1249019 九、發明說明： 【發明所屬之技術領域】 本發明係關於一種應用在主動式3D量測系統之彈性化校 正方法，详言之，係關於一種利用一具有複數個圓孔且已 知尺寸之校正平板以求得較精確之系統校正參數。 【先前技術】 由於光電技術、微機電技術及精密機械技術的快速發 展，其應用的層面也愈來愈廣泛，此一趨勢，使得各種元 件表面量測的範圍朝更微小、更精準、更快速的方向發展。 在精微元件表面的量測及檢測應用上，目前的量測方式常 因人為判斷程序過多以及設備精度不足，造成所量測出的 結果無法完全顯示真實的表面狀態；傳統的接觸式量測因 過長的量測時間以及易對待測物造成損傷，使其應用空間 觉限’因此光學量測法應運而生。 過去在光學式扣量測方法主要是採用㈣式的量測架 構，將兩個或兩個以上的攝影機裝設於待測物上方，由雷 射光投射一光點或一直線至待測物表面形成標示區，並在 各攝影機所操取之影像上形成—匹配點群（晴叫。滅_ P_t cl(Hld)，透過計算可以找出各攝影機間的匹配點，瘦 由三角測量原理計算出待測物的高度'然後，利用移動待 測物或是工作平台來量取待測物之整個外形，此種方式之 缺點為需要較長之量測時間，同時因為量測時需要移動機 構，造成精度下降。 而主動式的量測方法主要是將其中—台攝影機及雷射光 97302.doc 1249019 源以投射裝置取代，利用投射裝置投射出具有平面結構性 之光學圖案，同時光學圖案可以依時序作變化，增加光學 圖案的變化性並有利於後續之匹配運算。 參考圖1 ’顯不習用主動式3D量測系統之示意圖。該習用 主動式3 D里測糸統1包括一平台丨丨、一光源產生器12、一透 鏡13、一處理單元14、一投射裝置（包括一數位式反射裝置 15及一投射鏡頭16)及一影像擷取裝置（例如一 CCD攝影機 17)。該平台11係用以置放一待測元件。該光源產生器^係 用以產生一光線，經由該透鏡13而射至該數位式反射裝置 15。該處理單元14(例如一電腦），其產生一可以彈性編輯與 調變之圖系文。該數位式反射裝置15包# 一數位式光源處理 器（Digital Light Processor，DLP)151及一數位式微鏡片裝置 (Digital MiCr〇-mirrors Device，DMD)152，該數位式微鏡片 裝置152係由1024x768個微鏡片所組成，該處理單元14係透 過该數位式光源處理器15 1而控制每一微鏡片之轉動，決定 微鏡片之ON(反射光至待測元件）或〇FF(反射光至背景）之 位置，同時藉由對微鏡片0N/0FF時間長短的控制，可以決 定投射點的光場強度，所以透過微鏡片之控制可將來自該 光源產生器12之光線反射成與該圖紋相對應之光學圖案。 該投射鏡頭16將該光學圖案縮小至所需要的量測範圍之 後投射至該待測元件之表面。該CCD攝影機17係用以擷取 來自該平台11上該待測元件之影像，且將該擷取影像傳送 至該處理單元14。該處理單元14比對該擷取影像與該圖^ 之變化’以得到該待測元件之表面狀態。 97302.doc 1249019 然而’接著最重要的任務就是量測系統的校正，校正程 序與校正品質的優劣，直接影響到量測系統的效能及精 度:因為光學系統的校正主要是求得系統的内部參數（光學 中心、有效焦距、失真率、影像中心）、外部參數（投射單元、 參考平面及CCD各座標系統間之旋轉矩陣及平移向量），以 提供後續量測之幾何運算。習知之校正方法均將實際之三 維系統簡化成二維模式，也就是必須假設該投射裝置Μ 该CCD攝影機17之光學中心之連線，必須與該平心上之 參考平面平行，或是假設該投射裝置15與該CCD攝影機17 光學中心以及量測點處於共面，再運用:機 丹連用一維之二角幾何關 ^刀析㈣❹數與㈣m由於三維量測被 間化成二維系統’雖然公式的推導被簡化，但是相對的校 ^精度較低，僅錢詩精度要求不高的條件下，而且在 貫際的系統安裝，不易達到所謂的二維且共面之模型。另 二：習知之校正方法中均直接假設該投射裝置15與該 :景，：17之光學中心與幾何中心重合，但是實際上其 先學中心與幾何中心並未重合。 因此，有必要提供一種創新且具進步性的校正方法，以 解決上述問題。 【發明内容】 本發明之主要目的係提供一種彈性化校正方法，呈對於 :射^學中心與影像類取裝置光學中心的位置並無特定限 =:::有原來全域式量測的特點外，在應用時更具有 較大之译性’可縮短系統架設時間，同時因與實際物理架 97302.doc 1249019 構吻合，且不豐狡ι 、 卜而移動任何機構，因此具有較高之量測精度。 ^達上述目的’本發明提供-種應用在主動式3D量測系 、洗之彈性彳b 正法，該主動式3]3量測系、统包括_平台、 一投射,置及-影像掏取裝置，該校正方法包括： 0)提供一杈正平板，其具有複數個圓孔，該等圓孔之孔 徑及孔位距離係為已知； (b)將該校正平板置於該平台上； (0利用該影像擷取裝置擷取該等圓孔之影像； ⑷利用影像處理技術求出各該圓孔之圓心座標，並以最 J平方法求彳于该影像擷取裝置與該平台之座標轉換關係以 及-亥〜像#貞取裝置之有效焦距、失真率及光學中心座標； (e) 移開該校正平板； (f) 利用S亥投射裝置投射一光學圖案至該平台，其中該光 子圖案包括複數個圓點，該等圓點之圓心距與該等圓孔之 圓心距相同； (g) 利用該影像擷取裝置擷取該等圓點之影像；及 (h) 於垂直方向移動該平台，以取得更多之校正點’再以 最小平方法求得該投射裝置、該影像擷取裝置與該平台之 座標轉換關係。 【實施方式】 本發明係關於一種應用在一主動式3D量測系統之彈性化 校正方法，在本發明中，該主動式3D量測系統丨之硬體架構 係為習知，如圖1所揭示，包括一平台u、一光源產生器12、 一透鏡13、一處理單元14、一投射裝置（包括一數位式反射 97302.doc 1249019 裝置15及一投射鏡頭16)及一影像擷取裝置（例如一 CCD攝 影機17)。 參考圖2，顯示本發明所建立之該主動式3D量測系統1 之數學模型。該數學模型包括三個座標系統··固定於該 CCD攝影機1 7光學中心（9C之CCD座標系統（Camera Coordinate System，CCS)7、固定於該投射裝置光學中心 之投射座標系統（Projector Coordinate System，PCS)8 以 及固定於該平台11(定義為一參考平面（reference pi ane)4) 上之世界座標系統（World Coordinate System，WCS)9。假 如有一點= . /P1位於該數位式微鏡片裝置152圖案 (定義為一投射平面（projection plane)5)上之某一像素上， 一像素射線（Pixel ray) ，由他經Op射向待測工件表面一 點A 4]，該射線0在w尸產生散射並由該CCD攝影 機17擷取其光場強度後，成像於該CCD攝影機17之感測晶 片（定義為一影像平面（image plane)6)上之％ % /c]， 與形成一匹配對（correspondence pair)，此匹配對可 用來求解之空間座標。因此，整個系統除了投射裝置 可以投射光學圖案至待測工件表面外，亦可視為一立體視 覺系統（Stereo Vision System)，另外，由於光具可逆性， 像素射線由ββ/射向在方向上是可逆的，也就是可 以將％視為7之成像點。