基于Kirsch方向模板的SAR相干斑噪声抑制方法

A method of SAR speckle suppression based on Kirsch direction template

提出一种抑制SAR相干斑噪声的加权平滑滤波方法。该方法利用 8个 3× 3的Kirsch方向模板分别与SAR图像进行卷积运算 ,取其中的最大值作为窗口中心像素的边缘强度值 ,中心像素灰度值的滤波输出为它的 8个邻域像素灰度值的加权平均 ,加权系数为每个像素的边缘强度值。给出了滤波核方程的推导过程 ,分析了滤波原理。将该方法用于SAR图像相干斑噪声抑制 ,并与Lee滤波器的去斑效果进行了比较。实验结果表明 ,该方法具有良好的抑制相干斑噪声效果和较强的边缘保持能力。

A weighted smoothing filtering method for SAR speckle suppression is presented. Eight 3×3 Kirsch direction templates are convoluted with a SAR image respectively and the maximum of convoluting results is used as the edge strength value of the central pixel in the window. The output gray level of the central pixel is the weighted average of gray levels of its eight neighboring pixels, and the weighting coefficients are corresponding edge strength values. The derivation of the filter's kernel equation is given, and the filtering principle is analyzed. Experiments are made on SAR images by using the method, and the results are compared with those of the Lee filter. Experimental results show that the proposed method can suppress speckle and preserve edges better.