轨道缺陷自动检测系统的研究与应用
|
摘要
铁路轨道在使用过程中,由于行车载荷以及自然因素的作用,会使轨道表面产生各种缺陷,对列车运行的安全构成威胁。由于轨道路线长,地形复杂,人工检查和测量比较危险和困难;要对数量如此庞大的轨道表面资源进行有效管理和维护,必须迅速提高检测技术的整体水平。轨道表面检测技术要求由人工检测向自动化检测技术发展,以使轨道表面质量的监测、评估和病害分析更加快捷,养护更加合理和经济。
本论文在理解图像检测系统相关技术的基础上,提出了轨道表面缺陷自动检测系统设计方案。该系统利用车载摄像机,在合适的光照条件下,自动采集轨道表面图像,实时地传入计算机或存储系统中。论文参照项目提出的实际要求,通过研究与比较,选择适合轨道缺陷检测的设备组成硬件系统。在图像的处理过程中,系统通过软件设计与实现,自动将缺陷从图像中标注出来,并将检测结果储存在文件当中。在图像处理算法上,本文提出三种不同的方案,分析了各自的优缺点,并对它们的处理结果进行多方面比较,筛选出最优方法。
在对现场采集的轨道表面图像的实验研究中,该方法准确、快速地检测出轨道表面的缺陷情况,满足项目对系统提出的准确性和实时性等技术要求,为检修提供了数据依据,具有一定的参考价值。
In the use of railway, the surface can be damaged by the train or the factor of nature. Surface defect test is difficult and dangerous as the railway is very long and the environment is complex for human. To manage and service the entire road surface resource and data as they are so large in quantities, we must improve the level of test as swift as it is possible. So it is useful of automatic testing to the surface of railway, not by people. Only using this method, the test of railway surface can be efficient.
This paper studies the detect technology of the railway surface on the base of image detection system. We put forward the test system of railway surface's damage, which collecting the road surface's image, using camera in the condition of fit illumination, and save the image to the disk, then process them online. With reference to the project's requirements, this paper chooses devices which are suitable for railway testing to assemble the hardware system by comparing their functions. In the procedure of image processing, the software calculates the area of the damage and marks the defects of the railway surface, then picks up the damage region from the image and stores the result to text files. We put forward three different algorithms, analyze their advantages and disadvantages, compare their results in several aspects and choose the best one.
In the application of pictures collected in worksite, this chosen method is rapid and accurate. It successfully gives the damage information of road surface which meets the requirement of this project. The result is just the evidence by which we repair the road surface. We can test the damage of road surface in this way usually, and then solve the problem in time.
本论文在理解图像检测系统相关技术的基础上,提出了轨道表面缺陷自动检测系统设计方案。该系统利用车载摄像机,在合适的光照条件下,自动采集轨道表面图像,实时地传入计算机或存储系统中。论文参照项目提出的实际要求,通过研究与比较,选择适合轨道缺陷检测的设备组成硬件系统。在图像的处理过程中,系统通过软件设计与实现,自动将缺陷从图像中标注出来,并将检测结果储存在文件当中。在图像处理算法上,本文提出三种不同的方案,分析了各自的优缺点,并对它们的处理结果进行多方面比较,筛选出最优方法。
在对现场采集的轨道表面图像的实验研究中,该方法准确、快速地检测出轨道表面的缺陷情况,满足项目对系统提出的准确性和实时性等技术要求,为检修提供了数据依据,具有一定的参考价值。
In the use of railway, the surface can be damaged by the train or the factor of nature. Surface defect test is difficult and dangerous as the railway is very long and the environment is complex for human. To manage and service the entire road surface resource and data as they are so large in quantities, we must improve the level of test as swift as it is possible. So it is useful of automatic testing to the surface of railway, not by people. Only using this method, the test of railway surface can be efficient.
This paper studies the detect technology of the railway surface on the base of image detection system. We put forward the test system of railway surface's damage, which collecting the road surface's image, using camera in the condition of fit illumination, and save the image to the disk, then process them online. With reference to the project's requirements, this paper chooses devices which are suitable for railway testing to assemble the hardware system by comparing their functions. In the procedure of image processing, the software calculates the area of the damage and marks the defects of the railway surface, then picks up the damage region from the image and stores the result to text files. We put forward three different algorithms, analyze their advantages and disadvantages, compare their results in several aspects and choose the best one.
