常用图像去噪方法探析与比较

摘要

均值滤波、中值滤波、维纳滤波和基于小波变换的软阈值法和硬阈值法是常用的几种图像去噪方法，通过对添加了不同噪声的图像进行了仿真实验。实验结果进一步验证了均值滤波在去除噪声的同时造成了图像边缘一定程度的模糊，中值滤波能相对较好的保留了图像的边缘，维纳滤波对高斯白噪声去噪效果较好但计算量大，相比较小波阈值法去噪是目前较好的的去噪方法，通过选取合适的阚值可以得到较为理想的去噪效果。%Average filter, median filter, Wiener filter, and soft and hard threshold method based on wavelet transform are most commonly of several denoising methods . A simulated experiment was carried by adding different noise images. The experimental results demonstrate that the average filter can remove noise but it cause the image edge blurred in a certain degree,the median filter can give a relatively fine retained image edge, and the Wiener filter require much of calculation. Comparatively the threshold denoising in wavelet domain is a better denoising method at present. It can be obtained an ideal effect in denoising by choosing appropriate threshold val- ues.