Application of feature enhancement method based on channel number optimization in small object detection

Small object detection remains a challenging problem in target detection due to ambiguous feature representation, susceptibility to background interference, and high requirements for localization accuracy. These challenges are further amplified in small object detection from unmanned aerial vehicle (UAV) perspectives. YOLOv8 is an effective algorithm for small object detection, and YOLOv8n-p2 is a variant that introduces a smaller object detection layer (P2 layer). In this paper, we further improve the model by optimizing the channel numbers of relevant layers based on YOLOv8n-p2 to enhance small object detection performance in UAV scenarios. Experimental results show that all four designed improved models outperform the original YOLOv8n-p2 model. Among them, the YOLOv8n-p2-p model demonstrates the most significant performance improvement, achieving an $8.3 \%$ increase in precision, a $4.9 \%$ increase in recall, a $7 \%$ increase in mAP@0.5, and a $4.8 \%$ increase in mAP@ [0.5:0.95]. These results validate the effectiveness of channel number optimization in further enhancing model performance.