The cluster routine uses a two-pass sequential clustering algorithm. In the first pass, the program reads through the entire data set, and sequentially builds clusters (groups of points in spectral space) based on parameters selected by the user, and computes the mean value for each cluster. These clusters become the signatures used to assign classes in the output GIS file. The second pass classifies each pixel in the data set according to a minimum distance classifier. The algorithm calculates the spectral distance between the candidate pixel and the mean value for every cluster, using the mean values that were computed in the first pass

%B Geocarto International %V 3 %P 37 -40 %G eng %M KNZ00199 %R 10.1080/10106048809354174