NAME

g.gui.iclass - Tool for supervised classification of imagery data.
Generates spectral signatures for an image by allowing the user to outline regions of interest.

KEYWORDS

general, GUI, classification, supervised classification, signatures

SYNOPSIS

g.gui.iclass
g.gui.iclass --help
g.gui.iclass [-m] [group=name] [subgroup=name] [map=name] [trainingmap=name] [--help] [--verbose] [--quiet] [--ui]

Flags:

-m: Maximize window
--help: Print usage summary
--verbose: Verbose module output
--quiet: Quiet module output
--ui: Force launching GUI dialog

Parameters:

group=name: Name of input imagery group
subgroup=name: Name of input imagery subgroup
map=name: Name of raster map to load
trainingmap=name: Ground truth training map to load

KEYWORDS

display, GUI, imagery, classification, supervised classification

DESCRIPTION

Supervised Classification Tool (wxIClass) is a wxGUI compoment which allows the user to create training areas and generate spectral signatures. The resulting signature file can be used as input for i.maxlik or as a seed signature file for i.cluster. WxIClass can be launched from the Layer Manager menu Imagery → Classify image → Interactive input for supervised classification or via command line as g.gui.iclass.

wxIClass currently allows you to:

create training areas (using customized vector digitizer)
show histograms for each band and class (category)
show coincidence plots for each band
show raster cells that match training areas (within the number of standard deviations specified)
specify color of class
write signature file
import vector map
export vector map with attribute table

wxIClass performs the first pass in the GRASS two-pass supervised image classification process; the GRASS module i.maxlik executes the second pass. Both programs must be run to generate a classified map in GRASS raster format.

wxIClass is an interactive program that allows the user to create multiple training areas for multiple classes and calculate the spectral signatures based on the cells that are within the training areas. During this process the user will be shown histograms for each image band. The user can also display the cells of the image bands which fall within a user-specified number of standard deviations from the means in the spectral signature. By doing this, the user can see how much of the image is likely to be put into the class associated with the signature.

The spectral signatures are composed of region means and covariance matrices. These region means and covariance matrices are used in the second pass (i.maxlik) to classify the image.

Alternatively, the spectral signatures generated by wxIClass can be used for seed means for the clusters in i.cluster.

AUTHOR

Anna Kratochvilova, Czech Technical University in Prague, Czech Republic
Vaclav Petras, Czech Technical University in Prague, Czech Republic

SOURCE CODE

Available at: wxGUI Supervised Classification Tool source code (history)