Detection of road surface conditions in winter using CCTV camera images

As vehicle driving is heavily influenced by road surface conditions, especially in the winter season, it is essential for ensuring traffic safety in this season to properly detect the condition of road surfaces. Though various methods to detect road surface conditions have been proposed, a method using visible cameras to detect conditions at nighttime has not been established. This paper proposes the method using images acquired through CCTV cameras on the roads in order to detect road surface conditions in winter both during daytime and nighttime. Texture features and the Mahalanobis distance are used in this method to detect the condition of road surfaces divided into blocks. It was concluded that the method can effectively detect dry, wet and snowy conditions.