运动模糊图像边缘振铃效应的抑制方法  

Algorithm on border ringing effects reduction of motion-blurred image

作　　者：王靖[1] 施刚[1] 张磊[1] 韩斌[1] 黄子娟[1] 

机构地区：[1]南京人口管理干部学院信息科学系,南京210042

出　　处：《电子测量技术》2013年第5期62-67,共6页Electronic Measurement Technology

基　　金：江苏省高校自然科学研究(09KJD510006);江苏省大学生创新训练计划(2011JSCXXL12)资助项目

摘　　要：边缘振铃效应是影响图像复原效果的重要因素之一,为此提出一种基于边缘恢复的运动模糊图像复原算法。首先将原运动模糊图像的边缘按图像点扩散函数在其扩散方向上移动的像素值进行二维边缘延拓,形成一个新的延拓图像;然后用最优窗函数对新图像进行加窗预处理;最后用维纳滤波算法对加窗处理结果进行复原。实验结果表明,提出的算法对边缘区域梯度变化较大的运动模糊图像的复原效果较好,可以有效地抑制复原图像边界处的振铃效应。Border ringing effect is one of the important factors that impact image restoration quality,so a motion-blurred image restoration approach based on image borders is proposed. Firstly, the borders of an original motion-blurred image are extended according to the number of pixels in the spreading directions of the two dimensional point spread function of a motion-blurred image to a new extensional image. Furthermore,an optimal window function is applied to the image. Then,the windowed image is restored with Wiener filtering algorithm. Experimental results show that the proposed restoration approach has an obvious effect on the motion blurred image with shape gradient changes on the area near to image edges,which can effectively suppress the ringing effects on the image edge area.

关 键 词：运动模糊图像 维纳滤波 振铃效应 边缘延拓 最优窗预处理 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]