| 题名 | 焊接机器人视觉测量与控制研究 |
| 作者 | 鄢治国 |
| 学位类别 | 工学博士 |
| 答辩日期 | 2009-05-18 |
| 授予单位 | 中国科学院研究生院 |
| 授予地点 | 中国科学院自动化研究所 |
| 导师 | 徐德 |
| 关键词 | 焊接机器人 视觉测量 视觉控制 焊缝跟踪 welding robot visual measurement visual control weld seam tracking |
| 其他题名 | Research on visual measurement and visual control of welding robot |
| 学位专业 | 控制理论与控制工程 |
| 中文摘要 | 焊接机器人具有精度高、可靠性好等优点，能够提高焊接效率和焊接质量，减轻工人劳动强度，在焊接领域得到广泛应用。目前，大多数焊接机器人工作在示教再现方式下，这种方式缺乏自适应性，对焊件的安装以及焊接环境的要求较高，焊接质量易受多种因素干扰。开发较高自动化和智能化水平的焊接机器人已经成为焊接自动化领域的一个研究热点。 视觉传感器由于其无接触、信息量大、精度高等优点，成为最常用的焊接信息测量传感器。本文围绕焊接机器人的视觉测量与控制开展了研究工作。 第一，提出了一种自然光照下坡口焊缝图像的图像处理算法。通过增强焊缝特征、自适应确定图像阈值以及噪声消除三个步骤，使算法具有较强的自适应性，对不同环境下焊缝图像都有较好的处理结果。针对集装箱拼板窄焊缝图像的特点，提出了一种高可靠性的图像处理算法，能够较好地克服焊接中弧光、飞溅等因素的干扰，稳定地提取焊缝特征。 第二，为提高传统示教再现焊接机器人的智能化水平，提出了一种基于示教与视觉纠偏的自动焊接方法。将结构光焊缝跟踪传感器应用于能接受外界信号的示教再现机器人，使机器人具有焊缝纠偏功能。 第三，根据焊枪与焊缝起始点在图像空间及笛卡尔空间的相对关系，提出了一种无标定视觉控制方法。该方法采用双目视觉和基于图像的视觉伺服策略，实现了焊枪与焊缝起始点的自动对准。 第四，为提高大型工件的焊接效率和质量，设计了一个9关节、6自由度的焊接机器人。提出了分层控制结构和宏微运动控制方法。该机器人具有示教再现、焊缝跟踪、焊接起始点对准等功能，可实现大范围、高精度的焊接作业。 第五，针对集装箱拼板窄焊缝，开发了一种基于PLC与智能像机的焊缝跟踪系统。该系统具有体积小、操作方便、可靠性高等特点，可提高集装箱拼板的焊接效率和焊接质量。 最后，对本文的研究成果进行了总结，并指出了下一步的研究方向。 |
| 英文摘要 | Welding robot has many merits, such as high accuracy and reliability. It can improve the welding quality and welding efficiency, while reducing labor intensity of workers. It is widely used in the welding field. At present, most of the welding robots work in teaching-playback mode. This mode has poor adaptability, and has high requirement for the installation of workpieces and the welding environment. At the same time, its welding quality may be disturbed by many factors. The development of welding robot with high automation and intelligent level has become a research focus in welding automation field. Vision sensors are widely used in the detection of welding seam because of its non-contact, informative and high precision. Visual measurement and control of welding robot are focused in this thesis. Firstly, an image processing method for groove welding seam under natural light is proposed. The method contains three steps: enhance the image feature of welding seam, determine the threshold adaptively and eliminate noises. Good results can be achieved by using the method to process welding images under different environments. An image processing algorithm for narrow weld in container manufacture is presented. Noises, such as arc light and splash, can be eliminated and the feature of the welding seam can be extracted reliably by using the algorithm. Secondly, to improve the intelligence level of the traditional teaching-playback welding robot, an automated robotic welding method based on teaching and visual correcting is presented. It applies laser-based vision sensor to traditional teaching-playback welding robot to realize the seam tracking function. Thirdly, an uncalibrated visual control method is proposed according to the relations between the welding torch and the welding seam in image space and Cartesian space. Binocular vision and image based visual servoing strategy are employed in the control method. The welding torch can be aligned to the initial welding position successfully by using the method. Fourthly, to improve the welding efficiency and quality of large scale workpieces, a welding robot with 9 joints and 6 degrees of freedom is designed. Hierarchical control structure and macro and micro motion control are presented. The robot has fuctions such as teaching-playback, seam tracking, initial welding position guiding, and so on. It can achieve high positioning precision in large scale work space. Fifthly, a compact visual system, which is... |
| 语种 | 中文 |
| 其他标识符 | 200618014628019 |
| 内容类型 | 学位论文 |
| 源URL | [http://ir.ia.ac.cn/handle/173211/6146]  |
| 专题 | 毕业生_博士学位论文 |
| 推荐引用方式 GB/T 7714 | 鄢治国. 焊接机器人视觉测量与控制研究[D]. 中国科学院自动化研究所. 中国科学院研究生院. 2009. |
| 个性服务 |
| 查看访问统计 |
| 相关权益政策 |
| 暂无数据 |
| 收藏/分享 |
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。
修改评论