基于NSCT阈值和自适应扩散的图像降噪  

Image Denoising Based on NSCT Threshold and Adaptive Diffusion

作　　者：陈建军[1] 田逢春[1] 李灿[1] 

机构地区：[1]重庆大学通信工程学院,重庆400044

出　　处：《计算机工程》2011年第12期204-205,共2页Computer Engineering

基　　金：重庆市自然科学基金资助项目(2009BB2359);中国博士后科学基金资助项目(20080430096)

摘　　要：提出一种基于非下采样Contourlet变换(NSCT)阈值和P-Laplace扩散的图像降噪方法。通过NSCT阈值的平移不变性得到初步降噪图像,保留阈值变换后需要置0的高频部分,利用P-Laplace算子对其进行扩散,并将得到的扩散图像融合到阈值降噪图像中,从而提高降噪图像的质量。数值结果表明,该方法能有效保持原图像的纹理细节,减少图像降噪的吉普斯震荡现象。This paper presents a image denoising method based on Non Subsampled Contourlet Transform（NSCT） threshold and P-Laplace diffusion. The denoising image by threshold method of NSCT translational invariant is obtained. The NSCT coefficients which are set to zero by the threshold procedure are retained. Then the P-Laplace diffusion method is applied directly with the reconstruction image from the retained NSCT coefficients. The diffusion image is fused with the denoising image which by threshold method. Numerical results show that this method can eff＇ectively keep the original image texture details, and reduce the Gibbs shaken phenomenon of image noise.

关 键 词：图像降噪 非下采样CONTOURLET变换 非线性 P—Laplace扩散 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]