【作者】 肖伟；

【导师】 马子骥； 李彦鹏；

【作者基本信息】 湖南大学 ， 电子与通信工程（专业学位）， 2020， 硕士

【摘要】 我国的国土面积为960万平方千米,人口14亿,人口的流动与资源的运输变得尤为重要。随着我国科技技术的不断发展,传统的人工效率在各行业都面临着改进和淘汰,现阶段我国正从制造大国向创造大国转型。我国的基建发展相当迅速,已经跻身于世界前列,特别是铁路的发展,高铁动车技术已非常成熟,不仅在国内替代了原有的绿皮火车,还出口到了世界各地。在此背景下,与铁路相关的配套服务也亟待推陈出新。道钉的缺失与损坏,都需要大量的人力物力和时间去排除隐患,包括对道钉的更换和安装,在之前的传统人工技术里面,效率太低,工程量太大,并且受到外界天气,光照和工人专业能力等因素的制约。利用计算机视觉系统可以降低测试人员的劳动强度,提高工作效率与工作安全,并且钢轨维护的全面自动化检测设备是未来发展的重要方向。本课题研究的是铁轨道钉的中心定位系统设计,论文的主要工作概括如下:(1)论述了轨道交通的发展以及道钉在轨道安全中的作用,对比了常用的机器视觉的定位方法,并选用线阵CCD相机作为该定位系统的图像采集元器件。(2)设计了系统总体方案,搭建了硬件平台以及设计了软件操作的人机交互界面,实现了数据图像采集的实时显示与中心定位的图像标注。(3)根据本文所提出的图像处理的中心定位算法,实现了对被测元件的中心定位,采用模块匹配搜寻道钉区域并提取,再使用canny算子进行轮廓筛选后用最小二乘法进行道钉螺纹中心柱的圆形耦合实现道钉中心粗定位,最后再根据区域纹理特征找到螺纹中心柱的中心位置并标注。(4)根据系统设计的方案,搭建了实验小车平台,并在汨罗轨道线进行了实地测试,主要测试了系统的有效性、鲁棒性以及噪声实验。实验论证结果表明本文所设计的道钉中心定位系统设计符合基本预期要求。更多还原

【Abstract】 China has a land area of 9.6 million square kilometers and a population of 1.4billion.The flow of population and the transportation of resources have become particularly important.With the continuous development of science and technology in China,the traditional labor efficiency is facing improvement and elimination in various industries.At this stage,China is transforming from a manufacturing country to a creating country.China’s infrastructure development is quite rapid,has been in the forefront of the world,especially the development of railway,high-speed rail car technology has been very mature,not only in the domestic alternative to the original green train,but also exported to all over the world.In this context,railway related supporting services also need to be innovated.The lack and damage of spikes need a lot of manpower,material resources and time to eliminate the hidden dangers,including the replacement and installation of spikes.In the previous traditional artificial technology,the efficiency is too low,the work volume is too large,and it is restricted by external weather,light and professional ability of workers.The use of computer vision system can reduce the labor intensity of testing personnel,improve work efficiency and work safety,and rail maintenance of comprehensive automatic testing equipment is an important direction of future development.The main work of this paper is summarized as follows:(1)This paper discusses the development of rail transit and the role of spikes in rail safety,compares the commonly used machine vision positioning methods,and selects linear CCD camera as the image acquisition component of the positioning system.(2)The overall scheme of the system is designed,the hardware platform and the human-computer interface of software operation are built,and the real-time display of data image acquisition and the image annotation of center positioning are realized.(3)According to the center location algorithm of image processing proposed in this paper,the center location of the tested component is realized.The module matching is used to search and extract the area of the spike,and then the Canny operator is used to filter the contour.Then the least square method is used for thecircular coupling of the center pillar of the spike thread to realize the coarse location of the center pillar of the spike,and finally the center pillar of the screw thread is found according to the texture characteristics of the area Center position and dimension.(4)According to the scheme of the system design,the experimental car platform is built,and field test is carried out in the Miluo track line,mainly testing the effectiveness,robustness and noise of the system.The experimental results show that the design of the center positioning system meets the basic requirements.更多还原