Automatic Number Plate Recognition Using OpenCV And Machine Learning

Abstract - A well-known application for detecting cars in complex security systems, parking lots, and traffic controls is the Automatic Number Plate Recognition System (ANPRS). ANPR is a relatively new field in image processing that employs a wide range of algorithms, methodologies, and methods. This study describes the creation of an ANPR system using Python, a free and dynamic programming language that works in conjunction with Open Computer Vision Libraries. The main goal is to combine and employ many morphological procedures in order to properly recognize and translate the license plate of a certain car. This is based on picture improvement, grayscale transformation, Bilateral Filtering edge detection, and getting the vehicle's license plate from a photograph, among other procedures. After the previous processes have been completed, the segmentation approach is utilized to detect the text on the number plate utilizing template matching and OCR. From an image of a vehicle, our technology can precisely and quickly recognize the license number. This system achieved 91.5 % accuracy, 91.1 % precision, and 91.8 % recall