车轮踏面缺陷检测系统的研究
|
摘要
随着铁路系统大提速,机车运行的安全程度越来越受到关注。车轮作为机车车辆的重要部件,其车轮踏面故障的测量是铁路部门一直密切关注和不断研究改进的课题,因为踏面的擦伤是影响运行安全性、平稳性和运用经济性的重要因素。
论文主要由硬件和软件两部分构成。
第一部分主要是借鉴常用的机器视觉系统模型,设计出适合本系统的在线检测硬件构成,并针对系统所要求达到的精度进行了相应的参数计算,选择合适的元器件。
第二部分主要是图像处理、畸变图像的标定部分。根据车轮踏面缺陷图像的特点,利用数字图像处理技术对采集到的踏面图像进行处理并提取出缺陷,同时对缺陷图像进行标定,最后输出缺陷图像的实际尺寸。图像处理的主要步骤包括预处理、检测区域定位、边缘检测、边缘连接、目标缺陷定位。以下所提出的算法对准确地检测和分析缺陷具有极大的影响。
1.先对获取的畸变图像进行标定。
2.为避免车轮背景干扰缺陷的检测,减少运算量,在缺陷分析之前首先对踏面区域进行定位,利用垂直边缘检测算子强化踏面左右边界,再利用投影法确定踏面位置并将其剪切下来。
3.缺陷的边缘具有不规则性,为了对缺陷形状进行准确提取,论文中选择Canny算子来完成边缘检测,提取效果好,可以有效地保护边缘信息。
4.改进了边缘生长算法,对得到的边缘图像进行连接,达到一定效果。增强了目标缺陷的连通性。
5.通过对模板的标定,得到单个像素的大小,并以此来计算已定位的缺陷的实际大小。
With the development of the railway, defect inspection of rails becomes more and more important to ensure the safety. The problem that how to measure the fault of the tread profile is concerned all along and perpetual studied and improved. That is because the irregularities of tread profile are the key factor of influencing the safety, stabilization and economy of running.
The paper is composed of a hardware part and a software part.
The first part mainly designs the hardware frame of the system using for reference universal the model of machine vision system. It aims at the system precision's demand and chooses the right components and computers each component' parameters.
The second part is mainly about image processing. According to the surface defect image's traits and using digital image processing it calibrates the distorted images and abstracts the defects from the tread images. The key steps include pretreatment, tread location determination, edge detection, edge link, defect orientation. Under-mentioned arithmetic has a great influence on inspecting and analyzing defects exactly.
1. In order to avoid the disturbance of the wheel, tread region is oriented by using vertical edge detection. The operation can intensify rail's borderline and cut it down by projection.
2. The edge of the defect is abnormity. In order to abstract the defect with high precision, canny method is applied to carry on edge detection.
3. Improved edge growing arithmetic is advanced to link discontinuous edge, which increases connectedness.
4. According to the calibration stencil-plate, we get the size of one pixel; then after Zhang Zhengyou image calibration, we get the tread defect's real size.
论文主要由硬件和软件两部分构成。
第一部分主要是借鉴常用的机器视觉系统模型,设计出适合本系统的在线检测硬件构成,并针对系统所要求达到的精度进行了相应的参数计算,选择合适的元器件。
第二部分主要是图像处理、畸变图像的标定部分。根据车轮踏面缺陷图像的特点,利用数字图像处理技术对采集到的踏面图像进行处理并提取出缺陷,同时对缺陷图像进行标定,最后输出缺陷图像的实际尺寸。图像处理的主要步骤包括预处理、检测区域定位、边缘检测、边缘连接、目标缺陷定位。以下所提出的算法对准确地检测和分析缺陷具有极大的影响。
1.先对获取的畸变图像进行标定。
2.为避免车轮背景干扰缺陷的检测,减少运算量,在缺陷分析之前首先对踏面区域进行定位,利用垂直边缘检测算子强化踏面左右边界,再利用投影法确定踏面位置并将其剪切下来。
3.缺陷的边缘具有不规则性,为了对缺陷形状进行准确提取,论文中选择Canny算子来完成边缘检测,提取效果好,可以有效地保护边缘信息。
4.改进了边缘生长算法,对得到的边缘图像进行连接,达到一定效果。增强了目标缺陷的连通性。
5.通过对模板的标定,得到单个像素的大小,并以此来计算已定位的缺陷的实际大小。
With the development of the railway, defect inspection of rails becomes more and more important to ensure the safety. The problem that how to measure the fault of the tread profile is concerned all along and perpetual studied and improved. That is because the irregularities of tread profile are the key factor of influencing the safety, stabilization and economy of running.
The paper is composed of a hardware part and a software part.
The first part mainly designs the hardware frame of the system using for reference universal the model of machine vision system. It aims at the system precision's demand and chooses the right components and computers each component' parameters.
The second part is mainly about image processing. According to the surface defect image's traits and using digital image processing it calibrates the distorted images and abstracts the defects from the tread images. The key steps include pretreatment, tread location determination, edge detection, edge link, defect orientation. Under-mentioned arithmetic has a great influence on inspecting and analyzing defects exactly.
1. In order to avoid the disturbance of the wheel, tread region is oriented by using vertical edge detection. The operation can intensify rail's borderline and cut it down by projection.
2. The edge of the defect is abnormity. In order to abstract the defect with high precision, canny method is applied to carry on edge detection.
3. Improved edge growing arithmetic is advanced to link discontinuous edge, which increases connectedness.
4. According to the calibration stencil-plate, we get the size of one pixel; then after Zhang Zhengyou image calibration, we get the tread defect's real size.
引文
[1]徐之海.现代成像系统.国防工业出版社.
