Image Classification
and its Applications
Introduction – What is Image
Classification?
• It is the task of extracting and categorizing useful information from a
multi-band raster image. This extracted information is further used to
create thematic maps.
• The image is classified according to its visual content.
• The classification scheme/criteria depends on the need of the end
user.
Why Classify Images?
• The intent of the classification process is to categorize all pixels in a
digital image into one of several land cover classes, or "themes“.
• This categorized data may then be used to produce thematic maps of
the land cover present in an image.
• The objective of image classification is to identify and portray, as a
unique grey level (or colour), the features occurring in an image in
terms of the object or type of land cover these features actually
represent on the ground.
Types of Image Classification
• Unsupervised Classification: Unsupervised classification is a method
which examines a large number of unknown pixels and divides into a
number of classes based on natural groupings present in the image
values. Unsupervised classification does not require analyst-specified
training data. The basic premise is that values within a given cover
type should be close together in the measurement space (i.e. have
similar grey levels), whereas data in different classes should be
comparatively well separated (i.e. have very different grey levels).
Types of Image Classification
• Supervised Classification: With supervised classification, we identify
examples of the Information classes (i.e., land cover type) of interest
in the image. These are called "training sites". The image processing
software system is then used to develop a statistical characterization
of the reflectance for each information class.
• Once a statistical characterization has been achieved for each
information class, the image is then classified by examining the
reflectance for each pixel and making a decision about which of the
signatures it resembles most.
Supervised Classification - Steps
Supervised Classification –
Some Advanced Techniques and
Concepts
• Neural Networks: Use flexible functions to partition the spectral
space.
• Contextual Classifiers: Incorporate spatial or temporal conditions in
classification.
• Linear Regression: Instead of discrete classes, apply proportionate
value of classes to each pixel.
Maximum Likelihood
Classification
• Maximum likelihood Classification is a statistical decision criterion to
assist in the classification of overlapping signatures; pixels are
assigned to the class of highest probability.
• The maximum likelihood classifier is considered to give more accurate
results. However, it assumes that classes in the input data have a
Gaussian distribution and that signatures were well selected; this is
not always a safe assumption.
Applications of Image
Classification
• Application 1
Remote Sensing: Used in classifying raw satellite image of land area,
taken from space, into useful segregators based on geography.
STEPS
1) Acquisition of suitable satellite imagery.
2) Visual Interpretation of satellite image.
3) Radiometric and Geometric correction.
4) Stretch and filter logarithm
5) Supervised classification: Every object class in the image will be
coordinated to reference areas called ‘training areas’, and this
enhances the statistical classification.
Sample Satellite Image
Classification
Applications of Image Classification
• Application 2
Infrared Thermal Imagery
• All objects emit infrared energy (heat) as a function of their
temperature. This is called ‘heat signature’.
• The hotter an object is, the more radiation it emits.
• A thermal camera captures heat signatures and classifies the image of
an object or area based on differences in temperature.
• The distinction between areas with high and low heat signatures is
clearly visible in the image. This is used in many popular real-world
scenarios such as in cricket to detect if the ball has touched the bat or
not for contentious catch appeals.
Image of IR Camera used in
Cricket
