Research On Tire Identification System

The tire surface identification characters refer to the characters used by tire manufacturers to record tire parameter information on the side surface of the tire using the mold.For manufacturers,these parameters play a vital role in tire classification management and sales.If the misprinting of the logo leads to the wrong use of the tire and an accident occurs,it will cause irreparable losses to both the manufacturer and the customer.In order to avoid such accidents,when mass-producing a certain type of tire,manufacturers will first produce a sample tire,and then verify its parameter information.In the vast majority of tire manufacturing companies,the inspection of sample tire logo characters is done manually.However,since the time of manual detection is too long,the accuracy is difficult to control,and it is difficult to match the large-scale production requirements.Therefore,the automatic verification technology of tire identification characters is an urgent problem that tire manufacturers need to solve.The automatic verification technology of tire logo is divided into three parts: logo area positioning,logo segmentation and logo recognition.The content of the subject research is mainly concentrated in the following aspects: First,according to the actual tire characteristics produced by the tire production enterprise,through calculation,select the appropriate camera,lens,light source,motor and other hardware equipment to build a machine vision imaging system.Through repeated adjustment and testing,the best imaging scheme is determined,and then the image acquisition is completed and the tire gallery is established.After that,the background of the tire side surface and the characteristics of the mark to be recognized are analyzed.Using Halcon image processing software,the positioning of the mark area to be recognized is first achieved through shape template matching.Then,based on dynamic threshold,double peak threshold and other methods to achieve the segmentation of the logo.The feature extracted from the sample identification is sent to the svm-based training model for training,and then when the identification in the test set is identified,the trained classifier is called for identification.At the same time,a template matching method based on correlation is used to create a template from the image within the smallest circumscribed rectangle of the logo in the training set.When recognizing the logo,the template matching method is used for secondary recognition to ensure the recognition accuracy.Record the accuracy of the recognition results,and adjust the training and matching process repeatedly based on the test results until the recognition accuracy reaches the standard.Finally,on the Visual Studio 2012 development platform,a good software architecture was designed using C ++ language,which enabled motion control,image processing and other functions to cooperate with each other.Finally,according to the tire characteristics and production line requirements,an online identification system for tire identification integrating image acquisition,image processing,data display,and motion control was designed to achieve real-time online identification of the side surface identification of conventional radial tire products.Tested by the experimental platform,the current recognition accuracy rate of the current image library is 99.315%,and the recognition time of the single-sided tire tire is about 9.7s,which can basically meet the requirements of enterprises for recognition accuracy and efficiency.