【作者】 卓红俞；

【导师】 王黎；

【作者基本信息】 西南交通大学 ， 光学工程， 2005， 硕士

【摘要】 随着经济的发展,机车的高速、重载、高密度运行,对于轨道质量的要求越来越高,轨道的状态是影响铁路行车安全的重要因素,因此确保线路质量成了工务段面临的一项艰巨任务,而与这直接相关的是钢轨状态。 本文首先介绍了该项目的开发背景、意义;介绍了钢轨的分类、几何尺寸,钢轨磨耗的定义、成因、机理、伤损标准及轨距标准等;介绍了国内外钢轨检测的发展,包括检测方法、检测原理和现有产品等;分析和比较了国内外现有检测原理、方法的优缺点和检测精度等;在此基础之上,设计了本系统的检测原理,即基于CCD的非接触光截法。根据该检测原理,进行了系统总体方案设计,包括检测功能、检测原理、技术指标、系统软件、系统硬件、系统机械。本系统重点在于图像采集、处理、分析部分。其中,采集部分包括最佳光截曲线设计、CCD参数设计、线光源参数设计、光学系统和双通道图像采集系统设计,并进行了多次现场图像采集实验。图像处理部分,对数字图像处理方法进行了理论的研究、分析和探讨,利用C++编程,处理采集的钢轨曲线图像,提取钢轨目标曲线,进行图像畸变校正、坐标变换等,对各种算法处理的结果进行了比较,选出最适合本系统的图像处理算法。在图像分析中,对图像基准点判别,磨耗和轨距计算,误差分析等问题进行了研究和算法编程。 本文对现场检测的软件进行了设计,该软件界面简单,易操作,包括检测数据显示、报警、人工判别等功能。软件自动保存检测数据,操作者可将检测数据由USB接口转移到地面分析软件进行地面分析、处理、报表生成和打印等。 本系统以非接触式CCD图像测量技术为基础,集光学技术、光电技术、数字图像处理技术和软件工程等多种先进技术于一体,可动态检测钢轨技术参数:垂直磨耗、水平磨耗、总磨耗和轨距。系统体积小、重量轻、精度高、实时处理、易操作,对工务段线路维修、维护和行车安全等起到积极作用。更多还原

【Abstract】 With the development of economy, and the operation of high speeds over loading, high density, the request of the railway quality is becoming more and more high, the railway condition is important factor to the security of operation.so it is a difficult job to ensure the railway quality well to the railway department, but the condition of railway is directly contacted with rail quality.First of all, the background and meaning of research and development is introduced in the thesis; the rail’s classify and geometry dimension, and the rail wear’s definition, formation cause, mechanism, wear standard and gauge standard are introduced; the domestic and overseas development of rail detection is introduced, including detecting method, detecting principle and detecting products; the advantage and disadvantage of domestic and overseas rail detecting principle, detecting method and detecting precision are analysed and compared. Then, CCD Non-contact light section method is confirmed to be the detecting principle to rail profile in the thesis. The whole project is designed on the base of the detecting principle, including: detecting function, detecting principle, technology index, system software, system hardware and system machine. The main work of this thesis is image collecting, image processing and image analysing. The image collection include the design of best optical curve, the design of CCD parameter, the design of line laser parameter, the design of optical system and double channel image collecting system, and several locale test. Image processing theory is researched, analysed and discussed in image processing part, and the rail curve image is processed by C++ coding, pick-uping rail object curve, aberration emendation, and coordinate transformation, Every arithmetic is compared and the best image processing arithmetic is choosed to the thesis. Image datum mark distinguishing, wear counting and gauge, error analysing are researched and the arithmetic are coded in image analysing part.The locale detecting software is designed, the interface of software is simple, operating easily, and including display of detecting data, warning, manual distinguishing and so on. Software save detecting data automatically, the detecting data could be moved to ground analysing software by the interface USB that could analyse, process, make report and print.The system is based on the CCD Non-contact image detecting technology, including the optical technology, photoelectricity technology, digital image processing technology and software engineering and so on, the system could detect the technology parameter of rail: vertical wear, horizontal wear, whole wear and gauge. Small bulk, light weight, high precision, real time processing, easily operating, so it have positive effect on the railway maintain,security of railway and so on.更多还原