Virtual Toll Collection System using YOLOv8 and OCR

This research work proposes a real-time Automatic Toll Detection System that is designed to simplify the processes of toll management. It uses advanced machine learning and cloud technology to enhance efficiency and accuracy. The system offers a modern solution for seamless toll operations. Using real-time video inputs, the suggested system precisely detects and locates vehicle license plates using the YOLOv8 object identification model. Tesseract OCR is then used to extract alphanumeric text from the identified license plate photos, transforming the visual input into legible car registration numbers. The system incorporates Firebase, a cloud-based database that stores and retrieves information including vehicle type, registration number plates, and associated toll costs, to manage toll information and car-related data. Unique characteristics of the system include 24-hour two-way toll concessions, which allow for effective cost management for cars making several trips in a predetermined amount of time. To give administrators an easy-to-use interface for visualizing toll-related data, such as the total toll collected, the number of cars processed, and a breakdown of vehicles by type, a real-time Node-RED dashboard is also created. Transparency is guaranteed, monitoring. The proposed solution automates toll collection, reducing delays, errors, and inefficiencies associated with manual methods. Additionally, the dashboard provides real-time insights for efficient toll operations. High accuracy, shorter processing times, and increased toll management efficiency are all achieved by the system’s integration of contemporary IoT technologies with machine learning (YOLOv8) and OCR tools. With possible uses in parking lots, roads, and smart city infrastructure, this work presents a scalable, dependable, and affordable real-time toll detection method.