ADA-YOLO: An Adaptive Dynamic Aggregation Network for Small Object Detection in UAV Imagery.

Area of Science:

• Computer Vision
• Artificial Intelligence
• Remote Sensing

Background:

• Unmanned Aerial Vehicle (UAV) image object detection is crucial for applications like traffic monitoring and disaster rescue.
• Standard detectors struggle with UAV imagery due to small targets, dense distribution, occlusions, and complex backgrounds.

Purpose of the Study:

• To develop an efficient and accurate object detection model specifically for UAV small-object detection.
• To enhance the performance of YOLOv8 for complex aerial imaging scenarios.

Main Methods:

• Proposed the Adaptive Dynamic Aggregation YOLO (ADA-YOLO) network, building upon YOLOv8.
• Incorporated a high-resolution P2 detection branch for P2-P5 multi-scale prediction.
• Introduced the DySample dynamic upsampling module and Adaptive Spatial Feature Fusion (ASFF) for improved feature fusion.

Main Results:

• ADA-YOLO achieved an 11.3% increase in mAP@0.5 and an 8.2% increase in mAP@0.5:0.95 on the VisDrone2019 dataset compared to YOLOv8n.
• The model demonstrated performance gains with minimal parameter increase and acceptable computational cost.
• Ablation studies confirmed the effectiveness of individual modules and their combined synergistic effect.

Conclusions:

• ADA-YOLO effectively addresses the challenges of small-object detection in UAV imagery.
• The proposed network architecture offers a significant improvement in accuracy and robustness for aerial surveillance applications.
• The adaptive feature fusion and dynamic upsampling contribute to superior performance in complex environments.