CCD雙目立體視覺測(cè)量系統(tǒng)的理論研究 外文翻譯

1、 畢業(yè)設(shè)計(jì)外文資料翻譯題 目: CCD stereo vision measurement system theory 院系名稱：電氣工程學(xué)院 專業(yè)班級(jí)： 自動(dòng)化 學(xué)生姓名： 學(xué) 號(hào)： 指導(dǎo)教師： 教師職稱： 起止日期：2011-2-262011-3-14 地點(diǎn): 附 件： 1.外文資料翻譯譯文；2.外文原文。 指導(dǎo)教師評(píng)語：簽名： 年 月 日附件1：外文資料翻譯譯文CCD雙目立體視覺測(cè)量系統(tǒng)的理論研究摘要: 利用幾何成像原理建立起 CCD 雙目立體視覺測(cè)量系統(tǒng)的數(shù)學(xué)模型 ,從提高系統(tǒng)測(cè)量精度出發(fā) ,在理論上重點(diǎn)對(duì)系統(tǒng)結(jié)構(gòu)參數(shù)、 圖像識(shí)別誤差與系統(tǒng)測(cè)量精度的關(guān)系進(jìn)行了深入的分析和探討 ,并通過

2、實(shí)驗(yàn)對(duì)結(jié)論進(jìn)行了驗(yàn)證。研究內(nèi)容對(duì)實(shí)際建立該測(cè)量系統(tǒng)具有很強(qiáng)的指導(dǎo)作用。關(guān)鍵詞: 立體視覺; CCD ; 測(cè)量精度; 圖像識(shí)別; 系統(tǒng)測(cè)量引言雙目立體視覺測(cè)量技術(shù)是計(jì)算機(jī)視覺中的一個(gè)重要分支,一直是計(jì)算機(jī)視覺研究的重點(diǎn)和熱點(diǎn)之一。由于其近似于人眼視覺系統(tǒng) ,具有較高的測(cè)量精度和速度 ,并具有結(jié)構(gòu)簡單 ,便于使用等優(yōu)點(diǎn) ,所以被廣泛應(yīng)用于工業(yè)檢測(cè)、 物體識(shí)別、 工件定位、 機(jī)器人自導(dǎo)引等諸多領(lǐng)域。近年來許多學(xué)者對(duì)此進(jìn)行了大量的研究工作1 - 4 。其中大量的工作集中在對(duì)視覺測(cè)量系統(tǒng)的數(shù)學(xué)模型、 系統(tǒng)的定標(biāo)方法5 - 7 以及目標(biāo)特征點(diǎn)匹配算法8 - 9 的研究上 ,而對(duì)系統(tǒng)的結(jié)構(gòu)參數(shù)(兩個(gè) CCD

3、之間的距離、 光軸夾角等研究得卻很少。文獻(xiàn)10 對(duì)立體視覺結(jié)構(gòu)參數(shù)進(jìn)行了相應(yīng)的理論研究 ,但它是從觀看物體時(shí)的深度感出發(fā)研究CCD與物體之間的距離、 兩個(gè)CCD間距和觀看距離3個(gè)參數(shù)之間的關(guān)系 ,沒有涉及到結(jié)構(gòu)參數(shù)對(duì)系統(tǒng)測(cè)量精度的影響。而實(shí)踐證明系統(tǒng)的結(jié)構(gòu)參數(shù)設(shè)置在實(shí)際應(yīng)用中對(duì)于系統(tǒng)的測(cè)量精度是至關(guān)重要的。此外 ,從立體視覺測(cè)量原理中 ,可以看出圖像識(shí)別誤差是另一個(gè)對(duì)系統(tǒng)測(cè)量精度產(chǎn)生直接影響的重要因素。綜合以上考慮 ,從理論上對(duì)系統(tǒng)的結(jié)構(gòu)參數(shù)設(shè)置和圖像識(shí)別誤差對(duì)系統(tǒng)測(cè)量精度的影響進(jìn)行了深入的分析和研究。結(jié)合系統(tǒng)結(jié)構(gòu)參數(shù)對(duì)攝像機(jī)定標(biāo)精度的影響 ,給出了實(shí)際應(yīng)用中組建雙目立體視覺測(cè)量系統(tǒng)的設(shè)計(jì)方案

