Satellite image classification of different resolution images using cluster ensemble techniques

The classification of Satellite image is an imperative system utilized to retrieve information in remote sensing. Primary data of extraordinary significance to several difficulties can be acquired straightforwardly from Land-cover observing. As it is required to discuss about the issue of supervised Land-cover classification of multispectral satellite images in the perspective of cluster ensemble and self learning. Different information partitions inferred by several clustering methods which are gathered into a better solution by cluster ensembles. supervised iterative Expectation-Maximization (EM) method can be initialized by cluster ensemble based strategy which will be examined in the paper. This will deliver better approximation of cluster parameters. Here definition of Land-cover classes is vital. Another method for producing suitable labeling model for each and every clustering of the consensus is introduced for cluster ensembles. The upgraded parameter set acquired from the EM step is trained by maximum likely-hood classifier to classify the rest of the pixels. The effect of data overlapping from several clusters can be reduced by the self learning classifier. Comparison is made on the performance of the clustering between the proposed method and individual clustering of the ensemble for medium resolution and a very high spatial resolution images.