Vision-based UAV Detection for Air-to-Air Neutralization

The widespread availability of Unmanned Aerial Vehicles (UAVs) poses potential threats for people and properties on the ground, and other airspace users. This work introduces the design, development and validation of a UAV neutralization system that is based on another UAV with a capture device. The operation is fully autonomous, and only relies on data captured by two cameras onboard the captor UAV: one for long-range detections up to 40m, and another one for short-range accurate estimations prior to the actual capture. The approach has been extensively validated in field experiments, proving robustness and computational efficiency.