基于Contourlet域持续性和聚集性的合成孔径雷达图像融合分割算法  

Fusion Segmentation Algorithm for SAR Images Based on the Persistence and Clustering in the Contourlet Domain

作　　者：吴艳[1] 肖平 王昌明[1] 李明[3] 

机构地区：[1]西安电子科技大学电子工程学院,陕西西安710071 [2]陕西测绘局遥感院,陕西西安710054 [3]西安电子科技大学雷达信号处理国家重点实验室,陕西西安710071

出　　处：《光学学报》2010年第7期1977-1983,共7页Acta Optica Sinica

基　　金：国家自然科学基金(60872137);国防预研基金(914CDO10307100C01);航空科学基金(20080181002)资助课题

摘　　要：针对合成孔径雷达(SAR)图像含有大量斑点噪声的特点,基于Contourlet的多尺度、局部化、方向性和各向异性等优点,并结合隐马尔科夫树(HMT)模型和隐马尔科夫场(MRF),提出了一种基于Contourlet域持续性和聚集性的SAR图像模糊融合分割算法。该算法有效捕获了Contourlet子带的持续性和聚集性,并分别用HMT和MRF来刻画,再依据模糊测度,将多尺度HMT和MRF有机融合,建立Contourlet域HMT-MRF融合模型,并导出新模型下的最大后验概率(MAP)分割公式。对实测SAR图像进行了仿真,仿真结果和分析表明:与小波域上的HMT-MRF融合分割及Contourlet域上HMT和MRF分割算法相比,该算法在抑制斑点噪声的同时,有效地提高了SAR图像的分割精度。In view of the speckle noise in the synthetic aperture radar（SAR） images,and based on the Contourlet′s advantages of multiscale,localization,directionality,and anisotropy,a new SAR image fusion segmentation algorithm based on the persistence and clustering in the Contourlet domain is proposed.The algorithm captures the persistence and clustering of the Contourlet transform,which is modeled by hidden Markov tree（HMT） and Markov random field（MRF）,respectively.Then,these two models are fused by fuzzy logic,resulting in a Contourlet domain HMT-MRF fusion model.Finally,the maximum a posterior（MAP） segmentation equation for the new fusion model is deduced.The algorithm is used to emulate the real SAR images.Simulation results and analysis indicate that the proposed algorithm effectively reduces the influence of multiplicative speckle noise,improves the segmentation accuracy and provides a better visual quality for SAR images over the algorithms based on HMT-MRF in the wavelet domain,HMT and MRF in the Contourlet domain,respectly.

关 键 词：图像处理 合成孔径雷达图像分割 CONTOURLET变换 持续性 聚集性 模糊逻辑融合 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]