4、。1 雙目立體視覺測(cè)量原理及數(shù)學(xué)模型1. 1 攝像機(jī)成像模型攝像機(jī)的成像模型,是光學(xué)成像系統(tǒng)幾何關(guān)系的數(shù)學(xué)表示。目前在攝像機(jī)標(biāo)定中應(yīng)用的攝像機(jī)成像模型主要有針孔成像模型、 雙平面模型和人工神經(jīng)網(wǎng)絡(luò)模型等。其中針孔成像模型是目前大量采用的一種成像模型,它反映的是一種理想的線性映射關(guān)系,如圖1 所示。其中, Oc為攝像機(jī)的光心, Oc XcYc Zc為攝像機(jī)坐標(biāo)系, OXYZ 為世界坐標(biāo)系, oxy 為攝像機(jī)成像平面物理坐標(biāo)系。P( X , Y , Z 為空間一物點(diǎn), p ( x , y是其在圖像平面上的投影點(diǎn)。根據(jù)該成像模型,空間物點(diǎn) P( X , Y , Z與像點(diǎn)之間的關(guān)系可表示為當(dāng)兩個(gè)或兩個(gè)

5、以上攝像機(jī)進(jìn)行交會(huì)時(shí),可以得到2 i ( i 2個(gè)方程所組成的超定方程組,因此可以用最小二乘法對(duì)方程組求解以確定空間物點(diǎn)的坐標(biāo)。1. 2 雙目立體視覺測(cè)量系統(tǒng)的數(shù)學(xué)模型雙目立體視覺系統(tǒng)通常由兩臺(tái)結(jié)構(gòu)和性能完全相同的CCD組成 ,并且兩個(gè)CCD擺放位置對(duì)稱?；谏鲜鰯z像機(jī)成像模型 ,由式(2可以推出如下超定方程組:其中 r , t , r , t 分別為兩攝像機(jī)坐標(biāo)系相對(duì)于世界坐標(biāo)系的平移和旋轉(zhuǎn)矩陣,可以通過攝像機(jī)標(biāo)定得到。( x , y , ( x , y 分別為空間物點(diǎn)( X , Y , Z在兩CCD圖像平面的投影的物理坐標(biāo)。圖2 雙目立體視覺簡易模型為了對(duì)系統(tǒng)結(jié)構(gòu)參數(shù)進(jìn)行分析,這里構(gòu)建如圖

6、2所示的立體視覺系統(tǒng)。在該結(jié)構(gòu)中,三坐標(biāo)系處于同一平面內(nèi),其中 Y 軸垂直紙面向里,世界坐標(biāo)系 OXYZ 與左攝像機(jī)坐標(biāo)系Oc Xc Yc Zc原點(diǎn)重合,兩攝像機(jī)間距為 L ,兩光軸夾角為 2。根據(jù)上述結(jié)構(gòu),可以確定兩攝像機(jī)組成的超定方程組為則由式(4可得: 2 圖像識(shí)別誤差對(duì)系統(tǒng)測(cè)量精度的影響立體視覺系統(tǒng)中 ,空間物點(diǎn)在兩個(gè)攝像機(jī)圖像平面上的位置是通過像素坐標(biāo)來表示的 ,而面陣CCD攝像機(jī)像素具有一定的物理尺寸 ,這就使得空間物點(diǎn)在圖像上的真實(shí)物理坐標(biāo)無法得到準(zhǔn)確的表達(dá) ,從根本上造成了圖像識(shí)別誤差。如圖3所示。圖像平面上像素坐標(biāo)系與物理坐標(biāo)系有如下關(guān)系:其中: ( u , v是以像素為單位

7、的圖像坐標(biāo)系的坐標(biāo),dx , dy 分別為每一個(gè)像素在x 軸和 y 軸方向上的物理尺寸, ( u0 , v0為 oxy 坐標(biāo)系原點(diǎn)o在o0 uv 坐標(biāo)系中的坐標(biāo)3 。如圖 4 所示,假設(shè)圖像識(shí)別精度達(dá)到0. 5個(gè)像素級(jí)。對(duì)于一空間物點(diǎn),設(shè)其投影到圖像平面上的第 i 行,第 j 列的像素中,則此時(shí)該點(diǎn)的物理坐標(biāo)為 x = ( i - u0 ×dxy = ( j - v0 ×dy即只要該點(diǎn)落在該像素內(nèi),其坐標(biāo)值是一個(gè)定值。而理想情況下的坐標(biāo)應(yīng)分別在一定的范圍內(nèi):( i - u0 - 0. 5 ×dx < x < ( i - u0 + 0. 5 ×

