Overview Motivation State of the Art Implementation Challenges Results References

Automatic License Plate Recognition


Ben Farley, Carter Steffen, Chan Hyeok Yun

        Our project for CS 534 centers on implementing an Automatic License Plate Recognition (ALPR) system. Current ALPR technology is often narrow in scope, expensive, and is not always effective. Our goal was to implement an ALPR system that could be used in a variety of different locations at low cost. The solution we developed, using Matlab, is mostly successful in this endeavor. It accepts a wide range of images with a quality near or above 720p. It is successful with different types of lighting, license plate locations within the image, and license plate types. At its core it uses edge detection and then filtering to find the license plate within the image. From this, it uses text recognition to determine the letters and numbers on the plate. From our selected filtering, it inherently fails given an angled photo of the plate, a plate with an atypical border, or a plate of a dark color. Future improvements involve adjusting the filtering methods to incorporate these scenarios.