button. You can establish the
relationship be-
also find the state and survey tween the COMP
area ID for the desired area at this table and the
site, which you need for the attribute table
SSURGO download. After from the im-
ported vector
downloading the MUIR data, file (from E00,
go back to the soils web site and the PAT table).
choose to download SSURGO.

page 7
Precision Farming

Activity Layers: Heads Up Digitizing

As soon as you have digi- By this point you should have accumulated a num-
tized your fences and fields ber of background layers at little or no cost. The
over a georeferenced image
and generated standard DOQQs are all that are really needed before you can
attributes, measurements of start creating data that applies to your individual
lengths and areas for se- property, such as fences, field borders, building
lected elements can be locations, and so on. It is easiest to create separate
viewed in the standard at-
tributes table. If you want to vector objects for different themes: the amount you
know the total length of have to learn to assign different drawing styles to the
fences or total area of fields, elements representing different themes, such as fences
simply add Sum statistics to and fields, is minimized by this approach.
your standard attributes
table. Heads up digitizing means you are digitizing data
while looking at the computer screen rather than at
a digitizing tablet. The Spatial Data Editor in
TNTmips lets you create and edit raster, vector,
CAD, and TIN data with your mouse or digitizer.
Ultimately, you want your field
boundaries in vector format with
polygonal topology so you can
get accurate, non-overlapping
4th polygon
in progress
measurements and generate
your management grid. If you
are covering a large area that has
many fields or other areas you
are interpreting, you may want
to start in CAD format and con-
vert to vector after your draw-
ing is refined. This ap-
proach speeds creation of
your object because topol-
ogy does not have to be
field boundaries maintained as you add each
complete line. Consult the Getting
Started booklets entitled
Editing Vector Geodata and
Editing CAD Geodata for
basic information on object
creation in the Spatial Data
Editor.

page 8
 Precision Farming

Activity Layers: FSA Slides

The Farm Services Administration acquires
natural color, section centered slides of
agricultural areas each year in mid to late
July for the purpose of verifying acreages
participating in various farm programs. You 1999 FSA slide
can obtain copies of these slides for $1 each
from your local extension office and have
them transferred to photo-CD. These im-
ages can then be imported by TNTmips and
georeferenced from the DOQQs you have
acquired.
The resolution of your imported raster will
depend on the resolution at which your
slides are scanned. Scanning the slides at
1536 x 1024 pixels produces rasters with an
approximate 2-meter cell size. Higher reso-
lution scanning is available but may be such 2000 FSA slide
that you begin to see defects in the film, as
well as creating a much larger raster file.
The primary benefit of acquiring FSA slides
for your farm is their timelinessthe DOQQs
you obtained are likely to be several years
old. The FSA slides allow you to update
your field boundaries and other features on
a yearly basis. If you can obtain the latest FSA slides The slides above from two
consecutive years show
before digitizing your field boundaries and other the difference between a
features, you will be a step ahead in the process wet year (1999) and a dry
(unless the DOQQs for your area are quite recent). year (2000). They also
clearly show how they can
You can also use the imported FSA slides, which are be used as a reference
24-bit color images after scanning, as input to one of layer to update your exist-
ing vectors; the road that
TNTmips classification processes. One property to curves through the upper
look for is canopy cover. This property often corre- right is under construction
lates well with yield. If you assign class values that in 1999 and paved in 2000.
correspond to quantitative yield factors (the yield Differences in field bound-
aries because of changing
for class 2 is twice as high as for class 1, and so on), stream margins or other
you can use these class values in your management factors should be as appar-
formulas. Of course, a season is required to correlate ent when your vectors are
the classes you define with actual yields to validate overlaid.
this approach.

