A computer-vision approach to traffic analysis over intersections

In recent years, there has been an interest in detailed monitoring of road traffic, particularly in intersections, in order to obtain a statistical model of the flow of vehicles through them. These models aid in the optimization of traffic management and allow for smarter transportation systems. While conventional methods sensors at each of the intersections entrances/exits allow for counting, are limited in the sense that it is impossible to track a vehicle from origin to destination. A solution is presented using computer vision algorithms to footage obtained from drones floating over an intersection, in order to identify and track vehicles so that a statistical model may be extracted. This model is based on said association of an origin and a destination. The algorithm employs background subtraction and vehicle tracking with Kalman filters. The most significant challenge was how to compensate for camera movement. Based on the implementation which was developed, this the approach proved useful for tracking cars, buses, and trucks under some conditions. This paper is based upon research work done for a masters dissertation project.