Contrast model for three-dimensional vehicles in natural lighting and search performance analysis

Ground vehicles in natural lighting tend to have significant and systematic variation in luminance through the presented area. This arises, in large p art, from the vehicle surfaces having different orientations and shadowing relative to the source of illumination and the position of the observer. Th ese systematic differences create the appearance of a structured 3-D object . The 3-D appearance is an important factor in search, figure-ground segreg ation, and object recognition. We present a contrast metric to predict sear ch and detection performance that accounts for the 3-D structure. The appro ach first computes the contrast of the front (or rear), side, and top surfa ces. The vehicle contrast metric is the area-weighted sum of the absolute v alues of the contrasts of the component surfaces. The 3-D structure contras t metric, together with target height, account for more than 80% of the var iance in probability of detection and 75% of the variance in search time. W hen false alarm effects are discounted, they account for 89% of the varianc e in probability of detection and 95% of the variance in search time. The p redictive power of the signature metric, when calibrated to half the data a nd evaluated against the other half, is 90% of the explanatory power. (C) 2 001 Society of Photo-Optical Instrumentation Engineers.