基于Richardson-Lucy的图像去模糊新算法  

Novel image deblurring algorithm based on Richardson-Lucy

作　　者：赵博[1] 张文生[1] 丁欢[1] 

机构地区：[1]中国科学院自动化研究所,北京100190

出　　处：《计算机工程与应用》2011年第34期1-4,77,共5页Computer Engineering and Applications

基　　金：国家自然科学基金(No.90924026;No.81071141);国家基础研究"十二五"规划项目(No.0101050302)~~

摘　　要：针对传统的Richardson-Lucy(RL)算法存在振铃效应和噪声放大等问题,提出了一种基于RL的图像去模糊新算法,通过在迭代过程中引入增益图(GainMap),降低了复原图像的对比度,阻止迭代过程振铃效应的传播和噪声的进一步放大。详细讨论了引入的增益图的不同参数选择对去模糊结果的影响,通过选取合理的参数得到最佳的结果。实验结果验证了该算法抑制振铃效应的有效性,同时它很好地保留了图像细节,对含有噪声的模糊图像也有很好的复原效果。A novel image deblurring algorithm based on Richardson-Lucy（RL） is proposed to overcome ringing artifacts and noise amplification problems of original Richardson-Lucy algorithm.The Gain Map is introduced into the iterations of RL,thus the contrast of the restored image can be reduced.Consequently,the proposed method can prevent propagation of ringing artifacts and further amplification of image noise.Influences on the deblurring results by different choice of Gain Map parameters are discussed.It chooses reasonable parameters to achieve the best results.Experimental results demonstrate the effective of the proposed algorithm for suppressing ringing artifacts and retaining detail information.Moreover,it can obtain notable improvements for noisy blurred images using this method.

关 键 词：图像复原 去模糊 Richardson-Lucy(RL)算法 增益图 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]