外文翻譯--對(duì)輪副直徑參數(shù)的聯(lián)機(jī)測(cè)量辦法與測(cè)量系統(tǒng)

1、<p>　　ONLINE MEASURING METHOD AND SYSTEM FOR DIAMETER PARAMETERS OF WHEEL SET </p><p>　　WU Kaihua, ZHU Feng, ZHUANG Fei, YAN Kuang </p><p>　　School of Automation, Hangzhou Dianzi University

2、, Hangzhou 310018, P.R.China, Email: wukaihua@hdu.edu.cn </p><p>　　Keywords: Wheel set diameter, online measurement, optoelectronic measurement, automatic measuring system. </p><p><b>　　Ab

3、stract </b></p><p>　　The measuring of diameter parameters of wheel set is an important step for the safety of train vehicle running. The wear degree of wheel increases with the speedup of train. The pa

4、per introduced an online measuring method of diameter parameters of running wheel set based on optoelectronic detecting technique while the train is running in low-speed within 5-10km/h. The method used precision laser d

5、isplace-ment sensor, high speed and high resolution CCD and digital image processing technology to r</p><p>　　1 Introduction </p><p>　　In order to ensure the safety of a running train, it's

6、very important to detect the status of the wheel set automatically and regularly. The wearing degree of the wheel set is one of the main factors that influence the safety and stability of running train. In China, the mea

7、surement of wheel set wear is still static and by handwork mostly which limits the accuracy and reliability. The difference between different operators is often too big. The detecting efficiency is also very low. In rece

8、nt ye</p><p>　　The wear parameters of wheel set include geometrical parameters and surface defects. The main geometrical parameters include flange thickness, flange height, wheel diameter, roundness and diam

9、eter difference etc. The surface defects include scotching and flaking on wheel tread. These parameters are the main factors that influence the safety and stability of vehicle. The paper mainly introduced the online meas

10、uring method and system of diameter parameters. </p><p>　　Up to now, most of the maintaining factories were still using special mechanic tools to measure the geometrical parameters by manual work, especially

11、 in China. These tools, including special vernier caliper, wheel diameter ruler and rim inside distance ruler, will wear down. The status of these tools will influence the measuring accuracy. The manual work manner had s

12、trong labor intensity and low efficiency. The man-made factors produced by different operators perhaps bring different and variable</p><p>　　The automatic measuring methods and systems had been researched in

13、 recent years[1,2,3,4]. Yet these methods were only suitable for maintaining factory and the vehicle must be disassembled. Several countries, such as Germany[5], Japan, Italy and the United States, have developed online

14、automatic detecting systems based on different principles. These systems are very expensive and have not been widely used. In China, there are also research on this field[6], but no practical equipment had been dev</p

15、><p>　　Regular measurement and repair cannot fill the requirements of maintenance. The online measuring of profile parameters is important to the distinguishing and forecast of failures while the train is runni

16、ng in low-speed. The preventive repair will reduce the maintenance cost greatly. The paper introduced a new non-contact online measuring method and system of diameter parameters. </p><p>　　2 Measuring method

17、 </p><p>　　2.1 Definition of diameter parameters </p><p>　　Train vehicle wheel set is a rolling parts produced by pressing two wheels into an axle. Fig. 1 is the profile of a cargo wheel section

18、. Tread is the contacting part with the rail. The right plane is rim inside plane. The plane won’t contact with the rail and has no wear. The part between rim inside plane and tread is called flange. The taping point is

19、a base point on the tread and has 70mm distance to rim inside plane[7]. </p><p>　　The rolling circle is a circle passing through the taping point on the tread . The average diameter of rolling circle equals

20、wheel diameter. The roundness can be derived from diameters at different positions on the rolling circle. The diameter difference is the difference between left and right wheel. These parameters belong to geometric param

21、eters. </p><p>　　2.2 Measuring principle </p><p>　　The measuring principle of diameter parameters is based on the projection imaging of structured light beam. Fig.3 is the schematic diagram. Acc

22、ording to the definition of diameter parameters, the diameter can be deduced if the rolling circle was obtained. The left-right movement was unavoidable while the vehicle was running. So the position of rolling circle wa

