蚁群算法下焊接机器人焊缝表面图像裂纹检测  

Image Crack Detection of Welding Robot under Ant Colony Algorithm

作　　者：刘宇 赵辉[1] LIU Yu;ZHAO Hui(Xinjiang Institute of Engineering,College of Mechanical and Electrical Engineering,Urumqi Xinjiang 830023,China;School of Materials Engineering,Jilin University,Changchun Jilin 130015,China)

机构地区：[1]新疆工程学院机电工程学院,新疆乌鲁木齐830023 [2]吉林大学材料工程学院,吉林长春130015

出　　处：《计算机仿真》2024年第11期215-219,共5页Computer Simulation

摘　　要：焊接机器人在焊接时,由于电流过大、焊接速度过快等因素,导致焊接过热,从而产生裂纹。焊缝表面裂纹会导致焊接部位的强度大幅降低,严重影响设备的正常工作和使用寿命。为了精准检测出焊接机器人焊缝表面裂纹,提出一种基于蚁群算法的焊接机器人焊缝表面裂纹检测方法。采用小波和双边滤波相结合的方法,对焊接机器人焊缝表面图像去噪。根据去噪后的图像,采用蚁群算法结合模糊C-均值聚类算法,通过蚁群算法从去噪后的图像中获取初始聚类个数和聚类中心,将其作为模糊聚类的初始参数,同时对焊接机器人焊缝表面展开分割,提取焊缝表面裂纹特征,实现焊接机器人焊缝表面裂纹检测。实验结果表明,所提方法可以准确检测焊缝表面裂纹,且检测600张焊缝表面裂纹图像用时低于60s。During welding,the welding robot may cause cracks due to excessive current and high welding speed,which can significantly reduce the strength of the welded joint and affect the normal operation and service life of the equipment.In order to accurately detect the cracks on the welding surface,this paper presented a method for detecting the surface cracks of welding robots based on ant colony algorithm.Firstly,wavelet was combined with bilateral filter to denoise the welding surface images of welding robots.According to the denoised image,the ant colony algorithm was combined with the fuzzy C-means clustering algorithm to obtain the initial number of clusters and cluster centers from the denoised image as the initial parameters of fuzzy clustering.At the same time,the welding surface was seg-mented to extract the features of cracks,thus realizing the detection of welding surface cracks of welding robots.Experimental results show that the proposed method can accurately detect cracks,and it takes less than 60 seconds to detect 600 images of weld surface cracks.

关 键 词：蚁群算法 焊接机器人 焊缝表面 裂纹 模糊聚类 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]