Research Of Character Recognition Algorithm Based On Template Matching And Structural Characteristic

As a vital part of Intelligent Transportation System, License Plate Recognition is widely used in the management of cars in parking lot, the video monitoring on superhigh-way and electronic police, etc. Therefore, License Plate Recognition is of great research and practical value. The system includes license plate location, character segmentation and character recognition. This paper focuses on character recognition part of license plate recognition system.This paper first describes three character recognition methods:template matching method, structural characteristic method, neural network method, and proposes some improved algorithms effectively. The paper proposes a method of combining template matching and structural characteristic after a careful study of the license plate letters and numbers. First, using template matching method classifies characters into certain types so that high similarity of characters are assigned in the same group. Then the characters are classified according to a recognition of whether the characters have the same features such as horizontal lines, vertical lines and closed regions. In this paper, the final recognition rate and recognition speed are improved with an improvement of the algorithm on the basis of the existing character recognition algorithms.