Video-based traffic management is easier to implement than ever before, and provides solutions for a wide variety of applications. Nestor Intelligent Sensors has applied neural network technology to develop its Traffic Vision family of advanced video-detection products. They `see' vehicles, and provide extensive vehicle and traffic data. The Traffic Vision system uses programmed rules and its Ni1000 Recognition Accelerator to extract video content for freeway and junction traffic monitoring. It processes pixel values in the entry zone to detect and acquire new targets, and provides an image composed of 640 x 480 pixels. It can generate a set of traffic sensor outputs, including: (1) vehicle flow rate; (2) vehicle classification; (3) lane changes; (4) queue count; (5) spatial headway; (6) average vehicle speed; and (7) vehicle occupancy. Traffic Vision was demonstrated at the official opening of the new Traffic Operations Centre (TOC) in Rhode Island State in the USA on 18 February 1998. It provides real-time traffic management capabilities that will help Department of Transportation (DOT) agencies to monitor traffic flow. The RI TOC provides its personnel with live video displays and ongoing real-time data acquisition.
Samenvatting