A COMPARATIVE STUDY OF LINEAR MIXTURE MODELLING AND MAXIMUM LIKELIHOOD CLASSIFIER FOR SUB-PIXEL CLASSIFICATION

Abstract

Classification of images, dominated by mixed pixels, using conventional hard classification techniques may produce erroneous results and thus, may lead to unrealistic representation of land cover. This may affect the efficient and economic planning, management and monitoring of natural resources, Consequently, soft classification techniques which provide sub-pixel land cover information may have to be utilized. Of the various soft classification techniques available, for example, Artificial Neural Network and Fuzzy techniques, are relatively new while the Liner Mixture Modelling technique has been widely used by remote sensing community.