基于二次滤波的对接焊缝路径识别

Butt Weld Path Recognition Based on Secondary Filtering

当前视觉传感器广泛应用于焊接机器人的轨迹规划。针对焊接图像噪声干扰引起的焊缝路径提取精度差的问题,本文提出了一种基于二次滤波的方法实现焊缝的路径识别。通过对采集的焊接图像依次进行灰度处理,滤波处理,边缘提取和填充,将焊缝特征通过二值图像体现出来。同时,为了去除二值图像上噪声的干扰,再次进行滤波处理,并通过灰度重心法提取整张图像上的焊缝坐标。最后,在对接焊缝上验证了本方法识别焊缝路径的平均误差在1像素以内,满足工业的要求。

Currently,vision sensors are widely used in trajectory planning of welding robots.Aiming at the problem of inadequate precision of weld path extraction caused by welding image noise interference,this paper proposes a method based on secondary filtering to realize the path identification of the weld seam.By successively performing gray scale processing,filtering processing,edge extraction and filling on the acquired welding images,the weld features are reflected by the binary image.At the same time,in order to remove noise interference on the binary image,the filtering process is performed again,and the weld coordinates on the entire image are extracted by the gray center of gravity method.Finally,it is verified that the identification average error of the method trajectory is within1 pixels on the butt weld,which meets the requirements of industry.