8、dx( j - v0 - 0. 5 ×dy < y < ( i - v0 + 0. 5 ×dy綜合上式,則對(duì)應(yīng)該點(diǎn)的圖像識(shí)別誤差為 ex = 0. 5 - ( x - x ey = 0. 5 - ( y - y 其中: x , y 分別表示對(duì) x , y 取整。由于圖像識(shí)別誤差的存在,則實(shí)際像點(diǎn)坐標(biāo)與理想像點(diǎn)坐標(biāo)有如下關(guān)系: x = x + ex , y = y + ey , ( x , y為實(shí)際像點(diǎn)物理坐標(biāo), ( x , y 為理想像點(diǎn)物理坐標(biāo)。設(shè)被測(cè)物點(diǎn)坐標(biāo) P = ( X , Y , Z ,則由公式(1計(jì)算出該點(diǎn)在兩攝像機(jī)投影的理想像點(diǎn)坐標(biāo)( x , y和

9、( x , y ,考慮圖像識(shí)別誤差,根據(jù)式(8得出實(shí)際像點(diǎn)坐標(biāo)( x ,y , ( x , y ,將其代入(4式可得被測(cè)點(diǎn)的空間坐標(biāo)( X , Y , Z ,則被測(cè)物點(diǎn)的測(cè)量誤差可表示為eX = X - X , eY = Y - Y , eZ = Z - Z (93 系統(tǒng)結(jié)構(gòu)參數(shù)分析及實(shí)驗(yàn)結(jié)果3. 1 結(jié)構(gòu)參數(shù)與系統(tǒng)測(cè)量精度的關(guān)系根據(jù)上述系統(tǒng)模型及數(shù)學(xué)推導(dǎo)過程,得出了系統(tǒng)結(jié)構(gòu)參數(shù)與系統(tǒng)測(cè)量誤差之間的關(guān)系圖。圖5為特定物點(diǎn)的測(cè)量誤差與兩攝像機(jī)夾角2之間的關(guān)系圖。從圖中可以看出, eX 變化波動(dòng)不大, eY 隨 2的增大呈緩慢的上升趨勢(shì), eZ 則變化比較劇烈,隨2的增大大幅提升。圖5 特定空間點(diǎn)

10、測(cè)量誤差與兩軸夾角的關(guān)系圖圖6為一定范圍內(nèi)的物點(diǎn)與世界坐標(biāo)系原點(diǎn)的平均測(cè)量誤差與兩攝像機(jī)夾角 2和兩攝像機(jī)間距離L 之間的關(guān)系圖?？梢钥闯?當(dāng)兩攝像機(jī)間距離一定時(shí),平均誤差隨 2的增大而增大,當(dāng)兩攝像機(jī)夾角一定時(shí),平均誤差隨其距離的增加而不斷增大。圖6 平均誤差與兩軸夾角、 兩攝像機(jī)距離的關(guān)系圖總體上講,結(jié)構(gòu)參數(shù)與系統(tǒng)測(cè)量精度是一個(gè)較為復(fù)雜的函數(shù)關(guān)系,可以總結(jié)如下:1 攝像機(jī)之間的間距大小與系統(tǒng)測(cè)量誤差成正比關(guān)系,間距越小,誤差越小;2 當(dāng)攝像機(jī)之間的夾角不大于130° 時(shí),測(cè)量誤差較小,反之較大。有極高的可信性。3. 2 結(jié)構(gòu)參數(shù)對(duì)攝像機(jī)定標(biāo)精度的影響在實(shí)際應(yīng)用中,需要建立兩個(gè) C

11、CD 坐標(biāo)系之間的聯(lián)系 ,這就需要對(duì)兩 CCD 進(jìn)行立體定標(biāo) ,以求取兩 CCD之間的旋轉(zhuǎn)矩陣和平移矩陣。事實(shí)證明攝像機(jī)定標(biāo)精度與系統(tǒng)結(jié)構(gòu)參數(shù)設(shè)置同樣有著非常緊密的關(guān)系。定標(biāo)方法采用的是基于單平面模板定標(biāo)策略 ,精度評(píng)估采用基于棋盤格長度的評(píng)估方法。該方法亦可以作為系統(tǒng)測(cè)量誤差的測(cè)量方法。對(duì)定標(biāo)模板上的10個(gè)長為 50 mm的棋盤格進(jìn)行了反復(fù)實(shí)驗(yàn) ,實(shí)驗(yàn)結(jié)果見表1。表1 結(jié)構(gòu)參數(shù)對(duì)系統(tǒng)定標(biāo)平均誤差的影響從表1中可以看出 ,當(dāng)攝像機(jī)距離一定時(shí) ,定標(biāo)誤差隨光軸夾角的增加而不斷增加 ,當(dāng)光軸夾角固定不變時(shí) ,定標(biāo)誤差隨攝像機(jī)距離的增加而不斷增加。大量實(shí)驗(yàn)證明 ,當(dāng)兩攝像機(jī)間距離不超過 500mm

