SAR图像各向异性扩散滤波算法

为了抑制SAR图像中固有的乘性相干斑噪声，本文提出了一种基于各向异性扩散滤波的新算法。该算法首先将自蛇扩散引入经典的SRAD与DPAD算法各向异性扩散方程中，形成了一种带边缘增强功能的各向异性扩散新方程；然后利用改进Frost滤波系数作为扩散方程的新扩散函数，形成了一种既能有效抑制相干斑，又能较好保护边缘、抑制块效应现象的新扩散函数。实验表明，新扩散滤波算法在相干斑抑制、边缘保护、块效应抑制方面均比传统各向异性扩散滤波算法更优异。

Anisotropic diffusion filtering algorithm for SAR images

A new anisotropic diffusion filtering algorithm is proposed to reduce multiplicative speckle noise in SAR images. A new anisotropic diffusion equation that can strengthen edge is established by introducing the self snake diffusion into the diffusion equation of the classic SRAD and DPAD algorithms. Then, a new diffusion function that is estimated by improved Frost filtering coefficients is introduced into the new anisotropic diffusion equation, and it can effectively smooth speckle noise, preserve edges, and reduce blocking artifacts. The experimental results show that the proposed algorithm can outperform classic anisotropic diffusion equation in despeckling, preserving edges and reducing blocking artifacts.