page 9
Precision Farming

Activity Layers: Management Boundaries

Precision farming involves managing different parts
A GPS unit is es- of each field differently. To achieve this end, you
sential for preci-
sion farming and divide each field into smaller subfields, or manage-
can be shared ment zones. You could approach the desired size of
between your your management zones theoretically, but a practi-
different cal approach is more reasonable. The practical
equipment as
can the palm approach takes into consideration the accuracy of
computer you use to your GPS device, the response time of your variable
program application rates. rate application equipment, what you can afford in
The controller is likely to be terms of sampling, and even the width of your
equipment specific.
normal application equipment. The size of your
management zone determines your
management resolution. When you
first begin precision farming, you
can start at a coarse resolution. As
you see results, you can increase the
resolution by reducing the manage-
ment zone size.
TNTmips Polygon Grid process
generates polygons representing
management zones of the size you
specify. There are a number of differ-
ent shapes, or cell types, to choose
from for your zones. Hexagons are
probably best for management zones because they
better represent the average values for attributes
This field is associated with an area. This process has an orien-
divided into tation tool so you can set the direction of the grid
0.25 acre
hexagons, cells to align with a field boundary.
which is high
Once your grid is generated, the same process will
resolution for
a manage- generate sample points within the grid. These points
ment zone. can be at the cell center, placed randomly, system-
atically unaligned, or the latter two constrained to
be within a specified distance of the cell center. The
point coordinates are in decimal degrees. Use these
coordinates to direct your collection of soil samples,
yields, and other data.

page 10
 Precision Farming

Current Layers: Surfaces

A set of point data, such as soil tests or yield, can be
used to produce a surface raster that represents varia-
tions in the selected attribute value of the points.
Using a query to select the
These points can be at the sample point locations points for surface fitting,
determined for your grid or much denser, such as such as (Import.PStatus ==
yield points collected by your combine. 1) and (Import.Flow > .2),
cleans up the yield data
TNTmips offers a number of different choices for eliminating turn arounds
surface fitting to point data. For dense data, such as and points where there is
yield points, the inverse distance method is a good no flow through the com-
bine. (The PStatus indi-
choice. For sparse data sets, such as soil tests, cates whether the combine
minimum curvature or Kriging may produce better header is up or down and
results. cutting. This status is one
of the data values routinely
Data collected from a yield monitor, planter, or reported by a yield moni-
sprayer, as opposed to data collected at individually tor.)
specified sample points, may need to be cleaned up
prior to generating the sur-
face. You can readily re-
move points where the
combine header was up or
there was minimal flow by
using a query to select
points for surface fitting.
The surface created will
generally extend beyond
the data points and can be
clipped to field boundaries
using the Raster Extract
process.
dense surface surface
all yield generated extracted
points points from using field
points bound-
aries

query
selected
points

page 11
Precision Farming

Current Layers: More Surfaces

Soil test data is an example of point data that may
have multiple attributes you would like to create
soil
sample
surfaces from. Standard soil tests include measure-
points ments of pH, buffer pH, calcium, magnesium, potas-
sium, and phosphorus. They may include organic
matter and micronutrients, such as boron, copper,
iron, manganese, molybdenum, and zinc. Recom-
field
boundary mendations for application of phosphorus and po-
tassium may also be included.
When you have a variety of different attributes for
each point, you probably want to construct a surface
pH level for each attribute. It would be quite tedious to run
the surface modeling process over and over again
followed by raster extraction for each result to clip
it to the field boundaries. Instead, you can adapt a
script already prepared in TNTmips Spatial Ma-
nipulation Language (SML) to create surfaces for all
the attributes of interest from the points in a selected
vector object and to trim them all using a selected
Calcium
boundary vector. You just run the script once and
all your surfaces are created.
As with many sample SML scripts, it is
unlikely that the script will work with your
data without some modification. The table
and field names are stated explicitly in the
script so if your table and/or field names are
not the same or you do not have all the fields
specified for generating surfaces, you will
Magne-
sium need to modify the script. You may also
want to change the search distance or the
tension value in this Minimum Curvature surface
fitting script.
Potassium Once you have a series of related surfaces, you can
use them to generate raster expression regions to
answer questions such as where pH is low and
calcium levels are high. Because you are creating a
organic region, you can immediately see how many acres
matter fulfill the expression youve written (see the later
discussion on Raster Expression Regions).

