A Road Visibility Information System (RVIS) was developed that uses images transmitted by multiple closed-circuit television (CCTV) cameras as road visibility information. Feasibility studies in the daytime during the winters of 2005-2006 and 2006-2007 were performed on National Routes 230 and 231. During those two winters, the RVIS calculated visibility scale (VS) and road visibility index (RVI) automatically, accurately, continuously and in a timely manner from road images recorded by multiple CCTV cameras. The RVIS processed 90% of the images transmitted by fourteen CCTV cameras. The RVIS provided area-wide visibility information in many formats and through many kinds of media and devices. Based on stored road visibility data, we found a high correspondence between the RVI and the subjective visibility evaluation for the same road images. In addition, this study demonstrates that the RVI monitoring website and e-mail alerts can speedily and reliably inform road administrators of snowstorm-induced poor visibility. The RVIS has the potential to reliably collect data on road visibility during daytime using images transmitted by multiple CCTV cameras. However, it is necessary to improve data collection at night and to capture still images accurately and quickly.
Abstract