SAR Image Segmentation Based on Diffusion Equations and MRF
Jia Ya-fei①② Zhao Feng-jun② Yu Wei-dong① Ai Jia-qiu①②
①(Institute of Electronics, Chinese Academy of Sciences, Beijing 100190, China) ②(Graduate University of the Chinese Academy of Sciences, Beijing 100039, China)
Abstract：A novel image diffusion equations and Markov Random Field (MRF)-based segmentation method for SAR images is proposed. The method involves the image diffusion based on the traditional MRF method. It is used to smooth the noise components in the SAR image, protect the desired object boundaries, and reduce the iterative times. The input SAR image is initially smoothed. Estimates of posterior probabilities are then obtained using MRF. The posterior probability is also smoothed using diffusion. Comparing with the traditional MRF method, this method wipes off the inaccuracy classified blocks, and reduces the cost of time.
2011, Vol. 33 
