Surface Defect Detection of Industrial Parts Based on YOLOv5

Industrial product quality inspection, a crucial procedure in industrial production, is assuring yield. Product safety and inspections on assembly lines are predominantly manual, there currently dearth of safe dependable inspection techniques. An improved surface defect detection approach based YOLOv5 proposed for the problem flaws components order to improve effect production parts. To dense object detection, image features extracted by convolutional network enhanced coordinate attention. BiFPN utilized fuse multi-scale lower rate missed false small target samples. The detectors from Transformer structure added complex fine-grained predictability challenging occurrences. According experimental findings, dataset parts defects, increases recall original algorithm abnormal classes 5.3%, reaching 91.6%. Its inference speed can 95FPS, indicating an real-time performance.