基于马尔科夫随机场的SAR图像超像素分割算法  

A Superpixel Algorithm for SAR Image Segmentation Based on Markov Random Field

作　　者：宋文青 董锦[1] 相飞[1] 张俊 SONG Wenqing;DONG Jin;XIANG Fei;ZHANG Jun(Xi'an Electronic Engineering Research Institute, Xi'an 710100)

机构地区：[1]西安电子工程研究所,西安710100

出　　处：《火控雷达技术》2020年第1期1-6,18,共7页Fire Control Radar Technology

摘　　要：针对SAR图像超像素分割易受图像固有相干斑噪声影响的问题,本文提出一种基于马尔科夫随机场(Markov Random Field,MRF)的超像素分割算法。算法中采用基于马尔科夫随机场的最大后验概率优化模型对SAR图像超像素分割问题进行建模。相较于传统SAR图像超像素提取方法,该方法通过对像素的分割标号在空间邻域上进行马尔科夫特性约束,进而有效抑制相干斑噪声的影响。实验结果表明,与传统方法相比,本文算法的噪声抑制效果更好,分割精度更高。A superpixel algorithm based on Markov Random Field(MRF)for SAR image segmentation is proposed aiming to reduce the influence of the inherent speckle noises on superpixel SAR image segmentation.A MRF-based maximum posterior probability optimization model is used to model superpixel SAR image segmentation.The proposed algorithm effectively suppresses the influence of the speckle noises by constraining the pixel labels within their spatial neighborhoods under the assumption of Markov property.The test results show that the proposed algorithm has better noise suppression effect and higher segmentation precision over traditional superpixel SAR image segmentation methods.

关 键 词：SAR图像 超像素分割 相干斑噪声 马尔科夫随机场 

分 类 号：TN958.94[电子电信—信号与信息处理]