Digital Image Processing

1
Digital Image Processing

• A wide variety of tools that we use to make
  remote sensing data provide even more
  information.
• Tools include rectification, resampling,
  smoothing, edge enhancement, stretching etc.
• The first tool considered is classification

2
Classification

• Patterns in digital numbers
• Patterns on the landscape
• These tools use statistical analysis of
  multi-spectral images to enhance the information
  content provided by RS data.
• Creating a thematic map

3
Multispectral image classification depends on the
the fact that surface materials have different
spectral reflectance patterns. Different
spectral signatures.
4
Supervised vs. Unsupervised

• In supervised classification the interpreter
  provides information about the classes he expects
  (or wants) to find.
• Training Sites are selected on the image to
  identify the patterns in spectral space of
  classes/features that are to be identified
• Unsupervised Classification patterns inherent in
  the spectral data drive the classification
  process.

5
Unsupervised classification (contd.)

• Unsupervised classification can often produce
  information that is not obvious to visual
  inspection.
• Very useful for areas where ground truth data
  is difficult to obtain
• Purely spectral pattern recognition
• The critical issue in ALL image classification is
  to equate spectral class to informational
  class!!

6

• The trick then becomes one of trying to relate
  the different clusters to meaningful ground
  categories. We do this by either being adequately
  familiar with the major classes expected in the
  scene, or, where feasible, by visiting the scene
  (ground truthing) and visually correlating map
  patterns to their ground counterparts

7
The process.

• Step one cluster analysis (Identifying clusters
  in the data)
• Step twoClassification of pixels into classes
  based on cluster centers

8
Simple X Y example
(if it were only this simple in reality.)
9
A 3D version of spectral clusters can easily
be extended to n dimensional spectral space.
10
The clustering process

• Virtually all programs use the identical
  algorithm ISODATA
• Iterative Self-Organizing Data Analysis (ISODATA)
  Tou and Gonzalez 1974)
• Begins by assigning class centriods in
  statistical space (random assignment or some
  variation)

11
Cluster analysis.

• The input parameters always requested by ISODATA
  include
• The initial number of classes
• A class separation distance (a lumping threshold)
• And the number of iterations (or statistical
  threshold) that will define the end of the process

12
(No Transcript)
13
The first stage.cluster analysis

• At the end of the first stage nothing exists
  but a set of spectral coordinates in n
  dimensional space (where n is the number of
  spectral bands used in the classification)
• Clusters have been defined based on the number of
  cluster centers you start with and the lumping
  threshold which defines the distance between
  centers in spectral space
• ERDAS reports this as a .sig file

14
(No Transcript)
15
The right number of classes?

• How does one select the correct or natural or
  right number of classes?
• The goal is INFORMATION CLASSES not spectral
  classes
• Expert Assessment and Visual Comparison (it
  just looks better!)
• Statistical tools?

16
Stage 2 putting all pixels into classes

• There are three primary methods for assigning
  image pixels to classes
• Minimum Distance to Means (mindist)
• Parallellpiped
• Maximum likelihood classification (maxlik)

17
The simplest classification Minimum Distance to
Means (MINDIST)
Pixels are assigned to a class based only on the
minimum Euclidian distance to the closest cluster
center. Quick and easy but doesnt consider
variability in the data (the density of the
cluster)
18
Parallepiped classification defines rectangular
decision boundaries around classes. The size of
the rectangular decision boundary is defined by
the variability in the spectral data.
19
Decision boundaries are defined by variability of
the cluster in each dimension
20
Misidentified pixels. A common problem
21
Maximum Likelihood Classification

• This classification is very common
• Used the variance and covariance of the data to
  define a probability density function or
  probability surface.
• (this assumes a normal distribution of the data)

22
A probability density surface for the sample data
set.. Based on the variability in the cluster,
how likely is the inclusion of a given pixel
23
Probability contours for classes in 2 dimensional
space these statistical clouds extend in n
dimensions.
24
Supervised Classification

• Creating statistical clusters based on a priori
  information
• The interpreter knows what he wants to find and
  creates signature files (cluster centers) from
  training sites on the image.

25
(No Transcript)
26
Choosing training sites

• Every class has to be fully identified

27

• Training sites should be chosen from all across
  the image
• Training sites should avoid edges where mixed
  pixels can add uncertainty to the classified
  image
• A tool to accurately classify mixed pixels
  or highly heterogeneous areas is to choose
  training sites within the mixed area the
  spectral signature for this class can be worked
  with independently.

28

• Training sites should include 10 to 100 times as
  many pixels as the total number of bands being
  used in the classification e.g. for 7 TM bands
  training sites for each class ought to contain at
  least 70 700 pixels.
• In agricultural applications not uncommon to have
  100 training sites / class
• Polygons vs. seeds. Rather than delineate the
  entire polygon, software can be used to grow a
  training site