摘要
提出一种基于KummerU分布的SAR图像统计建模方法。首先在SAR图像乘积模型框架下,假设SAR图像的相干斑和雷达截面积分量分别服从Gamma分布和Fisher分布,导出了SAR图像的KummerU分布。接着,基于Mellin变换得到了KummerU分布的对数累积量参数估计方法。最后,利用KummerU分布对TerraSAR-X和AIRSAR系统获取的不同场景的SAR图像数据进行统计建模。定性和定量分析的实验结果表明,与经典SAR图像分布模型相比,KummerU分布具有更优的SAR图像统计建模能力。
A statistical modeling method for SAR images is proposed based on the KummerU distribution. Firstly, under the framework of multiplicative model of SAR images, it is assumed that the speckle and the Radar Cross Section （RCS） factors of SAR images follow the Gamma and Fisher distributions respectively, and then the Kum- merU distribution for SAR images is derived. Secondly, based on the Mellin transform, the parameter estimation method for KummerU distribution using log-cumulants is presented. Finally, the KummerU distribution is employed to the statistical modeling of the SAR images of different scenes acquired by the TerraSAR-X and AIRSAR systems. The experimental results with qualitative and quantitative analysis demonstrate that the KummerU distribution is superior to the classical SAR image distributions for the statistical modeling of SAR images.
出处
《科学技术与工程》
北大核心
2016年第28期235-240,共6页
Science Technology and Engineering
