DEVELOPMENT OF A VEHICLE MANAGEMENT SYSTEM FOR CAMPUS SECURITY USING DEEP LEARNING

ABSTRACT

This project aims to present the development and implementation of an extensive system that enables automated license plate detection, text recognition, and vehicle speed estimation through the use of advanced deep-learning techniques namely YOLOv8 and EasyOCR. The system leverages YOLOv8 for precise license plate detection, EasyOCR for robust text recognition, and a customized vehicle detection model, coupled with motion tracking algorithms, for accurate speed estimation. The system's performance has been thoroughly evaluated using meticulously collected datasets that include various conditions, such as diverse lighting and backgrounds. The results demonstrate that the system has remarkable performance in detecting and recognizing license plates. Furthermore, the integration of vehicle tracking across frames enables accurate speed calculation, making the system highly applicable in real-world vehicle monitoring scenarios. The study also addresses challenges encountered during implementation, such as dataset quality and algorithm performance, and proposes avenues for future research and improvement. This developed system represents a significant advance in computer vision with potential applications in vehicle surveillance, law enforcement, and road safety enhancement.