Improved AADT Estimation by Combining Information in Image- and Ground-Based Traffic Data

Abstract

On most segments in a state highway system, annual average daily traffic (AADT) in the current year is estimated by applying an applicable growth factor to coverage count data obtained in an earlier year. The accuracy of AADT estimates could conceivably be improved by exploiting existing imagery of highway segments containing vehicles. When both earlier year coverage counts and a current year image containing traffic information are available, a weighted combination of both pieces of information is proposed for AADT estimation. The weights can be easily estimated from available information and data regularly collected by state departments of transportation. An empirical study was conducted using 122 highway segments to evaluate the improved accuracy in estimating AADT that would have resulted from using this weighted method with a single contemporary image for a 10-year period from 1994 and 2003. The accuracy was markedly improved and stable over a large range of important input values. The demonstrated improvements in accuracy and the ease of using this method with existing data are great enough that field testing should now be considered.