SAR image segmentation based on multi-scale Markov random field
一种基于多尺度马尔可夫随机场的SAR图像分割算法

The multi-scale edge information was detected by useing module maxima of wavelet-transform.Edge position could be precisely localized due to the conception of clique introduced to MRF model.Then the SAR image was devided by useing Markov Random Field(MRF) model.Experiment results indicate that the method can improve the quality of image segmentation.