基于能量谱函数相关系数的极化SAR数据滤波方法研究

摘要

有效的极化SAR相干斑滤波器需要在抑制噪声的同时避免图像细节信息的模糊.然而固定大小窗口的滤波器很难做到两方面都好的效果.由于极化总功率(Span)图像可以有效地反映地物极化散射特征的变化,提出根据Span图像局部能量谱函数的相关系数值自适应地选择窗口大小,能够在均质地区选择较大窗口而在异质地区选择较小窗口,再利用方向窗选择像素进行最小均方误差(MMSE)滤波.实验结果表明,该方法能在滤除噪声的同时,有效地保持对地物细节的分辨能力,为极化SAR数据的后期处理和应用打下基础.%Effective PolSAR filters should not only inhibit speckles, but also preserve resolution of SAR image.However, it is difficult to achieve effectiveness in bother aspects if a fixed size filtering window is used. The energy spectrum function of the total polarimetric power ( Span) image is used to calculate homogeneity of targets and thus choose filtering window size adaptively. Then edge-aligned window is used and minimum mean square error ( MMSE) is applied. According to experimental results, the proposed filter is effective in retaining SAR resolution while reducing speckles, which lays a sound foundation for post-processing of PolSAR data.