Application of Image Processing to the Analysis of Congested Traffic

Traffic congestion has become a significant problem in all developed countries. This is mainly due to the increasing number of vehicles but also to the fact that the infrastructures are usually not designed to take over all this traffic. As a result of this increasing number of vehicles on the roads, bridges are becoming serious strangulation points in the transport system. This issue is more important because most of bridges are either approaching or have surpassed their expected design life and traffic data traditionally collected with inductive loops detectors do not provide information about congested traffic situation. Due to this drawback, it needs a better solution for traffic monitoring. The aim of this paper is to explore the capabilities of using images for applications on transport, especially for traffic monitoring, to extract information about traffic such as gaps between cars, cars and trucks, or trucks. In that sense, a high resolution camera will be used in this work in order to capture aerial images of congested traffic. These images will be processed to distinguish all vehicles as different objects on the road, to identify the type of vehicles (regular cars or trucks) and to measure the length for each vehicle. In order to achieve that result, an algorithm able to detect and count the vehicles on the road as separated objects will be firstly applied, enclosing each object within a rectangle.