23、s uncertain. A laser displacement sensor was used to acquire the position of rim inside plane in real time. So the position of rolling ci</p><p>　　If a CCD sensor captured the strips image at a certain angle

24、 relative to the incident optical plane, the profile information was recorded in the image . Then the geometric parameters can be calculated by digital image processing method. </p><p>　　2.3 Position the rim

25、 inside plane </p><p>　　The position of taping point should be obtained before measuring the diameter. It means that the position of rim inside plane should be determined firstly according to Fig.1. The posi

26、tion of rim inside plane was varied because different wheel set had different rim inside distance. The measuring accuracy of diameter was directly influenced by the position accuracy of rim inside plane. A precision lase

27、r displacement sensor was used to acquire the position of rim inside plane accurately based on th</p><p>　　2.4 Image acquisition and processing </p><p>　　For image acquiring, the key problem is

28、how to obtain the clear tread profile at certain position. So the relative position of light source, CCD and wheel detecting sensor is very important. Because the movement of the train will results in overlap of consecut

29、ive two field images, the CCD cannot work in mode of frame. Only the single field image will be acquired or the CCD supports the progressive scan mode. In order to obtain clear image, the exposure time must very small an

30、d the CCD and image </p><p>　　The natural light had influence on the captured original image. In order to improve the original image quality and obtain high image SNR, a set of narrow band-pass optical filte

31、rs which wavelength matched with laser source were added in front of CCD lens. Besides, the electronic exposure time of CCD was controlled to avoid the image blur. The resolution of CCD was 1024×768. The central wav

32、elength of laser source was 650nm. The band width was 15nm. The electronic exposure time was set to 1/10000s</p><p>　　The original profile curve image was diffused because the laser line had 1.0mm width. The

33、 one pixel width central line of the profile curve was obtained by applying (1)geometric correction, (2)window adjacent average smoothing, (3)threshold segmentation, (4) thinning, (5)edge tracing, and (6)curve fitting al

34、gorithm. Fig.4 showed the digital image processing results (photography condition: speed: 5.6km/h, power of laser source: 20mw, resolution of CCD: 1024x768, shutter time: 1/10000s, light wave</p><p>　　2.5 Me

35、asurement of wheel diameter </p><p>　　Based on Fig.4 and the position of rim inside plane decided by laser displacement sensor, a rolling circle was fitted according to the relative position between two line

36、s and base point. The object-image relation and magnifying ratio were determined by the optical imaging system and will be used in the calculation process. </p><p>　　Diameter of a circle was certain if 3 poi

37、nts’ coordinates were known on an arc. A serial of diameters at different positions could be deduced based on Fig.5. If image was acquired while vehicle rolled to different position, many other diameter values could be c

38、oncluded, then wheel diameter was the average of these diameters. The roundness could also be deduced according to the distribution of these diameters. The diameter difference is the maximum difference between left and r

39、ight wheel diameter</p><p>　　3 Measuring system </p><p>　　The schematic diagram of measuring system based on the above detecting principle showed in Fig. </p><p>　　The system consi

40、sted of wheel detector, precision laser displacement sensor, laser line source, CCD and lens, optical narrow-band filter, signal processing and isolated amplifying circuits, image acquisition circuits, interface circuits

41、, control circuits and industrial controlled computer. The arrangement of measuring units was symmetrical. Two to three measuring units every side were necessary in order to ensure the precision and reliability. </p&g

42、t;<p>　　All of the initial data and measured parameters were transferred to database and could be transmitted to central control computer by Internet for further processing and analyzing. </p><p>　　In

43、 the detecting process, the laser source, CCD and laser displacement sensor didn’t contact with the wheel set. This method had high position accuracy and simplified the image acquisition equipment. </p><p>　

44、　The non-contact measuring manner was easy to maintaining and had characteristics of high reliability, high accuracy and repeatability. </p><p>　　4 Error analysis </p><p>　　The measuring precisi

45、on was influenced by many factors, such as lens aberrance, CCD resolution, trigger error produced by wheel detector, movement error, location error of rim inside plane, error produced by the vibration, image acquisition

46、and processing error etc. The measuring units were fixed and vibrated with the railway and wheel, so the vibration error can be ignored. In these factors, the movement and trigger error were the main error source. </p

