基于图像处理技术的机器人焊接焊缝跟踪研究  

Research on Robot Welding Seam Tracking Based on Image Processing Technology

作　　者：杨群 YANG Qun(Wuhu Institute of Technology,Wuhu,Anhui 241006,China)

机构地区：[1]芜湖职业技术学院智能制造学院,安徽芜湖241006

出　　处：《九江学院学报（自然科学版）》2024年第3期27-32,共6页Journal of Jiujiang University：Natural Science Edition

基　　金：2023年度芜湖职业技术学院校级科学研究项目(编号wzyzrzd202315);2022年度安徽省高校科学研究项目(编号2022AH052196)的研究成果之一。

摘　　要：为解决焊接机器人面对不同焊接工况时,焊缝跟踪精度低的问题,文章提出了基于图像处理技术的机器人焊接焊缝跟踪方法。利用CCD传感器采集机器人焊接焊缝图像,基于点操作的图像增强方法,灰度变换机器人焊接焊缝图像。利用SegFormer图像语义分割模型,通过Transformer编码器与多层感知机解码器,采用递归空洞自注意力机制,从机器人焊接焊缝的灰度图像中提取焊缝跟踪路径。利用卡尔曼滤波算法消除焊缝跟踪路径中的非焊接点,通过机器人焊接焊缝图像的坐标系转换,更新焊接焊缝轨迹堆栈,实现机器人焊接焊缝跟踪。实验结果表明,该方法能够依据机器人焊接焊缝图像处理结果,精准跟踪焊接焊缝,圆弧型与S型焊缝的跟踪误差低于3mm。Here was the research on a robot welding seam tracking method based on image processing technology to solve the problem of low seam tracking accuracy when welding robots face different welding conditions.Using CCD sensors to capture images of robot welding seams,using point operation based image enhancement methods,we could get grayscale transformation of robot welding seam images.Using the SegFormer image semantic segmentation model,a Transformer encoder and a multi-layer perceptron decoder were used to extract the weld seam tracking path from the grayscale image of the robot welding seam using a recursive cavity self attention mechanism.Using Kalman filtering algorithm to eliminate non welding points in the weld seam tracking path,updating the welding seam trajectory stack through coordinate system transformation of robot welding seam images,we could get robot welding seam tracking.The experimental results showed that this method could accurately track welding seams based on the image processing results of robot welding seams,with tracking errors of less than 3mm for arc and S-shaped welds.

关 键 词：图像处理技术 机器人 焊接焊缝跟踪 灰度变换 图像语义分割 卡尔曼滤波 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]