多小波域上各向异性扩散在纹理图像去噪中的应用  

The application of the anisotropic diffusion in multi-wavelet domain for texture image denoising

作　　者：伍尤富[1] 

机构地区：[1]韶关学院物理与机电工程学院,广东韶关512005

出　　处：《电路与系统学报》2010年第3期63-67,共5页Journal of Circuits and Systems

摘　　要：去噪是图像处理中的一个重要技术,一般的去噪算法会造成图像边缘信息被平滑,为了有效地抑制噪声而同时又保护好边缘信息,在多小波变换的基础上,提出了一种新的去噪算法,它结合了多小波变化和各向异性扩散(P-M扩散)两者的优点,利用多小波变换把纹理图像分解为高频子带和低频子带,然后根据子带图的特点分别采用不同的各向异性扩散方法,实验结果表明,该算法去噪效果好,改善了图像的峰值信噪比(PSNR)和最小均方误差(MSE),同时更好地保留了图像的纹理和细节。Denoising is one of important technology in image processing, but image edge information might be smoothed with the common denoising methods. In order to reduce noise and protect the edge effectively at the same time, a new algorithm is proposed in the paper for texture image denoising which is based on an anisotropic diffusion in multi-wavelet transform domain. The denoising scheme combines the strongpoints of multi-wavelet transform and anisotropic diffusion（P-M diffusion）. The noised image is decomposed to the high frequency sub-image and the low frequency sub-image by multi-wavelet transform, then according to the characteristics of each sub-images, different ways of anisotropic diffusion are used to the image denoising. The experimental results show that the denoising effect of the presented algorithm is better, and this algorithm can improve the PSNR and the MSE of the image, and keep the textures and details of images better.

关 键 词：图像去噪 各向异性扩散 多小波变换 偏微分方程 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]