A Simplified Traffic Flow Monitoring System Using Computer Vision Techniques

Millions of Muslims from all over the world visit the holy city of Makkah to perform Umrah throughout the year, but especially in the month of Ramadan the quantity of pilgrims increases significantly. Same conditions can be observed in the month of Dhu al-Hijjah during the Hajj season. To transfer the pilgrims from one place to another various many vehicles are used. It is very important to monitor the traffic flow during these days to avoid any congestion which may lead to any mishap. Therefore, there is a need for a traffic flow monitoring system that can be used to count vehicles, and classify them according to their size to facilitate traffic management. A computer vision based system is proposed in this research that uses MATLAB vision tool to identify and count the vehicles. The novelty of the system is its simplicity that uses a camera as a video source, then suitable algorithms are implemented on it for processing. A portion of a video is used for system training that first obtains a binary foreground from the initial frame, then removes noise and then counts the vehicles present in the frame. The functionality of the proposed system has been verified by a number of experiments with the accuracy of 90%. It can be easily implemented in various environments for vehicles counting that can be used for better traffic management.