毕业论文

打赏
当前位置: 毕业论文 > 外文文献翻译 >

机器视觉三坐标测量英文文献和中文翻译(2)

时间:2019-04-16 21:25来源:毕业论文
1) The basic technique of the measuring system can bedescribed as follows: The touch-trigger probe of thelarge-scale CMM is replaced by a single CCD camera,and then, the noncontact measurement can be


1) The basic technique of the measuring system can bedescribed as follows: The touch-trigger probe of thelarge-scale CMM is replaced by a single CCD camera,and then, the noncontact measurement can be realized.Meanwhile, the CCD camera can be moved along theXYZ directions of the large-scale CMM instead of thetouch-trigger probe; thus, the large-scale noncontact 3-Dcoordinate vision measuring system of the feature pointscan be realized. (It means that the camera transversalmoving-type binocular stereo vision measuring systemcan be realized. Note that the optical axes of the cameraat two measuring positions are parallel.)The measuring system synthesizes the advantages ofnoncontact, high efficiency, high precision, and shortmeasuring range of the vision measuring method, as wellas the advantages of large measuring range, high preci-sion, and flexibility of the large-scale CMM. The disad-vantages of short measuring range are as follows: 1)Mea-surement precision will decrease as the measuring rangeof the fixed-type machine vision system increases; and2) the difficulty of camera calibration of the intersectionaldirection and positional parameters of each camera in amulticamera vision measuring system can be avoided.2) In the proposed vision measuring system, the featurepoint is the intersection point of cross-cutting lines, whichare cut at the measured position on the surface of aworkpiece.3) In the monocular vision method, because it is difficult toaccurately calibrate the out-of-focus model, the depth-from-defocus method cannot directly be used in thehigh-precision measurement [13], [14]. The measurementprecision of the depth-from-focus method is related to thedepth of focus of the camera optical system [15], [16] andthe precision of the 3-D driving device of the camera.It is difficult to determine whether the measured fea-ture point quickly and accurately locates in the rangeof depth of focus by eye observation. Meanwhile, ifthe eye-observation method is adopted, the measurementerror will be introduced because of the different artificialevaluation criteria. Because the imaging system is not amicrographic system, the depth of focus is too large forthe requirement of measurement precision, and it willintroduce a measurement error to the depth-from-focusmonocular vision method.4) The main difficulty of the binocular stereo vision methodis stereo matching [17]. Because the intersection pointsof cross-cutting lines act as feature points and can beextracted by the image processing algorithm, then stereomatching can be avoided. When the binocular stereo vi-sion method is used, the positions of the measured featurepoints should be ensured within the depth of focus ofthe camera; otherwise, the blurred images will introducedifficulties for image processing. Although the camera islocated in the range of depth of focus, the image points’positions of feature points on the image plane of theCCD camera will also change as the different positionsof the camera. The aforementioned two aspects will allintroduce measurement errors.Because the large-scale CMM is selected as the 3-Ddriving device in the measuring system, the short visionmeasuring range is adopted to increase the vision mea-surement precision. The focal length of the camera andthe object distance are also short. (An imaging opticalsystem is a reduced system.) The aforementioned variousaspects are synthetically considered; thus, the depth offocus of the camera is not long, and it is not suitable forbinocular stereo vision measurement.5) Based on the aforementioned advisements, monocularvision [18] and binocular stereo vision are combinedin the proposed measuring system to measure the 3-Dcoordinates of feature points.In the general planning of the measuring system, the goalof the monocular vision method is to locate the position ofthe camera near the focusing position relative to the measured feature point.
The depth-from-defocus and depth-from-focusmethods are integrated together to find the focusing position.First, the position of the camera can be located near the fo-cusing position that is relative to the feature point by meansof the depth-from-defocus method. Second, the accurate focus-ing position can be calculated by means of the depth-from-focus method. A new kind of image focus measure functionof energy-spectrum entropy is proposed, and the experimentalresults indicate that it is more accurate, reliable, and stablethan other image focus measure functions. Then, the focusingposition of the camera relative to the feature point can quicklyand accurately be acquired. Because the camera optical systemhas a certain depth of focus, the calculated focusing positionalso has an error, but the calculated focusing position accordingto the aforementioned algorithm will extremely approach theideal focusing position in the certain error range, and this willsatisfy the depth measuring condition of the binocular stereovision method.The purpose of the aforementioned method is to eliminate theinfluence of unstable human judgments of the focusing positionand to automatically find the focusing position according to theunified evaluation criterion. Then, the measurement precisionand reliability can be improved. Ahuja and Abott [19] andDalmia and Trivedi [20] presented similar ideas as the proposedmeasuring system, but only the general concepts and qualitativeexplanations are given, the detailed and complete theoreticaldeductions are not presented in [19], and the measurementprecision of experimental results in [20] is lower than that ofthe proposed measuring system.The measuring system can be represented by four coordi-nate systems, namely, the large-scale CMM coordinate systemOWXWYWZW, the camera coordinate system OCXCYCZC,the CCD image plane coordinate system OP XP YP , and thecomputer image coordinate system OIXI YI . The large-scaleCMM coordinate system acts as the world coordinate system.The original point of the camera coordinate system is the prin-ciple point of the object lens of the camera, and the direction ofthe ZC axis is the optical axis of the object lens of the camera.The CCD image plane coordinate system is set up on the CCDimage plane, and the original point of the CCD image planecoordinate system is the left top corner of the CCD image plane.The computer image coordinate system is set up on the imagedata array, and the original point of the computer image coor-dinate system is the pixel position corresponding to the originalpoint of the CCD image plane coordinate system. (XP0,YP0) isthe intersection point of the optical axis of the object lens withthe CCD image plane, and it is the geometrical center and idealprinciple point of the CCD image plane. (XI0,YI0) is the pixelposition corresponding to (XP0,YP0) in the computer imagecoordinate system. The geometrical model and four coordinatesystems of the measuring system are shown in Fig. 1.The basic principle of the measuring system is shown inFig. 2.The main measurement procedures of the measuring systemare given in the list that follows.1) According to the combination of position-from-defocusand position-from-focus methods, the focusing positionsof the camera relative to the measured feature points canbe located in the range of depth of focus of the cameraoptical system. 机器视觉三坐标测量英文文献和中文翻译(2):http://www.751com.cn/fanyi/lunwen_32194.html
------分隔线----------------------------
推荐内容