SUBPIXEL CLASSIFICATION OF BALD CYPRESS AND TUPELO GUM TREES IN THEMATIC MAPPER IMAGERY

A subpixel spectral analytical process was used to classify Bald Cypre ss and Tupelo Gum wetland in Landsat Thematic Mapper imagery in Georgi a and South Carolina. The subpixel process enabled the defection of Cy press and Tupelo trees in mixed pixels. Two-hundred pixels were field verified for each tree species to independently measure errors of omis sion and commission. The cypress total accuracy was 89 percent and the tupelo total accuracy was 91 percent. Field investigations revealed t hat both cypress and tupelo frees were successfully classified when th ey occurred both as pure stands and when mixed with other tree species and water. In a comparison with traditional classification techniques (ISO-DATA clustering, maximum likelihood, and minimum distance) the s ubpixel classification of cypress and tupelo yielded improved results. Large areas of wetland where cypress was heavily mixed with other tre e species were correctly classified by the subpixel process and not cl assified by the traditional classifiers.