A novel traffic light recognition method for traffic monitoring systems

This paper proposes a traffic light recognition method for traffic monitoring system at road intersections. In the traffic light detection algorithm, the Y value statistic incorporated with the color analysis in the YUV space is employed to detect the traffic light coarsely in the recognition region. The constraint strategy is used to refine the detection result. In the experiment, the test videos were recorded in advance by the recorders of actual E-police systems at several different city intersections. Based on all experimental results, the average probability of false alarm is 0.073, and the average probability of missed plates is 0.071. Moreover, a set of traffic examples are provided to demonstrate the effectiveness of the proposed system.