Tire Body Defect Detection: From the Perspective of Industrial Applications

Radial tires have a large market share in the tire market due to their better wear resistance and puncture resistance. However, the complexity of the production process makes impurities and irregular cord spacing defects often appear in the body cord area. In this paper, a method relying on X-ray images to detect impurities and irregular cord spacing defects is proposed. The detection problems of these two types of defects are transformed into calculating cord pixel spacing and background connected domains. Firstly, the tire crown, tire ring, and tire body regions are segmented by a novel semantic segmentation network (SSN). Then the background and the cord are separated by an adaptive binarization method. Finally, the irregular cord spacing defects are detected through the refinement and column statistics. The impurities of tire body are located by the marks of the connected domains. The experimental results of the X-ray images show that this method can meet the positioning requirements of irregular shaped impurities. A idea of setting the threshold column effectively improves the detection speed of irregular cord spacing. In addition, the detection accuracy rates for both types of defects are higher than 90\(\%\), which is helpful for further research on various types of tire defects and the design of an automatic tire defect identification system.