高分辨率合成孔径雷达图像的Gamma分布下最大后验概率降斑算法

针对传统的Gamma分布下最大后验概率降斑算法不能有效保留均匀区域的点目标, 不能有效保留弱边缘以及不能有效滤除强边缘区域的斑点等问题, 提出了基于第二类统计量的先验参数估计的高分辨率合成孔径雷达图像Gamma 分布下最大后验概率降斑算法. 使用Mellin卷积和斑点的乘性模型, Gamma先验分布的参数可由观察图像的前两阶对数累积量精确估计.所提算法具有解析的滤波输出, 便于实现.农田和城区的高分辨率合成孔径雷达图像的降斑实验表明, 与传统的Gamma分布下最大后验概率降斑算法相比, 所提算法既能有效保留均匀区域的点目标, 又能有效保留弱边缘, 还能有效滤除强边缘区域的斑点.关键词：高分辨率合成孔径雷达图像Gamma分布下最大后验概率降斑算法第二类统计量对数累积量

Gamma-distributed maximum a posteriori despeckling algorithm of high-resolution synthetic aperture radar images

In order to solve the problem that the traditional Gamma-distributed maximum a posteriori despeckling algorithm cannot effectively preserve the point target in the homogeneous region, nor effectively keep the weak edge, and nor efficiently suppress the speckle in the strong edge region, the Gamma-distributed maximum a posteriori despeckling algorithm with prior parameter estimation based on second-kind statistics is proposed for high-resolution synthetic aperture radar images. Using the Mellin convolution and the multiplicative model of speckle, the parameters of the Gamma prior distribution are accurately estimated from the first two log-cumulants of the observed image. The proposed algorithm has the analytical filtering output, and it is easy to implement. Despeckling experiments on high-resolution synthetic aperture radar images of agricultural field and urban region demonstrate that compared with the traditional Gamma-distributed maximum a posteriori despeckling algorithm, the proposed one can effectively preserve the point target in the homogenous region, effectively retain the weak edge, and efficiently suppress the speckle in the strong edge region.
Keywords：high-resolution synthetic aperture radar imagesGamma-distributed maximum a posteriori despeckling algorithmsecond-kind statisticslog-cumulants