Designing human friendly road information board for highway drivers from the viewpoint of multiple spatial frequency channel model of human visual system

Since the first road information board in Japan in 1966, many more have been introduced and have become an indispensable means of providing drivers with information relevant to changing transport conditions. Studies on the visibility of road information boards have been mostly based on empirical approaches. We implement a model of the human visual system to establish a more rational method of designing information boards. We report on a model focused on spatial frequency characteristics. While driving on highways, drivers may be visually stimulated by scenes consisting of a somewhat constant spatial frequency component with some specific orientations. Drivers may be subject to visual adaptation specific to the spatial frequency as well as its orientation. In order to cope with such perceptual phenomena, we simulate the human visual system with a multiple frequency channel model. The model was proposed by Cambell and Robson (1968). The envelope of the set of frequency characteristics constitutes the well-known modulation transfer function. As a preliminary version, each channel is modeled with a set of four Laplacian-of-Gaussian type filters functioning in parallel whose band-pass widths are rather wide, and non-specific to orientation. The model was implemented and applied to Kanji characters in a conventional information board to simulate how they look to a human driver at certain realistic viewing distances. In addition, we examine the spatial frequency characteristics with a set of grating patterns.