基于双色光源视觉的标准焊接工件搜寻与焊缝定位  

Standard welding workpiece search and weld positioning based on bicolor light source vision

作　　者：皮冬冬 闫志鸿[1] 卢振洋[1] 刘子虓 罗明睿 PI Dongdong,YAN Zhihong, LU Zhenyang, LIU Zixiao, LUO Mingrui(Department of Mechanical Engineering ＆ Applied Electronics Technology,Beijing University of Technology, Beijing 100124, China)

机构地区：[1]北京工业大学机械工程及应用电子技术学院,北京100124

出　　处：《电焊机》2018年第6期73-78,共6页Electric Welding Machine

基　　金：国家自然科学基金(51475009)

摘　　要：近年来,机器人焊接在越来越多的领域替代了人工焊接,并逐渐向着带有感知的智能化方向发展。形状、尺寸一致的标准化工件的焊接,如筋板、连接件、冲压件、铸件等的焊接是焊接结构中常遇到的焊接问题,这类工件的特点是尺寸较小、形状一致且带有一些几何特征。针对该类工件开发了一套基于双色光源视觉的机器人焊接定位系统,图像处理算法采用焊接工件的初步定位和焊缝的精确定位两个步骤来实现,双色光源有效提高了识别目标与背景之间的区分度,分步图像处理算法提高了定位算法的可靠性。In recent years,robot welding has replaced artificial welding in more and more fields,and robotic welding has been gradually moving towards the intelligent direction with perception. Welding of standardized workpieces of uniform shape and size, such as tendons, fittings, stampings,castings, etc.,is a type of welding problem frequently encountered in welding constructions. This type of workpiece is characterized by its small size,uniform shape and some geometric features. This paper develops a set of robotic welding positioning system based on two-color light source vision for this kind of workpieces. The image processing algorithm uses two steps of the initial positioning of the welding workpiece and the precise positioning of the welding seam. The two-color light source effectively improves the discrimination between the target and the background,step-by-step image processing algorithm improves the reliabiliU of positioning algorithm.

关 键 词：焊接 机器视觉 标准化工件 图像处理 定位 

分 类 号：TG409[金属学及工艺—焊接]