47、><p>　　4.1 Movement error </p><p>　　If the running speed was 10km/h and the exposure time was set to 1/10000 second, the movement error e will be 0.278mm. </p><p>　　4.2 Trigger error &

48、lt;/p><p>　　The trigger moment may delay or ahead of the designed time and position. So the image acquisition was also different to the scheduled time. If the response speed of wheel detector was 100 kHz, the p

49、osition trigger error will be less than 0.1 mm. </p><p>　　4.3 General error </p><p>　　The position accuracy of laser displacement sensor was 0.03mm. The algorithm error of image processing was o

50、ne element. Movement and trigger error were 0.278 mm and 0.1mm relatively. Experimental results showed the general error of diameter measuring was within 1.0 mm. </p><p>　　5 Conclusion </p><p>　

51、　A new method for wheel set diameter measuring was introduced in running condition. The method combined optoelectronic measuring and digital image processing technology together and realized the non-contact automatic mea

52、suring of diameter parameters. The dynamic measuring system was designed. The measuring accuracy of diameter parameters was less than 1.0mm. The repeatability and accuracy of the system can meet the demand of online whee

53、l set maintaining. </p><p>　　Acknowledgements </p><p>　　The paper was supported by (1) Zhejiang Provincial Natural Science Foundation of China (Y104578), (2) Education Department of Zhejiang Pro

54、vince of China (20040446) and (3) Science and Technology Department of Zhejiang Province of China (2005C31064). </p><p>　　References </p><p>　　[1] Zuo Jianyong, Zhou Wenxiang, Zeng Jing, et al,

55、“Experimental research on measuring wheel rim profile using laser sensor”, Railway Vehicle, 40(2): 11-13, 2002. </p><p>　　[2] Zheng Fenfang, Liu Ji, Fan Peixin, “Measuring wheel set tread profile using digit

56、al camera”, Railway Vehicle, 40(1): 19-22, 2002. </p><p>　　[3] Wu Kaihua, Yan Kuang, “Research on the method of measuring defects of wheel set tread using optoelectronic technique”, Optical Technique, 31(3):

57、 465-467, 2005. </p><p>　　[4] Wu Kaihua, Zhang Jianhua, Yan Kuang, Jiang Peng, “Optoelectronic Automatic Measuring System for Wheel Set Parameters”, Chinese Journal of Scientific Instrument, 27(3): 298-301,

58、306, 2006. </p><p>　　[5] Wang Hao, Wang Li, Gao Xiaorong, “Application of electro-magnetic and ultrasonic technology in the detection of wheel and the processing of detection signals”, Locomotive & Rolli

59、ng Stock Technology, 6, 34-36, 2004. </p><p>　　[6] Yan Kuang, Wu Kaihua, Wang Ruirong, Jiang Peng, “Theoretical research on the measuring method of the running wheel set tread defects based on optoelectronic

60、 technique”, Proc. of SPIE, 6150, 61502w-1-61502w-7, 2006. </p><p>　　[7] Train Vehicle [1998] No.2, “Assemble and Regulation of Cargo Wheel set and Rolling Axle”, Beijing: China Railing Press, 1998. Fig.7: m

61、easurement flow chart. I/O control Image acquisition Data acquisition Industrial Computer Database Image processing measuring unitTrigger control Data processing</p><p>　　對(duì)輪副直徑參數(shù)的聯(lián)機(jī)測(cè)量辦法與測(cè)量系統(tǒng)</p><p

62、>　　WU Kaihua, ZHU Feng, ZHUANG Fei, YAN Kuang</p><p>　　杭州電子大學(xué)自動(dòng)化學(xué)院 中國(guó)杭州 310018 </p><p>　　Email: wukaihua@hdu.edu.cn</p><p>　　關(guān)鍵詞：輪輻直徑，聯(lián)機(jī)測(cè)量，光電測(cè)量，自動(dòng)測(cè)量系統(tǒng)</p><p>　　摘要：

