基于机器视觉的跨座式单轨轨道梁晃动检测系统
|
摘要
跨座式单轨交通-PC轨道梁交通制式技术,是通过单根轨道来稳定、支承和导向,车体骑跨在轨道梁上运行的交通制式,其技术特点是:以梁代轨,这种桥梁不但具有传统桥梁的承载功能,更重要的是还具有轻轨列车行驶的轨道功能。轨道梁安装在墩台上,单轨车辆通过时承载着巨大的运动载荷,经过长期运营后,轨道梁可能会发生晃动,影响轻轨列车的安全运行。
目前,轨道梁的晃动检测方法为人工撬动轨道梁,依靠位移传感器来测量梁体的晃动幅度值。使用人工方法检测,工人的劳动强度大,安全性差,并且检测精度较低。为解决上述问题,本文提出了一种基于机器视觉的检测方法。基于机器视觉的检测具有速度快、精度高、在线实时、非接触、现场抗干扰能力强等优点。在所提方法的基础上,本文设计了一套车载轨道梁晃动幅度检测系统。为了实现该系统,本文主要做了如下几方面的研究工作:
①通过大量试验和对现场环境的考察,搭建了系统的硬件平台。其中包括系统的框架设计,摄像机、镜头、光源等硬件设备的选择与装配等。
②完成图像采集软件,对相机进行标定和参数调节,达到测量精度。
③对多种边缘特征提取方法进行了实验比较,选择了合适的边缘检测算法;对传统的Otsu阈值分割算法做了改进,加快了图像处理速度;选择经典的模板匹配算法来计算晃动幅值,并提出了自动模板匹配方法,提高检测的自动化程度。
④现场安装、调试系统,使整个系统成功地在实际中运用,对重庆轻轨2号线全线轨道梁进行了检测,检测出晃动幅度超过1mm的梁33榀,经过人工检测方法的验证,检测结果准确。
本文设计的基于机器视觉的轨道梁晃动检测系统,能够准确地检测出轨道梁的晃动幅度值,另外,它的检测效率是原来的人工检测方法的几十倍。本系统的成功研制实现了轨道梁晃动幅度的自动测量,为跨座式轨道交通的安全运营提供了保障,现已成功应用于重庆市跨座式单轨交通轨道梁晃动检测项目中。
Straddle-type Monorail PC Track Beam Transportation System is such a kind of transportation system in which bodywork is running by riding on the track beam, through monorail track beam supporting, stabilizing and be oriented to the train. The main features of its technology: this track beam not only has such a traditional function of the load, but more importantly also has a roadway function when the train is running.
There is huge moving load on track beam set on pier while the monorail vehicle is passing on it, after long time operation, substantial transverse swing of the track beam may be produced and the resultant serious threat to safety.
Currently, a solution to the measurement of the swing amplitude of the track beam relying on artificially prizing the track beam and recording the swing amplitude with the displacement sensor is dominated, which however has the disadvantages of huge intensity of labor, lack of security and low measuring precision. A new solution based on Machine Vision method is made to successfully achieve the goal of rapid measurement, non-contact and high measurement precision. Based on the solution, detection system of the swing amplitude of the track beam is designed inside the paper.
In this paper, for accomplishing the detection system the major work is as follows:
Firstly, a hardware platform is built through many tests and investigation of spot environment, including frame design of the system, choice of the hardware such as light source, camera, lens, computer and so on
Secondly, the software of image acquisition has been completed. According to the accuracy of measurement requirements, it is needed to calibrate the CMOS camera and set its parameters.
Thirdly, a comparison is made among the results of different edge detection methods and the best one is selected. In order for accelerating the speed of threshold segmentation, the methods of Otsu has been improved. In addition, classical template matching is selected to calculate the amplitude of track beam body. An automatic matching method is also made to improve the automation level of detection.
Finally, after installing and debugging in the scene, now the whole system can be applied successfully. The detection toward all the track beam of Light Rail in ChongQing has been made and there are thirty-three places where the swing amplitude of tract beam in excess of one millimeter. the accuracy is a hundred percent, compared with the manual results.
In this paper, the detection method based on MV system can detect swing amplitude of track beam body accurately. In addition, it provides several time higher efficiency in comparison to original manual detection method. This system has successfully achieved the goal of self-measurement of swing amplitude of track beam and ensured the safety of the track working and it is applied in swing detection items of Light Rail in ChongQing.
