多视SAR图像K分布特征函数与统计量研究

K分布是一种广泛应用于描述SAR图像杂波统计特性的概率模型，然而对于其特征函数及各类统计量的研究却少见于报道。本文基于Fourier变换和Mellin变换，详细推导了多视K分布第一类和第二类第一、第二特征函数的解析表达式，获得了原点矩、累积量、对数矩和对数累计量的计算公式，在此基础上给出了基于原点矩的参数估计器和基于对数累积量的参数估计器。最后利用Kendall近似展开法，理论分析了两种估计器的正规化偏差与方差。结果表明，对于小样本不均匀SAR图像场景，对数累积量估计器的性能明显优于矩估计器。 

Characteristic Functions and Statistics of K-Distribution in Multi-look SAR Image

K-distribution has been proved to be a good model for describing the statistical properties of clutter in multi-look synthetic aperture radar (SAR) images. However, the form of characteristic functions and the statistics are rare in reports. The close form s of the first and second characteristic function of the first and second kind are derived using the Fourier and Mellin transform, respectively. The formulas of moments, cumulants, log-moments and log-cumulants are also deduced, and also the method of log-cumulants (MOLC) and the method of moments (MOM) parameter estimators. Finally, the normalized biases and variances are calculated through Kendall's approach of approximate expansion. Theoretical analysis results show that MOLC has better performance in comparison with MOM when the true order parameter is small and the number of SAR imagery samples is relatively low.