Metal surface defect detection based on deep learning
Due to various factors in the production of metal products,some surface defects may exist on metal workpieces.This can reduce the strength of the material,shorten the life of the workpiece,and increase the safety risk.Therefore,it is necessary to carry out quality inspection on the surface of metal products,which is also a key link to ensure the quality of industrial production.Compared with the traditional manual inspection,the machine vision-based surface defect detection method has the advantages of high speed and high accuracy.An improved YOLOv5 algorithm is proposed for metal surface defect detection research,which replaces the spatial pyramid pooling structure SPP with SPPCSPC on the basis of the original YOLOv5 algorithm to improve the model's ability to detect defects on metal surfaces.In order to verify the effectiveness of the algorithm,a comparative test is conducted on 1 800 samples of metal surface defects using YOLOv3,YOLOv4,YOLOv5,and an enhanced YOLOv5 algorithm.The results show that compared with the original algorithms of YOLOv3,YOLOv4,and YOLOv5,the mean average target detection accuracy of the improved YOLOv5 algorithm has been improved by 4.3％,3.3％,and 2％,respectively.Abetter accuracy rate can be obtained by learning from a large number of images.
metal surface defectsspatial pyramidal pooling structuremachine visiontarget detection
万方数据
维普
