WHE-YOLO: A Lightweight PCB Surface Defect Detection Model Based on YOLOv8n

Detecting defects in printed circuit boards (PCBs) is crucial for ensuring the quality and reliability of electronic products. However, existing methods often struggle with complex processing methods and excessive parameters, which can lead to suboptimal performance in real-world applications. This paper proposes a novel approach, WHE-YOLO, which builds upon YOLOv8n by introducing a more efficient feature extraction mechanism (SPPELAN) and a lightweight neck structure (HSFPN). These innovations result in a 29 % reduction in model parameters and a 13 % decrease in computational costs while achieving a 3.4% improvement in mean Average Precision (mAP50). This work provides a more efficient and reliable scheme for PCB surface defect detection, with potential applications in real-time quality control systems.