63、輪輻直徑參數(shù)的測(cè)量是確保車(chē)輛系安全運(yùn)行的一個(gè)重要環(huán)節(jié)。車(chē)輪的磨損程度與列車(chē)增加的速度成正比。這篇論文介紹了當(dāng)列車(chē)在以5-10千米/時(shí)的低速運(yùn)行時(shí)，運(yùn)用光電探測(cè)技術(shù)的一種對(duì)運(yùn)行中的齒輪的直徑參數(shù)聯(lián)機(jī)測(cè)量的方法。此法利用精密激光位移傳感器，高速高分解CCD和數(shù)字圖象處理技術(shù)實(shí)現(xiàn)輪輻參數(shù)的無(wú)接觸自動(dòng)測(cè)量。這些參數(shù)主要包括直徑，圓度及直徑偏差。測(cè)量的精確度的主要影響方面是對(duì)機(jī)械裝置，震動(dòng)裝置和啟動(dòng)裝置的誤差的分析。自動(dòng)測(cè)量系統(tǒng)因此而生。此系統(tǒng)包

64、含車(chē)輪探測(cè)器，激光位移傳感器，半導(dǎo)體激光器來(lái)源，CCD，聯(lián)機(jī)圖象搜索器處理電路。理論與實(shí)踐的結(jié)果表明，直徑參數(shù)的測(cè)量精確度在1.0毫米內(nèi)。此精確度與聯(lián)機(jī)直徑測(cè)量的要求正好符合。</p><p><b>　　1 緒論</b></p><p>　　為了確保運(yùn)行中的列車(chē)的安全，自動(dòng)地規(guī)律地探測(cè)輪輻的狀況很重要。輪輻的磨損度是影響列車(chē)運(yùn)行的安全性，穩(wěn)定性的幾個(gè)主要方面之一。在

65、中國(guó)，輪輻磨損度的測(cè)量方法仍是靜止的，人工的，而這限制了它的準(zhǔn)確性與可靠性。不同的操作人員得出的結(jié)果往往大不相同，探測(cè)效率也很低。近年來(lái)，其運(yùn)行速度已達(dá)到160-200千米/時(shí)。但其磨損的程度與速度比以前快了很多。靜止的人工的測(cè)量方法已不能適應(yīng)高速列車(chē)的發(fā)展。探索自動(dòng)聯(lián)機(jī)的方法是運(yùn)輸與維修部門(mén)的緊急需要。</p><p>　　輪輻的磨損參數(shù)包括幾何參數(shù)與表面缺陷。主要的幾何參數(shù)包括邊緣厚度，邊緣高度，車(chē)輪直徑，圓

66、度，直徑差異等等。表面缺陷包括車(chē)輪支撐面上的擦傷，發(fā)裂。這些參數(shù)是影響車(chē)輛安全性，穩(wěn)定性的主要方面。這篇論文主要介紹聯(lián)機(jī)直徑參數(shù)的測(cè)量方法與測(cè)量系統(tǒng)。</p><p>　　至今為止，尤其在中國(guó)，大多數(shù)維修工廠(chǎng)仍通過(guò)手動(dòng)使用專(zhuān)門(mén)的機(jī)械工具來(lái)測(cè)量幾何參數(shù)。這些工具包括專(zhuān)業(yè)游標(biāo)彎腳規(guī)，車(chē)輪直徑尺，內(nèi)邊緣距離尺，會(huì)磨損壞掉。這些工具的使用程度會(huì)影響測(cè)量效果。手動(dòng)人工的工作方式帶來(lái)的是高強(qiáng)度勞動(dòng)量和低工作效率。不同操作員引

67、起的人為因素可能會(huì)產(chǎn)生各種各樣的不同的誤差，其測(cè)量精密度，可靠性，可重復(fù)性不符合輪輻維修部門(mén)的要求。</p><p>　　自動(dòng)測(cè)量方法與系統(tǒng)已被研究多年[1,2,3,4]。而這些方法僅僅適合于維修工廠(chǎng)，而且車(chē)輛要被分解。一些國(guó)家，例如德國(guó)[5]，日本，意大利和美國(guó)，已不同原則地發(fā)展了聯(lián)機(jī)檢測(cè)系統(tǒng)。這些系統(tǒng)非常昂貴，還沒(méi)有被廣泛使用。在中國(guó)，這個(gè)領(lǐng)域也有研究[6]，但沒(méi)有實(shí)際的設(shè)備發(fā)展。</p>&l

