Etudiant : MOUSSDDIK HAJAR

Filière : Master Informatique Décisionnelle et Vision Intelligente (MIDVI)

Encadrant : Pr. YAHYAOUYI ALI

Annèe : 2023

Résumé : Traffic violations in Morocco have witnessed a considerable increase in recent years, posing serious concerns for road safety and public well-being. To overcame this issue, this project proposes the implementation of two systems: License Plate recognition and speed estimation. While there is no universal Automatic License Plate Recognition (ALPR) system that covers all countries due to various factors such as: country-specific layouts, fonts, languages, colors, and environmental conditions. We made a real-time ALPR system specifically designed for detecting and recognizing Moroccan license plates using the You Only Look Once 8Th Version (YOLOv8) model. The system includes two deep convolutional neural networks (CNN) which were trained using a diverse set of images. Through evaluations, our proposed system demonstrates superior robustness in detecting and recognizing Moroccan license plates. Furthermore, our system effectively utilized YOLOv8 for speed estimation and achieved good results. Keywords : Deep Learning, CNN, YOLO, YOLOv8, Detection, Recognition, Automatic license plate recognition, Moroccan License Plates, Speed Estimation, Traffic Violation.