ABSTRACT
Classification techniques of remotely sensed data are of the most powerful tools to acquire land use and land cover information in a cost effective means. When applied to a temporal data set, it becomes extremely useful to determine landuse/ Iandcover changes and their trends.
In this paper a features (bands) selection algorithm has been developed to determine the optimal bands that can be used in the classification process . Also, a classification scheme is developed using the information extracted from the classified images, visual interpretation and the available maps. A Database is designed using the proposed classification scheme .