An Approach Research to Segmentation of the SAR Image Based on the Markov Random Field

摘要

Markov random field method is a very active research field in image segmentation. In this paper, a new approach to the segmentation of SAR target chip image based on the Markov random field (MRF) and Maximum a Posteriori (MAP) rule is proposed. By statistical analysis of the different regions in the SAR target chip image, the valid segmentation algorithm for choosing the iteration initial value is obtained. Compared with other segmentation algorithms such as Simulate Annealing algorithm and the Iterative Conditional Model (ICM) algorithm by the synthetic image and operational SAR image, the proposed approach in the paper can accelerate the convergence speed, reduce the iterative numbers and improve the segmentation speed.