In the application of pictures collected in worksite, this chosen method is rapid and accurate. It successfully gives the damage information of road surface which meets the requirement of this project. The result is just the evidence by which we repair the road surface. We can test the damage of road surface in this way usually, and then solve the problem in time.
引文
[1]刘圣.图像检测系统的关键技术.计算机应用研究.2007.4.26-28
[2]王晓勇.基于决策层信息融合的边缘检测算法及其在混合图像滤波中的应用.计算机工程与科学.2007.3.11-16
[3]高成丽.基于机器视觉的工件检测技术研究.海洋测绘.2006.4.45-52
[4]李贤帅.图像检测识别技术应用研究.计算机工程与科学.2005.11.22-31
[5]夏培容.石中锁等.汽车牌照自动识别系统.微计算机信息.2003.12.43-52
[6]PKyoung-Su Park,.PSangjin Hong Jan.Spectral content characterization for efficient image detection algorithm design.EURASIP Journal.2007.35-41
[7]Julien Law-to.Video and image copy detection demo.ACM international conference.Jul.2007.66-78
[8]Toshiyuki Amano.Correlation Based Image Defect Detection 18th International Conference on Pattern Recognition.Aug.2006.33-45
[9]Cem Unsalan.Kim L.Boyer.A system to detect houses and residential street networks in multispectral satellite images.Computer Vision and Image Understanding.Jun.2005.77-89
[10]Duan-sheng Chen.Zheng-kai Liu.A Novel Approach to Detect and Correct Highlighted Face Region in Color Image IEEE Conference.Jul.2005.153-155
[11]Jiebo Luo.Matthew Boutell.Automatic Image Orientation Detection via Confidence-Based Integration of Low-Level and Semantic Cues.IEEE.May.2005.99-114
[12]Jingran Sun.Peihua Qiu.Edge detection,image segmentation and their applications in microarray image analysis.Doctoral Thesis:University of Minnesota.Jan.2005.66-76
[13]Zhiguo Cheng.Mang Chen.Yuncai Liu.A robust algorithm for image principal curve detection Pattern Recognition Letters.Aug.2004.36-45
[14]David R.Martin.Charless C.Fowlkes Learning to Detect Natural Image Boundaries Using Local Brightness,Color,and Texture Cues.IEEE.May.2004.81-92
[15]Daniel Keren.Margarita Osadchy.Antifaces:A Novel,Fast Method for Image Detection IEEE.Jul.2001
[16]J Pynn.A Wright and R Lodge.Automatic identifacation of cracks in road surfaces.Proceedings of 7th International Congress on Image Processing and its Applications.vol.2.IEE 1999.p671-675
[17]Hu GongZhu.Gorge Stonman.3D surface solution using structured light and constraint propagation.IEEE-PAMI,1995,11(4),225231.332-338
[18]大桓图像.DH-HV1300UM数字摄像机使用说明书.3-4
[19]H.D.Cheng.M.Miyojim.Automatic pavement.distress detection system.Journal of Information Sciences 108.pp.219-240,1998.22-26
[20]The Benefits of LED Light Sources.http://www.machinevisiononline.org/public/articles/NER_The Benefits of LEDLightSources.pdf.54-66
[21]袁祥辉.固体图像传感器及其应用.计算机工程与科学.2004.1.98-102
[22]冈萨雷斯(Gonzalez R.C.).数字图像处理(第二版)[M].北京:电子工业出版社.2003.453.-455
[23]汪同庆.叶俊勇等.锚固螺杆健康监测系统实施方案.计算机应州研究.2003.8.79-81
[24](美)卡斯尔曼(Castleman K.R.)著.朱志刚等译.数字图像处理[M].北京:电子工业出版社.2003.165-177
[25]Milan Sonka.Vaclav Hlavac.Roger Boyle著.艾海舟.武勃等译.图像处理、分析与机器视觉[M].北京:人民邮电出版社.2003.2-5
[26]机器视觉中的光源.http://www.evenfit.com.cn/info/091652559_2.htm.56-59
[27]Kane,JS.贡树行.光学设计是机器视觉系统的关键.红外.1999(8):37-39.
