License Plate Character Recognition Algorithm Based on Shape Feature Vector

Abstract

A simple character recognition algorithm based on Shape Feature Vector (SFV) is proposed. As we know, the character set of vehicle license plate is shaped according to the standard GA36-2014 in China, that is, the same character in different vehicle license plates has the same shape. So Shape Feature can be used to identify characters. Some concepts, such as Shape Feature Vector (SFV), are proposed after examining the character set used by vehicle license plate and the feasibility of using SFV to identify vehicle license plate characters is proven theoretically. Then, the character recognition algorithm is evaluated by simulation. The result of the simulation shows that SFV is qualified to recognize vehicle License Plate characters and the possibility of recognizing the right character is up to 95%. Besides, the recognition algorithm doesn’t need training and a lot of labelled training data, can be programmed in a simple way and is a light-weight vehicle License character recognition algorithm.