Skip to Main Content
In this paper, a multiscale Markov random field method for segmentation of the synthetic aperture radar (SAR) images is proposed. A classifier which inherits the strongpoint of the Markov random field (MRF) and the multiscale autoregressive (MAR) model is designed. The MAR models are utilized to extract the multiscale feature of SAR image, which is used to train the MRF with the proposed algorithm, and then the SAR images is segmented by the trained random field. The experimental result demonstrates the effectiveness and efficiency of the proposed method.