Automatic Recognition of Non-standard Number Plates using YOLOv8

A critical feature of vehicle movement applications is the detection and identification of a vehicle’s License Plate (LP). Despite technological and algorithm developments, differences in LP properties by nation, such as number system, colors, character language, fonts, and size, necessitate more study to enhance detection and recognition. Despite substantial study, many systems function in well-defined contexts, necessitating the use of expensive equipment to capture photos from slow-moving vehicles or generate high-quality shots. This study describes a deep learning-based automated approach for LP detection and recognition, which is further separated into detection and character recognition. Yolov8 (You Only Look Once Version-8) was utilized for automatic pixel separation and gradient computation, making complicated model training easier. Using deep learning, Easy OCR (Optical Character Recognition) was utilized to retrieve text from number plates. Easy OCR (Optical Character Recognition) was utilized to extract text from license plates, with deep learning algorithms employed to effectively read text in difficult settings. For image and video processing, OpenCV (Open-Source Computer Vision Library) was employed. The project titled ‘Automatic Number Plate Using YOLOv8 Model’ leverages the advanced capabilities of the YOLOv8 model to develop a robust system for automated number plate recognition, serving as the core component for real-time object detection and number plate localization within images and video streams.