NdFeB-YOLOv10: An Intelligent Detection Network for Sintered NdFeB Raw Defects

Abstract NdFeB magnets, composed of neodymium, iron, and boron, are essential components in various technological applications due to their superior magnetic properties. In the production of rare earth permanent magnets, surface defects in NdFeB significantly impact their application, resulting in substantial economic losses. Given the high precision and large volume required for NdFeB surface defect detection (32,000 pieces per day), accurately detecting and sorting sintered NdFeB raw online is a challenging task. This paper proposes a fast and accurate surface defect detection model for NdFeB, which can allows real-time sorting based on defect detection results. Our model uses an attention module to improve detection accuracy and enable real-time control. Compared to the original model, the proposed model increases detection accuracy by almost 6% and achieves a detection speed of 29 FPS, meeting real-time performance requirements. This study provides new insights into the application of computer vision in the construction of intelligent factories.