一种异性扩散的自适应平滑滤波算法

固定尺度的高斯滤波器对图像中的所有像素使用相同的加权系数，在平滑噪声的同时也模糊了边缘信息，影响了后期的边缘定位和提取，为此提出了一种异性扩散的自适应仿高斯滤波器，通过计算当前像素与周围像素的灰度梯度，预判该像素类型（噪声点或是边缘点），并根据像素点类型和梯度生成自适应的、各向异性的加权系数，通过多次平滑迭代实现权重的异性扩散。实验结果表明，该滤波算法具有平滑噪声的同时保留并锐化边缘效果。

Anisotropic diffusion and adaptive smoothing algorithm

Gaussian filter with fixed scale-parameter uses the fixed coefficients for every pixel to smooth image. The noise is smoothed out while the edge is distorted. The defect will influence the edge detection and extraction later. Therefore a kind of anisotropic diffusion smooth algorithm is proposed in this paper. In the algorithm, the type（noise or edge） of current pixel is predicted by calculating the gray-gradients of the pixels around the pixel, and according to the pixel type and gradients, the unique and anisotropic coefficients for the pixel are generated. The coefficients anisotropic diffusion is achieved by some times iteration. The experimental results show that, the algorithm can smooth noise while preserving and sharpening the edge.