Distinction of road surface conditions at nighttime using brightness co-occurrence matrix

In this paper, we propose a distinction method of road surface conditions by using only video information from a visible video camera. The feature of this method is simply to use the cars' headlights as the light source. In this method, using the probability function of brightness, we extract the road area that is having moderate quantities of light, touched by the headlights. Then, we differentiate two road surface conditions such as dry and wet, by using the amount of characteristic of the brightness co-occurrence matrix. As a result, by using this method, it became possible to distinguish the road surface conditions with high accuracy.