基于局部模板匹配的GMAW初始焊位识别及工艺研究  

Research on recognition of initial welding position based on partial template matching in GMAW and its welding process

作　　者：詹超[1] 林涛[1] 陈华斌[1] 陈善本[1] 刘信诠[2] 沙万华[2] 杨惠勤[2] 

机构地区：[1]上海交通大学材料科学与工程学院焊接所,200240 [2]上海锅炉厂有限公司,200245

出　　处：《焊接》2011年第9期22-25,70,共4页Welding & Joining

摘　　要：为解决超超临界锅炉悬吊管自动焊初始焊位导引问题,提出了一种基于模板匹配的机器人GMAW初始焊位识别的新方法。对焊接机器人单目CCD视觉传感器的手-眼关系进行快速标定,对待匹配图像和模板图像进行了预处理,利用翅片角点特征,采用搜索策略和匹配算法对工件进行初始焊位识别,计算初始焊位偏差值并发送机器人控制器进行焊缝纠偏。分析了异形悬吊管工件的初始焊位识别方法,获取了待焊工件的初始焊接位置并计算了原始起焊点的位置偏差量并进行了焊接工艺试验。研究结果表明,该算法对硬件资源要求低,识别速度快,构建的弧焊机器人系统CCD纠偏精度可达1m,焊缝成形良好。In order to solve the recognition and guidance of the initial welding position in automation welding of boiler suspension tube,a new method to recognize the initial welding position in GMAW(gas metal arc welding)robot was developed based on template matching.A rapid calibration of hand-eye relationship of welding robot with monocular vision was rapidly calibrated,and the template and the target image were preprocessed.The the initial welding position of the workpiece was identified by using matching algorithm and searching strategy according to the feature of fin.The deviation of the initial welding position was calculated and sent to robot controller for correction.The method to recognize the initial welding position of anomalous suspension tube was analyzed,and got the position of the initial welding position as well as the deviation value and welding technical parameters.The research result proved that this method needs only a small quantity of hardware resources with high recognizing speed,and the system deviation remains less than 1 mm as well as good formation of welding.

关 键 词：视觉传感 焊位识别 模板匹配 OPENCV 

分 类 号：TG409[金属学及工艺—焊接]