目前,轨道梁的晃动检测方法为人工撬动轨道梁,依靠位移传感器来测量梁体的晃动幅度值。使用人工方法检测,工人的劳动强度大,安全性差,并且检测精度较低。为解决上述问题,本文提出了一种基于机器视觉的检测方法。基于机器视觉的检测具有速度快、精度高、在线实时、非接触、现场抗干扰能力强等优点。在所提方法的基础上,本文设计了一套车载轨道梁晃动幅度检测系统。为了实现该系统,本文主要做了如下几方面的研究工作:
①通过大量试验和对现场环境的考察,搭建了系统的硬件平台。其中包括系统的框架设计,摄像机、镜头、光源等硬件设备的选择与装配等。
②完成图像采集软件,对相机进行标定和参数调节,达到测量精度。
③对多种边缘特征提取方法进行了实验比较,选择了合适的边缘检测算法;对传统的Otsu阈值分割算法做了改进,加快了图像处理速度;选择经典的模板匹配算法来计算晃动幅值,并提出了自动模板匹配方法,提高检测的自动化程度。
④现场安装、调试系统,使整个系统成功地在实际中运用,对重庆轻轨2号线全线轨道梁进行了检测,检测出晃动幅度超过1mm的梁33榀,经过人工检测方法的验证,检测结果准确。
本文设计的基于机器视觉的轨道梁晃动检测系统,能够准确地检测出轨道梁的晃动幅度值,另外,它的检测效率是原来的人工检测方法的几十倍。本系统的成功研制实现了轨道梁晃动幅度的自动测量,为跨座式轨道交通的安全运营提供了保障,现已成功应用于重庆市跨座式单轨交通轨道梁晃动检测项目中。
Straddle-type Monorail PC Track Beam Transportation System is such a kind of transportation system in which bodywork is running by riding on the track beam, through monorail track beam supporting, stabilizing and be oriented to the train. The main features of its technology: this track beam not only has such a traditional function of the load, but more importantly also has a roadway function when the train is running.
There is huge moving load on track beam set on pier while the monorail vehicle is passing on it, after long time operation, substantial transverse swing of the track beam may be produced and the resultant serious threat to safety.
Currently, a solution to the measurement of the swing amplitude of the track beam relying on artificially prizing the track beam and recording the swing amplitude with the displacement sensor is dominated, which however has the disadvantages of huge intensity of labor, lack of security and low measuring precision. A new solution based on Machine Vision method is made to successfully achieve the goal of rapid measurement, non-contact and high measurement precision. Based on the solution, detection system of the swing amplitude of the track beam is designed inside the paper.
In this paper, for accomplishing the detection system the major work is as follows:
Firstly, a hardware platform is built through many tests and investigation of spot environment, including frame design of the system, choice of the hardware such as light source, camera, lens, computer and so on
Secondly, the software of image acquisition has been completed. According to the accuracy of measurement requirements, it is needed to calibrate the CMOS camera and set its parameters.
Thirdly, a comparison is made among the results of different edge detection methods and the best one is selected. In order for accelerating the speed of threshold segmentation, the methods of Otsu has been improved. In addition, classical template matching is selected to calculate the amplitude of track beam body. An automatic matching method is also made to improve the automation level of detection.
Finally, after installing and debugging in the scene, now the whole system can be applied successfully. The detection toward all the track beam of Light Rail in ChongQing has been made and there are thirty-three places where the swing amplitude of tract beam in excess of one millimeter. the accuracy is a hundred percent, compared with the manual results.
In this paper, the detection method based on MV system can detect swing amplitude of track beam body accurately. In addition, it provides several time higher efficiency in comparison to original manual detection method. This system has successfully achieved the goal of self-measurement of swing amplitude of track beam and ensured the safety of the track working and it is applied in swing detection items of Light Rail in ChongQing.
引文
[1]罗秀云,蒲云.我国城市交通发展的研究与思考[J].城市轨道交通,2004,26(7):27-29.
[2]孟佳.钢轨表面缺陷识别系统的设计与研究[D].西南交通大学硕士论文,2005.
