基于存在概率图的圆检测方法

Existence Probability Map Based Circle Detection Method

提出了一种基于存在概率图的圆检测方法,将图像边缘点共圆结构信息变换为圆存在概率图中的峰值,再通过峰值检测确定圆的参数。该方法能有效地检出边缘不清晰、不完整的圆形轮廓,具有占用内存少,阈值物理意义明确的特点,较好地解决了标准HT圆检测方法中占用内存空间大,计算效率低的问题。实验结果表明圆存在概率图具有较强的稳定性和分辨能力。

A novel circle detection method based on probability of existence map is proposed.First,the method transforms the information of edge points lying on the same circle into peaks on the probability of existence map,then,parameters of circles being detected are determined by using peak detection.h can effectively detect circles with unclear and irregular contour and occupies less memory space.Experimental results show that probability of existence map has properties of steadiness and high resolving power.