Sistema de control de acceso vehicular mediante reconocimiento de placas para el personal de la Pontificia Universidad Católica del Ecuador Sede Ibarra
No Thumbnail Available
Date
2025
Journal Title
Journal ISSN
Volume Title
Publisher
PUCE - Ibarra
Abstract
Este informe detalla el desarrollo del prototipo de un sistema de control de acceso vehicular basado en el reconocimiento automático de placas mediante visión por computadora e inteligencia artificial. El sistema está diseñado para capturar imágenes de los vehículos mediante una cámara de seguridad, procesarlas con OpenCV para la detección y segmentación de la placa, y emplear EasyOCR y TensorFlow para la extracción y reconocimiento de caracteres. Una vez identificada la matrícula, la información es comparada con una base de datos alojada en MySQL, donde se registran los vehículos autorizados y los eventos de ingreso y salida. Para la visualización y gestión de los registros en tiempo real, se implementó una interfaz web desarrollada en Quasar Framework, la cual permite a los usuarios monitorear el acceso vehicular de manera remota desde cualquier dispositivo con conexión a la red. El sistema también incorpora un ESP32 y un módulo de relé, los cuales permiten la automatización del mecanismo de apertura y cierre de una barrera o portón eléctrico, activándose únicamente cuando se reconoce una matrícula autorizada. Esta integración de visión por computadora, aprendizaje profundo con TensorFlow e IoT proporciona una solución eficiente y de bajo costo para optimizar la seguridad en estacionamientos y zonas restringidas. Las pruebas realizadas con el prototipo muestran que el sistema es funcional y capaz de reconocer matrículas en diferentes condiciones de iluminación y ángulos de captura. Estos resultados permiten evaluar su desempeño y plantear mejoras para futuras versiones con el fin de aumentar su precisión y adaptabilidad en entornos reales.
This report details the development of a prototype for a vehicle access control system based on automatic license plate recognition using computer vision and artificial intelligence. The system captures vehicle images through a security camera, processes them with OpenCV for plate detection and segmentation, and utilizes EasyOCR and TensorFlow for character extraction and recognition. Once the license plate is identified, the information is compared with a database stored in MySQL, where authorized vehicles and access events are recorded. For real-time monitoring and data management, a web interface developed with Quasar Framework was implemented, allowing users to remotely supervise vehicle access from any network connected device. The system also integrates an ESP32 microcontroller and a relay module, enabling the automation of a barrier or electric gate, which is activated only when an authorized license plate is recognized. This combination of computer vision, deep learning with TensorFlow, and IoT provides an efficient and cost-effective solution to enhance security in parking lots and restricted areas. Tests conducted with the prototype demonstrate that the system is functional and capable of recognizing license plates under different lighting conditions and viewing angles. These results allow for performance evaluation and potential improvements in future versions to increase accuracy and adaptability in real-world environments.
This report details the development of a prototype for a vehicle access control system based on automatic license plate recognition using computer vision and artificial intelligence. The system captures vehicle images through a security camera, processes them with OpenCV for plate detection and segmentation, and utilizes EasyOCR and TensorFlow for character extraction and recognition. Once the license plate is identified, the information is compared with a database stored in MySQL, where authorized vehicles and access events are recorded. For real-time monitoring and data management, a web interface developed with Quasar Framework was implemented, allowing users to remotely supervise vehicle access from any network connected device. The system also integrates an ESP32 microcontroller and a relay module, enabling the automation of a barrier or electric gate, which is activated only when an authorized license plate is recognized. This combination of computer vision, deep learning with TensorFlow, and IoT provides an efficient and cost-effective solution to enhance security in parking lots and restricted areas. Tests conducted with the prototype demonstrate that the system is functional and capable of recognizing license plates under different lighting conditions and viewing angles. These results allow for performance evaluation and potential improvements in future versions to increase accuracy and adaptability in real-world environments.
Description
Keywords
Reconocimiento de placas, Opencv, Easyocr, Tensorflow, Control de acceso, Esp32, Quasar framework, Visión por computadora, Iot