[3]梁育贵.半径75 m曲线PC轨道梁制作[J].铁道标准设计,2004(9):89-91.
[4]刘浪.基于重庆轻轨2号线的单轨交通结构形式适应性分析[J].城市道桥与防洪,2009(5):165-167.
[5]张殿业,金键,杨京帅.城市轨道交通安全研究体系[J].都市快轨交通,2004(4):1-4.
[6] Kannan G,Martinez J C.Simulating wireless radio communications in earthwork construction [J]. Proceedings of Congress on Computing in Civil Engineering,1998,10(13): 53-60.
[7]常太华.检测技术与应用[M].北京:中国电力出版社,2003.
[8]沈中城.检测技术与仪器[M].北京:高等教育出版社,2003.
[9]潘长胜.桥梁检测技术及其发展趋势简述[J].黑龙江交通科技,2009(4):106-108.
[10]刘正光.人造卫星定位系统在桥梁结构健康监测系统中的应用[J].现代交通技术,2008,5(6):22-27.
[11]胡志伟,朱明达.光纤传感技术在桥梁检测中的应用[J].交通标准化,2008(8):125-127.
[12]高文,陈熙霖.计算机视觉[M].北京:清华大学出版社, 1999.
[13]许巧游.基于机器视觉系统的零件识别与检测的研究[D].南京航空航天大学硕士论文,2006.
[14]吴珂.基于机器视觉的弹片质量在线检测系统[D].重庆大学硕士学位论文,2010.
[15]蒋昀赟.基于机器视觉的PC轨道梁在役锚固螺杆检测的自动定位系统研究[D].重庆大学硕士学位论文,2007.
[16]黄正福.面向尺寸检测的机器视觉系统研究[D].浙江工业大学硕士论文,2004.
[17]王志凌.基于双CCD的大锻件尺寸测量的实验研究[D].燕山大学硕士论文.2005.
[18]施雄林,李朝晖等.曲轴锻件计算机视觉检测系统[J].重庆工学院学报. 2006, 20(8):28-31.
[19] Ghosal S, Mehorotrar. Orthogonal Moment Operators for Subpixel Edge Detection[J].Pattern Recognition,1993,9(3):203-210.
[20]廖强,周忆等.机器视觉在精密测量中的应用[J].重庆大学学报(自然科学版), 2002, 25(6):1-3.
[21]贾云得.机器视觉[M].北京:科学出版社, 2000.
[22]凌云光视数字图像处理公司.CCD&CMOS图像和机器视觉产品手册[Z].
[23] J.C.Nunes, Y.Bouaoune, E.Delechelle. Image analysis by bidimensional empirical mode decomposition[J]. Image and Vision Computing, 2003, 21(12):1019-1026.
[24] M. M Park, R. C. Singer, A. J. Plueddemann, R Weller.High-speed, real-time data acquisition for vector measuring current dynammeters[J].Proc. IEEE Fourth Working Conf. Current Meas , 1990(4):146-153.
[25]北京嘉恒中自图像技术有限公司.机器视觉系统[Z].
[26] Kim Jung-Yong, Cho Yoon-Ho. Property of Images of Asphalt Pavement and Enhancement Algorithm to Reduce Noise[J]. Proceedings of SPIE-The International Society for Optical Engineering, 2003,20(2):728-739.
[27]王永刚,卫保国.基于数码相机的图像采集系统[M].北京:北京工业大学信号与信息处理研究室,2000.
[28]刘丁,毛德柱,王云飞.USB在数据采集系统中的应用[J].电子技术与应用,2000(4):43-48.
[29]潘锋.一种基于直线校正的通用摄像机标定技术研究[C].第二届全国智能视视觉监控学术会议.北京:科学出版社, 2003.
[30] Guiducci, Antonio. camera calibration for road applications[J]. computer vision and image under stanging. 2000, 79(2):250-266.
[31]夏良正,李久贤.数字图像处理[M].南京:东南大学出版社,2006.
[32]尹兰.基于数字图像处理技术的混凝土表面裂缝特征测量合分析[D].东南大学硕士论文,2006.
[33]荣士杰,吴春镕,孙正兴.计算机图像学[M].北京:电子工业部出版社,1997.
[34]赵京东,赵景秀.改进的基于梯度幅值的自适应边缘检测算法[J].光电子技术, 2009,29(2):98-102.
