Motion estimation from sequential image using correlation

In this paper, we take aim at a flexible and active vehicle following system for autonomous vehicles. To realize the system, we use a license plate instead of the devices that receive and send signals. In order to gather information from a license plate, we propose three methods; one that continuously tracks a license plate, one that extracts the characters under conditions of changeable and uneven illumination, and one that calibrates the difference between the reference template image and the image selected by correlation. From these characters in a license plate, we can determine the relative distance and angle. The effectiveness of this system is shown by the experimental results.