[28]Kenneth R.Castleman著.朱志刚.林学.石定机等译.数字图像处理[M].北京:电子工业出版社.2002.18-22
[2]王晓勇.基于决策层信息融合的边缘检测算法及其在混合图像滤波中的应用.计算机工程与科学.2007.3.11-16
[3]高成丽.基于机器视觉的工件检测技术研究.海洋测绘.2006.4.45-52
[4]李贤帅.图像检测识别技术应用研究.计算机工程与科学.2005.11.22-31
[5]夏培容.石中锁等.汽车牌照自动识别系统.微计算机信息.2003.12.43-52
[6]PKyoung-Su Park,.PSangjin Hong Jan.Spectral content characterization for efficient image detection algorithm design.EURASIP Journal.2007.35-41
[7]Julien Law-to.Video and image copy detection demo.ACM international conference.Jul.2007.66-78
[8]Toshiyuki Amano.Correlation Based Image Defect Detection 18th International Conference on Pattern Recognition.Aug.2006.33-45
[9]Cem Unsalan.Kim L.Boyer.A system to detect houses and residential street networks in multispectral satellite images.Computer Vision and Image Understanding.Jun.2005.77-89
[10]Duan-sheng Chen.Zheng-kai Liu.A Novel Approach to Detect and Correct Highlighted Face Region in Color Image IEEE Conference.Jul.2005.153-155
[11]Jiebo Luo.Matthew Boutell.Automatic Image Orientation Detection via Confidence-Based Integration of Low-Level and Semantic Cues.IEEE.May.2005.99-114
[12]Jingran Sun.Peihua Qiu.Edge detection,image segmentation and their applications in microarray image analysis.Doctoral Thesis:University of Minnesota.Jan.2005.66-76
[13]Zhiguo Cheng.Mang Chen.Yuncai Liu.A robust algorithm for image principal curve detection Pattern Recognition Letters.Aug.2004.36-45
[14]David R.Martin.Charless C.Fowlkes Learning to Detect Natural Image Boundaries Using Local Brightness,Color,and Texture Cues.IEEE.May.2004.81-92
[15]Daniel Keren.Margarita Osadchy.Antifaces:A Novel,Fast Method for Image Detection IEEE.Jul.2001
[16]J Pynn.A Wright and R Lodge.Automatic identifacation of cracks in road surfaces.Proceedings of 7th International Congress on Image Processing and its Applications.vol.2.IEE 1999.p671-675
[17]Hu GongZhu.Gorge Stonman.3D surface solution using structured light and constraint propagation.IEEE-PAMI,1995,11(4),225231.332-338
[18]大桓图像.DH-HV1300UM数字摄像机使用说明书.3-4
[19]H.D.Cheng.M.Miyojim.Automatic pavement.distress detection system.Journal of Information Sciences 108.pp.219-240,1998.22-26
[20]The Benefits of LED Light Sources.http://www.machinevisiononline.org/public/articles/NER_The Benefits of LEDLightSources.pdf.54-66
[21]袁祥辉.固体图像传感器及其应用.计算机工程与科学.2004.1.98-102
[22]冈萨雷斯(Gonzalez R.C.).数字图像处理(第二版)[M].北京:电子工业出版社.2003.453.-455
[23]汪同庆.叶俊勇等.锚固螺杆健康监测系统实施方案.计算机应州研究.2003.8.79-81
[24](美)卡斯尔曼(Castleman K.R.)著.朱志刚等译.数字图像处理[M].北京:电子工业出版社.2003.165-177
[25]Milan Sonka.Vaclav Hlavac.Roger Boyle著.艾海舟.武勃等译.图像处理、分析与机器视觉[M].北京:人民邮电出版社.2003.2-5
[26]机器视觉中的光源.http://www.evenfit.com.cn/info/091652559_2.htm.56-59
[27]Kane,JS.贡树行.光学设计是机器视觉系统的关键.红外.1999(8):37-39.
[28]Kenneth R.Castleman著.朱志刚.林学.石定机等译.数字图像处理[M].北京:电子工业出版社.2002.18-22