Determining Human Perceptual Envelope in Fixed Wing UAV Surveillance

Detecting and recognizing targets from unmanned aerial vehicle (UAV) video is a common task for operators controlling, supervising, or monitoring UAVs. While operators everywhere rely on visual interpretation of UAV videos to perform their tasks, we are unaware of prior work quantifying the difficulty of these subject for this domain particularly at the extremes. While target detection and recognition may pose few problems for low and slow flying UAVs, more demanding applications requiring rapid surveillance of larger regions may approach and exceed the limitations of human performance. Detecting and recognizing targets from a fast moving, low flying platform such as a Group2 fixed wing UAV poses two complementary problems: 1) as altitude increases apparent target size decreases making it more difficult to discriminate targets; and 2) when altitude decreases, targets appear larger and are easier to discriminate, whereas targets remain on screen for less time shortening the opportunity to search for them. To comprehensively investigate humanUAV perception at various altitudes, in our work, operator detection and recognition performance was tested crossing 3 UAV speeds, 5 altitudes, and 3 levels of target conspicuity. While performance favored lower altitudes and slower speeds, detection and recognition performance remained robust within the envelope tested.