Fuzzy Active Contour Model With Markov Random Field for Change Detection

The traditional active contour models are sensitive to the speckle noise in synthetic aperture radar (SAR) images. In this paper, Markov random field (MRF) theory is incorporated into fuzzy model detect changes of multitemporal SAR proposed method, neighboring information considered modify pointwise prior probability for exploiting mutual and spatial information. addition, we incorporate MRF get resulting MRF-based energy function. Finally, drive associated first variation function compute membership. Due introduction MRF, robust images can achieve accurate change detection results. Experiments on four image datasets demonstrate that able accurately segment difference has better performance comparison with other techniques.