弧焊机器人线激光传感器焊缝识别研究

Research on weld recognition of arc welding robot line laser sensors

基于模块化的思想,搭建了焊接机器人系统的硬件平台,介绍了基于激光视觉的焊接机器人软硬件组成和其工作原理.根据三角测距原理,推导出激光的矩阵转化关系,将二维的像素点映射到三维相机坐标.采用二维标定法,利用matlab软件对相机完成标定.针对相机获取图像的ROI区域,使用ADAP及VISUS对该区域进行处理,最终使得图像处理过的特征点与实际焊缝根部之间的误差降至0.2 mm以下,实时跟踪误差范围由2.7 mm降至0.39 mm.试板焊接结果表明,该传感器具有良好的可行性和有效性.

Based on the idea of applying the unit system,which is a hardware platform for a welding robot system,we describe the installation of devices and software of a welding robot based on laser vision and the way it works.According to the principle of triangle ranging,the matrix transformation relation of laser is deduced,and 2-dimensional pixels are mapped to 3-dimensional camera coordinates.By the 2D calibration method,the camera is used to complete calibration with matlab software.The ROI area from which the image was obtained by the camera is processed using ADAP and VISUS,and finally the error between the feature points after image processing and the actual weld root is reduced to within 0.2 mm,and the real-time tracking error range is reduced from 2.7 mm to 0.39 mm.The test board welding results show that the sensor has desired feasibility and effectiveness.