Investigation of the Possibility to Prepare Supervised Classification Map of Gully Erosion by RS and GIS

This study investigates the possibility providing gully erosion map by the supervised classification of satellite images (ETM) in two mountainous and plain land types. These land types were the part of Varamin plain, Tehran province, and Roodbar subbasin, Guilan province, as plain and mountain land types, respectively. The position of 652 and 124 ground control points were recorded by GPS respectively in mountain and plain land types. Soil gully erosion, land uses or plant covers were investigated in these points. Regarding ground control points and auxiliary points, training points of gully erosion and other surface features were introduced to software (Ilwis 3.3 Academic). The supervised classified map of gully erosion was prepared by maximum likelihood method and then, overall accuracy of this map was computed. Results showed that the possibility supervised classification of gully erosion isn’t possible, although it need more studies for results generalization to other mountainous regions. Also, with increasing land uses and other surface features in plain physiography, it decreases the classification of accuracy. Keywords—Supervised classification, Gully erosion, Map.