OPTIMIZING HELMET AND VEHICLE NUMBER PLATE DETECTION WITH ADVANCED YOLO MODELS

Abstract

5 ABSTRACT The goal of this project is to enlarge dependable and structured system for identifying number plates using YOLOv5 and YOLOv8, two popular iterations of algorithm for identifying objects YOLO ,you only look once. For applications like automatic tolling, traffic tracking, and the authorities, this system's ability to automatically recognize and extract license plates from images of two wheelers vehicle is important. A specially set up dataset of labelled vehicle images is used to train the YOLOv5 and YOLOv8 models for project. The models are then assessed and contrasted according to their durability in different environmental conditions, detection accuracy, and diagnosis speed. Also, using optical character recognition ,OCR to fetch text from recognized license plates enhances the effectiveness of the system. The comparison's result show that YOLOv8 is more beneficial for real-time performance than YOLOv5 because of its improved architecture, which provides higher accuracy and faster inference rates. The effectiveness of DL techniques in solving problems related to license plate recognition is mentioned in this work, which offers a workable and scalable solution for automated transportation systems.