微细对接焊缝初始点定位方法

发布时间：2018-04-28 01:09

本文选题：焊缝初始点 + 归一化相关系数　；参考：《光学精密工程》2017年03期

【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。
[Abstract]:In order to solve the problem of automatic localization and welding of thin weld seam by welding robot, a high precision initial point location method based on vision was studied. In this paper, a two-step accurate localization method in local range is proposed. In the first step, the images of workpiece at different scales and rotation angles are obtained, the typical template matching library is established, and the initial position area of weld seam is extracted by template matching method. The Shi-Tomasi algorithm is used to detect the corner in this local range, and the sub-pixel corner position is calculated and the initial location is accurately located. Through the initial point detection of 50 images with different heights and rotation angles, all images were detected accurately. The experimental results show that the two-step localization method is robust and can accurately locate the initial point of the weld seam with regular and irregular boundaries. The problem of locating the initial point of the micro-butt weld is solved and the expected results are achieved.
【作者单位】：山东科技大学电子通信与物理学院;吉林大学烟台研究院;
【基金】：国家863项目子课题:矿井灾害动态演化技术验证平台(No.2015AA016404-4) 山东科技大学领军人才计划项目
【分类号】：TP242;TP391.41;TG409

【相似文献】