Automatic vehicle identification system using machine learning and robot operating system (ROS)

Abstract

The Automatic Vehicle Identification (AVI) is the process of detecting and recognizing vehicle on the road automatically. In this work, we propose a design of an automatic vehicle identification system using machine learning approach with Robot Operating System (ROS). The work is based on detecting and recognizing license plate number of any moving vehicle by processing images from a camera. In this work, we trained and used three different classifiers. Two Haar cascade classifier is used for detecting and recognizing license plates of cars. Another classifier trained using K-nearest neighbor technique is used for recognizing the characters of the license plate. The purpose of the work is to create the whole system with separate different parts and each part can be easily integrated with other systems with the similar purpose, which is why the Robot Operating System is used for building the system. Each of the three classifiers is tested in different conditions and each of them has an overall efficiency of over 90 percent.