[2]王庆有.CCD应用技术.天津大学出版社.2000
[3]刘文耀.光电图像处理.电子工业出版社.2002
[4]凌云光子图像公司说明书
[5]贾永红.数字图像处理.武汉大学出版社.
[6]Milan Sonka.图像处理、分析与机器视觉.人民邮电出版社
[7]冈萨雷斯.数字图像处理(第二版).电子工业出版社.
[8]朱学芳,智文广.计算机图像处理导论.北京科学技术文献出版社.2003
[9]章毓晋.图像处理和分析基础.高等教育出版社.2002
[10]徐瑞梅等.铁轨缺陷自动检测系统的研究与发展.大连海事大学硕士论文.2003
[11]T.Poggio,H.Voorhees and A.Yuille,"A Regularized Solution to Edge Detection," Tech.Rep.MA,Rep.AIM-833,MIT Artificial Intell.Lab.,May 1985
[12]杨振亚等.自适应Canny边缘检测算法.上海海运学院学报.2003(12)
[13]杨振亚,王仙道.自适应图像边缘检测算法.计算机应用.2003(5)
[14]赵业等.基于边界点方向的边缘检测方法.东南大学学报.1994(11)
[15]周丽娟等.图像边界自动跟踪的边缘提取算法.哈尔滨理工大学学报.1999(10)
[16]Vishvjit S.Nalwa and Thomas O.Binford,"On Detecting Edged,"IEEE Trans.Pattern Analysis and Machine Intelligence,Vol.PAMI-8,No.6,pp.699-714,Nov.1986。
[17]屈彬.一种基于区域生长规则的快速边缘跟踪算法.四川大学学报.2002(2)
[18]苏菡等.基于DIS边缘检测和自适应边缘生长的图像分割方法.哈尔滨工 程大学学报.2004(3)
[19]李春颖.机器视觉在钢球表面缺陷检测中的应用.计算机与现代化.2000(10)
[20]孙朝明,伏德贵.面向对象的边缘检测方法.机电一体化.2003(6):25-27
[21]PC2-VISION MANUAL
[22]邱钊,朱庆生,卢霞,伍星.基于边缘信息的工业CT图像分割法.计算机工程.2004(8):159-161
[25]Zhengyou Zhang.A Flexible New Technique for Camera Calibration[J].IEEE2004.2.
[23]DongHun Lee.A Simple and Effective Correction Method for Distorted Image from Camera Lens System.SICE-ICASE International Joint Conference[J].2006.
[24]XU Haixiang,ZHU Guangxi.Image Segmentation Based on Support Vector Machine.Journal of eletronic science and technology of china[J].2005.Vol3.
[26]郑南宁.计算机视觉与模式识别.国防工业出版社.1998
[27]Martin T.Hagan.Neural Network Design.北京机械工业出版社.
[28]刘相滨等.基于八邻域边界跟踪的标号算法.计算机工程.2001(23)
[2]王庆有.CCD应用技术.天津大学出版社.2000
[3]刘文耀.光电图像处理.电子工业出版社.2002
[4]凌云光子图像公司说明书
[5]贾永红.数字图像处理.武汉大学出版社.
[6]Milan Sonka.图像处理、分析与机器视觉.人民邮电出版社
[7]冈萨雷斯.数字图像处理(第二版).电子工业出版社.
[8]朱学芳,智文广.计算机图像处理导论.北京科学技术文献出版社.2003
[9]章毓晋.图像处理和分析基础.高等教育出版社.2002
[10]徐瑞梅等.铁轨缺陷自动检测系统的研究与发展.大连海事大学硕士论文.2003
[11]T.Poggio,H.Voorhees and A.Yuille,"A Regularized Solution to Edge Detection," Tech.Rep.MA,Rep.AIM-833,MIT Artificial Intell.Lab.,May 1985
[12]杨振亚等.自适应Canny边缘检测算法.上海海运学院学报.2003(12)
[13]杨振亚,王仙道.自适应图像边缘检测算法.计算机应用.2003(5)
[14]赵业等.基于边界点方向的边缘检测方法.东南大学学报.1994(11)
[15]周丽娟等.图像边界自动跟踪的边缘提取算法.哈尔滨理工大学学报.1999(10)
[16]Vishvjit S.Nalwa and Thomas O.Binford,"On Detecting Edged,"IEEE Trans.Pattern Analysis and Machine Intelligence,Vol.PAMI-8,No.6,pp.699-714,Nov.1986。
[17]屈彬.一种基于区域生长规则的快速边缘跟踪算法.四川大学学报.2002(2)
[18]苏菡等.基于DIS边缘检测和自适应边缘生长的图像分割方法.哈尔滨工 程大学学报.2004(3)
[19]李春颖.机器视觉在钢球表面缺陷检测中的应用.计算机与现代化.2000(10)
[20]孙朝明,伏德贵.面向对象的边缘检测方法.机电一体化.2003(6):25-27
[21]PC2-VISION MANUAL
[22]邱钊,朱庆生,卢霞,伍星.基于边缘信息的工业CT图像分割法.计算机工程.2004(8):159-161
[25]Zhengyou Zhang.A Flexible New Technique for Camera Calibration[J].IEEE2004.2.
[23]DongHun Lee.A Simple and Effective Correction Method for Distorted Image from Camera Lens System.SICE-ICASE International Joint Conference[J].2006.
[24]XU Haixiang,ZHU Guangxi.Image Segmentation Based on Support Vector Machine.Journal of eletronic science and technology of china[J].2005.Vol3.
[26]郑南宁.计算机视觉与模式识别.国防工业出版社.1998
[27]Martin T.Hagan.Neural Network Design.北京机械工业出版社.
[28]刘相滨等.基于八邻域边界跟踪的标号算法.计算机工程.2001(23)