Super-ADF: An enhanced algorithm for metal surface defect detection

Due to the aging of industrial equipment, safety accidents have become increasingly frequent. The prolonged use of mechanical equipment often leads to corrosion and rust on metal surfaces, resulting in defects and cracks. Therefore, metal surface defect detection plays a crucial role in quality control within the industrial and manufacturing sectors. To address this issue, this paper proposes a metal surface defect detection method based on an improved YOLOv9 model, named Super-ADF. Extensive experiments demonstrate that Super-ADF achieves an mAP of $84.0 \%$ on the NEU-DET dataset, representing an $8.5 \%$ improvement over the baseline YOLOv9 model. The results highlight that the proposed Super-ADF significantly outperforms traditional YOLOv9 in defect detection tasks, particularly in boundary detection and complex background scenarios. This represents a critical breakthrough for addressing metal surface defect detection challenges in industrial applications.