一种简易Hough变换圆检测的原木端面识别方法

A Simple Hough Transformation Circle Detection Method for Logs Cross Section

提出一种简易Hough变换的圆检测方法,其参数空间为二维累加器,在没有嵌套圆的情况下大大提高了圆的检测效率。基于K-means聚类分割算法下,采用随机寻找到原木边缘上的一点,根据其梯度角信息搜索同一原木边缘上的其它两点来寻找潜在原木图像上的类圆,根据潜在类圆的圆心参数累积和原木边缘验证,进而判定潜在原木是否真实存在。

A method of circle detection based on simple Hough transform is proposed. The parameter space is a two-dimensional accumulator,which greatly improves the detection efficiency of circles without nested circles. Based on the K-means clustering algorithm,the method searches randomly for a point on the edge of the log,and searches for two other points on the same log edge according to the gradient angle information,center of the cumulative parameters and log edge validation,and then determines the potential existence of logs.