68、t;p>　　正常的測(cè)量與維修不能滿(mǎn)足維修的要求。輪廓參數(shù)的聯(lián)機(jī)測(cè)量在低速運(yùn)行的列車(chē)對(duì)于區(qū)別與預(yù)測(cè)失誤很重要。預(yù)防的維修會(huì)大大降低維修成本。這篇論文介紹一種新的無(wú)接觸直徑參數(shù)聯(lián)機(jī)測(cè)量方法與系統(tǒng)。</p><p><b>　　2 測(cè)量方法</b></p><p>　　2.1 直徑參數(shù)的定義</p><p>　　車(chē)輛系的輪輻是通過(guò)擠壓兩個(gè)車(chē)輪形

69、成一個(gè)輪軸的卷型部分。圖.1是載物輪區(qū)域的輪廓（PROFILE）?；瑒?dòng)面（TREAD）是指與鐵軌接觸的部分。右平面是指內(nèi)邊緣平面。平面與鐵軌不接觸，沒(méi)有磨損。內(nèi)邊緣平面與滑動(dòng)面（TREAD）之間的部分叫軌底。絕緣繞阻點(diǎn)是指在滑動(dòng)面上的一個(gè)基本點(diǎn)，距離內(nèi)邊緣平面70毫米[7]。</p><p>　　齒輪的滾動(dòng)圓是指過(guò)滑動(dòng)面上絕緣繞阻點(diǎn)的一個(gè)圓圈。滾動(dòng)圓的平均直徑等于車(chē)輪直徑。圓度得自不同位置滾動(dòng)圓的直徑。直徑差異是

70、指左右車(chē)輪的差異。這些參數(shù)屬于幾何參數(shù)。</p><p><b>　　2.2測(cè)量原則</b></p><p>　　直徑參數(shù)的測(cè)量原則是基于由光束組成的圖象設(shè)計(jì)。根據(jù)直徑參數(shù)的定義，如果可以獲得滾動(dòng)圓，直徑可以推斷出來(lái)。當(dāng)車(chē)輛運(yùn)行時(shí)，左右機(jī)械裝置不可避免的，所以滾動(dòng)圓的位置是無(wú)法確定的。激光位移傳感器被用來(lái)獲取內(nèi)邊緣平面的位置，這樣滾動(dòng)圓的位置就可以確定了。兩道來(lái)自半導(dǎo)

71、體激光器的光束沿著滾動(dòng)圓的滾動(dòng)方向照亮滑動(dòng)面表面。滾動(dòng)圓在兩道光束之間。兩道光束間的空隙的設(shè)計(jì)是為了確保滾動(dòng)圓總在這里。外來(lái)的視覺(jué)光源平面并列滾動(dòng)圓平面,兩道亮光帶在滑動(dòng)面上形成。較窄的光帶記錄車(chē)輪的直徑信息。</p><p>　　若CCD感應(yīng)器在與外來(lái)視覺(jué)平面相關(guān)的某一角度獲取了圖像帶，輪廓信息圖像中就可以記錄輪廓信息。然后幾何參數(shù)就能夠通過(guò)數(shù)字圖像處理方法記錄下來(lái)。</p><p>　

72、　2.3內(nèi)邊緣平面的位置</p><p>　　在測(cè)量直徑之前絕緣繞阻點(diǎn)的位置應(yīng)該確定下來(lái)。這表明內(nèi)邊緣平面的位置應(yīng)該由圖.1先確定下來(lái)。內(nèi)邊緣平面的位置是多種多樣的，因?yàn)椴煌妮嗇椨胁煌膬?nèi)邊距。內(nèi)邊緣平面的位置的精確度直接影響直徑測(cè)量的效果。用三角測(cè)量原則，用精密激光位移傳感器精確獲取內(nèi)邊緣平面的位置。感應(yīng)器擁有色彩補(bǔ)償功能，適應(yīng)表面色彩的改變。位置的精確度是0.03毫米，反應(yīng)時(shí)間低于1秒鐘。這個(gè)特點(diǎn)能夠滿(mǎn)足聯(lián)

