The Research And Design Of Tyre-code Recognition System

Tyre is an important component of cars,and its quality has an direct influence on the performance and security features of cars. Currently, tyre manufacturers are making efforts to set up an information management system of tyre-code for indexing in order to improve the quality and standardize production management.Since the traditional method which is artificial copy and keyboard-input for establishing the information management system exists less efficient, human error and other factors,it is imperative to design a set of tyre-code recognition system for the factories.In the design of the system, an automatic image acquisition device is constructed which is consisted of auxiliary illuminating sources,CCD camera,image acquisition card and a computer.On the basis of the theory and methods of the digital image processing and character recognition,a practical recognition scheme is put forward.The recognition process is mainly comprised of image preprocessing, feature extraction and recognition.Image preprocessing and feature extraction methods are some of the gray conversion,gradient sharpening, characters slope adjustment,single character segmentation,normalization,feature selection and extraction;Whereas the neural network has a good anti-noise and fault-tolerant capability,in the design BP networks is selected as a classifier and its algorithm is improved to enhance the correct rate of tyre-code recognition .Software Implementation of the system is based on the Visual C++ 6.0 environment with object-oriented programming. The results show that the entire system is able to complete the functions of image pre-processing and character recognition.