Target Conspicuity for Human-UAV Visual Perception

With the burgeoning application of unmanned aerial vehicles (UAVs), UAV-based target detection and recognition have become a central task for operators who control, supervise, and monitor UAVs. Unfortunately, probabilities of target detection and recognition are complex functions of physical properties of UAV video as well as their interactions with human visual and perceptual systems. Due to lateral masking, conspicuity, the degree to which objects stand out against their background is a crucial feature in visual search. To advance the study of human-UAV interaction, we have extended and validated conspicuity measures to accommodate altitudes and target sizes relevant for UAV surveillance. A fortunate characteristic of this measure, after passing a near distance threshold, is the constant relationship between altitude and conspicuity that allows conspicuity to be measured once and predicted across a range of altitudes. Using software that allows rapid psychophysical measurement of conspicuity, large simulation scenarios and testing environments can now be populated with targets of known conspicuity.