机器人多层多道焊缝激光视觉焊道的识别  

Laser-based visual recognition of multi-pass seam in robot arc welding

作　　者：张华军[1,2] 张广军[1] 蔡春波[2] 高洪明[1] 吴林[1] 

机构地区：[1]哈尔滨工业大学现代焊接生产技术国家重点实验室,哈尔滨150001 [2]哈尔滨理工大学材料科学与工程学院,哈尔滨150040

出　　处：《焊接学报》2009年第4期105-108,共4页Transactions of The China Welding Institution

摘　　要：厚板多层多道焊缝形状特征的自动识别,对保证焊接质量和实现机器人自动化焊接生产具有重要意义.文中建立了基于条形激光源CCD视觉传感多道焊缝自动检测系统,通过对多道焊缝激光条纹图像特征的分析,提出了合理的图形处理识别流程,即经过图像平滑处理、纵向灰度梯度法识别条纹中心、再经过二重斜率平滑处理、多个峰值搜索等处理流程,可获得焊缝形状的四个拐点,为焊缝跟踪、弧长控制和焊枪位姿调整提供了重要的信息,其图像处理结果与试验结果相吻合.It is very necessary for the improvement of weld quality and the realization of robot automatic production to detect multi-pass weld profile parameters automatically.Automatic detective system of multi-pass weld profile parameters by band laser visual sensor was developed.The procedure of image process is as follows: smoothing,center recognition by longitudinal grey gradient,twice slope smooth process,and multiple peaks search by laser seam image analysis.It is very important for seam tracking,arc length control and torch attitude adjustment to obtain four character points of weld seam profile.And the recognition results agree with the experimental results.

关 键 词：多层多道焊 激光视觉 二重斜率平滑处理 图像处理 弧焊机器人 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]