Vision-Based Traffic Density Calculation System Based on Area Detection

Abstract

With the rapid development of times, various intelligent traffic systems are trying to replace the traditional timing traffic system. One type of intelligent traffic system detects the number of vehicles to determine the priority of traffic lights. This kind of traffic control system has some shortcomings, which lead to some errors in the calculation of traffic density. In order to overcome this problem, a vision-based traffic light density calculation system by calculate area of vehicle as traffic density parameter is proposed. The system is developed using image processing technique and a smart traffic density calculation system graphical user interface is built for visualization purposes. As the result of simulation in MATLAB, it shows is more appropriate and logical to calculate the traffic density as the parameter than the number of vehicles. When calculating number of vehicles as a parameter for determining the traffic condition may occur some errors that lead the inaccurate od traffic density. For example, two vehicles are very close distance may cause the system recognize as one vehicle. The final suggestion is that when installing the system of this project, it is recommended to find the appropriate camera angle according to the local road section and adjust some parameters in the system code according to the current environment.