73、機(jī)測(cè)量的需要。</p><p>　　2.4圖象的獲取與處理</p><p>　　要獲得圖象，關(guān)鍵問(wèn)題是如何獲得某一位置清晰的滑動(dòng)面輪廓。所以光源的相關(guān)位置，CCD和車(chē)輪檢測(cè)器都非常重要，因?yàn)榱熊?chē)的運(yùn)動(dòng)會(huì)使兩個(gè)局域圖象的連續(xù)交搭，CCD不能在結(jié)構(gòu)方式中工作。只有單一局域的圖象能被獲得，而CCD支持先進(jìn)的瀏覽方式。為了獲得清晰的圖象，曝光時(shí)間一定要短，CCD和圖象獲得卡必須支持異步復(fù)位方式。取

74、得的一瞬要由車(chē)輪探測(cè)感應(yīng)器控制。用高速圖象獲取卡通過(guò)DSP方式能夠獲得圖象。圖2：輪輻</p><p>　　自然光對(duì)于獲取的原始圖象是有影響的。為了提高原始圖象的質(zhì)量，獲得高質(zhì)量的圖象SNR，我們要在CCD透鏡前加上一個(gè)和激光源匹配波長(zhǎng)的窄帶視覺(jué)過(guò)濾器。除此以外，CCD的電子曝光時(shí)間要控制住，以免圖象模糊。CCD的分辨率是1024×768。激光光源的中央波長(zhǎng)是650納米，波寬是15納米。電子曝光時(shí)間是1

75、/10000秒。</p><p>　　因?yàn)榧す馐?.0納米寬，所以原始輪廓彎曲的圖象會(huì)被擴(kuò)散，輪廓彎曲中央線(xiàn)的寬度就能通過(guò)下列得出：（1）幾何矯正，（2）觀(guān)察口交界處均勻校平（3）開(kāi)端分割（4）修磨（5）邊緣處理（6）曲線(xiàn)適合的算法</p><p>　　圖4表明了數(shù)字圖象處理的結(jié)果（圖象數(shù)據(jù)：速度：5.6千米/時(shí)激光源的功率：20MW，CCD分辨率：1024×768，曝光時(shí)間：1

76、/10000秒，光波長(zhǎng)度：15納米，兩光束距離：5納米，光波寬度：1.0毫米）。</p><p>　　2.5 車(chē)輪直徑的測(cè)量</p><p>　　按照?qǐng)D.4和由激光位移傳感器決定的內(nèi)邊緣平面的位置，根據(jù)光束和基點(diǎn)之間的相關(guān)位置，滾動(dòng)圓可以配合上見(jiàn)。視覺(jué)圖象系統(tǒng)決定了目標(biāo)圖象放大率，而目標(biāo)圖象放大率會(huì)在計(jì)算處理中使用。如果知道在弧上的3個(gè)點(diǎn)的坐標(biāo)，圓的直徑就能確定。不同位置的圓的直徑能夠在圖

77、.5的基礎(chǔ)上推斷出來(lái)。當(dāng)車(chē)輛滾動(dòng)到不同位置，如果獲得了圖象知道了其它直徑的數(shù)值，車(chē)輪的直徑等于這些直徑的平均值。根據(jù)這些直徑，圓度也可以算出來(lái)。直徑的差異是左右輪直徑的最大不同所在。</p><p><b>　　3 測(cè)量系統(tǒng)</b></p><p>　　此系統(tǒng)由車(chē)輪探測(cè)器，精密激光移植傳感器，激光光源，CCD以及透鏡，視覺(jué)窄帶過(guò)濾器，信號(hào)處理，隔離擴(kuò)大電路，圖象獲取電

78、路，接口電路，控制電路和工業(yè)控制計(jì)算機(jī)組成。測(cè)量器的布局是勻稱(chēng)的。為了確保精確性與可靠性，每邊兩到三個(gè)測(cè)量器是必要的。</p><p>　　所有的原始數(shù)據(jù)和測(cè)量出的參數(shù)都會(huì)傳送到數(shù)據(jù)庫(kù)，然后通過(guò)網(wǎng)絡(luò)傳到中央控制計(jì)算機(jī)上做進(jìn)一步的處理與分析。</p><p>　　在探測(cè)處理中，激光光源，CCD和激光移植感應(yīng)器和輪輻并不接觸。這種方法擁有高的位置精確度和簡(jiǎn)化的圖象獲取裝備。 </p>

