A novel matting algorithm applied to SAS image segmentation

Poisson´ equation has been used to mat in digital image widely, but there is less report using it in acoustic image, such as SAS image. We proposed a novel matting algorithm of Poisson´ equation applied to SAS image, in the proposed algorithm, the statistics of image were studied firstly and then Markov Random Field (MRF) theory was applied to initialized the image which can mark the interest regions, at last Poisson´ equation is used to mat in the region. The simulation results have proved the new algorithm could be used in the real SAS images.