结合Contourlet变换的各向异性扩散图像去噪  

Image denoising algorithm integrating anisotropic diffusion with contourlet transform

作　　者：丁亮[1,2] 汪烈军[1] 贾振红[1] 

机构地区：[1]新疆大学信息科学与工程学院 [2]69018部队

出　　处：《激光杂志》2013年第3期16-18,共3页Laser Journal

基　　金：国家自然科学基金项目(项目编号:61261036)

摘　　要：提出了一种结合Contourlet变换的各向异性扩散图像去噪算法。结合Contourlet变换和各向异性扩散的优点,通过Contourlet变换得到图像不同尺度的Contourlet系数矩阵,实现了建立在对图像多尺度分析基础上新的各向异性扩散算法。实验表明,新算法能够获得良好的视觉效果并且有效地提高了去噪图像的PSNR值,同时避免了传统的各向异性扩散出现的阶梯效应,更好地保留了图像的纹理和细节。An image denoising algorithm integrating anisotropie diffusion with Contourlet transform was propesed which integrated the advantage of Contouflet transform with anisotropie diffusion. Different scale matrixes of Contourlet eoefficient of the image can be acquired through Contourlet transform, therefore, the improved image denoising algorithm made it possible to realize the new anisotropie diffusion denoising based on multi - scale analysis of the image. The experiment results have demonstrated that the proposed method can significantly improve the PSNR value, visual effect and avert the staircasing effect in the traditional anisotropie diffusion effectively while preserving image textures and details.

关 键 词：图像去噪算法 CONTOURLET变换 各向异性扩散 多尺度分析 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TN248.1[自动化与计算机技术—计算机科学与技术]