基于机器视觉的焊缝自动跟踪系统研究

发布时间：2018-11-17 13:17

【摘要】：当前,焊接结构件在各行各业均有着广泛的应用。但是,由于人工焊接劳动强度大、工作环境恶劣、效率低下、焊接热变形所引起的焊缝位置不可控以及操作者人为因素等问题,焊接加工质量往往难以保证。近年来,随着人工智能技术的发展,实时、自适应焊接智能化生产成为了可能。而焊缝跟踪技术作为焊接路径控制中最有效的解决方案之一,是国内外焊接界瞩目的研究热点。然而要实现焊缝自动跟踪,首先应解决焊缝自动跟踪的传感方式和焊缝特征提取问题,而这正是本文的主要研究内容。本文研究获得了苏州市产业前瞻性应用研究(工业)指导性计划项目“基于视觉的弧焊机器人智能焊接关键技术研究”项目的资助。在基于机器视觉的焊缝跟踪系统中,视觉传感系统是整个系统的核心部分。本文对视觉传感系统硬件,即光源、视觉传感器、光路系统等进行了综合分析研究,针对焊接工作状态下电弧光的特征、结构光的光谱特征与工业相机的成像特点,确定了工业相机、光源和工件在光路系统中相对位置,在保证视觉传感器能够清楚采集到焊缝位置并尽可能减少电弧光对图像采集干扰的前提下,设计了基于以太网的数字型CMOS工业相机和结构光法进行图像采集分析的视觉传感硬件系统方案。结合系统特点,整合了视觉系统标定与机器人—摄像头系统标定技术,得到了焊缝实际坐标、焊缝图像坐标和机器人坐标三者关系关联矩阵;分析图像特征,采用中值滤波平滑去噪、灰度拉伸增强对图像进行预处理;然后采用Canny算子提取图像边缘,骨骼化方法获得图像单像素骨骼线,最小二乘法对骨骼线上的点进行拟合,并通过直线求交点的方法得到表征实际焊缝特点的5个特征点。构建基于机器视觉的机器人焊缝自动跟踪系统试验平台。视觉传感器、工控机、焊接机器人三大硬件通过以太网进行连接,采用TCP/IP协议通讯;在集成了摄像头、机器人通讯软件二次开发包基础之上,设计了焊缝图像处理、焊缝PID纠偏控制以及机器人通讯检测等功能模块;最后,分别在短路过渡、射滴过渡和射流过渡工艺条件下,根据偏差“由小到大”,示教点“先疏后密”对该系统进行测试。试验结果表明,该系统能够克服常规焊接工艺过程中不同强度电弧光的干扰,并能对0.5mm-3mm之间的偏差进行纠偏处理。
[Abstract]:At present, welded structural parts are widely used in various industries. However, the welding quality is often difficult to guarantee because of the large labor intensity of manual welding, poor working environment, low efficiency, uncontrollable welding seam position caused by welding heat deformation and human factors of operator. In recent years, with the development of artificial intelligence technology, real-time intelligent production of adaptive welding has become possible. As one of the most effective solutions in welding path control, welding seam tracking technology is a hot research topic in welding field at home and abroad. However, in order to realize automatic seam tracking, it is necessary to solve the problem of sensing mode and feature extraction of weld seam automatic tracking, which is the main research content of this paper. In this paper, the project of "Vision based intelligent welding key technology research for arc welding robot" is supported by Suzhou industry prospective application research (industrial) guidance project. In the weld seam tracking system based on machine vision, the vision sensing system is the core of the whole system. In this paper, the hardware of the vision sensing system, namely light source, vision sensor, optical circuit system and so on, are comprehensively analyzed and studied. Aiming at the characteristics of arc light, the spectral characteristics of structure light and the imaging characteristics of industrial camera, the characteristics of arc light, structure light and industrial camera are analyzed and studied in this paper. The relative position of the industrial camera, the light source and the workpiece in the optical path system is determined. On the premise of ensuring that the visual sensor can clearly capture the weld seam position and minimize the interference of arc light to the image acquisition, This paper designs a hardware system of visual sensor based on Ethernet digital CMOS industrial camera and structured light method for image acquisition and analysis. Combined with the characteristics of the system, the visual system calibration and the robot-camera system calibration technology are integrated, and the actual welding seam coordinates, weld image coordinates and robot coordinates are obtained. The image feature is analyzed and the image is preprocessed by median filter smoothing denoising and gray-scale extension enhancement. Then Canny operator is used to extract the edge of the image and the skeleton method is used to obtain the single pixel skeleton line of the image. The points on the bone line are fitted by the least square method and five feature points representing the characteristics of the actual weld are obtained by the method of straight line intersection. A robot welding seam automatic tracking system test platform based on machine vision is constructed. Vision sensor, industrial control computer and welding robot are connected by Ethernet, and TCP/IP protocol is used to communicate. Based on the integration of the camera and the secondary development kit of robot communication software, the function modules of weld image processing, weld PID correction control and robot communication detection are designed. Finally, under the conditions of short-circuit transfer, droplet transfer and jet transfer, the system is tested according to the deviation "from small to large". The experimental results show that the system can overcome the interference of arc light with different intensity in conventional welding process and can correct the deviation between 0.5mm-3mm.
【学位授予单位】：中国矿业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41;TG409

【参考文献】