基于hough变换圆检测算法的改进  

Improvement of Circle Detection Algorithm Based on Hough Transform

作　　者：孙方敏[1] 乔艳涛[1] 孙祥国[1] 陈伟[1] 刘纪鹏[1] 

机构地区：[1]四川理工学院机械工程学院,四川自贡643000

出　　处：《机械工程师》2016年第10期29-31,共3页Mechanical Engineer

基　　金：四川省教育厅青年基金项目(10ZB097);四川理工学院人才引进项目(2009XJKRL002);过程装备与控制工程四川省重点实验室项目(GK201101)

摘　　要：传统的hough变换在进行圆的检测时,是对整幅图像进行计算的,因此计算量大,运行速度慢,而且不适用于复杂的图像。在用三坐标测量机的测头部分进行相机标定时,除三坐标测量机的圆形测头外,图像中还有多个圆形固定孔,严重影响了圆形测头的检测,因此传统的hough变换圆检测算法不能准确测出测头的圆心。针对以上问题,文中介绍了选定感兴趣区域的改进方案。感兴趣区域包含了三坐标测量机测头的大体位置,只对感兴趣区域进行hough变换可以大大减少计算量,提高运行速度,还能忽略其他相似图形的干扰,将三坐标测量机的测头部分快速检测出,并得到其在像素坐标系中的坐标。The traditional Hough transform circle detection not only cannot calculate whole image, which may lead to large amount of computation and slow speed, but also cannot calculate complex image. When the probe of three coordinate measuring machine is used to calibrate camera, several fixed circular holes affect the measurement accuracy seriously. So the traditional Hough transform circle detection algorithm cannot accurately measure center of the probe. To solve the above problems, this paper introduces a Hough transform method based on selecting the region of interest, which contains the probe position of three coordinate measuring machine. The Hough transform method can not only greatly reduce the computation, improve the operating speed, but also neglect other similar graphics interference, and the three coordinate measuring machine to detect the first part of the machine, and get the coordinates in the pixel coordinates.

关 键 词：相机标定 三坐标测量机 HOUGH变换 测头部分 

分 类 号：TP391.7[自动化与计算机技术—计算机应用技术]