Application Research on Surface Defect Detection of Guide Claw of Automobile Safety System Based on YOLOv5
ZHANG Qianqian
HE Shirong
ZHAO Shiyu
ZHANG Haoyang
Abstract：Aiming at the stamping part of the guide claw of the automobile safety system,a target detection algorithm based on YOLOv5 is proposed to detect the surface defects of the guide claw of the automobile safety system.Firstly,the original data is fil-tered to remove invalid data.Secondly,to solve the problem of unbalanced samples,some data are amplified.Then the LabelImg tool is used to label the expanded data set.Then the labeled data set is put into the model perform training.Finally the optimal weights obtained from training are used for testing.Experiments show that the detection effect of YOLOv5 is significantly better than the three target detection models of Faster R-CNN,SSD,and RetinaNet.Its mAP can reach 98.4%,and the average detection time is 6.2 ms.The defect detection method based on YOLOv5 can meet the detection speed and accuracy requirements of the actual pro-duction line,and has important guiding significance for the defect detection of stamping parts.
Keywords：stamping partsguide claw of automobile safety systemYOLOv5target detection
Publication Date：2025-01-19
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)
Pages：7( 77-83 )