Supervised Classification
O.Conrad (c) 2005
Supervised Classification: Minimum Distance, Maximum Likelihood.
Parameters
Grid system
Grid system
Grid system
Grids
Input Grid list
Training Areas
Input Shapes
Class Identifier
Table field
Class Information
Output Table
Classification
Output Grid
Quality
Output Grid
Dependent on chosen method, these are distances or probabilities.
Method
Choice
Available choices: parallelepiped, minimum distance, mahalanobis distance, maximum likelihood, spectral angle mapping, binary encoding
Normalise
Boolean
Distance Threshold
Floating point
Let pixel stay unclassified, if minimum or mahalanobis distance is greater than threshold.
Probability Threshold (Percent)
Floating point
Let pixel stay unclassified, if maximum likelihood probability is less than threshold.
Maximum: 100.0
Probability Reference
Choice
Available choices: absolute, relative
Spectral Angle Threshold (Degree)
Floating point
Let pixel stay unclassified, if spectral angle distance is greater than threshold.
Maximum: 90.0