DEVELOPMENT OF ARTIFICIAL INTELLIGENCE-BASED VEHICLE LICENSE PLATE RECOGNITION SYSTEM FOR MULTI LANE FREE FLOW TOLL E-COLLECTION

• Main Author: Ihsan Rasyidin, Ahadi
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/75457
• Institution: Institut Teknologi Bandung

Efficient transportation facilities are vital to support community mobility. As the number of vehicles continues to grow, it necessitates the development of appropriate and efficient transportation infrastructure. Toll roads have emerged as a significant effort to enhance transportation efficiency. However, conventional toll gate systems hamper toll efficiency, requiring drivers to stop their vehicles for toll fee payment. To overcome this challenge, the development of Multi Lane Free Flow (MLFF) has emerged as a promising solution. MLFF introduces contactless electronic payment technology, automatically charging the driver's account and eliminating the need to stop at toll gates, thereby reducing transaction time. MLFF relies on the accurate identification of vehicles to enable automatic toll fee billing. In this paper, we propose an advanced system utilizing artificial intelligence for license plate recognition as vehicle identity on MLFF. The system utilizes the YOLOv8 model for precise license plate detection and employs OCR algorithms for license plate character recognition. The OpenCV library is utilized for effective pre-processing of license plate data. Through rigorous testing, the developed system achieved an impressive accuracy of 91.2%. The proposed approach demonstrates the potential to significantly enhance toll road efficiency, providing a seamless travel experience for road users in Indonesia. The position of capturing the license plate image and the presence of other objects on the license plate influence the system's accuracy.