基于逐行扫描法的焊缝识别研究  

Research on weld seam identification based on line-by-line scanning

作　　者：邓洁 张文志[1] 赵志强 陈楷 李鑫 DENG Jie;ZHANG Wenzhi;ZHAO Zhiqiang;CHEN Kai;LI Xin(School of Mechanical Engineering,Inner Mongolia University of Technology,Hohhot 010051,China)

机构地区：[1]内蒙古工业大学机械工程学院,呼和浩特010051

出　　处：《内蒙古工业大学学报（自然科学版）》2024年第4期343-350,共8页Journal of Inner Mongolia University of Technology：Natural Science Edition

基　　金：内蒙古自治区科技计划项目(2021GG0260);内蒙古自治区自然科学基金项目(2020LH06003)。

摘　　要：爬壁机器人通过跟踪焊缝以完成对大型压力容器密封性的检测。压力容器上的焊缝主要存在直线型和T型两种焊缝,传统的焊缝识别方法在正常环境下可以准确地识别焊缝,但压力容器上的焊缝长时间暴露在空气中会产生许多干扰信息。针对原始焊缝图像存在腐蚀、锈斑等问题,导致传统方法难以准确识别,提出了一种基于逐行扫描的焊缝识别方法。在经过灰度拉伸变化、二值化、滤波及形态学处理等预处理的基础上,利用Canny算法结合修正技术进行边缘检测,采用逐行扫描获取中点的方法提取中心线,使得检测机器人能准确识别焊缝。通过爬壁机器人搭载相机对焊缝进行实时检测实验,实验结果表明该方法可帮助机器人稳定可靠地完成对大型压力容器上的焊缝识别。The wall-climbing robot completes the sealing detection of large pressure vessels by tracking weld seams.There are mainly two types of weld seams on pressure vessels:linear and T-shaped.Traditional methods of weld seam identification can accurately identify weld seams in normal environments.However,weld seams on pressure vessels exposed to air for a long time will produce a lot of interference information.Due to issues such as corrosion and rust spots in original weld seam images,traditional methods have difficulty in accurate identification.A weld seam recognition method based on line-by-line scanning is proposed.After preprocessing steps such as grayscale stretching,binarization,filtering,and morphological processing,edge detection is performed using the Canny algorithm combined with correction techniques.The method uses line-by-line scanning to extract the centerline through midpoint acquisition,allowing the detection robot to accurately identify weld seams.Real-time detection experiments on weld seams were carried out using a wall-climbing robot equipped with a camera.The experimental results show that this method can help the robot stably and reliably complete the identification of weld seams on large pressure vessels.

关 键 词：焊缝识别 图像处理 爬壁机器人 逐行扫描 

分 类 号：TP751.2[自动化与计算机技术—检测技术与自动化装置] TP242.2[自动化与计算机技术—控制科学与工程]