Road Weather Management Program
photos of lightning, trucks plowing snow, an empty road before a storm, and an evacuation
Office of Operations 21st century operations using 21st century technologies

Best Practices for Road Weather Management Version 2.0

Title:

Development of a Visibility Estimation Model Based on Visibility Information from Road Images Captured in Winter

Abstract:

This study proposes a visibility estimation model in winter that uses multiple regression analysis. The Road Visibility Information System (RVIS) can provide only visibility information at the present time. Road maintenance offices and drivers need to know road visibility information a few hours in advance. The project used a database of visibility information stored by the RVIS and 1-km-mesh meteorological data recorded during the winter of 2005/2006 to develop the visibility estimation model. Four kinds of multiple regression analysis were established with WIPS values as the dependent variable and 1-km-mesh meteorological data as independent variables. The best of the four models has a 60 percent hit rate, an 80 percent extended hit rate and a 2 percent undetected rate. The study revealed that visibility information based on road images captured in winter can be estimated from 1-km-mesh meteorological data using a multiple regression analysis. Implementation will require increased accuracy under poor visibility conditions, expansion to fog-induced poor visibility adding relative humidity as a factor, and overcoming inaccuracy of 1-km-mesh meteorological data under poor visibility conditions using the advanced method.

Source(s):

87th Transportation Research Board (TRB) Annual Meeting, Hokkaido University, Japan Weather Association and Hokkaido Development Engineering Center. For an electronic copy of this resource, please direct your request to WeatherFeedback@dot.gov.

Date: 2007

Author:

Minami, et al

Keywords:


Winter storm
Closed Circuit Television (CCTV)
Air temperature
Wind
Snow

PDF files can be viewed with the Acrobat® Reader®.

Office of Operations