Vehicle Detection Based on Video for Traffic Surveillance on road

Abstract

Today most of the cities of the world have intelligent transport system which is equipped with electronics devices to communicate about the traffic condition with the moving vehicle and also monitor the traffic rules and regulation. Today most of the vehicles are equipped with on-board automotive driver assistance system aiming to alert drivers about driving environments, and possible collision with other vehicles has attracted a lot of attention lately. In these systems, robust and reliable vehicle detection is a critical step. This paper presents a review of recent vision-based on-road vehicle detection systems. Our focus is on systems where the camera is mounted on the vehicle and also being fixed such as in traffic/driveway monitoring systems. First, we discuss the problem of on-road vehicle detection using optical sensors followed by a brief review of intelligent vehicle research worldwide. Then, we discuss active and passive sensors to set the stage for vision based vehicle detection. Methods aiming to quickly hypothesize the location of vehicles in an image as well as to verify the hypothesized locations are reviewed next. Tracking of vehicle is also reviewed to illustrate the benefits of exploiting temporal continuity for vehicle detection.