基于机器视觉的钢轨表面面型缺陷分类实验设计  

Experimental Design of Rail Surface Defect Classification Based on Machine Vision

作　　者：李珂嘉 张璐薇[1] 马跃洋 尹昱东 杨帆[3] 张璐[1] LI Kejia;ZHANG Luwei;MA Yueyang;YIN Yudong;YANG Fan;ZHANG Lu(School of Instrument Science and Technology,Xi’an Jiaotong University,Xi’an 710049,China;School of Mechanical Engineering,Xi’an Jiaotong University,Xi’an 710049,China;School of Chemistry Engineering,Xi’an Jiaotong University,Xi’an 710049,China;School of Optoelectronics,Xi’an Technological University,Xi’an 710021,China)

机构地区：[1]西安交通大学仪器科学与技术学院,西安710049 [2]西安交通大学机械工程学院,西安710049 [3]西安交通大学化学学院,西安710049 [4]西安工业大学光电工程学院,西安710021

出　　处：《实验室研究与探索》2024年第3期122-127,134,共7页Research and Exploration In Laboratory

基　　金：陕西省自然科学基础研究计划项目(2023-JC-QN-0649)。

摘　　要：随着城市轨道交通的飞速发展,实现钢轨表面缺陷实时检测对铁路行业稳步发展意义重大。如何实时检测钢轨表面缺陷是保障铁路运行安全亟须解决的一个关键问题。鉴于此,设计了一套基于机器视觉的钢轨表面缺陷检测实验仿真方法。搭建图像采集、图像预处理和缺陷分类等模块;提出自拟合亮度调整算法完成像素值统计,得到清晰的缺陷特征图像;用750组数据训练网络权值,实现缺陷分类预测;经过数据分析和误差评估,识别准确率在90%以上,相关系数高达0.96,单幅图像平均耗时1.267 s,测试表明,所提方法能准确、高效地实现钢轨表面缺陷信息的缺陷分类与识别。With the rapid development of urban rail transit,the realization of real-time detection of rail surface defects is of great significance to the steady development of the railway industry.Detection of rail surface defects in real time is a key issue,and needs to be solved to ensure the safety of railway operations.In view of this,firstly,a simulation experimental method for rail surface defect detection based on machine vision is designed,and an image acquisition module,image preprocessing module and defect classification module are built.Secondly,a self-fitting brightness adjustment algorithm is proposed to complete the pixel value statistics,and finally obtain the image with clear defect features.The testing result shows that the method can quickly identify and classify the surface defect information of the classified rail.

关 键 词：钢轨表面缺陷检测 机器视觉 图像处理 缺陷分类 

分 类 号：G642[文化科学—高等教育学]