Automatic Number Plate Detection Using Deep Learning Techniques

Abstract

Automatic Number Plate Recognition (ANPR) is a technique that reads a vehicle's license plate using photographs given to the Deep Learning model as its input. Existing ways to find and identify license plates on cars are severely limited by the low resolution and significant loss of edge data in plate photos. The process of automatic identification of license plates necessitates a high level of accuracy when cars are moving rapidly, and number plate abstraction is a difficult task due to number arrangement and environmental factors. The primary objective of this work is to create an Automatic Number Plate Detector that aims to recognize the text present in the number plates of cars, vans, bikes, trucks, and other vehicles that are either stationary or in motion with good accuracy. The recognition model was created by training it on a large dataset. The deep learning models Faster-RCNN and Yolo were employed to fetch the Region of Interest, followed by OCR for character extraction.