Satellite Image Classification Using Self Organizing Map And Ensemble Classifiers

Satellite picture request process incorporates gathering the image pixel regards into significant classes. A few satellite picture characterized strategies and systems are accessible. In existing k-medoid clustering technique is used for clustering the satellite data, with this method not able to cluster accurately all the classes. In our proposed method self-organizing map as a clustering technique is used. Self-organizing maps clustering the data based on similarity, topologies, with a preferences of appointing the same value to each classes. Self-organizing map clustering are used to reduces a dimensionality of data and to cluster the data. These are motivated by the tactile and engine mappings in the vertebrate cerebrum, which additionally appears to consequently organizing out data topologically. Our propose technique is Ensemble clustering with subspace discriminant algorithm for classification of satellite data into water, Agriculture, Barren land, Green Land. The proposed method of self-organising map clustering and ensemble classifier with subspace discriminant is given best result compared to existing ones.