混合Gamma拖尾Rayleigh分布的高分辨SAR图像建模  

SAR Image Modeling Based on the Mixture Gamma Heavy-tailed Rayleigh Distribution

作　　者：王灿[1] 苏卫民[1] 顾红[1] 邵华[1] 

机构地区：[1]南京理工大学电光学院,江苏南京210094

出　　处：《测绘学报》2014年第2期151-157,共7页Acta Geodaetica et Cartographica Sinica

基　　金：总装备部预研部基金(9140A13030211BQ02);教育部博士点基金(20113219110018);南京理工大学研究基金(2010ZDJH05;2011ZDJH13);江苏省创新计划(CXLX11_0252)

摘　　要：提出一种基于乘积模型的统计模型,称为混合Gamma拖尾Rayleigh分布模型。在该模型中,利用拖尾Rayleigh分布对相干斑进行建模,使模型可以精确地拟合高分辨率合成孔径雷达SAR图像相干斑的尖峰和拖尾的特征;同时引入混合Gamma分布对高分辨SAR图像雷达散射截面积(radar cross section,RCS)复杂起伏特性进行表征。基于Mellin变换,推导出混合Gamma拖尾Rayleigh分布对数累计量参数估计公式,提高了参数估计精度,从而实现了对高分辨率合成孔径雷达SAR图像的精确建模。最后通过真实SAR图像对本文提出的模型与已有模型进行比较。试验结果表明,本文提出的模型能够对不同的高分辨率合成孔径雷达SAR图像进行统计建模,并且具有较高的拟合精度。A product based model called mixture Gamma heavy-tailed Rayleigh distribution is proposed. By using the Heavy-tailed Rayleigh density, it modeled the speckle to describe the characteristic of sharp peak and heavy toil of the high resolution SAR image. By introducing the mixture Gamma density, it choracterized the complex fluctuation characteristics of the RCS of the high resolution SAR image. For the sake of accurate modeling of the high resolution SAR image, it derived mixture Gamma heavy Rayleigh distribution Iog-cumulant Parameter estimation formula based on Mellin transformation to improve the accuracy of parameter estimation. Furthermore, experimental results on several actual SAR images ore given and the method preposed is compared with some existing methods. Experimentol results demonstrate that the modeling proposed in this paper is available for various high resolution SAR images with relatively high fitting accuracy.

关 键 词：合成孔径雷达图像 乘积模型 拖尾Rayleigh分布 混合Gamma分布 Mellin变换 

分 类 号：P236[天文地球—摄影测量与遥感]