因此，CCD攝影機17之模型亦 可以適用於投射裝置模型，其表示式如下所示， 1249019 P = pwRwp+：t ⑽ Qu: J”p pDZ m Qd(l + KprP2)=p Qu (lc)1249019 IX. DESCRIPTION OF THE INVENTION: TECHNICAL FIELD The present invention relates to an elastic correction method applied to an active 3D measurement system, and more particularly to a known size having a plurality of circular holes Correct the plate to obtain more accurate system calibration parameters. [Prior Art] Due to the rapid development of optoelectronic technology, micro-electromechanical technology and precision mechanical technology, the application level is becoming more and more extensive. This trend makes the surface measurement range of various components smaller, more precise and faster. The direction of development. In the measurement and detection applications of the surface of fine components, the current measurement methods are often due to excessive human judgment procedures and insufficient precision of the equipment, so that the measured results cannot fully display the true surface state; the traditional contact measurement results Excessive measurement time and damage to the object to be measured, making it a space-limited threshold, so the optical measurement method came into being. In the past, the optical buckle measurement method mainly adopted the measurement structure of (4) type, and two or more cameras were installed above the object to be tested, and a light spot or a straight line was projected by the laser light to form a surface of the object to be tested. Marking area, and forming a matching point group on the images taken by each camera (clearing, extinguishing_P_t cl(Hld), through calculation, can find the matching points between the cameras, and the thinness is calculated by the triangulation principle Measuring the height of the object' Then, using the moving object or the working platform to measure the entire shape of the object to be tested, the disadvantage of this method is that it requires a long measurement time, and because the moving mechanism is required during the measurement, The accuracy is reduced. The active measurement method mainly replaces the source camera and the laser light 97302.doc 1249019 with a projection device, and uses the projection device to project a planar structural optical pattern, and the optical pattern can be timed. Change, increase the variability of the optical pattern and facilitate the subsequent matching operation. Refer to Figure 1 for a schematic diagram of the active 3D measurement system. The active active 3 D The measuring system 1 includes a platform, a light source generator 12, a lens 13, a processing unit 14, a projection device (including a digital reflecting device 15 and a projection lens 16), and an image capturing device ( For example, a CCD camera 17) is used for placing a component to be tested. The light source generator is configured to generate a light beam, and the lens 13 is incident on the digital reflecting device 15. The processing unit 14 (for example, a computer), which produces a graphic system that can be flexibly edited and modulated. The digital reflecting device 15 includes a digital light processor (DLP) 151 and a digital microlens device (Digital). MiCr〇-mirrors Device (DMD) 152, the digital microlens device 152 is composed of 1024 x 768 microlenses, and the processing unit 14 controls the rotation of each microlens through the digital light source processor 15 1 to determine the micro The ON of the lens (reflecting light to the component to be tested) or 〇FF (reflecting light to the background), and by controlling the length of the microlens 0N/0FF, the intensity of the light field of the projection point can be determined, so The control of the sheet can reflect the light from the light source generator 12 into an optical pattern corresponding to the pattern. The projection lens 16 reduces the optical pattern to a desired measurement range and then projects onto the surface of the device under test. The CCD camera 17 is configured to capture an image from the component to be tested on the platform 11 and transmit the captured image to the processing unit 14. The processing unit 14 compares the captured image with the image. Change 'to get the surface state of the component under test. 97302.doc 1249019 However, the most important task is to measure the calibration of the system, the accuracy of the calibration procedure and the calibration quality, directly affecting the performance and accuracy of the measurement system: because The correction of the optical system is mainly to obtain the internal parameters of the system (optical center, effective focal length, distortion rate, image center), external parameters (rotation matrix and translation vector between the projection unit, reference plane and CCD coordinate system) to provide Geometric operations for subsequent measurements. The conventional correction method simplifies the actual three-dimensional system into a two-dimensional mode, that is, it must be assumed that the connection of the projection device to the optical center of the CCD camera 17 must be parallel to the reference plane on the center, or it is assumed that The projection device 15 is coplanar with the optical center and the measuring point of the CCD camera 17, and then uses: the machine uses a one-dimensional two-dimensional geometrical tool to analyze (4) the number of turns and (4) m is three-dimensionally measured as a two-dimensional system. The derivation of the formula is simplified, but the relative calibration accuracy is low, only under the condition that the accuracy of the money poem is not high, and in the continuous system installation, it is not easy to reach the so-called two-dimensional and coplanar model. The other two: the conventional correction method directly assumes that the projection device 15 and the optical center of the scene: 17 coincide with the geometric center, but in fact the center of the learning and the geometric center do not coincide. Therefore, it is necessary to provide an innovative and progressive correction method to solve the above problems. SUMMARY OF THE INVENTION The main object of the present invention is to provide an elasticization correction method, which has no specific limitation on the position of the optical center of the radiology center and