12、,兩光軸夾角不超過60° 時(shí) ,定標(biāo)誤差較小。4 結(jié)束語綜合系統(tǒng)結(jié)構(gòu)參數(shù)對(duì)測(cè)量精度及定標(biāo)精度的影響 ,在建立立體視覺系統(tǒng)時(shí) ,兩攝像機(jī)光軸夾角和兩攝像機(jī)間距應(yīng)盡可能小 ,但在實(shí)際應(yīng)用中 ,考慮到便于目標(biāo)特征點(diǎn)視覺匹配 ,尤其是對(duì)運(yùn)動(dòng)目標(biāo)進(jìn)行大范圍實(shí)時(shí)跟蹤測(cè)量 ,有目標(biāo)被遮擋的情況發(fā)生時(shí) ,兩攝像機(jī)光軸夾角應(yīng)選擇在 30° 60° 之間。此外 ,目標(biāo)特征點(diǎn)圖像識(shí)別精度應(yīng)盡可能達(dá)到亞像素精度 ,盡量避免 “失之毫厘 ,差之千里” 的現(xiàn)象。根據(jù)以上推導(dǎo)搭建了立體視覺系統(tǒng),取得了較好的實(shí)驗(yàn)結(jié)果,證明本文的結(jié)論對(duì)實(shí)踐具有較大的指導(dǎo)意義,為進(jìn)一步開展深入研究打下了堅(jiān)實(shí)的基礎(chǔ)。

13、附件2：外文原文（復(fù)印件）CCD stereo vision measurement system theoryAbstract: Using the principles of geometrical imaging CCD stereo vision to build a mathematical model of measurement systems, improve the precision of the system from the start, in theory, focus on system parameters, image recognition errors an

14、d the relationship between the precision of the system has in-depth analysis and Explore and experiment on the conclusion was verified. Research on the actual establishment of this system has a strong guiding role.Key words: stereo vision; CCD; measurement accuracy; image recognition; system measure

15、mentsIntroduction Stereoscopic measurement in computer vision technology is an important branch of computer vision has been the focus of the study and the hot spots. Due to its similar to the human visual system with high precision and speed, and has a simple structure, easy to use, etc., they were

16、widely used in industrial inspection, object recognition, workpiece positioning, robot homing, and many other fields . In recent years many scholars have done a lot of research work 1 - 4. A lot of work which focused on the mathematical model of vision measurement system, the system calibration meth

17、od 5 - 7 and the target feature point matching algorithm 8 - 9, while the structural parameters of the system (of two CCD distance between the optical axis angle, etc. are rarely studied. 10 structure parameters of stereoscopic vision corresponding theoretical research, but it is viewing objects fro

18、m the start of the depth of feeling between the CCD and the object distance and viewing distance of the two CCD spacing between 3 parameters , did not address the structural parameters of the system measurement accuracy. The practice proved that the structure of the system parameters in practical ap

19、plications of the system's measurement accuracy is essential. In addition, from the stereo vision measurement principle, we can see the error image recognition system accuracy is another important factor in a direct impact. Based on the above considerations, the structure from the theoretical pa

20、rameters of the system and image recognition errors on the precision of the system of in-depth analysis and research. Combination of system parameters on the accuracy of the camera calibration given set of practical applications, stereo vision measurement system design1 Binocular stereo vision measu

21、rement principle and mathematical model1.1 Camera imaging modelCamera imaging model, the geometric relationship between the optical imaging system of mathematics said. At present the application of camera calibration are pinhole camera imaging model imaging model, two-plane model and artificial neur

22、al network model. In which a large number of pinhole imaging model is used in an imaging model, which reflects the relationship between an ideal linear map shown in Figure 1. One, Oc for the camera optical center, Oc XcYc Zc for the camera coordinate system, OXYZ the world coordinate system, oxy for

23、 the camera imaging plane physical coordinates.P (X, Y, Z for the space of a material point, p '(x, y is its projection point on the image plane. According to the imaging model, the space object point P (X, Y, Z and the relationship between the image point can be expressed asWhen two or more int

24、ersection cameras, you can get 2 i (i 2 consisting of equations overdetermined equations, so the least squares method can be used to solve equations to determine the spatial coordinates of object points.1.2 stereo vision measurement system modelBinocular stereo vision system usually consists of two

25、identical CCD structure and properties of the composition and placement of the two CCD symmetry. Based on the above camera imaging model, from (2 can introduce the following overdeterminedEquations:Where r, t, r ', t' are the two camera coordinate system relative to the world coordinate syst

