一种基于相关性和局部标准差的图像失焦模糊区域检测方法  

A METHOD FOR DETECTING IMAGE DEFOCUS BLURRED REGION BASED ON CORRELATION COEFFICIENT AND LOCAL STANDARD DEVIATION

作　　者：咸兆勇 甘金明[2] 玉振明[2] 李陶深[1] 

机构地区：[1]广西大学计算机与电子信息学院,广西南宁530004 [2]梧州学院,广西梧州543002

出　　处：《计算机应用与软件》2014年第9期198-200,275,共4页Computer Applications and Software

基　　金：国家自然科学基金项目(61062014)

摘　　要：针对局部失焦图像的模糊区域检测问题,提出一种基于相关性系数和局部标准差的简单有效的检测方法。该方法通过计算模糊图像中各像素的邻域在频域内的对数相关性系数以及在空域内的局部标准差作为描述该像素模糊程度的评价指标,将它们同预设的阈值进行比较以实现失焦模糊区域的检测。实验结果表明,该方法能有效检测出失焦模糊图像中的模糊区域。As for detecting the blurred region of partially defocus blurred images,we present a simple and efficient detection method whichis based on correlation coefficient and local standard deviation.Through computing the logarithmic correlation coefficients of the neighbour-hoods of pixels in blurred image in their frequency domains and the local standard deviation in spatial domain,the method uses them as theevaluation indices in describing blur magnitude of each pixel,and then compares them with the preset thresholds to realise the detection ofdefocus blurred region.Experimental results show that this method can detect the blurred regions in defocus blurred images efficiently.

关 键 词：失焦模糊 模糊特征 相关系数 标准差 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]