Combination of fuzzy clustering algorithms for change detection in remote sensing images

Here we propose a methodology to combine the output of fuzzy clusterings to detect changes in remote sensing images. In this regard we select two fuzzy clustering algorithms, namely fuzzy c-means (FCM) and Gustafson Kessel clustering (GKC). For clustering purpose various image features are extracted using the neighborhood information of pixels from the difference image (DI). To assign a pixel-pattern to either of the two groups (for changed and unchanged regions of the DI) maximum of the two membership-values (given by FCM and by GKC for the same pattern for the same cluster) is considered. It has been observed experimentally that the changesare detected more efficiently using the proposed ensemble-based procedure. To show the effectiveness of the proposed technique, experiments are conducted on two multispectral and multitemporal remote sensing images. Results are compared with those of existing stand-alone fuzzy clustering based techniques, Markov random field (MRF) & neural network based algorithms and found to be superior.