[35]朱晓林,高诚辉等.基于Canny算子的机械零件图像分块自适应边缘检测[J].江南大学学报(自然科学版),2010,9(3):304-307.
[36]曲昆鹏,郑丽颖.基于目标、背景比例的灰度图像自动阈值选取法[J].应用科技,2010,37(2):52-54.
[37]刘鑫.基于投影的运动目标跟踪研究[J].电子测试,2009,3:1-4.
[38] Otsu N. A threshold selection method from gray-level histograms[J].IEEE Trans on Systems,Man,and Cybernetics.1979,9(1):62-66.
[39]乔万波,曹银杰.一种改进的灰度图像二值化方法[J].电子科技,2008,21(11):63-64,71.
[40]马建红,沈西挺.Visual C++程序设计与软件技术基础[M].北京:中国水利水电出版社,2002.
[41]王勇智.数字图像的二值化技术探究[J].湖南理工学院学报,2005,18(1):31-33.
[42] Rathank,, Karuk, Chen S Y,et al. A real-time matching system for large ingerprint databases[J]. IEEE Transactions on pattern analysis and machine intelligence, 1996, 18(8): 799-813.
[43]王广伟,李铸国,吴毅雄.基于特征点提取和旋转不变技术的锁片基体视觉检测[J].中国机械工程,2007,18(6):643-645.
[44] You J,Bhattacharya P A.Wavelet-based Coarse to fine Image Matching Scheme in a Parallel Virtual Machine Environment[J].IEEE Trans on Image Processing,2000,9(9):1547-1559.
[45]田涌涛,李霞等.基于二维三阶多项式拟合的阈值曲面分割法[J].计算机工程, 2003, 29(4):127-129.
[46] Steger C,Ulrich M.Machine Vision Algorithms and Applications[M].北京:清华大学出版社, 2008.
[47]刘波,叶俊勇,让晓勇.基于机器视觉的轨道梁晃动量检测方法[J].计算机系统应用,2009(4):202-204.
[2]孟佳.钢轨表面缺陷识别系统的设计与研究[D].西南交通大学硕士论文,2005.
[3]梁育贵.半径75 m曲线PC轨道梁制作[J].铁道标准设计,2004(9):89-91.
[4]刘浪.基于重庆轻轨2号线的单轨交通结构形式适应性分析[J].城市道桥与防洪,2009(5):165-167.
[5]张殿业,金键,杨京帅.城市轨道交通安全研究体系[J].都市快轨交通,2004(4):1-4.
[6] Kannan G,Martinez J C.Simulating wireless radio communications in earthwork construction [J]. Proceedings of Congress on Computing in Civil Engineering,1998,10(13): 53-60.
[7]常太华.检测技术与应用[M].北京:中国电力出版社,2003.
[8]沈中城.检测技术与仪器[M].北京:高等教育出版社,2003.
[9]潘长胜.桥梁检测技术及其发展趋势简述[J].黑龙江交通科技,2009(4):106-108.
[10]刘正光.人造卫星定位系统在桥梁结构健康监测系统中的应用[J].现代交通技术,2008,5(6):22-27.
[11]胡志伟,朱明达.光纤传感技术在桥梁检测中的应用[J].交通标准化,2008(8):125-127.
[12]高文,陈熙霖.计算机视觉[M].北京:清华大学出版社, 1999.
[13]许巧游.基于机器视觉系统的零件识别与检测的研究[D].南京航空航天大学硕士论文,2006.
[14]吴珂.基于机器视觉的弹片质量在线检测系统[D].重庆大学硕士学位论文,2010.
[15]蒋昀赟.基于机器视觉的PC轨道梁在役锚固螺杆检测的自动定位系统研究[D].重庆大学硕士学位论文,2007.
[16]黄正福.面向尺寸检测的机器视觉系统研究[D].浙江工业大学硕士论文,2004.
[17]王志凌.基于双CCD的大锻件尺寸测量的实验研究[D].燕山大学硕士论文.2005.
[18]施雄林,李朝晖等.曲轴锻件计算机视觉检测系统[J].重庆工学院学报. 2006, 20(8):28-31.
