DESIGN AND DEVELOPMENT OF VEHICLE LICENSE PLATE RECOGNITION FOR INTELLIGENT PARKING PAYMENT SYSTEM USING OBJECT DETECTION, IMAGE PROCESSING, AND OPTICAL CHARACTER RECOGNITION

• Main Author: Alvindo Riandova, Ilham
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/74791
• Institution: Institut Teknologi Bandung

License Plate Recognition (LPR) is an important technology in various applications, including smart parking systems. This final project focuses on building a smart parking payment system using machine learning algorithms. There are five main subsystems developed: machine learning, blockchain, backend, Internet of Things (IoT), and web application. The research primarily focuses on the modeling of machine learning algorithms used for license plate recognition. License plate recognition enables payment based on the vehicle's license plate number. In this parking system, drivers only need to pass through the parking gate equipped with a camera for license plate recognition. The license plate recognition system is developed using YOLOv5 for license plate detection, image processing to enhance image quality, and Tesseract OCR for reading the license plate numbers. Based on the experimental results, it is found that YOLO performs well under sufficient lighting conditions, with an average confidence score of 95.7%. On the other hand, Tesseract OCR is also effective in reading Indonesian license plate characters, achieving an accuracy of 90.52 for Indonesian plate numbers. In terms of integration testing, the smart parking payment system has an average processing time of 10.4 seconds for check-in and 11.1 seconds for check-out.