基于视觉的焊缝路径识别与轨迹规划

发布时间：2019-02-26 20:41

【摘要】：我国制造业未来十几年要向自动化、智能化的方向发展。焊接技术在现代制造业中占据非常重要的地位,因此焊接过程的自动化和智能化是焊接技术未来发展的方向。目前,焊接机器人大多通过示教再现的方式进行焊接作业,当外部环境发生变化时就需要对机器人进行重新示教,效率低下,自动化程度也达不到要求。把视觉技术应用到焊接中,利用视觉传感器识别焊缝路径,再对焊缝路径进行轨迹规划,可以在一定程度上提高焊接机器人的自动化程度。本文采用CMOS摄像机与6自由度机器人形成“Eye-to-hand”的手眼关系,搭建图像采集实验平台获取平面曲线焊缝图像。对视觉识别系统进行建模,介绍了摄像机的标定模型和标定方法以及手眼标定原理,利用平面棋盘格作为靶标对摄像机进行了标定实验和手眼标定实验,得到了摄像机的内外参数和手眼关系矩阵。对采集到的平面曲线焊缝进行图像预处理和特征信息提取,提出了一种基于局部阈值分割的图像二值化方法,并利用求平均值的方法得到了焊缝路径中心线的坐标,最后经过拟合得到了焊缝路径中心线在图像坐标系下的方程,再经过坐标转换得到了焊缝路径中心线在世界坐标系下的方程。介绍了一种焊缝特征坐标系建立的方法,并建立了平面曲线焊缝的特征坐标系;介绍了6自由度机器人的运动学模型,分析了其正运动学方程和逆运动学方程;提出了一种改进的等间距直线逼近法并利用该方法生成了平面曲线焊缝路径轨迹,并进行了运动学仿真。开发了基于视觉的焊缝路径识别与轨迹规划的系统软件,进行了焊接实验,并对实验结果进行了误差分析。
[Abstract]:China's manufacturing industry in the next ten years to automation, intelligent direction of development. Welding technology occupies a very important position in modern manufacturing industry, so the automation and intelligence of welding process is the direction of future development of welding technology. At present, the welding robot mostly carries on the welding operation through the teaching and reappearance way, when the external environment changes, it needs to teach the robot again, the efficiency is low, and the degree of automation is not up to the requirement. The application of vision technology to welding can improve the degree of automation of welding robot to a certain extent by using vision sensor to identify weld path and then carry on trajectory planning of weld path. In this paper, the CMOS camera and the 6-DOF robot are used to form the "Eye-to-hand" hand-eye relationship, and an image acquisition experiment platform is set up to obtain the plane curve weld image. The model of vision recognition system is modeled, the calibration model and method of camera and the principle of hand-eye calibration are introduced. The camera calibration experiment and hand-eye calibration experiment are carried out by using plane chessboard as target. The inner and outer parameters of the camera and the hand-eye relation matrix are obtained. After image preprocessing and feature information extraction, a binary image segmentation method based on local threshold segmentation is proposed, and the coordinates of the weld path centerline are obtained by means of the average value method, which is based on the image pre-processing and feature information extraction of the plane curve weld seam collected from the image, which is based on the local threshold segmentation. Finally, the equation of weld path centerline in image coordinate system is obtained by fitting, and then the equation of weld path centerline in world coordinate system is obtained by coordinate transformation. This paper introduces a method of establishing the characteristic coordinate system of weld seam, and establishes the characteristic coordinate system of plane curve weld, introduces the kinematics model of 6-DOF robot, analyzes its forward kinematics equation and inverse kinematics equation, and analyzes its forward kinematics equation and inverse kinematics equation. In this paper, an improved linear approximation method with equal spacing is proposed and used to generate the path trajectory of planar curve weld, and the kinematics simulation is carried out. The system software of welding path recognition and trajectory planning based on vision is developed. The welding experiment is carried out and the error analysis of the experimental results is carried out.
【学位授予单位】：山东科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TG441.7

【参考文献】