Research On Defect Detection Method Of Tire X-ray Image

Tire as one of the key components of motor vehicles,its quality safety is the most important to ensure the safety of motor vehicles.At present,domestic tire manufacturers mainly buy foreign tire X-ray testing equipment.Although there are a few companies abroad has developed an tires of defect detection X-ray detection equipment,but are not very well in the domestic tire production application,most of domestic tire manufacturers still rely on manual testing tire defects,low detection efficiency,so using the method of machine vision research tire defect detection is of great significance.In this paper,the image characteristics of each cord area are analyzed firstly,and the location method of tire cord area is designed according to the cord crossing characteristics of tire X-ray image.According to the image characteristics of different cordon regions,kinds of algorithms,combining extreme point method,extreme point processing method and minimum point of minimum value method,are proposed to distinguish cordon points and wave points in noncross regions,and compared with the down-sampling method.Then the adjacent maximum point method and extended maximum point method are designed and compared.Finally,the peak width of absolute amplitude array,the minimum absolute amplitude,the gap between concave line segments and the gap between classes of amplitude array are compared and analyzed.Finally,the interclass spacing feature of amplitude array is used to locate tire cord area in X-ray image.Secondly,a tire cord detection method is designed to solve the problem of uneven distribution of image gray value.At present,a large number of tire cord image detection studies only focus on the horizontal cord without cross area,but in this paper,the horizontal and vertical cord can be separated separately.In this paper,adjacent maximum points are used to calculate the amplitude in horizontal cord detection,and adaptive relative amplitude thresholds are used to realize automatic segmentation of horizontal cord in different regions.In the smoothing stage of vertical cord detection,gaussian filter is used to carry out experiments.In order to solve the problems that gaussian filtering is not targeted and smoothing is not complete,this paper proposes a filling method and analyzes the experimental results.Then the automatic vertical cord segmentation is realized by using adaptive absolute amplitude threshold in the shallow crossing region.Finally,based on the analysis of pixel gray curve,this paper designed and implemented the tire X-ray image defect detection algorithm.In this paper,the thin line detection of horizontal cord image is performed using adaptive cord spacing threshold.In this paper,the maximum distance feature of convex segment is used to detect debris after removing the fluctuation point of tire image by extremum processing method.In this paper,threading method is used to detect the vertical cord image.In this paper,threading method is used to detect the open cord image.The experimental results show that the tire defect detection method proposed in this paper can accurately detect the defects and has certain practical value.