一种基于小波变换图像去噪的方法

提出了一种基于图像软阈值小波变换的高斯白噪声消除法。该算法根据含噪声图的特点，把信号分成信号象素与可能噪声象素两类，对于可能是噪声的象素，采用图像的小波软阈值去噪方法进行滤波，而对信号象素不产生影响，且能保留更多的图像细节。文中也给出了标准中值滤波，自适应维纳滤波算法和小波软阈值去噪的算法进行比较实验，结果表明用小波软阈值去噪的算法处理高度污染高斯白噪声的图像能力明显强于标准中值滤波，稍微优于自适应维纳滤波算法，且能够比较好保留图像的细节部分。

Image De-Noising Based on the Wavelet Transform

In this paper a method for image de-noising is based on the wavelet transform by thresholding. Each pixel is classified to be signal pixel or possible noise pixel according to the character of gaussian white noise. One of the types may be noise pixels, the wavelet soft-thresholding is adopted to remove the noise. It will not influence the signals, and what's more it can keep more image details. Comparing experiments are done using standard median filtering, adaptive winner filtering and wavelet thresholding filtering algorithm in this paper respectively. Simulation results show that de-noising by wavelet thresholding outperforms the standard mediate filtering algorithm especially for highly gaussian white noise corrupted images. And it also does little better than adaptive winner filtering algorithm ,keeping the details of the image more.