Time-independent detection of road surface conditions in winter using road surveillance cameras

The stopping distance of vehicles on wet or snowy roads is considerably longer than in dry conditions. For this reason, detection of road surface conditions is a critical issue for safe driving in winter. Conventional detection methods face many problems like deployment costs and detectable time zone. This paper proposes a method to detect road surface conditions in winter, which is time-independent and uses road surveillance cameras on the roads. The method detects road surface conditions based on texture features and the Mahalanobis distance. It has been concluded that road surface conditions can be identified as wet or snow using road surveillance cameras at daytime, twilight and night-time.