using contextual information for unsupervised change detection using multitempolar sar images based on clustering and level set methods

in this research, the framework is presented for unsupervised change detection using multitemporal sar images based on integration clustering and level set methods. spatial correlation between pixels were considered by using contextual information. also as proposed method was used integration of gustafson-kessel clustering techniques (gkc) and level set methods for change detection. using clustering techniques has caused production of initial curve for lsm with minimum convergence time and also use of lsm leads to increased accuracy of change map using iterative process. the results show that overall error rate of the proposed method reduced than otsu thresholding, gkc and fcm clustering, and em algorithm as much as 0.62, 1.78, 1.70 and 4.48. results showed that the high capability of the proposed method in unsupervised change detection of multitemporal satellite sar images.