A Non-learning-oriented Recognition Algorithm of Vehicle License Plate Characters

Abstract

Based on the fact that characters of vehicle license plate are standardized, a recognition algorithm of Vehicle License Plate Characters based on non-learning-oriented method is proposed. This method uses some concepts such as Shape Feature Vector (SFV, for short) to describe the shape of the to be recognized characters. Each character will be evaluated with a vector and the recognition is realized by calculating the similarity of vectors. This method achieves character recognition by vector arithmetic avoiding the learning/training process often used in some algorithms, such as artificial neural network, and SVM and so on. The recognition is simplified tremendously. The feasibility of using SFV to recognize vehicle license plate characters is proven theoretically, and then, the character recognition algorithm of vehicle license plate is evaluated by simulation. The result of the simulation shows that SFV can be used as license plate character recognition. And the license plate recognition algorithm based on SFV has an accuracy rate of 97.31%. This method is a helpful attempt of character recognition based on non-learning-oriented method.