Road Marking Contrast Threshold Revisited

Sufficient contrast between road surface and road markings is key for a safe and comfortable driving experience. This calls for a comprehensive and well established contrast (threshold) model, which ideally results in a single contrast threshold value independent of object angular size or road luminance. The contrast threshold model introduced by Adrian is still commonly used in road lighting. More recently, new contrast metrics that also predict supra-threshold contrast visibility have been proposed, but the corresponding visibility thresholds are not yet known. In the present study, participants are presented a rendering of a highway, including road marking arrows of various size and luminance and were asked to indicate the direction of the arrow. The luminance of the road surface, acting as background for the markings, was varied too. Due to the very low luminance values and the very small differences in luminance, measurement accuracy and calibration issues require special attention. The results show good agreement with Adrian's visibility model (R-2 = 0.75) in terms of luminance contrast, background luminance and size. In addition, we used our experimental data to define contrast thresholds for several other existing image based contrast models. Unfortunately, it seems to be impossible to state one unique threshold contrast value independent of object angular size and road luminance.