Small Target Detection Algorithm Based on Improved YOLOv8s

Aiming at the underperformance of YOLOv8s in small target detection, an improvement method incorporating multiple structural optimizations is proposed. By introducing the P2 detection layer, optimizing the feature extraction module, and integrating the lightweight attention mechanism with the multi-scale fusion strategy, the AP_S performance of the model on VisDrone, TinyPerson and other datasets is effectively improved. The experimental results show that the improved model maintains the inference efficiency while the small target detection accuracy is significantly improved, which has good practical application value.