Vehicle Number Plate Detection and Recognition Using Computer Vision and Deep Learning

Abstract

Vehicle Number Plate Detection and Recognition is an important application of computer vision and artificial intelligence that helps in identifying vehicles through their registration numbers. This paper presents an automated system that detects and recognizes vehicle number plates using input captured through a laptop camera. The system captures an image when a number plate is shown in front of the camera and processes it for recognition. The proposed system uses computer vision techniques through OpenCV to capture frames and identify the number plate region. Deep learning‑based object detection (YOLO) improves detection accuracy. After detection, preprocessing techniques such as grayscale conversion, noise removal, and thresholding enhance image quality. The processed image is then passed to the Tesseract OCR engine, which extracts the alphanumeric characters. The system achieves a detection accuracy of 94.2% and character recognition accuracy of 89.6% on a test dataset of 200 vehicle images captured under varying lighting conditions. The recognized vehicle number is displayed as output, making the system suitable for applications such as parking management, toll collection, and basic security systems.