图像自动识别算法在汽车轮胎无损检测中的应用研究

Application of Automatic Image Recognition Algorithm in Nondestructive Testing of Automobile Tires

针对汽车轮胎胎侧区生产过程中的各类缺陷,设计了无损检测图像的处理与识别方法。基于数字X射线摄影成像原理,分析了包括帘线开裂、帘线弯曲等在内的多种典型缺陷图像特征。提出一种结合均值修正的大津阈值图像分割算法和基于结构元素的迭代细化方法。通过构建不同类别的识别判据,实现了对包括异物夹杂、帘线断裂在内的4类微观和宏观缺陷的分类检测。算法整体缺陷识别概率达94.3%,漏检率低于8.7%。结果表明,所设计的图像处理与识别体系,可以有效实施轮胎胎侧区缺陷的自动化无损检测。该研究为保障汽车轮胎使用安全性提供了一定借鉴。

To address various defects in the production process of the tire sidewall area,an image processing and recognition method for non-destructive testing was designed.Based on the DR imaging principle,the image features of typical defects,including cord cracking and cord bending,were analyzed.A segmentation algorithm combining mean correction and Otsu thresholding,as well as an iterative refinement method based on structural elements,was proposed.By constructing different recognition criteria for different categories,the classification and detection of four types of microscopic and macroscopic defects,including foreign material inclusion and cord fracture,were realized.The algorithm achieved an overall defect recognition probability of 94.3%,with a missed detection rate below 8.7%.The results demonstrated that the designed image processing and recognition system can effectively implement automated non-destructive detection of tire sidewall defects.This study provided a valuable reference for ensuring the safety of automobile tires.