Multi-spectral image classification using spectral and spatial knowledge

Classification of land cover from multi-spectral remotely sensed imagery for large geographic regions requires complex algorithms and feature selection techniques. Statistical classification methods have been widely applied in remote sensing, especially the maximum likelihood classifier. Easy availability of spatial databases has opened several new opportunities to improve traditional classifiers and develop new classification systems that can incorporate these spatial databases into the decision process. In this paper we present a new classification approach which combines knowledge based systems and maximum likelihood classifier. This classification fusion approach has yielded an overall classification of accuracy of about 85% in a fairly complex geographic setting.