AUTOMATED LICENSE PLATE DETECTION AND RECOGNITION IN DYNAMIC ENVIRONMENT USING DEEP LEARNING

Abstract

This project presents an automated vehicle license plate detection system using the YOLOv8 deep learning model, designed and implemented in a cloud-based environment. The system focuses on accurately identifying and localizing license plates from vehicle images and video streams. The methodology involves dataset preprocessing, annotation parsing, and conversion into YOLO-compatible
format, followed by training a custom YOLOv8 model on labeled license plate images. The model is optimized for real-time detection with high precision and recall.
The implementation leverages Python-based libraries such as OpenCV, NumPy, and the Ultralytics YOLO framework within a Google Colab environment. As observed in the dataset preparation