改进的Hough变换实现圆检测  

Circle detection through improved Hough transform

作　　者：段黎明[1,2] 汪威[1,2] 张霞[1,2] 

机构地区：[1]重庆大学光电技术及系统教育部重点实验室ICT研究中心,重庆400030 [2]重庆大学机械工程学院,重庆400030

出　　处：《计算机集成制造系统》2013年第9期2148-2152,共5页Computer Integrated Manufacturing Systems

基　　金：重庆市科技攻关资助项目(CSTC2009AC3047)~~

摘　　要：为了自动检测出图像中的圆并精确测量其参数,提出一种改进的点Hough变换圆检测方法。首先对提取边缘后的轮廓进行一次筛选,得到所有的连续边缘轮廓;然后进行二次筛选,排除那些明显不可能为圆的图形,获得最终候选边缘轮廓。对候选边缘点组进行快速点Hough变换圆检测,计算圆的直径值和圆心坐标值,在检测过程中,采用自适应的点选择步长值和累加器判断阈值。利用VC++6.0开发了圆检测系统实验软件,并进行了对比实验,结果表明该方法可以有效解决原算法中固定值导致检测精度不高、误检测和漏检测的问题。To detect circles automatically and to obtain the parameters accurately in images,an improved point Hough transform was proposed.All continuous contours were selected through screening contours of edge exaction.After the second screening,the obvious impossible images were excluded and the final candidate-edge contour was obtained.Point Hough transform was operated in candidate-edge point groups to obtain the diameters and circle center coordinate values.Adaptive point-selected step value and judge threshold for radius accumulator were used in the proposed method.The experiment software for detecting circles was developed by VC＋ ＋ 6.0,and the comparative experiment showed that the proposed method could solve problems of low precision,error-detection and missing-detection.

关 键 词：圆 筛选 自适应 点Hough变换 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]