page 12
 Precision Farming

Current Layers: Classification

Classification has a different focus in terms of preci-
sion farming than it does in general. Classification
in general has the goal of identifying different types
of ground cover. You do not need to classify imagery
of the area you farm to determine what is planted
where; you already know what you planted. For
precision farming you look at areas with relatively
uniform ground cover, such as corn or soybeans, and
classify the imagery to provide a relative canopy
cover rating. This rating can become part of your
management formulas.
TNTmips provides two separate classification pro-
cesses: Automatic Classification and Feature Map-
ping. Feature Mapping is an interactive process that
works on composite color or multiband data and is
suitable for lower quality data, such as air video.
Automatic classification requires multiband data of
fairly high quality. If you have only a composite Feature Mapping result
color image, you can separate the image into its red,
green, and blue components using TNTmips color
conversion process. TNTmips offers a number of
unsupervised and supervised automatic classifica-
tion methods. Unsupervised classification is appro-
priate for determining relative canopy cover for a
single crop type. The ISODATA classification method
is well suited for precision farming; it is similar to K
Means but incorporates procedures for splitting,
combining, and discarding trial classes in order to
obtain an optimal set of output classes.
The results of classification using Feature Mapping
and ISODATA classification with an equal number
of classes defined are shown at the right. The class ISODATA classification
distribution is fairly similar, although you can see result
that human guidance in this case results in larger
uniform patches than found by purely automatic
classification.

page 13
Precision Farming

Current Layers: Quantify Class Rasters

The cell values shown are Both Feature Mapping and Automatic Classifica-
the result of the ISODATA tion produce rasters with arbitrary cell values. To
classification processes.
The cells with a value of 1 use these classes in a meaningful way in manage-
actually have the densest ment formulas, the cell values must be quantitative,
canopy while those with a which means that a cell value of 2 is equivalent to
value of 6 have the least twice the canopy cover as a cell value of 1 and a value
dense canopy. These cells
were assigned values of 15 of 10 represents five times the canopy cover as 2. In
and 2, respectively, in the order to make the transition from arbitrary values to
training set editor. quantitative values, you need to use the Training Set
Initial Cell
Value = 1; Editor, which is part of the automatic classification
Quantified process.
Value = 15
Initial Cell In general, the Training Set Editor is designed so that
Value = 6; you can identify areas known to contain a particular
Quantified
Value = 2 type of ground cover or surface material. One or
more training areas are identified for each feature
class to be defined. This information can then be
used in the supervised classification processes to
identify areas with statistical properties similar to
the training areas.
The Training Set Editor can also be used to make
classification raster cell values quantitative.
You simply select the class raster saved in
either Feature Mapping or Automatic Clas-
sification as the training set raster and change
the cell value for each class from its original
arbitrary number to the desired quantitative
value, which is entered as the tag value.
When you apply the tag values and save the
original
training set raster, you will have a raster
cell val-
ues with the appropriate cell values for use in
your management formulas. The magni-
tude of the values is relatively unimpor-
tantyou can use a scaling factor in your
assigned management formulas to adjust the values.
quantita-
tive cell The relative value is what is important; a
values value of 10 should represent twice the
canopy cover as a value of 5.

page 14
 Precision Farming

Current Layers: Productivity Potential

You may also want to have your management
vector input
formulas take into account the productivity
potential for a particular crop of the soil types on
the fields you farm. Vector soil maps often have
an associated productivity potential table with
numeric values. You can use Vector to Raster
Conversion to create a productivity potential
raster for the desired crop.
If your soil polygons have productivity for dif-
ferent crops separated into different tables, you
can choose to set the value by attribute and
select the desired Table.Field (for example,
COMPYLD.NIRRG). If, however, each polygon has
multiple records attached, as SSURGO data
does, some manipulation is required to get the
desired values in the output raster. One ap-
proach is to write a query that loops through all
the attached records until finding the record for
the desired crop type. Another approach is to select
only the records for the
desired crop type (first
sort on the crop type
field) then to create a new
table with only these Rather than con-
records attached to the polygons. You can then use structing a complicated
this table to specify the desired Table.Field. query to pull out the values
you want when multiple
records are at-
raster output tached, create a
new table that
only has the
single record of
interest for each
polygon at-
tached. You
then use that
table and the
desired field for
conversion. All
of these steps
can be done in
the conversion
process.

