Intelligent Visual Monitoring Terminal for Low-Altitude UAV Detection
摘要
In recent years, the consumer small UAV (Unmanned Aerial Vehicle) industry has surged, but unauthorized UAV entries into airport clearance zones pose safety risks to civil aviation. Traditional air defense methods like radar are costly and labor-intensive. Hence, developing low-cost, low-altitude UAV detection terminals is crucial. In this paper, a low-altitude UAV visual monitoring terminal based on the YOLOv5 model is introduced, which aims to achieve rapid, accurate and real-time UAV target detection and track multiple targets in combination with the StrongSOTR model, and identify the unique ID number of the target, which is trained on the Ascend 910NPU using the UAVSwarm dataset, and integrated into NVIDIA Jetson Nano for real-time UAV detection and verification. When a target is found, the terminal can immediately alert and transmit information to the server, and at the same time, the camera can automatically adjust to the best viewing angle through the gimbal control function.