The Defects Detection in Steel Coil End Face Based on SCED-Net

In the steel coil manufacturing industry, the quality of the steel coil end face is not only related to the product itself, but also directly affects the equipment safety of downstream production lines. Therefore, the detection of steel coil end face defects has a pivotal position. This paper proposes a SCED-Net (Steel Coil End face Detection Network) based on the latest YOLO series masterpiece YoloX. In this paper, by supplementing the data augmentation method, combining the proposed multi-information fusion module, while introducing focal loss and soft-nms, SCED-Net has achieved comparativly better results. Compared with the networks used in the field of steel coil end face detection in recent years and some excellent classic object detection networks, our proposed method solves the steel coil end face detection problem with outstanding effectiveness.