Enhancing Automated Vehicle License Plate Recognition with YOLOv8 and EasyOCR

This research focuses on the development of an automatic system for vehicle license plate recognition using YOLOv8, EasyOCR, and CNN methods for object classification. The main issue raised is the need for an accurate and efficient system for recognizing vehicle license plates in real-time in dynamic environments, especially in urban areas with high traffic levels. The method used in this study involves resizing the input image to 416x416 pixels to standardize the data, analyzing the YOLO architecture that divides the image into a 7x7 grid, and using the Convolutional Neural Network (CNN) algorithm for feature extraction and object classification. Object detection uses the YOLOv8 method which is tasked with recognizing license plates using a previously trained YOLO (pretrained model) model then implemented and tested using video with 4k quality to ensure its effectiveness in detecting vehicle license plate objects, followed by the Optical Character Recognition (OCR) process with the EasyOCR method to read text on license plates and tested to ensure its effectiveness in reading characters on license plates vehicle number. The purpose of this research is to develop a system that can improve accuracy and efficiency in vehicle license plate recognition. The results show that the accuracy, precision, recall and F1-Score for object detection reach 100% and the average percentage of detected text conformity is 74.66%, which shows that this system is reliable in real applications and contributes to the development of automatic license plate recognition technology.