26、em translation and rotation matrix, can be obtained through camera calibration. (X, y, (x ', y', respectively, for the space object point (X, Y, Z in the two CCD image plane projection of the physical coordinates.Figure 2 Simple model of binocular stereo visionIn order to analyze structural

27、parameters of the system, where the building shown in Figure 2 stereo vision system. In this structure, the coordinate system in the same plane, which for the Y axis perpendicular to the paper, the world coordinate system OXYZ with the left camera coordinate origin coincides Oc Xc Yc Zc, two camera

28、spacing L, the angle between the two axis 2. According to the above structure, the camera can determine the composition of the two equations for the overdeterminedBy equation (4 yields:2 Image recognition errors on the accuracy of the systemStereo vision system, the space object point in two camera

29、image plane position is represented by pixel coordinates, and pixel area array CCD camera has a certain physical size, which makes the space object point in the image on the real physical coordinates Not the exact expression, a fundamental cause of the image recognition errors.Figure 3.Image plane p

30、ixel coordinates and physical coordinates the following relations:Where: (u, v the image in pixels coordinate system, dx, dy, respectively, for each pixel in the x-axis and y axis on the physical size, (u0, v0 o the origin of the coordinate system for the oxy O0 uv coordinate system in the coordinat

31、es 3. Shown in Figure 4, assuming that the image recognition accuracy of 0.5 pixel. For a space object point, set the projection to the image plane on the i-line, the first j columns of pixels, then this time the physical coordinates of the point x = ( i - u0 ×dxy = ( j - v0 ×dyAs long as

32、the point falls within the pixel, the coordinate value is a constant. Ideally, the coordinates of which should be within a certain range, respectively:( i - u0 - 0. 5 ×dx < x < ( i - u0 + 0. 5 ×dx( j - v0 - 0. 5 ×dy < y < ( i - v0 + 0. 5 ×dyIntegrated on the type, the

33、n the error should point to Image Recognition ex = 0. 5 - ( x - x ey = 0. 5 - ( y - y Where: x, y, respectively, for x, y rounded. Since the existence of image recognition errors, the actual coordinates of image point coordinates and the ideal image point has the following relationship: x =? X + ex,

34、 y =? Y + ey, (x, y as the point of actual physical coordinates, ( ? x,? y as the ideal image point of physical coordinates. Let the measured object point coordinates P = ( X, Y, Z, by equation (1 to calculate the projection of the points in the two cameras the ideal image point coordinates (? X,? Y

35、 and (? ? x ',? y', consider the image recognition error, according to equation (8 come to the actual pixel coordinates (x, y, (x ', y', be substituted into (4, we have been space coordinates of measuring point (X, Y, Z, the measurement error of the measured object point can be expre

36、ssed as eX = X - X, eY = Y - Y, eZ = Z - Z (93 System parameters analysis and experimental results3. 1 Structural parameters of the relationship with the precision of the system and mathematical model based on the above derivation process, come to the system structure and system parameters of the re

37、lationship between measurement error map. Figure 5 is a specific object point measurement error and the camera angle 2 between the two diagrams. It can be seen from the figure, eX volatility is not changing, eY 2 increases with the rising trend was slow, eZ is more violent change, with the increase

38、of 2 increased significantly.Figure 5 The measurement error is a specific spatial point the angle between the two axis graphFigure 6 for a certain object point within the coordinate origin with the world average measurement error and the two cameras and two camera angle 2 between the distance betwee

39、n the L map. It can be seen, when a certain distance between the two cameras, the average error with the increase of 2, when the angle between the two cameras is fixed, the average error increases with the distance increasing.Figure 6 Average error and the angle between two axes, the distance betwee

40、n the two cameras chartOverall, structural parameters and precision of the system is a more complex function can be summarized as follows:1 Spacing between the camera and the measurement error is directly proportional to the smaller distance, the smaller the error;2 When the camera is not greater th

41、an the angle between the 130 °, the measurement error is small, whereas larger. Has high credibility.3. 2 Structural parameters on the accuracy of camera calibrationIn practice, the need to create two links between CCD coordinates, which need to be two-dimensional CCD calibration, in order to o

42、btain the rotation between the two CCD matrix and translation matrix. Proved the precision of camera calibration parameters and system structure also has a very close relationship. Calibration method uses a template based on single-plane calibration strategy, the accuracy was assessed using an assessment based on the length of the checkerboard method. The method also can be used as measurement error of the measurement method. Calibration template of 10 to 50 mm length of the checkerboard was repeated experimental result