改进的Hough变换检测圆方法①.PDFVIP

2015 年 第 24 卷 第 8 期 计 算 机 系 统 应 用 ① 改进的Hough 变换检测圆方法 1 1,2 2 陈小艳 , 王 强 , 李柏林 1(西南交通大学 机械工程学院, 成都 610031) 2(成都工业学院 机械系, 成都 610031) 摘 要: 计算机视觉目前已经得到广泛的应用, 通过图像检测圆在工程应用中具有重要的意义. 为了在实时系统 中自动检测图像中的圆并且精确得到圆心和半径, 提出了一种改进的Hough 变换检测圆的方法. 先对图像边界进 行过滤, 得到连续边界, 然后对每条连续边界进行排序并等分, 选择三点进行圆心和半径计算, 将传统 Hough 变 换检测圆的三维参数降到一维参数来统计, 在降低了计算时间和内存资源占用的同时, 提高了检测圆的准确性. 实验表明, 该方法能够满足实际检测的需要. 同时, 在检测时间上, 可以满足实时性的要求. 关键词: Hough 变换; 圆检测; 实时; 计算机视觉; 图像处理 Improved Hough Algorithm for Circle Detection 1 1,2 2 CHEN Xiao-Yan , WANG Qiang , LI Bo-Lin 1(School of Mechanical Engineering, South West Jiaotong University, Chengdu 610031, China) 2(Department of Mechanics, Chengdu Technological University, Chengdu 610031, China ) Abstract: Computer vision has already been widely used in engineering. The circle detection using image processing means significantly in engineering. In order to auto-detect the circle and accurately extract the centre and radius, an improved Hough transform algorithm for circle detection is proposed. The image border is filtered to get the continuous borders. And then the continuous borders are sequenced and equal divided. Only three points are chose for computation and three dimension parameters will be reduced to one dimension, which saves the computational time, memory sources and improves the accuracy. Experiments show the proposed method can realize real time detection and meet the reality needs. Key words: hough transform; circle detection; real time; computer vision; image process 现实生活或工程应用中, 很多物