the image capturing device =::: has the characteristics of the original global measurement. , more translatability in application 'can shorten the system setup time, and at the same time, because it is consistent with the actual physical frame 97302.doc 1249019, and does not move any mechanism, so it has a higher measurement. Precision. ^To achieve the above purpose 'The present invention provides a method for applying the active 3D measurement system to the elastic 彳b positive method, the active 3]3 measurement system, including the _platform, a projection, and the image capture The apparatus includes: 0) providing a positive flat plate having a plurality of circular holes, wherein the apertures and the hole distances of the circular holes are known; (b) placing the calibration plate on the platform; (0) using the image capturing device to capture the image of the circular holes; (4) using the image processing technology to determine the center coordinates of each of the circular holes, and using the most flat method to obtain the image capturing device and the platform Coordinate conversion relationship and effective focal length, distortion rate and optical center coordinates of the -Hang~image capture device; (e) removing the calibration plate; (f) projecting an optical pattern onto the platform by using the Shai projection device, wherein The photon pattern includes a plurality of dots whose center-to-center distance is the same as the center-to-center distance of the circular holes; (g) capturing an image of the dots by the image capturing device; and (h) vertically Move the platform to get more correction points' and then ask for the least square method The projection device and the image capturing device are coupled to the coordinates of the platform. [Embodiment] The present invention relates to an elastic correction method applied to an active 3D measurement system. In the present invention, the active The hardware architecture of the 3D measurement system is conventional, as disclosed in FIG. 1, including a platform u, a light source generator 12, a lens 13, a processing unit 14, and a projection device (including a digital reflection 97302). .doc 1249019 Apparatus 15 and a projection lens 16) and an image capture device (e.g., a CCD camera 17). Referring to Figure 2, a mathematical model of the active 3D measurement system 1 established by the present invention is shown. The coordinate system (PCS) 8 is fixed to the optical center of the CCD camera (9C, the Camera Coordinate System (CCS) 7 and the Projector Coordinate System (PCS) 8 fixed to the optical center of the projection device. And a World Coordinate System (WCS) 9 fixed on the platform 11 (defined as a reference pi ane 4). If there is a point = . /P1 is located A pixel of a digital microlens device 152 (defined as a projection plane 5), a pixel ray (Pixel ray), which is directed by Op to a surface of the workpiece to be tested A 4], the ray 0 after the corpse is scattered and the light field intensity is captured by the CCD camera 17, and imaged on the sensing wafer (defined as an image plane 6) of the CCD camera 17 as % % /c], Form a matching pair that matches the space coordinates that can be solved. Therefore, the entire system can be regarded as a stereoscopic system (Stereo Vision System) in addition to the projection device that can project an optical pattern to the surface of the workpiece to be tested. In addition, due to the reversibility of the optical device, the pixel ray is directed by ββ/in the direction. Reversible, that is, you can think of % as the imaging point of 7. Therefore, the model of the CCD camera 17 can also be applied to the projection device model, and its expression is as follows, 1249019 P = pwRwp+: t (10) Qu: J"p pDZ m Qd(l + KprP2) = p Qu (lc)
其中，y及7為外部參數，其定義PCS與WCS之轉換關 係，/户是有效焦距（Effect Focal Length，EFL)，心 P ]為 π相對於PCS之座標，/Cp為投射鏡頭失真率，Q 為數位式微鏡片裝置152上之投射平面中心。 同樣地，該CCD攝影機17之模型亦可以用下列方程式表 示 r：P<RwP+iT (2α) (2b) ^ Qdi^ + Kcrc ) ^ Qu (2c) ruP fp-rlxPQpxhp rnp fp^rnpQpxhp rup fp^r^Qpxhp r2：fp-r,；Qpyhp r22pfp-r32pQpyhp r2/fp-rnpQpyhpWhere y and 7 are external parameters, which define the conversion relationship between PCS and WCS, / household is the effective focal length (Effect Focal Length, EFL), heart P] is the coordinate of π relative to PCS, /Cp is the projection lens distortion rate, Q is the center of the projection plane on the digital microlens device 152. Similarly, the model of the CCD camera 17 can also be expressed by the following equation: r < RwP + iT (2α) (2b) ^ Qdi^ + Kcrc ) ^ Qu (2c) ruP fp-rlxPQpxhp rnp fp^rnpQpxhp rup fp^ R^Qpxhp r2:fp-r,;Qpyhp r22pfp-r32pQpyhp r2/fp-rnpQpyhp
pv P *yy P L m ^ }'QPyhP ⑶ 其中中每一元t為7元 素’同理，從（2a)、（2b)及（2c)中，可推導出％與化之關 係如下， ruCfc -^QcA rnfc -r^QcA rnfc-r^QA r2lcfc - r2XcQcyhc r22cfc - r32cQcyhc r23c fc - r,；Qcyhc ^ k； P \\y lz Qaxhc -txcfc P L wz _ }z QCyhc (4) 利用pseudo-inverse法聯立求解由（3)及（4)式，可^〜 於在矩陣方程式（3)及（4)式中，包含量測 里刳糸統之外部表數、 及投射裝置與CCD攝影機17之内部表赵 ^ " 1多數，所以如何求得較 高精度的系統參數更顯得重要。 97302.doc -10- 1249019 在本發明中，對於量測系統參數的校正程序，共分成該 CCD攝影機17之參數校正及該投射裝置（包括該數位式反 射裝置15及該投射鏡頭16)之參數校正二步驟。 第一步驟是該CCD攝影機17之參數校正，此一校正的主 要目的是為了獲得-群校正點對於wcs及其在ccs的對應 y/像的關係、首先’將一加工完成具有i 〇χ】〇個標準圓孔η 之高精度校正平板20(如圖3所示）置於該平台丨丨上，該平台 11定義為該參考平面4。該校正平板加上圓㈣之孔徑及孔 位距離是以高精度之機械加卫程序完成，其係為已知，並 在校正時提供3〇之校正點，該校正平板2g被該平⑽分別 疋位在不同同度，並由該€(：：1)攝影機17擷取影像後計算出 每一張影像中孔位的幾何中心（即該等圓孔2丨之圓心座 標）’這些孔位中心可視為校正點的影像座標，並以最小平 方法求得該CCD攝影機17與該平台此座標轉換關係以及 該CCD攝影機17之有效焦距、失真率及光學中心座標。該 CCD攝影機17之校正流程，分成線性最佳化⑽w optimization)及非線性最佳化（n〇n_nnear 〇ptimizati〇n)兩部 分，線性最佳化過程，由影像中心（Cx，Cy)、校正點及其對 應影像求得Γ、/、r，我們使用校正參數與—評價函數 /(cost function)，將校正點γ映射回影像平面得％,，並中/ 定義為，非線性最佳化主要的目的是透過、一已 完善建立的數值分析方法將j最小化，並求得最佳的影像中 心（Cx，Cy) 〇 第二步驟是該投射裝置（包括該數位式反射裝置15及該 97302.doc -11 - 1249019 投射鏡頭16)的校正，其最佳化之過程非常相似前述之 CCD攝影機校正方法，唯一不同的地方是其校正點的產 生。首先，移開該校正平板20(亦即在此一步驟並不需要 該校正平板20)。接著，利用該投射裝置投射一光學圖案 30(如圖4所示）至該平台11，其中該光學圖案30包括複數 個圓點3 1，該等圓點3 1之圓心距與該校正平板20上之該 等圓孔2 1之圓心距相同。接著，利用該已經完成校正之 CCD攝影機17擷取該等圓點31之影像，並藉由校正過的 CCD攝影機17模式計算出X-Y座標。接著，於垂直方向移 動該平台11，使該平台11位於不同高度，例如：Z=2 mm， Z=0 mm，Z=_2 mm，Z=-3 mm等，以產生一群非共面的校 正點，如圖5所示。再以最小平方法求得該投射裝置、該 CCD攝影機17與該平台11之座標轉換關係。