基于边缘检测的自适应降噪处理方法

提出了一种分区处理的降噪方法,对图像边缘和非边缘区域分别采用自适应中值滤波和均值滤波的方法进行处理.论及的噪声区分高斯噪声和椒盐噪声两种,对含有混合噪声的图像首先滤除椒盐噪声,然后标定图像的边缘细节,在保留图像细节的前提下充分降低噪声.测试结果表明本算法有效降低噪声,改善了图像视觉效果,提高视频编码中压缩效率.

A Method of Adaptive Noise Reduction Based on Edge Detection

Mean filter and Median filter have different effect to Gaussian and pepper-and-salt noise respectively. It is difficult to find the balance of reducing mixed noise and keep details by using mono-filter. A method was proposed to keep the details of image and to reduce mixed noise as much as possible based on different zones with different filter. First, reduce pepper-salt noise with adaptive median filter, then mark the edge and non-edge area of the image. Finally, process the edge area with adaptive median filter while non-edge area with mean filter. The test results show that the algorithm efficiently reduces noise, improves the visual images and enhances the compression efficiency in video coding.