Research And Design On Tier Surface Mark Recognition System

In order to test whether the vulcanizated model match with the production drawing before the batch of car tire production, it need to manufacture a sample tire to identify the mark on each side of it. If it is suitable, the vulcanizated model is qualified and quantity production can be carried out. If not consistent, incorrect vulcanizated model must be changed. Once any tire marks mistakes, not only tires enterprise will face huge compensation, but also tires the users may also meet with accident. Presently, tire enterprise normally identify tire marks by using man power, it always takes long. The computer assistant identification will greatly shorten the labor time and labor intensity. This paper is mainly about the research on tire surface mark recognition on sample tire. It could avoid artificial mistake, increase recognition speed and save manpower and material resources.The car tire surface mark is distributed according to tire circle, therefor the centre of the circle is determined and the original arc image is stretched to rectangular image by using Gray interpolation method. Then Canny operator is adopted to detect the boder of the character image. Then the image is projected and tire mark area is located according to the wave crests and troughs of projection image. Thereafter, multipile lines is segmented by accurate location. Furthermore, the corresponding original image in the segmentation area is pretreated and binarized for the second time. Finally, after character segmentation based on sciagraphy, character feature matrix is extracted by using local block feature method. Recognition result is got by'and operation'between the feature matrix and the template. The experiment result shows that the above method can be used to recognize the tire surface mark.Gray interpolation method is used in this paper to stretch arc image, although the effect is good. But gray interpolation calculation time is long, the method need to be improved in the future.