Abstract:
All E-Grade expressways (commonly called highways) in Sri Lanka are levying a user charge
depending on the type of vehicle and currently the fee-levying operation is totally manually
handled. When a vehicle enters an expressway, an electronic ticket is issued at the entrance
interchange mentioning the type of vehicle and the name of the interchange entered. When the
vehicle reaches the exit point, the ticket received at the point of entry is handed over to the
exit toll gate and the user charge based on the vehicle type is informed. The user has to make
the payment only in cash to the toll gate and a receipt is issued for the payment. After
finishing the payment, the barrier gate is opened and the vehicle can pass through the lane.
The current system of operation has led to wastages of time both at the entry and exit points of
the expressways. The intent of this research is to propose an automated expressway fee-levy
system using Computer Vision and Artificial Intelligence concepts. According to the proposed
fee-levy system, the vehicle registration number is detected and recognized using a digital
camera and Automatic License Plate Recognition (ALPR) software at the entrance, and the
date, time and point of entry are saved securely in a database. The payment can be made using
a credit/debit card and mobile cash while in transit. At the exit point, the registration number
of the vehicle is again recognized and the relevant entry details are traced. Exit toll gate
barriers open automatically for those vehicles that made the relevant payments while in
transit. This mode of operation will reduce waiting time at the entry and exit points
particularly during weekends, long holidays and festive seasons. For other non-regular
vehicles, once the user fee is paid at the exit, the toll gate barriers open automatically. The
system was implemented using MATLAB, OpenCV library and NewSQL. Experimental
results have shown that the system is capable of recognizing vehicle registration number
plates, in all standard formats provided by Department of Motor Traffic Sri Lanka, even in
noisy images. The system works satisfactorily even under wide variations of illumination in
real-time situations.
