基于模板匹配的法兰盘位姿估计研究

针对工业机器人自动分拣中当工件有较大的旋转角度时,普通的模板匹配法无法准确地得到抓取和放置工件正确位置的问题,提出一种基于内边界追踪粗定位和基于模板匹配精定位的两步视觉定位方法。针对谐波减速器输出法兰盘,采用八邻域内边界追踪法对工件旋转角度进行粗定位,通过霍夫圆检测技术得到工件的中心位置,并通过一种改进的模板匹配法得到工件精确的旋转角度。实验结果表明,位姿检测的平均误差不大于0.1°,该定位方法在合适条件下可以满足精度要求。

In order to solve the problem that the common template matching method can not get the correct position of grasping and placing workpiece accurately when the workpiece has a large rotation angle in the automatic sorting of industrial robot,a two-step visual localization method based on inner boundary tracking and fine positioning based on template matching is proposed.Aiming at the output flange of harmonic reducer,the eight-neighborhood inner boundary tracking method is used to roughly locate the rotation angle of the workpiece,and the central position of the workpiece is obtained by Hough circle detection technique.The accurate rotation angle of the workpiece is obtained by an improved template matching method.The experimental results show that the average error of pose detection is less than 0.1,and the positioning method can meet the accuracy requirements under suitable conditions.