基于高阶统计量的小波变换去噪算法

HOS-based Wavelet Denoising Algorithm

图像在获取和传输的过程中经常要受到噪声的污染。传统的去噪方法不仅滤出了图像的噪声,同时使图像细节变得模糊。本文提出一种基于双谱和小波变换的去噪算法。该方法是根据高斯噪声及椒盐噪声在小波变换下的不同特征,并结合双谱滤波、中值滤波的特点,在小波域内对高频子带进行双谱滤波,去除图像中的高斯噪声,然后进行中值滤波,去除图像中的椒盐噪声。高斯噪声的双谱为零,能够彻底的去除高斯噪声。该算法的实验结果表明不仅能滤出图像中高斯噪声和椒盐噪声,而且能较好的保留图像的边缘细节,其滤波效果优于传统的图像去噪方法。

The image is often corrupted by noise in its acquisition or transmission. The traditional denoise before analysis will cause the image detail change fuzzy when filter out image noise. In the paper, an efficient technique based on Bispectral and Wavelet transform is proposed. According to the characteristic of ＇ Gaussian＇ noise and ‘salt and pepper＇ noise under wavelet transform, the high-frequency subband images is filtered with the property of Bispectral filter, this remove the ＇ Gaussian＇ noise. Then the subband images is denoised by using median filter to remove the ‘salt and pepper＇ noise. The experiment results show that the method not only efficiently remove the mixed ＇salt and pepper＇ noise and ‘Gaussian＇ noise in image, but also preserve the image edge information. The filtering performance is better than that by traditional denoising method.