Determination of vehicle origins and destinations, calculation of travel times and measurement of traffic congestion are the common goals of traffic surveys. In the last few years the use of machine vision has been able to increase the accuracy of these measurements, the turn around time of traffic surveys and improve the breadth and depth of the information that can be capture regarding these traffic flow parameters. This paper describes several surveys that have been conducted using machine vision, in both the United States and Europe. The paper presents the preliminary results of field trials conducted in several United States cities to calculate travel times, as well as the use of machine vision to address certain `micro survey' data requirements. The paper discusses the experience gained in conducting these surveys.
Salvatore A. D Agostino,
"Machine vision-based traffic surveys", Proc. SPIE 2592, Collision Avoidance and Automated Traffic Management Sensors, (27 December 1995); doi: 10.1117/12.228907; https://doi.org/10.1117/12.228907