一种基于修正Frost核的SAR图像斑点噪声抑制方法

SAR Image De-speckling Based on Modified Frost Filter

针对合成孔径雷达(SAR)图像的斑点噪声抑制问题,给出了一种新的基于斑点噪声统计特性的滤波方法。该方法是对Frost滤波核的修正,综合考虑滤波窗口内的中心像元以及其他像元的局部统计信息,克服了传统Frost滤波器由于等权运算所带来的边缘等结构信息的模糊;同时也弥补了增强型Frost滤波器保留过多“假细节”的不足。实验结果表明,该方法在噪声抑制和细节保持方面取得了较好的折衷。

In this paper, a new filter method is proposed for SAR image de\|speckling. As a modified version of Frost filter structure, the given method combines the statistical information from both the central pixel and other pixels within the filter window, therefore it not only avoids over-smoothing of SAR image structural information caused by Frost (filter’s) iso-weighting operation, but also amends the deficiency of “false edges” by enhanced Frost filter. The experimental results confirm the new filter method can efficiently reduce speckle while preserve useful image formation details.