AN ITERATIVE CLASSIFICATION APPROACH FOR MAPPING NATURAL-RESOURCES FROM SATELLITE IMAGERY

This project explores an iterative classification process as an altern ative to two-stage classifications. In the iterative classification ap proach cover types are classified, one or two at a time, and the band selection process is repeated in each iteration, so that the combinati on of bands that provides the best separability among the classes that remain to be classified is selected. The optimum number of bands to p erform the classification is also determined for each iteration, so th at the classification of the area that is masked in each iteration ach ieves the highest possible accuracy. Spectral Pattern Analysis, and Sp ectral Separability Indices are used in the band selection process. GI S analysis is used to obtain prior probabilities, and to determine if variables such as elevation, slope, and aspect can result in a source of information for segmentation of the study area into more homogeneou s strata. The results of this study show that: (1) The proposed iterat ive classification approach is superior to traditional single-step sup ervised and unsupervised approaches with a 99 per cent confidence leve l, and (2) Prior probabilities improve band selection process only whe n the TD separability index is used, but do not improve the classifica tion process itself. GIS analysis of the study area may serve as a ver y useful tool for segmenting the study area into more homogeneous stra ta, but due to the large size of the study area, and the large number of classes (21 classes) being discriminated it did not help in the cla ssification performed in this study.