page 15
Precision Farming

Management Layers: Pull It All Together

Management layers are derived by combin-
ing the background, activity, and current
layers you have collected and generated.
You can also integrate university research,
industry trends, and personal experience
when you derive management formulas.
Your management boundaries (grid vector)
need to be combined with the other data you
have collected to be useful for making deci-
All selected rasters and the sions. The Raster Properties process handles this
vector to which the proper- task. The process creates one or two tables for each
ties are transferred are dis- input raster (the table that contains a histogram for
played in the process. You
each polygon is optional and not really useful for
can view the results without
exiting the process by using management formulas). The statistics table has one
the Layer Controls to open record for each polygon that contains the minimum
the newly created tables. and maximum values of the raster that fall within the
polygon and the mean, mode, median, and standard
deviation of raster values, the cell count for the
polygon, and a normalization factor. The cell count
for a polygon will differ from table to table unless
your input rasters have the same cell size.
There are a number of choices for how to count cells
that fall along the polygon boundary. A good
choice for this application is Include if largest
portion. This option includes a cell in a polygons
statistics if more than 50% of
the cell is within the polygon.
Two tables gen-
erated by Raster The NormalizeFactor field is
Properties are calculated by dividing the glo-
shown for the
bal mean for the polygons (the
same selected
polygon. means of all the means) by the
value for that polygon. It is
useful for comparing data for different years with
different crops in the same field. If the normalize
factor is consistently low from year to year, the
grid cell has a consistently high yield whether it
is corn (at 200 bushels an acre) or soybeans (at 50
bushels an acre).

page 16
 Precision Farming

Mgt. Formulas Using Computed Fields

Now that you have raster properties associated with
your grid polygons, you can derive your manage-
ment formulas and vi-
sualize the results. A
grid enumeration
table was generated
along with your vec-
tor grid. This table is
a good place to add
the computed fields expression
for Planting
that contain your man-
agement formulas unless polygon grid numbers are
not uniquely assigned. If not unique, you should
create a new table that has an implied one-to-one
attachment type with the first computed field taking
its value from the Polygon_ID table.
Computed fields let you easily vary your manage-
ment formulas and immediately visualize how
changes you make affect the intended action, such
as planting, fertilizer, and herbicide rates. Add a expression
computed field for each action you want to model. for Fertilize
If application rates are theme mapped, you get imme-
diate feedback when you redraw after changing your
management formulas.
With each new year or set of data collected, you can
add more raster properties tables. You can then
readily update your management formulas to
include this new data if they are kept together in
a single table, as shown above.
The expressions for two of the computed fields in
this application rates table are shown above. Note
that one of the expressions makes use of the mean
from the rasters properties generated from the
yield surface and the other uses the normalize
factor. These expressions are fairly simple, each
taking only one factor into account. You can
include as many factors as you want in your
management formulas.

page 17
Precision Farming

Management Layers: Action Maps

The maps created from your management for-
mulas, such as the theme map at the bottom of
the preceding page, are your action maps. If
exported to ArcView Shapefile format, these
maps can be used directly to control your plant-
ing or herbicide / fertilizer application rates in
conjunction with Farm Site Mate and variable
rate equipment. Farm Site Mate utilizes a palm
or handheld computer along with a GPS receiver for
precise application of products with most variable
rate controllers.
You get four files when you export to Shapefile
format, all with the same name but a different exten-
sion (.dbf, .prj, .shp, and .shx). You
can continue to refine and adjust your
management formulas in TNTmips.
When you decide you want to update
your planting or application rates
using your new formula results, you
do not need to create a new Shapefile,
you simply need to replace the database (.dbf) file,
which can be accomplished in the display process.
In order to have the Save As choice to create the
database file, you must have the relevant table open
in tabular view. You then elect to save all data
records in dBASE III format and overwrite your
previous .dbf file.
You can refine your model with the original data, but
you can also incorporate new data as it is received.
When you get yield data for a
new year or soil samples you
didnt have previously, incor-
porate this new information. Just
use Raster Properties to transfer
the new information to your vector grid then edit
your computed fields to include the new data in your
management formulas. You can also try different
management resolutions using the same manage-
ment formulas.

page 18
 Precision Farming

Mgt. Layers: Raster Expression Regions

Management analysis is not restricted to your vector
grid with associated raster properties. Raster expres-
sion regions let you identify areas that meet speci-
fied criteria in one or more rasters. You can readily
identify areas that satisfy a different query for each
raster of interest in a single step. The query can be set
up so that the area identified satisfies the expression
for all layers, for any one of the layers, or somewhere
in between. The query shown on this page identifies
areas where the calcium levels are greater than 1,000
parts per million and the pH is less than 7.
To create a raster expression region, click on the
Create Region icon for one of the raster layers you
want included in the expression, then click on the
Add Rasters button and select the other rasters.
Rasters must be listed in the Raster Expression
Region Generation window in order to be valid in
the expression. Rasters are included in the expres-
sion by name, which can be inserted in the expres- Query:
(Calcium > 1000) and
sion by typing or using Insert / Symbol and choosing
(pH_Levels < 7)
raster as the type to list the names of all rasters you
selected for use in the expression (see illustration at
right). You then use operators, such as >, <=, and ==,
to designate the cell values of interest. Enclose the
part of the expression that applies to each raster in
parentheses and link them with an and or an or. You
can also directly compare raster values (for example,
Magnesium > Manganese).
Once you have generated a raster expression region,
or any other region type, you can open the GeoTool-
box and immediately see the area and perimeter of
the region, along with other information. If you have
created multiple regions, the information will be for
the last region created. You can change the region
for which measurements are shown by clicking on
the Region tab and selecting the desired region from
the list. Default region names are the same as the
object they are created from.

