Number Plate Character Recognition System

Abstract

Number Plate Character Recognition System is a relatively well-explored issue with effective solutions. However, such solutions are usually tuned to a particular area due to differences in number plate features around the world. These features are based on algorithms written for number plate recognition and thus it would be difficult to realize a universal solution as the image  analytics techniques used to develop these algorithms can not boast hundred percent accuracy themselves. The subject of this paper is a proposed algorithm that is optimized for working with number plates of Indian vehicles. Written in Python with the OpenCV library, the algorithm makes use of edge detection and feature detection techniques combined with mathematical morphology to locate the plate. Then, the Tesseract OCR engine was used to classify the characters found on the plate.