基于改进YOLOv5s的汽车白车身焊点检测  

Solder Joint Detection of Automobile Body in White Based on Improved YOLOv5s

作　　者：从桂浩 杨志芳[1] 张强 CONG Guihao;YANG Zhifang;ZHANG Qiang(School of Electrical and Information Engineering,Wuhan Institute of Technology,Wuhan 430205,China)

机构地区：[1]武汉工程大学电气信息学院,武汉430205

出　　处：《自动化与仪表》2024年第10期113-117,130,共6页Automation & Instrumentation

基　　金：武汉工程大学研究生教育创新基金项目(CX2023560)。

摘　　要：针对传统的基于机器视觉的汽车焊点检测中焊点不规则及易受环境光线影响导致产生漏检错检的问题,该文提出一种改进YOLOv5s的算法YOLOv5s_CB_SI。该算法通过在主干网络Backbone引入CBAM卷积注意力机制以提高模型对图像重要区域信息的关注度和对目标缺陷的学习能力,并引入SIoU定位损失函数,加快模型的收敛速度,而且有效提升模型检测及定位的能力。将改进后的算法与YOLOv5s法在焊点数据集上作对比,实验结果表明,该文提出的算法与原始算法相比精确率和召回率分别提升了4.8%和2%,在有效地提升了在焊点检测过程中的准确率的同时减少了错检漏检率,证明了改进后模型的有效性。Aiming at the problem of missed detection and error detection caused by the irregularity of solder joints and the vulnerability to environmental light in the traditional machine vision-based automobile solder joint detection,this paper proposes an improved YOLOv5s algorithm YOLOv5s_CB_SI.The algorithm improves the model attention to the important area information of the image and the learning ability of the target defect by introducing the CBAM convolution attention mechanism in the backbone network Backbone,and introduces the SIoU positioning loss function to accelerate the convergence speed of the model,and effectively improve the model detection and positioning ability.The improved algorithm is compared with the YOLOv5s method on the solder joint data set.The experimental results show that the accuracy and recall rate of the proposed algorithm are improved by 4.8%and 2%respectively compared with the original algorithm,which effectively improves the accuracy in the solder joint detection process.At the same time,the false detection rate is reduced,which proves the effectiveness of the improved model.

关 键 词：目标检测 损失函数 YOLOv5s算法 注意力机制 焊点识别 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]