page 19
Precision Farming

Organize Data into Atlases

When you organize your data into an atlas using the
Otoe County Farms atlas
HyperIndex Linker tool in TNTmips, you create
Adobe Acrobat Search
point and click access to all your geodata and can
include links to external informa-
tion, such as weather, markets,
Precision Farming page
on MicroImages web site genetics, and so on. The data also
becomes portable when burned
to CD-ROM along with TNTatlas,
which means the atlas can be
viewed on any computerTNT-
mips need not be installed and
there is no license key required.
Install TNTatlas
When you create an atlas in TNT-
COOP web site mips, you can choose to bring up a graphic with
Iowa State University Install, Browse, and Exit options when the CD is
weed science page inserted. This information is stored in an autorun.inf
Nebraska Statewide atlas file. You can also create this file yourself in a text
editor and specify that some other program, such as
Adobe Acrobat Reader, run and open a specified file.
The upper illustration on this
page is of an atlas that uses a
graphic front end with a num-
ber of hyperlinked buttons
that let you install TNTatlas
if you havent before, launch
the atlas provided on the CD,
connect to a variety of re-
lated web sites, and use the
Acrobat search capabilities
This atlas opens when you click on the Farms button
for the PDF documents linked
to the atlas.
The home page of this atlas also has a number of link
buttons. In fact, the link to the more detailed farm
There are a number of Get- layout is from one of these buttons, not from the area
ting Started booklets (soon
to be six) that describe con-
of Otoe County where the farms are found, which
struction, design, and use of enables you to find the information without know-
atlases. ing its actual location. The links from the geodata
on the home page are for weather (by ZIP code) and
soil information (by soil type where you click).

page 20
 Precision Farming

Provide Access to All Materials

Use your atlas to organize all the data for the areas
you farm. Create an Acrobat document for each
farming unit and add all the cost related information
you have to it, such as seed and spraying costs. This
material can simply be scanned invoices. At the end
The report shown below is
of the season, add your yield data as the first page. for a different crop year
Create a similar file for each crop year and add it to than displayed in the atlas.
your atlas. Creating links to all materials
for a new crop year takes little time if you
create a directory structure for each year
with all files and directories named as
they were in the previous year. To create
the computed field to use for links by
attribute (links that vary according to
the attributes of the selected polygon),
you need only copy the expres-
sion from the previous year and
change the date to have the full
set of linked polygons.
JDmap uses the pro-
Links by attribute can also be prietary data from your
used to dynamically link to a yield monitor card to
create the color page
variety of web sites, such as local weather (from ZIP
below, which can
code polygons), soil details be saved as an
(from soil type polygons), and Acrobat document.
markets, genetics, and univer- The other
pages
sity research (from field bound-
were
aries and crop type), scanned.
You can either use existing poly-
gons, such as field boundaries,
or create new polygons that
dont cover your yield points,
as shown here, to link to the
Acrobat documents. As long as
the document contains one page
with searchable text (not simply
scanned), these documents can be lo-
cated from the front-end graphic for the
atlas by searching, as well as from their HyperIndex
links.

