A Lightweight Building Detection Algorithm Based on Attention Mechanism for UAV Aerial Images

With the advancement of massive data processing in cloud computing, building object detection technology for UAV aerial images exhibits high recognition rates at limited computational resources. Yet it remains challenging to recognize in complex backgrounds (e.g. cloud cover and shifting shadows) and small objects. To address these issues, we proposed a lightweight YOLO11n_CBAM algorithm to achieve highprecision detection. The CBAM module is utilized to effectively improve the generalization ability of the model, ensuring the sustained maintenance of high performance when confronted with novel tasks. Experiment results demonstrate that our proposed YOLO11n_CBAM algorithm achieves a 6.3 % and 5.4 % improvement in mAP@50 and mAP@50-95, respectively, on the VisDrone2019 dataset compared to the YOLO11n (baseline).