一种新的自适应边缘保持滤波算法

在进行图像去噪时，为解决兼顾降低噪声和保留细节的难题，提出了一种新的自适应边缘保持滤波算法——区域灰度值最小方差的均值滤波算法．算法在图像去噪过程中利用邻域区间中各像素点之间的相互关联性，在8个方向上对像素点进行处理，具有较强的相关性和自适应性．利用该算法对随机噪声和椒盐噪声进行仿真去噪试验，结果表明，与已有的去噪算法相比，所提算法在去除噪声的同时较好地保持了图像的边缘、细节信息．

A New Adaptive Edge-Preserving Filtering Algorithm

In order to resolve a problem of how to give consideration to denosing and retaining details while image denosing, this paper presents a new adaptive edge-preserving filtering algorithm-mean filtering algorithm of least variance of area gray-scale value. This algorithm, which makes use of correlation among pixels in the process of image denoising, disposes the pixels in 8 directions. So this algorithm has stronger relativity and adaptivity. By using the algorithm to simulate random noise and pepper-sah noise, the results show that the algorithm retains edge details better than others in image denosing.