基于双正交小波域的SAR图像滤波方法

将E-lee、E-Kuan、GammaMap、wiener等经典滤波算法和双正交小波变换相结合,提出了基于双正交小波变换域的局部统计特性SAR图像滤波方法,同时提出了一个运算量少,且是归一化的对数变换,它将乘性的Speckle噪声转为加性噪声。在小波域内建立了局部统计特性SAR图像滤波算法,使用多分辨率的手段,因为在每个方向上的小波系数都具有相同的特征,可以很好地处理图像的一些特性,使得图像边缘被模糊的相对少些。实验结果表明,此方法比经典算法的效果要好。

SAR Image Filtering Based on Biorthogonal Wavelet Transform

In combination of biorthogonal wavelet transform with the such classical filtering methods, as E-Lee, E-Kuan, Gamma Map and wiener, a local statistical SAR image filtering based on biorthogonal wavelet transform is proposed. At the same time, a normalized logarithmic transform is suggested, which could reduce the computation complexity, and transform the multiplicative Speckle noise into additive noise. Then the local statistical SAR image filtering is established in wavelet domain by using multi-resolution means. Some characteristics of the image could be easily treated because the wavelet coefficients in each direction have the same characteristics, thus resulting in correspondingly little fuzzy edge of image. The experiments indicate that this method is better than the classical filtering methods.