机器人焊缝视觉跟踪与控制系统研究

发布时间：2018-04-26 08:47

本文选题：焊缝磨抛 + 焊缝视觉跟踪　；参考：《吉林大学》2017年硕士论文

【摘要】：当前,我国已进入高铁建设的黄金时代。轨道车辆车体表面一般由分块的构件通过拼接后焊接成形,焊接过程中会产生很大的焊接应力,存在较大焊接应力会使构件焊接部位的强度和抗疲劳强度降低。为了消除焊接应力,提高服役寿命和表面美观度,为后续工艺提供高质量的焊接表面,焊接结束后,一般需对焊缝进行磨抛处理以消除内应力和实现焊缝区域表面光滑一致。目前,绝大多数大型结构件焊缝磨抛处理都是靠人工磨抛完成的,对焊接后的焊缝采用手工磨抛处理,生产效率低、劳动强度大,对工人的技术水平要求高,加工精度一致性难以保证,还会经常磨伤基本表面。本文以焊缝磨抛处理为背景,基于移动式磨抛机器人机械本体,开发磨抛机器人的基于位置的视觉伺服控制系统。实现机器人的焊缝跟踪、焊缝特征点三维测量和机器人自动作业控制功能。本文主要研究内容如下:(1)搭建了机器人双目视觉测量系统平台,对双目摄像机立体标定,建立特征点的三维坐标视觉测量模型。为实现对机器人工具系统导航,需将特征点坐标转换到机器人坐标系中,设计了一种基于三目标点的手眼变换关系标定方法,该标定方法可以简单、快速获取机器人手眼变换关系矩阵。(2)使用滤波片和滤波算法对图像处理,消除噪点干扰,得到高质量焊缝图像。提出一种基于灰度阈值提取的方法,实现对激光线区域的提取。对提取后激光区域使用形态学运算处理,填补了激光线区域的缺陷并平滑激光线区域边界。针对激光线中心线提取问题,应用区域骨架化处理算法提取激光中心线,获得准确、平滑的中心线,同时避免了中心线间断问题。对焊缝特征点搜寻方法改进,使搜寻路径沿中心线搜寻,提高双目特征点匹配的速度。对搜寻匹配到的特征点三维坐标重构,实现机器人对焊缝的识别、跟踪、测量功能。(3)设计了机器人控制系统的结构,确定上、下位机二级分布式结构。其中PC机作为上位机,负责整个机器人系统管理、各轴运动量计算以及焊缝图像处理等。下位机由运动控制卡与图像采集卡组成,分别控制机器人的运动系统和视觉系统图像采集。这种结构能明显提高控制系统工作速度和控制性能。设计了机器人多模式控制策略,编写机器人控制软件,使软件具有工艺参数设置、手动模式、自动模式控制功能。实验研究证明机器人视觉伺服控制系统稳定性和精度较高,满足磨抛机器人样机要求,为后续磨抛工艺制定提供有力支撑。
[Abstract]:At present, our country has entered the golden age of high-speed railway construction. Rail vehicle body surface is usually welded by spliced components, which will produce great welding stress during welding process, and the strength and fatigue strength of the welded parts will be reduced by the existence of large welding stress. In order to eliminate welding stress, improve service life and surface aesthetics, and provide high quality welding surface for subsequent process, it is generally necessary to grind and polishing weld seam after welding to eliminate internal stress and achieve smooth and consistent weld area surface. At present, most of the welding seam grinding and polishing treatment of large structural parts is accomplished by manual grinding and polishing. The welding seam after welding is treated by manual grinding and polishing. The production efficiency is low, the labor intensity is large, and the technical level of the workers is high. Processing accuracy consistency is difficult to ensure, but also often wear the basic surface. In this paper, the position based visual servo control system of the grinding and polishing robot is developed based on the mechanical body of the mobile grinding and polishing robot under the background of welding seam grinding and polishing. The welding seam tracking, the 3D measurement of weld characteristic point and the automatic operation control function of robot are realized. The main contents of this paper are as follows: (1) the platform of robot binocular vision measurement system is set up, the stereo calibration of binocular camera is carried out, and the 3D coordinate vision measurement model of feature points is established. In order to realize navigation of robot tool system, the coordinate of feature points should be transformed into robot coordinate system. A calibration method of hand-eye transformation relationship based on three-objective points is designed. The calibration method is simple. Fast acquisition of robot hand-eye transform matrix. 2) using filter slice and filter algorithm to process the image, eliminate noise interference, and get high quality weld image. A method based on gray threshold is proposed to extract laser line region. Morphological operation was used to deal with the extracted laser region, which filled the defects of the laser line region and smoothed the boundary of the laser line region. Aiming at the problem of laser centerline extraction, the region skeleton processing algorithm is applied to extract the laser centerline to obtain accurate and smooth centerline, while avoiding the problem of centerline discontinuity. The method of seam feature point search is improved to make the search path search along the center line, and the matching speed of binocular feature points is improved. The structure of the robot control system is designed to reconstruct the 3D coordinates of the matching feature points to realize the identification, tracking and measuring functions of the welding seam. The upper and lower level distributed structure of the robot control system is determined. PC, as the host computer, is responsible for the management of the whole robot system, the calculation of the motion of each axis and the image processing of the weld. The lower computer consists of a motion control card and an image acquisition card, which respectively controls the motion system and the visual system of the robot. This structure can obviously improve the working speed and control performance of the control system. The multi-mode control strategy of robot is designed, and the robot control software is written. The software has the functions of setting process parameters, manual mode and automatic mode control. The experimental results show that the robot visual servo control system has high stability and precision, which meets the requirements of the grinding and polishing robot prototype and provides a strong support for the subsequent grinding and polishing process development.
【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41;TP242

【相似文献】