Using Target Detection Probability to Evaluate Area Coverage by a UAV

A common task for unmanned aerial vehicles (UAVs) is wide area search using an onboard camera with an object detection model. However, constraints of flight time, camera optics, and onboard compute, particularly in time sensitive applications like search and rescue, requires tradeoffs in strategies that balance precision and speed. To address these needs, we propose a novel method for evaluating coverage path plans by estimating the probabilities of detection and false alarm for ground targets for a set of poses that the UAV can reach in the search domain. To demonstrate our method, we evaluate trajectories for various coverage path plans flown by a UAV in a high-fidelity simulation.