微细对接焊缝初始点定位方法

摘要

为了解决焊接机器人对薄板微细焊缝的自动定位和焊接问题,研究了一种基于视觉的高精度微细对接焊缝初始点定位方法.提出一种局部范围内两步精确定位方法,第一步,获取工件在不同尺度及旋转角度下的图像,建立典型的模板匹配库,利用模板匹配方法提取出焊缝初始点位置区域;第二步,采用Shi-Tomasi算法在此局部范围内进行角点检测,计算亚像素级角点位置并进行初始点精确定位.通过对现场拍摄的50幅不同高度和旋转角度的图像进行初始点检测,实现了对所有图像的准确检测.实验结果表明,两步定位方法鲁棒性强,可以精确定位出规则边界和不规则边界焊缝的初始点,解决了微细对接焊缝初始点定位问题,达到了期望结果.%A vision based method for initial point positioning of micro butt welding seam was studied in order to automatically position and weld the welding seam for the welding-robot.The method could be divided into two steps as follows.Firstly,the images of the workpiece at different scale and rotation angle were obtained and a typical template matching library was established,and then the initial point location area of the weld was extracted by using the template matching method.Secondly,the ShiTomasi algorithm was used to detect the corner points in the local area and to calculate the corner position at the sub pixel level.Fifty images at different heights and angles captured on the spot were used in the experiment,and all the initial points of the micro butt welding seams were detected correctly.Experimental results show that the two-step localization method has strong robustness,high positioning precision of a regular boundary or an irregular boundary.This method solved the problem of micro butt weld initial point positioning,resulting in a satisfactory result.