gvSIG allows to perform a raster supervised classification by three different methods: maximum likelihood, minimum distance and parallelepiped.
To open the classification tool, you have to use the remote sensing toolbar by selecting “Raster process” from the left button and “Classification” from the right button.
The following window will be displayed
Operations Panel
In (1) choose the raster to classify. You can choose only an image already loaded in the view. For the selected image, chose also the bands you use to perform the classification process.
In (2) choose the method you use to perform the classification, you can select between: maximum likelihood, minimum distance and parallelepiped.
In (3) edit the classes on which the classification will be performed. By default, when you select the image to classify, a number of classes equal to the number of regions of interest (ROIs) linked to image are loaded.
You can modify (see document of ROIs editing) class number and composition, according to your need, without modifying the original ROIs.
Options Panel
To setup the parameters and the output options, activate the Option panel
In (4) adjust the maximum standard deviation if the selected method is the parallelepiped one. In case of other methods, the panel will be empty.
In (5) setup the usual output parameters. Choose the output file name and path.
