改进的Hough变换检测圆方法  

Improved Hough Algorithm for Circle Detection

作　　者：陈小艳[1] 王强[1,2] 李柏林[2] 

机构地区：[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

出　　处：《计算机系统应用》2015年第8期197-201,共5页Computer Systems & Applications

基　　金：四川省科技支撑计划(2013GZ0032)

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.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.

关 键 词：HOUGH变换 圆检测 实时 计算机视觉 图像处理 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]