Publication

Abstract :

Automated License Plate Reader (LPR) technology finds significant use in traffic monitoring, vehicle identification, and security systems. In this paper, we propose a practical LPR system with the feature extraction capability of Convolutional Neural Networks (CNNs) and the You Only Look Once (YOLO) object detection system for real-time and high-speed recognition. Our approach not only enhances the detection rate but also effectively reduces processing time, which makes it deployable in real-time applications. Our model consistently delivers outstanding performance across various settings, achieving an impressive accuracy of 97.69%. This highlights its robustness and adaptability. In this paper, we detail the system architecture, methodology, and experimental results, focusing on how effective our approach remains even under challenging conditions. Our model innovatively combines YOLOv8, CNN and EasyOCR which ensures an ultra fast and effective system for license plate detection, feature extraction and character recognition in a layered system. This hybrid model also reduces the dependency on single algorithms, improving the reliability in real life scenarios.