An Integrated YOLO–OCR Architecture for Real-Time Automatic Number Plate Recognition (ANPR) in Intelligent Access Management

Abstract

This project presents a real-time Automated Number Plate Recognition (ANPR) system developed to support intelligent vehicle access management in controlled and security-sensitive environments. The proposed solution combines image processing with instant decision-making to enable a secure, contactless, and autonomous gate control mechanism. It responds to the increasing demand for efficient access systems, particularly in high-security locations such as institutional facilities and residential compounds. The framework is designed to be scalable and practical for deployment in real-world scenarios through a tight integration of hardware and software components.

The detection module relies on YOLOv11, which was fine-tuned to accurately localize license plates under varying conditions. For character recognition, EasyOCR was adopted to reliably read both Arabic and English plate formats. A PostgreSQL database serves as the backend layer, enabling structured storage of plate records and efficient log management. Development was carried out using Visual Studio Code for Python scripting, debugging, and project organization. The overall architecture integrates ESP32-CAM units for image acquisition, Python-based real-time processing, and a user-friendly dashboard for monitoring captured plates and reviewing logs.

To validate performance, the system was evaluated across diverse operational settings, including low-light environments and high-glare scenes, to ensure robustness and stability. The results demonstrated fast response times and low latency in both detection and recognition tasks, alongside smooth communication between the user interface and the database layer. Overall, the final implementation delivers a secure, efficient, and cost-effective access control solution that can be adapted for a wide range of public and private surveillance and entry-management applications.