基于改进YOLOv5的焊缝识别算法研究  

Research on weld identification algorithm based on improved YOLOv5

作　　者：周翌晨 虞旦[1] 李佳成 蔡春波 张华军[2] ZHOU Yichen;YU Dan;LI Jiacheng;CAI Chunbo;ZHANG Huajun(College of Logistics Engineering,Shanghai Maritime University,Shanghai 201306,CHN;School of Materials Science and Engineering,Shanghai Jiao Tong University,Shanghai 200240,CHN)

机构地区：[1]上海海事大学物流工程学院,上海201306 [2]上海交通大学材料科学与工程学院,上海200240

出　　处：《制造技术与机床》2024年第7期19-25,共7页Manufacturing Technology & Machine Tool

摘　　要：针对机器人自适应打磨焊缝的问题,文章提出一种基于YOLOv5改进的焊缝检测算法,实现焊缝的识别和初定位。使用焊接机器人制作各类不同参数和形貌的焊缝,自制一个包含3 996张焊缝图像的数据集用来深度学习。选用YOLOv5s模型进行训练,在Backbone中添加了GAM注意力机制模块;同时引入GhostNet,用GhostConv模块和C3Ghost模块替换原模型的Conv模块和C3模块。改进后的YOLOv5s-GhostNet-GAM模型的mAP@0.5达到了90.21%,相比原YOLOv5s模型提高了4.05%,同时参数量减少了5.64%,FLOPs降低了27.44%,检测速率为23.47 FPS,达到了机器人自适应打磨焊缝对识别精度与后期软件部署的要求。To solve the problem of robot adaptive weld grinding,this paper proposes a weld detection algorithm based on improved YOLOv5,which realizes the identification and initial positioning of the welds.The welding robot is used to produce weld seams with different parameters and shapes,and a dataset containing 3 996 weld images is self-made for deep learning.The YOLOv5s model is selected for training,and the GAM attention mechanism module is added to the backbone.At the same time,Ghost Net is introduced,and the Conv module and C3 module of the original model are respectively replaced by the Ghost Conv module and C3 Ghost module.The m AP@0.5 of the improved YOLOv5sGhost Net-GAM model reaches 90.21%,4.05% higher than that of the original YOLOv5s model.At the same time,the number of parameters is reduced by 5.64%,and the FLOPs are reduced by 27.44%.The detection rate is 23.47 FPS.It meets the requirements of robot adaptive weld grinding for identification accuracy and later software deployment.

关 键 词：机器人焊缝打磨 焊缝识别 深度学习 YOLOv5模型 GAM注意力机制 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]