基于图像边缘信息的多小波阈值去噪方法

Image Denoising in Wavelet Multi-thresholding by Edge Information

基于小波变换的图像去噪方法是小波应用较成功的一个方面,阈值大小的确定是该方法最终去噪效果好坏的一个决定性因素。基于图像边缘信息的多小波阈值去噪方法充分研究了信号与噪声在小波变换各分解层上的不同传播特性,在保留代表边缘信息的小波系数的基础上,对不同方向、不同分解层的小波系数分别选取最佳阈值处理。与Donoho等人提出的Visu shrink去噪方法相比,此方法提高了去噪后图像的峰值信噪比(PSNR),使图像更加清晰,去噪效果更好。

Image denoising via wavelet transform is a success for wavelet applications, where the most important case is how to obtain the optimal threshold. The selection of threshold decides the result of denoising directly. This paper proposes an image denoising in wavelet Multi-thresholding by edge information to select the optimal threshold based on different subbands and orientations. This method selects the threshold and keeps wavelet coefficients of edge inforemation through researching on the different characteristics of signal and noise wavelet coefficients. Compared with Donoho＇ s Visu shrink this method improves the peak signal-to-noise ratio, makes denoised image more clear, and thus acquires better denosing result.