79、;<p>　　利用這種無(wú)接觸測(cè)量方式，維修就變得很簡(jiǎn)單，而且它還有可靠性高，準(zhǔn)確性高，可重復(fù)性高的特點(diǎn)。</p><p><b>　　4 誤差分析</b></p><p>　　測(cè)量的精確度由很多方面影響，比如透鏡象差，CCD分辨率，由車(chē)輪探測(cè)器引起的啟動(dòng)裝置誤差，機(jī)械裝置誤差，內(nèi)邊緣平面定位裝置的誤差，由震動(dòng)引起的誤差和圖象獲取與處理誤差等。因?yàn)闇y(cè)量器由

80、鐵路和車(chē)輪確定，所以震動(dòng)裝置的誤差可以忽略。這些誤差當(dāng)中，機(jī)械裝置和啟動(dòng)裝置的誤差是主要誤差。</p><p>　　4.1 機(jī)械裝置中的誤差</p><p>　　如果運(yùn)行速度是10千米/時(shí)，曝光時(shí)間在1/10000秒，運(yùn)動(dòng)誤差e會(huì)在0.278毫米。</p><p>　　4.2啟動(dòng)裝置的誤差</p><p>　　啟動(dòng)的時(shí)刻可能會(huì)先于或后于設(shè)計(jì)的

81、時(shí)間和位置，所以圖象的獲取與預(yù)計(jì)的時(shí)間也會(huì)不同。如果車(chē)輪探測(cè)器的反應(yīng)速度是100 kHz，啟動(dòng)位置誤差會(huì)小于0.1毫米。</p><p><b>　　4.3常見(jiàn)誤差</b></p><p>　　激光位移傳感器的位置精確度是0.03毫米。圖象處理的算法誤差是一個(gè)要素。機(jī)械裝置和啟動(dòng)裝置的誤差是0.278毫米和0.1毫米。實(shí)驗(yàn)結(jié)果表明直徑測(cè)量的的常見(jiàn)誤差是在1.0毫米內(nèi)。

82、</p><p><b>　　5 總結(jié)</b></p><p>　　這里已經(jīng)介紹了一種新的測(cè)量輪輻直徑的方法。這種方法結(jié)合光電測(cè)量與數(shù)字圖象處理技術(shù)一起實(shí)現(xiàn)直徑參數(shù)的無(wú)接觸自動(dòng)測(cè)量。人們?cè)O(shè)計(jì)了這種動(dòng)態(tài)才測(cè)量系統(tǒng)。直徑參數(shù)的測(cè)量精確度在1.0毫米以?xún)?nèi)。它的可重復(fù)性與精確度能夠滿(mǎn)足聯(lián)機(jī)輪輻維修的要求。</p><p><b>　　參 考

83、文 獻(xiàn)</b></p><p>　　[1] Zuo Jianyong, Zhou Wenxiang, Zeng Jing等，關(guān)于用激光傳感器測(cè)量輪緣剖面的實(shí)驗(yàn)研究.鐵路交通，40(2)：11-13, 2002. </p><p>　　[2] Zheng Fenfang, Liu Ji, Fan Peixin,基于數(shù)碼相機(jī)的輪輻剖面測(cè)量. 鐵路交通， 40(1)：19-22,

84、 2002. </p><p>　　[3] Wu Kaihua, Yan Kuang，關(guān)于運(yùn)用光電子設(shè)備對(duì)輪副剖面測(cè)量的缺點(diǎn)研究.光學(xué)技術(shù)，31(3)：465-467, 2005. </p><p>　　[4] Wu Kaihua, Zhang Jianhua, Yan Kuang, Jiang Peng，輪副的光電子自動(dòng)測(cè)量系統(tǒng).中國(guó)科學(xué)儀表， 27(3)：298-301, 306,

85、 2006. </p><p>　　[5] Wang Hao, Wang Li, Gao Xiaorong, 電磁及超聲波技術(shù)在齒輪探傷中的應(yīng)用與探測(cè)信號(hào)處理，車(chē)頭及整車(chē)技術(shù)， 6, 34-36, 2004. </p><p>　　[6] Yan Kuang, Wu Kaihua, Wang Ruirong, Jiang Peng, 關(guān)于運(yùn)用光電子設(shè)備對(duì)運(yùn)動(dòng)輪副的測(cè)量缺點(diǎn)的理論研究，國(guó)