节点文献

基于轮胎质量检测成像与常见故障识别算法研究

Research on the Image of Tyres Quality Test and Common Fault Recognition Algorithms

【作者】 李杰

【导师】 原培新;

【作者基本信息】 东北大学 , 机械设计及理论, 2012, 硕士

【摘要】 子午线轮胎具有弹性大,耐磨性好,滚动阻力小,附着性能好,缓冲性能好,承载能力大,不易刺穿等优点。由于子午线轮胎生产制造工艺复杂,精度高,所以在生产过程中容易产生帘线断裂、帘线稀、钢丝弯、帘线排列不均、帘布重叠、帘线断、胎里气泡、脱层等质量问题,直接影响人车行驶的安全。因此,汽车子午线轮胎质量检测已经成为一个国内外众多轮胎生产厂家重视的问题,以便及时发现问题,提高轮胎的产品质量。本文首先分析轮胎子午线检测设备的工作原理、机械结构以及各组件的性能,为软件的设计打下基础。之后重点探讨了轮胎子午线检测系统所应用到的一些数字图像处理技术,并应用这些技术和方法对采集的轮胎图像进行处理和分析。其中主要研究内容包括图像增强、图像滤波、边缘检测、图像轮廓跟踪、种子填充及模板匹配。然后结合轮胎的实际图像,在大量实验的基础上,总结出了适合轮胎子午线检测的方法。此外,由于传统轮廓跟踪算法得到的图像存在“割角”问题,因此我们利用一种全新的轮廓跟踪算法,用其来提取子午线轮胎图像的轮廓。最后由于传统种子填充算法存在着像素重复入栈以及对一些连通区域不能进行正确填充的问题,对其进行了算法改进,最终得到了比较满意的结果。在课题研究过程中,用MATLAB7.1、VC++语言编程实现了各种算法并对检测图像进行了处理,通过大量实验确定一种比较完善的算法,并将其运用到实际项目中。最后,对图像匹配法进行了深入的研究,提出一种基于灰度图像模板匹配的快速搜索匹配方法。

【Abstract】 Radial tire is flexible, good abrasion resistance, low rolling resistance, good adhesion and buffer performance, great carrying capacity, and difficult to be pierce etc. Due to the manufacturing process of the radial tire is complex and high precision, so in the process of production there are mang qulity problems, such as, cord thread breakage, rare cord thread, bending steel wire, unevenness cord thread, cord fabric overlap, cord thread break, tire cavity bubble, and delamination etc. That impacts directly the safe driving of the vehicle and people. Therefore, the car meridian tyres quality test has become a problem that numerous domestic and international tyre manufacturers pay more attention to, in order to discover in time and improve the quality of products.This paper researched on X-ray detection method of radial tire defects. The main work include:understand and analys the operating principle of the meridian tire detection equipment, the mechanical structure and properties of the components, and lay the root for software design. Discuss the digital image processing technology that the meridian tires detection system applied to, and use the technology and method to dispose and analyse the collected tire image. The study includes mainly gray level transformation, value filter, edge examination, outline track, seed packing and template match. Combine with the tires actual images, on the basis of a large number of experiments, the suitable meridian tires detection method is summarized. The image obtained by traditional outline track algorithm exists a problem of "corner cutting", therefore, we explored a new outline tracking algorithm, which applys to extract the outline of the radial tire image. Moreover, because traditional seed filling algorithm exists pixels pushed into stack repeatly and some of the connected region can not be correctly filled, these problems have been improved and the quite satisfied results have been obtained finally. During the study course of the subject, we had programmed with MATLAB7.1、 VC++, and realized kinds of algorithm, dealed the inspection image with them, then confirmed a more perfect algorithm through a large number of experiments that is applied to the actual project. Finally, after researching to image matching deeply, we proposed the quick search strategy based on gray image template matching algorithm.

  • 【网络出版投稿人】 东北大学
  • 【网络出版年期】2015年 05期
  • 【分类号】TQ336.1;TP391.41
  • 【被引频次】8
  • 【下载频次】148
  • 攻读期成果
节点文献中: