基于整数DCT和块匹配的彩色图像去噪  

Image denoising based on integer DCT and block-matching

作　　者：姚乃萌 薛操[1] 狄明[1,2] 

机构地区：[1]天津大学电子信息工程学院,天津300072 [2]天津泰达有线网络电视有限公司,天津300072

出　　处：《信息技术》2015年第10期97-99,共3页Information Technology

摘　　要：降噪研究是图像和视频处理方面一个传统的研究领域,不同于传统的空域去噪,频域去噪算法取得了显著成就。文中提出基于块匹配的频域彩色图像降噪方法。首先对图像进行块匹配,完成后采用整型DCT变换对空域数据进行处理,实行频域去噪后进行整型DCT反变换得到空域结果。该算法测试整型DCT对降噪方法的影响,并且提高整体降噪方法的处理时间。文中提出的降噪方法在提高效率的同时,取得了同样优秀的降噪成果,在彩色图像处理方面取得显著效果。Image denoising is a traditional point in the field of image processing and video processing,and despite the traditional denoising in spatial domain,denoising methods in frequency domain have achieved significant results. The paper proposed a color image denosing method in spatial domain based on block-matching in frequency domain. Applying integer DCT on achieved matched blocks after blocking-matching is prepared for denoising in frequency. After denoising,through integer inverse DCT the results in spatial domain are presented. The method proposed gets efficient processing and achieves equal denoising results,which performs well on color images denoising.

关 键 词：彩色图像处理 整型DCT 块匹配 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]