Research on improved yolov5 quay crane defect detection algorithm

To solve the safety and work efficiency issues during the inspection process of quay crane, surface defects of quay cranes are prone to missed and false detections. A surface defect detection algorithm for quay cranes based on YOLOv5 improvement has been proposed. The CA attention mechanism has been introduced to more accurately locate defect targets on the surface of quay cranes, and the loss function CIoU has been replaced by NWD, the detection performance of the model has been significantly improved, with an average detection accuracy increase of 4.2%. Experimental data confirms that the improved model performs well in identifying surface defects of quay cranes, validating its effectiveness.