Automatic number plate recognition

Abstract

In today's world, there is a need for automated traffic monitoring and control due to the ever increasing number of cars. Such monitoring and control is needed in car park entrances and exits, at traffic stops and border crossings to name just a few. An effective way to do this is by reading the license plate on the vehicle. This can reliably be done due to the fact that a license plate number is a unique set of characters belonging to a vehicle which can easily be read by a computer. This Final Year Project is an implementation of an Automatic Number Plate Recognition System. The user gives the system a bitmap image and given control parameters, the system locates the characters in the number plate. By using a template matching based OCR algorithm, the characters can be recognized and transformed to data which is easily recognized by a computer, such as ASCII. This system investigates a template matching system for the OCR stage. Template matching makes use of images representing all possible characters which may turn up in a number plate. Then the template image which best matches a character in the number plate is assigned the character name of that template. Improvements to the system can be integrating the system with a database of vehicle number plates containing the owner's name, car type, and other data which may be useful to that particular situation or environment in which the system is installed. Other additions can be to make the system work with video images.