Hawkeye Target Tracking System --- An Air-Ground Distributed Computing Intelligent Aerial Robot Target Tracking System

In recent years, rotor drones have been increasingly applied in disaster rescue scenarios, yet there remains a significant shortage of skilled drone operators. Therefore, research on drone target tracking systems is of great significance for intelligent drone-based rescue operations. This paper proposes an intelligent aerial robot target tracking system based on air-grounddistributed computing (Hawkeye target tracking system), enabling drones to be applied toemergency rescue tasks in an automated, low-cost, and long-endurance manner. The systemuses YOLOv8 target detection algorithm on the ground station to search for targets and theKCF tracking algorithm on the drone for visual target tracking. Finally, visual servo tracking is performed based on the targets position in the image. By employing air-ground distributed computing, this method significantly reduces the computational demands on the airborne side, lowers costs, and improves drone endurance. Simulation and real flight tests demonstrate the reliability and cost-effectiveness of the system. Compared to the Jetson Nano, our systemachieves similar results with an 87% cost reduction and a 92% weight reduction.