基于边缘检测的Contourlet变换图像去噪  

Image Denoising with Contourlet Transform Based on Edge Detection

作　　者：于梅[1] 殷兵[1] 梁栋[1] 王璐[1] 

机构地区：[1]安徽大学计算智能与信号处理教育部重点实验室

出　　处：《计算机技术与发展》2008年第6期104-106,110,共4页Computer Technology and Development

基　　金：国家自然科学基金资助项目(60772121);安徽省高等学校自然科学研究项目(2005KJ005ZD);安徽省自然科学基金资助项目(070412065);安徽大学211工程学术创新团队资助

摘　　要：为了实现较好保留图像边缘特征的有效去噪,提出了一种基于边缘检测的Contourlet变换去噪方法。该方法先用LOG算子提取图像边缘,进而在Contourlet变换域上对图像的边缘部分和非边缘部分分别选取不同阈值进行最佳软阈值去噪处理。实验表明,与采用Donoho软阈值的Contourlet变换去噪方法相比,该方法可有效地保留图像的边缘信息,达到了更好的去噪效果。The edge information of image used to be destroyed while denoising. A new method of image denoising using contourlet transform is proposed to maintain more image＇s edge information. In this way, the edge of image was detected with LOG operator firstly. Then, the image edge and non- edge character in Contourlet coefficients were dealt with an optimal soft thresholding algorithm through selecting different denoising thresholding. Experiment results show that, compared with the commonly- used Donoho soft thresholding Contourlet transform denoising method, this method can get better results in image denoising and keep image＇s edge information more effectivdy.

关 键 词：CONTOURLET变换 图像去噪 边缘检测 最佳软阅值 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]