Year: 2024
Type: article
Abstract: The paper presents a highly accurate automated number plate recognition (ANPR) algorithm designed to correctly recognize Indian license plates with over 99.5% accuracy. The system utilizes OpenCV, Python, and machine learning models in combination to achieve this high level of precision. The algorithm captures and processes images to recognize and identify license plates, including the colours from the plates. Initial plate recognition is performed using Haar cascades, which are subsequently transferred to YOLO v3, enhancing both accuracy and speed. The system incorporates sophisticated image pre-processing techniques—including grayscale adjustment, thresholding, erosion, detail, and contour detection—to ensure images are optimized for character separation and recognition. This integrated approach not only increases recognition rates but also handles images more efficiently, particularly in scenarios where traditional systems may fail. As a result, it paves the way for robust ANPR implementation in dynamic environments.
Author: Gopal D. Upadhye
Institution: Vishwakarma Institute of Technology
Language: English
Subfield: Media Technology
Field: Engineering
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W4404389694