基于视觉的焊缝特征识别与自动跟踪研究

发布时间：2019-04-27 12:52

【摘要】：焊缝自动跟踪技术是实现焊接自动化的基础和关键。焊接传感技术又是实现焊缝自动跟踪的前提条件,相比电弧、接触、超声波等传感技术,结构光视觉传感凭着获取的信息量大、灵敏度和测量精度高、抗强电场和强磁场的干扰能力强、适用于不同形式焊缝等强大的优势,成为了焊缝跟踪研究者研究的热点对象。本文首先阐述了焊缝自动跟踪对实现焊接自动化的关键性作用,分析了不同传感器在焊接应用中的原理及优缺点,综合视觉焊缝自动跟踪技术的研究及应用现状,再结合虚拟仪器的特点和应用,设计了以相机作为传感器、线激光器作为辅助光源、NI系统为控制器、行走机构为执行机构的结构光视觉焊缝自动跟踪系统。其中,高速相机和线激光器封装在一起组成了视觉部分;NI系统包括图像采集卡、数据采集卡、LabVIEW软件等;行走机构由工控主机、伺服系统和一些传动机构组成。焊接工件固定放置在行走机构放置台上,激光线投射到工件焊缝上,通过视觉系统获取焊缝图像,将图像传送给NI主机,在LabVIEW软件平台上对图像进行灰度图转换、中值滤波、二值化、散点去除的处理,根据模板匹配方法对焊缝特征识别后,计算出焊缝中心位置与焊枪位置的坐标偏差值,偏差信息作为控制信号由NI系统传给行走机构的控制器,控制器根据接收的控制命令实时调整焊枪位置,最终实现焊缝的实时跟踪。本系统搭建完成后,先做了焊缝实时跟踪验证实验,用点激光模拟焊枪,寻找点激光跟踪路径,结果显示点激光所走路径基本保持在焊缝中心位置上。有了此验证基础,又对本系统使用微束等离子焊枪进行了焊缝跟踪焊接实验,结果表明,系统在焊接过程中也能清晰识别到焊缝并完成自动跟踪焊接。
[Abstract]:Welding seam automatic tracking technology is the foundation and key to realize welding automation. Welding sensing technology is also a prerequisite for automatic seam tracking. Compared with arc, contact, ultrasonic and other sensing technologies, structured light vision sensing has a large amount of information, high sensitivity and measurement accuracy. Because of its strong anti-interference ability of strong electric field and strong magnetic field, it is suitable for different types of weld seam and other powerful advantages, so it has become a hot research object of weld seam tracking researchers. In this paper, the key role of automatic seam tracking in welding automation is described, the principle, advantages and disadvantages of different sensors in welding application are analyzed, and the research and application status of visual automatic welding seam tracking technology are summarized. Combined with the characteristics and application of virtual instrument, a structured light vision welding seam automatic tracking system is designed, which uses camera as sensor, line laser as auxiliary light source, NI system as controller and walking mechanism as executive mechanism. Among them, the high-speed camera and line laser are encapsulated together to form the vision part; the NI system consists of image acquisition card, data acquisition card, LabVIEW software and so on; the walking mechanism is composed of industrial control host computer, servo system and some transmission mechanisms. The welding workpiece is fixed on the platform of the walking mechanism, the laser line is projected onto the weld seam of the workpiece, the weld image is obtained by the visual system, the image is transmitted to the NI host computer, and the gray-scale image is converted on the LabVIEW software platform, and the median filtering is carried out. After recognizing the weld seam features according to the template matching method, the coordinate deviation between the weld center position and the welding torch position is calculated, and the deviation information is transmitted to the controller of the walking mechanism by the NI system as the control signal. According to the received control command, the controller adjusts the position of welding torch in real-time, and finally realizes the real-time tracking of welding seam. After the system is built, the real-time welding seam tracking verification experiment is done firstly. The spot laser simulation torch is used to find the spot laser tracking path. The result shows that the point laser path is basically in the weld center position. On the basis of this verification, the welding seam tracking experiment is carried out by using the micro-beam plasma welding gun. The results show that the welding seam can be clearly recognized in the welding process and the automatic tracking welding can be completed by the system.
【学位授予单位】：昆明理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TG409;TP391.41

【参考文献】