Automatic license plate recognition using mobile device

Abstract

The rapid development of intelligent transport systems (ITS) brings us a safer and more convenient life. In this field, automatic license plate recognition (ALPR) plays an important role in many applications which have been deployed in reality such as stolen car detection, parking system management, and automatic transport charging system. The traditional ALPR methods usually need a high resolution camera to capture good quality images and a powerful computer to process the complex algorithms. However, with many great scientific breakthroughs in mobile device technology, we can easily buy smartphone/tablet which equipped strong central processing unit (CPU) and excellent camera to meet the requirements of ALPR. In this paper, we develop an Android program which processes the image captured by built-in camera of mobile device to have license plate number and save it into database for further applications. Open-source OpenCV libraries are imported into project for some image processing steps with purpose of programming time saving. We apply Tesseract engine and neural network which are two optical character recognition (OCR) methods to convert license plate image to machine-encoded text.