论文部分内容阅读
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 statistical model based on product model is proposed, which is called hybrid Gamma-tail Rayleigh distribution model. In this model, the trailing Rayleigh distribution is used to model the speckle, which makes the model fit the spikes and trailing features of the speckle of SAR images accurately. The mixed Gamma distribution is also used to distinguish the high resolution The complex ups and downs of radar cross section (RCS) of SAR images were characterized. Based on the Mellin transform, the parameter estimation formula of logarithmic cumulant of hybrid Gamma-tailing Rayleigh distribution is deduced, which improves the accuracy of parameter estimation and realizes accurate modeling of SAR image of high-resolution synthetic aperture radar. Finally, the real SAR image is compared with the existing model. The experimental results show that the proposed model can statistically model different high-resolution SAR images with high fitting accuracy.