Research on Surface Defect Detection of O-Rings in Flexible Production Lines Based on Deep Learning

Utilizing the YOLOv8 model, this study introduces a lightweight, rapid, and efficient improved method for detecting surface defects on O-rings in flexible production lines by incorporating AFPN, Focal Modulation Networks, and RT-DETR techniques. Experimental results demonstrate that the enhanced model outperforms traditional methods in both recognition accuracy and speed, especially when handling smaller defect targets and complex environments.