Unit 3

Introduction
 Cell represents
only the digital
number.
 We have convert
that data into
information.
 Can be done
manually by
What does cell image
represent? interpretation or
automatically by
classification.
 Image Classification
• The science of turning RS data into meaningful
categories representing surface conditions or classes
(feature extraction)
• Methods-
1. Spectral pattern recognition procedures classifies a
pixel based on its pattern of radiance measurements
in each band: more common and easy to use
2. Spatial pattern recognition classifies a pixel based on
its relationship to surrounding pixels: more complex
and difficult to implement
3. Temporal pattern recognition looks at changes in
pixels over time to assist in feature recognition
 Limitations of Training samples
• The sample data should be collected for the
entire study area.
• Need to have an idea of the area.
• Time consuming and expensive
• Same period of time should be followed for
the sample and the area to be analyzed.
 Supervised Classification
• Algorithms
1. Euclidean distance method
2. Parallelepiped Method
3. Maximum likelihood method
 Euclidean distance method

The simplest method is “minimum distance” in which a

theoretical center point of point cloud is plotted, based on
mean values, and an unknown point is assigned to the nearest
of these. That point is then assigned that cover class.
 Parallelepiped Method

In this method the variance of the class is considered. Based on

variance values, an unknown point is assigned to the nearest of
these. That point is then assigned that cover class.
 Maximum likelihood

In this method the probability distribution of each class is

taken into consideration. The unknown pixel is assigned by
considering the maximum probability of occurrence.
 Unsupervised Classification
 Assumes no prior knowledge
 Computer groups all pixels according to their spectral relationships
and looks for natural clustering's
 Assumes that data in different cover class will not belong to same
grouping
 Once created, the analyst assesses their utility and can adjust
clustering parameters
 After comparing the reclassified image (based on spectral classes)
to ground reference data, the analyst can determine which land
cover type the spectral class corresponds to
 Has advantage over supervised classification: the “classifier”
identifies the distinct spectral classes, many of which would not
have been apparent in supervised classification and, if there were
many classes, would have been difficult to train all of them. Not
required to make assumptions of what all the cover classes are
before classification.
 Nearest Neighbour Approach

In this method, odd number of neighboring pixels are chosen

and then the distance to the unknown pixel are calculated
from each pixel. The unknown pixel is assigned by considering
the nearest neighbor.