一种新的小波图像去噪方法

小波图像去噪已经成为目前图像去噪的主要方法之一，目前的研究主要集中于如何选取阈值使去噪达到较好的效果。边缘信息是图像最为有用的高频信息，在图像去噪的同时，应尽量保留图像的边缘信息，基于这一思想，提出一种新的小波图像去噪方法。用数学形态学算子对图像小波变换后的小波系数进行处理，以去除具有较小支持域的噪声，保留具有连续支持域的边缘。实验结果表明，与普通的小波阈值去噪方法相比，该方法不但可以保留图像的边缘信息，而且能提高去噪后图像的峰值信噪比2～5dB，提高信噪比6～10dB。

Novel wavelet image denoising method

Wavelet image denoising has been well acknowledged as an important method of image denoising. Recently, most of the research efforts about wavelet image denoising focus on how to select the threshold. Edge information is the most primary high frequency information of an image. Therefore maintenance of more edge information is more important in the denoising process. According to the idea of this paper, an image denoising method based on wavelet transform and mathematic morphology is proposed. The coefficients of wavelet transform of image are manipulated by morphological operator, so as to remove the noise whose support are small area or no area at all and preserve the edge and the small features whose support are large or consecutive area. The experimental results show that, compared with the usually used wavelet threshold denoising method and median filtered method, the proposed method can keep images edges from damaging and increase PSNR to 3～5 dB and SNR to 6～10 dB respectively.