Study On Recognition Technology Of Automobile Tire Marking Points

As an important part of car, the demand of auto tire is increasing greatly with the development of automobile industry. However,classification of automobile tire is always a difficult problem in tire manufacturing process. In order to improve the efficiency and accuracy of the tire classification,a method of recognition of auto tire marking points based on machine vision is proposed in this paper. How to distinguish those tire marking points which printed on side of the auto tire effectively is the foundation of classification of automobile tire The main part of the tire marking points recognition system including:tire marking points images acquisition system,image processing of tire marking points and recognition of tire marking points.The main research contents of this paper are as follows:(1) An image acquisition device for tire marking points based on PLC,the photoelectric switch and the rotary encoder was designed in this paper. Using this tire marking points images acquisition device we can get the tire marking points images accurately. In addition, the LED external light source of LED was used in this image acquisition device for tire marking points in order to eliminate the influence of natural light. At the same time, this also lay a foundation for the subsequent of image processing.(2) In the process of image processing for tire marking points, the image preprocessing methods i.e image denoising, threshold segmentation, edge detection and so on will be conducted on those collected tire marking points images. After that, the hue vectors and Fourier descriptors will be extracted from tire marking points images. More important, the normalization will be applied to those extracted Fourier descriptors in order to ensure those Fourier descriptors with rotation, shift and scaled invariance.(3) In the process of tire marking points recognition, the artificial intelligence algorithms-Support Vector Machine (SVM) was applied to construct the color and shape classifiers. Using those constructed color and shape classifiers to classify the hue vectors and Fourier descriptors into various groups and in this way we can recognition the tire marking points.(4) In order to realize the visualization of tire marking points recognition, the Microsoft Visual C++program language and the software of MATLAB were used to construct the image recognition platform for tire marking points.Using the recognition system of automobile tire marking points we can not only improve the efficiency of tire classification but also we can avoid the phenomenon of error classification manually.