Publication

Abstract : The rapid urbanization and increased vehicular population in metropolitan areas have presented a formidable obstacle to the manual identification of license plates. The surge in vehicular activity has led to a complex task, requiring efficient surveillance of each vehicle for vital functions such as theft prevention and traffic regulation. In response to this intricate challenge, the deployment of an Automatic License Plate Recognition (ALPR) system emerges as a crucial necessity for smart city transportation. This research aims to develop an efficient ALPR system employing YOLOv8, the latest iteration in the YOLO series of object detection models, for license plate localization, and utilizing Efficient Net B7 for character recognition on the detected plates. The YOLOv8 license plate detection module achieved a mean Average Precision (mAP 50) of 99.5%, ensuring robust and reliable license plate localization. The character recognition component, utilizing Efficient Net B7, achieved an accuracy of 98.22%, further enhancing the system's overall performance and determining the amalgamation of YOLOv8 and Efficient Net B7 as an optimal solution. To validate the efficacy of our ALPR system, a diverse set of test data comprising various license plate types is employed. The system demonstrates remarkable identification accuracy, achieving a recognition rate of 96.6 %. These results highlight the practical feasibility and efficacy of our proposed ALPR system in real world applications. © 2023 IEEE.