Visual Information Requirements for Dismounted Soldier Target Acquisition

Abstract

We conducted an empirical investigation of the visual information requirements for target detection and threat identification decisions in the dismounted soldier context. Forty soldiers viewed digital photographs of a person standing against a forested background. The soldiers made two-alternative detection decisions requiring them to determine whether the target was present in the scene, and two-alternative threat identification decisions that required discrimination of the objects held by the target, the clothing worn by the target, and target postures. The images were presented to subjects on a computer display, and variation in the apparent target distance was simulated through digital image magnification and by varying the viewing distance to the display. Image resolution was degraded progressively by spatial frequency filtering and we estimated the resolution threshold in each task. These threshold values were compared with the historical Johnson criteria for predicting imaging device performance. Our data are broadly consistent with the previously reported values, though our threat identification decisions required subjects to perceive information with a larger spatial scale than the Johnson criterion for identification of standing human targets. In a second experiment, we employed a four-alternative identification decision and found results that were consistent with those from Experiment 1. We also confirmed that the spatial scale of visual information used for target acquisition is highly task-specific, and provided a novel demonstration of changes in visual information requirements as a function of target range. These findings pose challenges for models of target acquisition with imaging devices.