【摘要】 介绍了一种基于随机Hough变换的圆检测的改进算法。本文解决了随机Hough变换检测圆中参数单元无效累积的问题,首先构造边缘点集的数据空间,采用搜索的方法将各连续曲线的边缘点顺序存储,然后从中选取最小点集,利用圆的性质求得圆参数,该算法计算量较小,并且避免了解方程组运算带来的误差。更多还原

【Abstract】 This paper introduces an improved method of circle detection which is based on Randomized Hough Transform. It solves the problem of the inefficacious parameter accumulation when using randomized Hough transform to detect circles. First, construct a data space including all the edge points; Second, use the searching method to store the edge points of every connected curve in sequence; Last, choose the minimum edge-points gather from the stored edge points and get the parameters of circle basing on the character of circle. The method needs less computing quantity, and it avoids the error when solving equation groups.更多还原