强噪声下的激光视觉焊缝跟踪图像处理研究现状及展望  

Research Status and Prospect of Laser Vision Weld Seam Tracking Image Processing Under Strong Noise

作　　者：董金枋 汤大赟 吴頔 张培磊[1,2] 于治水[1,2] 许燕玲 DONG Jinfang;TANG Dayun;WU Di;ZHANG Peilei;YU Zhishui;XU Yanling(School of Materials Engineering,Shanghai University of Engineering Science,Shanghai 201620,China;Shanghai Collaborative Innovation Center of Laser of Manufacturing Technology,Shanghai 201620,China;Special Equipment Safety Supervision and Inspection Institute of Jiangsu Procedure,Nanjing 210036,China;Shanghai Jiao Tong University,Shanghai 200240,China)

机构地区：[1]上海工程技术大学材料工程学院,上海201620 [2]上海市激光先进制造技术协同创新中心,上海201620 [3]江苏省特种设备安全监督检验研究院,江苏南京210036 [4]上海交通大学材料科学与工程学院,上海200240

出　　处：《电焊机》2022年第12期1-16,共16页Electric Welding Machine

基　　金：国家自然科学基金(51905333);中国博士后科学基金(2021M692039);上海市青年科技英才扬帆计划(19YF1418100)。

摘　　要：激光视觉焊缝跟踪系统已经成为机器人焊接必不可少的关键环节之一,但焊接过程中存在的金属飞溅、烟尘以及强弧光等常见的强噪声干扰,给激光视觉焊缝跟踪的稳定性和精确性带来极大挑战。分别对比了焊缝跟踪最为关键的图像预处理、激光中心线提取、特征点识别三个环节中传统算法与新兴算法的区别,总结归纳各类图像处理算法的检测精度、实时性以及适用场景。并对激光焊缝跟踪技术在焊接领域未来的发展方向进行了展望。Laser vision seam tracking system has become an indispensable key technical equipment for robot welding. Due to the inherent strong noise conditions such as spatter, narrow gap and strong arc light in welding processing, it brings great challenges to the stability and accuracy of laser vision seam tracking system. This paper compares the differences between traditional and emerging algorithms in the three key links of weld seam tracking system: image preprocessing, centerline extraction of weld line structured light stripe and feature point extraction, and summarizes the detection accuracy, real-time performance and applicable scenarios of various image processing algorithms. Finally, the future development direction of laser welding seam tracking technology in the field of welding is prospected.

关 键 词：强噪声 机器人焊接 激光视觉 焊缝跟踪 图像处理 

分 类 号：TP216[自动化与计算机技术—检测技术与自动化装置]