Improved YOLOv8-based defect detection model for hot rolled strip steel

Production defects in hot-rolled strip steel, arising from process design issues or variations in material properties, adversely affect both economic returns and production safety. Existing deep learning–based surface defect detection methods are often too slow and computationally heavy For time-critical industrial operations. This paper proposes an optimized YOLOv8s algorithm for defect detection on hot-rolled steel strips. By integrating Cloformer with the CBAM attention mechanism, the model effectively enhances its capability to capture spatial relationship features. Additionally, the combination of partial convolution and depthwise separable convolution significantly reduces computational load. The introduction of the SimSPPF module further improves inference speed. Experimental experimental data reveals that the improved YOLOv8s model achieves 80.4% mAP (mean average precision) and 182.4 FPS inference speed while substantially reducing model size compared to conventional approaches.