一种高效的小波变换去噪方法

A High-powered Image Denoising Method Based on Wavelet Transform

小波变换用于图像去噪的思想已经提出了很久,然而前人所提出的这种方法对于去噪的效果并不理想。图 像经这种小波变换去噪后,纹理特征被弱化,图像的边缘出现较明显的Gibbs效应,图像变模糊。针对以上问题,本文提 出了一种高效的小波变换去噪方法(HPID)。此去噪方法是基于小波变换的新方法,与经典的小波去噪方法不同,该方法不 依赖图像大小来判定去噪门限,不需方差信息,且适用于不同类型噪声。采用本方法处理的噪声图像与经典方法相比,不 仅消除了Gibbs效应,而且图像的边缘信息更清晰,纹理特征增强,去噪能力得到改善。

Image denoising with wavelet transform has come into use for a long time. However, the effect of this method is not as good as we desired in image denoising. With this method, texture characters of images are weakening, Gibbs Effect appears at edges of images and images is blurred. To solve these problems, a high-powered image denoising method based on wavelet transform is proposed, called HPID. This method is a new one based on wavelet transform. In this method, unlike classical wavelet denoising method, threshold doesn＇t depend on size of images and deviation is no more required. Moreover it is the same with all kinds of noise. Comparing with classical wavelet denoising method, HPID eliminates Gibbs effect, makes edges more clear, enhances texture characters and improve the ability of denoising.