Video image processing systems: applications in transportation

California Polytechnic State University, USA, completed an assessment of the potential of using a video image processing (VIP) system for traffic surveillance and detection on California freeways. The first phase of the project, sponsored by the California Department of Transportation, was completed in 1991. Eight turnkey and prototype systems were identified and tested under laboratory conditions. The results showed, that while no system tested was ready for deployment at that time, several systems warranted further testing under more realistic conditions. The second phase of the project began in June 1992 and was completed recently. Four VIP systems were chosen for extended field testing to measure accuracy in traffic counts and speeds, classification, reliability, cost, ease of set-up and operation, image quality, failure rate, and ease of recovery. These variables were assessed quantitatively and qualitatively under specific conditions that included variations of lighting, approaching and departing traffic, camera angle, weather, vibration, and presence of signal noise. This paper presents the Phase II field test results and recommendations for standardizing image processing systems with a view toward eventual statewide implementation for traffic detection.