基于混合Gamma建模与MRF的SAR图像分割方法

SAR image segmentation by combining gamma mixture model and MRF model

文中基于结合利用SAR图像统计信息和像素的空间约束信息的思路,提出了基于混合Gamma建模和MRF的分割方法,算法中利用混合Gamma模型对SAR图像进行统计建模,利用MRF模型对像素空间相关性进行建模。文中验证了混合Gamma模型参数估计的准确性以及自适应估计分割类别数目的有效性。最后用模拟SAR数据和实测SAR数据验证了文中所提出的分割算法的有效性。

Based on the idea of using statistic information of SAR image and neighborhood information of pixels,a method is proposed for segmenting SAR images,which the Gamma mixture model of statistic information is integrated with Markov random field model of spatial correlation. Then,the accuracy of estimating parameters of Gamma mixture model and the effectiveness of adaptive estimating number of segmentation categories are verified. Finally,the experiments with simulated and real SAR data indicate the algorithm is effective.