[19] Ghosal S, Mehorotrar. Orthogonal Moment Operators for Subpixel Edge Detection[J].Pattern Recognition,1993,9(3):203-210.
[20]廖强,周忆等.机器视觉在精密测量中的应用[J].重庆大学学报(自然科学版), 2002, 25(6):1-3.
[21]贾云得.机器视觉[M].北京:科学出版社, 2000.
[22]凌云光视数字图像处理公司.CCD&CMOS图像和机器视觉产品手册[Z].
[23] J.C.Nunes, Y.Bouaoune, E.Delechelle. Image analysis by bidimensional empirical mode decomposition[J]. Image and Vision Computing, 2003, 21(12):1019-1026.
[24] M. M Park, R. C. Singer, A. J. Plueddemann, R Weller.High-speed, real-time data acquisition for vector measuring current dynammeters[J].Proc. IEEE Fourth Working Conf. Current Meas , 1990(4):146-153.
[25]北京嘉恒中自图像技术有限公司.机器视觉系统[Z].
[26] Kim Jung-Yong, Cho Yoon-Ho. Property of Images of Asphalt Pavement and Enhancement Algorithm to Reduce Noise[J]. Proceedings of SPIE-The International Society for Optical Engineering, 2003,20(2):728-739.
[27]王永刚,卫保国.基于数码相机的图像采集系统[M].北京:北京工业大学信号与信息处理研究室,2000.
[28]刘丁,毛德柱,王云飞.USB在数据采集系统中的应用[J].电子技术与应用,2000(4):43-48.
[29]潘锋.一种基于直线校正的通用摄像机标定技术研究[C].第二届全国智能视视觉监控学术会议.北京:科学出版社, 2003.
[30] Guiducci, Antonio. camera calibration for road applications[J]. computer vision and image under stanging. 2000, 79(2):250-266.
[31]夏良正,李久贤.数字图像处理[M].南京:东南大学出版社,2006.
[32]尹兰.基于数字图像处理技术的混凝土表面裂缝特征测量合分析[D].东南大学硕士论文,2006.
[33]荣士杰,吴春镕,孙正兴.计算机图像学[M].北京:电子工业部出版社,1997.
[34]赵京东,赵景秀.改进的基于梯度幅值的自适应边缘检测算法[J].光电子技术, 2009,29(2):98-102.
[35]朱晓林,高诚辉等.基于Canny算子的机械零件图像分块自适应边缘检测[J].江南大学学报(自然科学版),2010,9(3):304-307.
[36]曲昆鹏,郑丽颖.基于目标、背景比例的灰度图像自动阈值选取法[J].应用科技,2010,37(2):52-54.
[37]刘鑫.基于投影的运动目标跟踪研究[J].电子测试,2009,3:1-4.
[38] Otsu N. A threshold selection method from gray-level histograms[J].IEEE Trans on Systems,Man,and Cybernetics.1979,9(1):62-66.
[39]乔万波,曹银杰.一种改进的灰度图像二值化方法[J].电子科技,2008,21(11):63-64,71.
[40]马建红,沈西挺.Visual C++程序设计与软件技术基础[M].北京:中国水利水电出版社,2002.
[41]王勇智.数字图像的二值化技术探究[J].湖南理工学院学报,2005,18(1):31-33.
[42] Rathank,, Karuk, Chen S Y,et al. A real-time matching system for large ingerprint databases[J]. IEEE Transactions on pattern analysis and machine intelligence, 1996, 18(8): 799-813.
[43]王广伟,李铸国,吴毅雄.基于特征点提取和旋转不变技术的锁片基体视觉检测[J].中国机械工程,2007,18(6):643-645.
[44] You J,Bhattacharya P A.Wavelet-based Coarse to fine Image Matching Scheme in a Parallel Virtual Machine Environment[J].IEEE Trans on Image Processing,2000,9(9):1547-1559.
[45]田涌涛,李霞等.基于二维三阶多项式拟合的阈值曲面分割法[J].计算机工程, 2003, 29(4):127-129.
[46] Steger C,Ulrich M.Machine Vision Algorithms and Applications[M].北京:清华大学出版社, 2008.
[47]刘波,叶俊勇,让晓勇.基于机器视觉的轨道梁晃动量检测方法[J].计算机系统应用,2009(4):202-204.
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |