A Lightweight UAV Image Object Detection Algorithm Based on Hybrid Attention Mechanism

With the wide application of Unmanned Aerial Vehicles (UAVs) in fields such as infrastructure monitoring and environmental perception, anomaly detection based on visual sensors has become an important research direction. Although existing object detection algorithms, such as the YOLO series, have made progress, they still face significant challenges in small object detection, computational complexity, and real-time performance. This study proposes a lightweight object detection algorithm based on a hybrid attention mechanism, which combines an enhanced multi-head self-attention mechanism and a linear attention mechanism, significantly improving the detection ability of small objects while reducing computational complexity. Furthermore, a lightweight backbone network is designed to address the hardware resource constraints of UAVs, improving detection efficiency and real-time processing. Experimental results demonstrate that the proposed EffiUAV-YOLO model outperforms traditional YOLO models in detection accuracy and real-time performance on the VisDrone2019 and UAVDT datasets, confirming its excellent adaptability and application potential in complex environments.