In this paper, four versions of the sequential maximum likelihood algorithm have been employed to classify LANDSAT data & their performance is compared with that of the maximum likelihood classifier. The sequential maximum likelihood algorithm is a faster, but slightly less accurate version of the conventional maximum likelihood algorithm. In the sequential case, the number of features (MSS bands) used to classify a given pixel depends upon the particular band values of that pixel, since on the average, the number of bands used will be less than the maximum there will be a serving in computational time.

Date of this Version