一种基于随机Hough变换圆检测的改进算法  

An Improved Randomized Hough Method of Circle Detection

作　　者：赵桂霞[1] 黄山[1] 

机构地区：[1]四川大学电气信息学院,四川成都610065

出　　处：《计算机技术与发展》2008年第4期77-79,共3页Computer Technology and Development

基　　金：2007年四川省重点技术创新项目计划(07XM020)

摘　　要：介绍了一种基于随机Hough变换(RHT)的圆检测的改进算法。该算法利用梯度方向信息来确定采样的三点是否进行累积,然后再利用确定候选圆范围的方法来缩小所要搜索的像素点的范围。此方法较好地解决了传统RHT中由于随机采样而造成的大量无效累积问题,并且改进后的算法使运行速度得到进一步的提高,检测性能也有较大的改善。该算法分别在加噪和不加噪的人工图像上做了实验,检测性能和处理速度方面都比传统的RHT有明显的改善和提高。Proposes an improved algorithm for detecting circles which is based on randomized Hough transform. It uses gradient direction information to determine whether the parameter should be accumulated or not. And it uses the range of circle to reduce the range of pixels searching. The problem of Useless accumulation in traditional RHT is well solved and the improved method has higher speed and better detection performance. Synthetic images with different level noises have been taken to test the detection performance and the speed. And this algorithm has been proved an efficient method.

关 键 词：随机HOUGH变换 圆检测 梯度方向信息 窗口 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]