page 21
Precision Farming

Publish Your Atlas

Once you have created an atlas, you can publish it
on CD-ROM. It is then easy to distribute among the
people you want to have access to the same infor-
mation, such as landlords, bankers, and crop
consultants. You can print and apply CD
labels for a more professional looking prod-
uct if desired. You can readily update atlases
for people with whom you share informa-
tionjust send them a new CD.
If your atlas has an Acrobat front-end as de-
scribed on the previous pages, people who use
your CD will have to have TNTatlas already in-
stalled to use the button that takes them to your atlas.
If the front-end for your atlas is also the atlas home
page, you have the option to have the TNTatlas
installer open automatically when the CD is inserted
or to have the atlas open automatically in TNTatlas
for Windows. If you choose the installer route, the
CD recipient installs TNTatlas then launches it from
their drive and opens the .atl file on the CD. If you
have the atlas open automatically in TNTatlas for
Windows, the program runs from the CD.
The TNTatlas Assembly Wizard is de-
signed to collect all the data for your atlas
into a single directory and to package
TNTatlas for installation if desired. You
can use the wizard to assemble your data
whether or not you want insertion of the
CD to launch the installer. Simply leave
The choices on this panel of blank the panel shown at the left if you do not want
the TNTatlas Assembly the installer to launch automatically. You will need
Wizard apply only if you to create your own .inf file if you want either Acrobat
want the CD to automati- or TNTatlas for Windows to run automatically.
cally run the installer when
inserted. If your atlas is of interest to a wider audience, you
may want to consider publishing it on the Internet.
Internet publishing requires an additional product,
TNTserver, to serve up the atlas to those that want
to view it.

page 22
 Precision Farming

Crop Production Networks

A Crop Production Network is an informa-
tion based COOP that uses TNTserver to
make that information available to all mem-
bers over the Internet. A Crop Production
Network allows groups to pool their re-
sources, such as data management tools,
agronomic expertise, background data lay-
ers, and wide area imagery. Crop production
modeling is easier when pooled over a larger area
and, thus, benefits from use by a Crop Production
Network.
An atlas can be available for viewing to anyone who
visits your web site, or it can be set up to require a
password for access. Atlas viewing over the web uses
one of the TNTclients: the HTML based client or one
of the two Java based clients. One of the Java clients
is downloaded every time you choose to view an
atlas and the other is downloaded once and installed
(Windows only). Use of TNTclient is free.
Your crop pedigree can be maintained and accessed
online. You can even provide direct links to web
sites for each seed variety. Information can be shared Use your atlas for crop
with food processors with trait and crop progress identity preservation. You
can even link by attribute to
information available as part of your atlas. A crops
the appropriate seed web
production history can be an important selling tool site.
and is readily available for viewing by all interested
parties.

page 23
 Advanced Software for Geospatial Analysis

MicroImages, Inc. publishes a complete line of professional software for advanced geospatial data
visualization, analysis, and publishing. Contact us or visit our web site for detailed product F
information. A
R
M
TNTmips TNTmips is a professional system for fully integrated GIS, image analysis, CAD,
I
TIN, desktop cartography, and geospatial database management. N
TNTedit TNTedit provides interactive tools to create, georeference, and edit vector, image, G
CAD, TIN, and relational database project materials in a wide variety of formats.
TNTview TNTview has the same powerful display features as TNTmips and is perfect for
those who do not need the technical processing and preparation features of TNTmips.
TNTatlas TNTatlas lets you publish and distribute your spatial project materials on CD-
ROM at low cost. TNTatlas CDs can be used on any popular computing platform.
TNTserver TNTserver lets you publish TNTatlases on the Internet or on your intranet.
Navigate through geodata atlases with your web browser and the TNTclient Java applet.
TNTlite TNTlite is a free version of TNTmips for students and professionals with small
projects. You can download TNTlite from MicroImages web site, or you can order
TNTlite on CD-ROM.

Index
action maps ............................................. 1 8 grid generation ....................................... 1 0
activity layers ................................. 3, 810 HyperIndex Linker ................................ 2 0
Atlas Assembly Wizard .......................... 2 2 management formulas ........................... 1 7
atlases ................................................ 2023 management layers ..................... 3, 1619
background layers ........................... 3, 47 MUIR ......................................................... 7
classification ............................................ 1 3 productivity potential ............................ 1 5
Crop Production Networks ................... 2 3 Public Land Survey ................................. 6
current layers .............................. 3, 1115 quantifying data ..................................... 1 4
DEMs ......................................................... 5 raster expression regions ....................... 1 9
digitizing ................................................... 8 raster properties ...................................... 1 6
DOQQs ................................................. 4, 9 soil tests .................................................... 1 2
DRGs .......................................................... 6 SSURGO .................................................... 7
external files ....................... 18, 20, 21, 23 surface fitting ................................... 1112
FSA slides .................................................. 9 TNTclient ................................................ 2 3
GPS .......................................................... 1 0 TNTserver ........................................ 22, 23

MicroImages, Inc.

Voice: (402)477-9554
www.microimages.com