在本發明之一 實施例中，該CCD攝影機17與投射裝置的校正結果如下表 —戶斤示0 表一、CCD攝影機與投射裝置的校正結果 投射裝置 CCD攝影機 旋轉矩陣 Γ-0.99696 -0.01998 0.075301 0.01537 -0.97544 -0.03913 [0.07603 -0.03877 0.99635J Γ-0.88017 -0.00834 -0.474591 0.00653 -0.97862 -0.00880 [-0.47450 -0.01107 0.88019J 平移向量Γ Γ-0.00237 Ί 0.00362 0.05947」 「0.00592 1 0.00183 -0.13754」 有效焦距/(m) 0.09567 0.19280 失真率κ 3.76e-8 1.20e-8 影像中心 (pixel) (28.93, 479.42) (831.0,420.2) 應用本發明校正方法之量測方法為，先分別計算出匹配 97302.doc -12- J249〇j9 對Ά與cQd，以及系統的棱夂 出 ；仅正參數，並代入式（4)尹，求解 旦"工件表面輪廓之三維資訊K之）。由於此主動式 :量測系統!可以一次投射整面的光學圖案，不但可以依待 =件的幾何形狀，投射不同的光學圖案進行量測作業， 同了可以在不需要移動機構的情形下，對待測工件進行全 域式的量冑’所以量測系統之結構較簡潔，而且且有較 的量測精度。 〃 同 准上述實施例僅為說明本發明之原理及其功效，而非用 以限制本發明。因此，習於此技狀人士可衫違背本發 明之精神對上述實施例進行修改及變化。本發明之權利範 圍應如後述之申請專利範圍所列。 【圖式簡單說明】 圖1顯示習用主動式3D量測系統之示意圖； 圖2顯示本發明所建立之該主動式3〇量測系統之數學模 圖3顯示應用在本發明之高精度校正平板； 圖4顯示本發明在該投射裝置校正步驟中所投射至該平 台之光學圖案；及 圖5顯示本發明中投射裝置之參數校正示意圖。 【主要元件符號說明】 1 習用主動式3D量測系統 4 參考平面 5 投射平面 6 影像平面 97302.doc -13- CCD座標系統 投射座標系統 世界座標系統 平台 光源產生器 透鏡 處理單元 數位式反射裝置 投射鏡頭 CCD攝影機 校正平板 圓孔 光學圖案 圓點 數位式光源處理器 數位式微鏡片裝置 -14-Pv P *yy PL m ^ }'QPyhP (3) where each element t is 7 elements' is the same. From (2a), (2b) and (2c), the relationship between % and chemistry can be derived as follows, ruCfc -^ QcA rnfc -r^QcA rnfc-r^QA r2lcfc - r2XcQcyhc r22cfc - r32cQcyhc r23c fc - r,;Qcyhc ^ k; P \\y lz Qaxhc -txcfc PL wz _ }z QCyhc (4) Using the pseudo-inverse method The solution is solved by equations (3) and (4), which can be used in the matrix equations (3) and (4), including the external number of the measurement system, and the inside of the projection device and the CCD camera 17. Table Zhao ^ " 1 majority, so how to find higher precision system parameters is more important. 97302.doc -10- 1249019 In the present invention, the calibration procedure for the measurement system parameters is divided into the parameter correction of the CCD camera 17 and the parameters of the projection device (including the digital reflection device 15 and the projection lens 16). Correct the two steps. The first step is the parameter correction of the CCD camera 17. The main purpose of this correction is to obtain the relationship between the group correction point for wcs and its corresponding y/image in ccs, and firstly 'have a processing completion i 〇χ 】 A high-precision correction plate 20 (shown in FIG. 3) of a standard circular hole η is placed on the platform raft, which is defined as the reference plane 4. The correction plate plus the circle (4) aperture and hole distance is completed by a high-precision mechanical reinforcement program, which is known and provides a correction point of 3 在 during calibration, and the correction plate 2g is respectively separated by the flat (10) The pupils are at different degrees, and the image is taken by the €(::1) camera 17 to calculate the geometric center of the hole positions in each image (ie, the center coordinates of the circular holes 2丨) The center can be regarded as the image coordinate of the calibration point, and the conversion relationship between the CCD camera 17 and the platform and the effective focal length, distortion rate and optical center coordinates of the CCD camera 17 can be obtained by the least square method. The calibration process of the CCD camera 17 is divided into two parts: linear optimization (10) w optimization and nonlinear optimization (n〇n_nnear 〇ptimizati〇n), linear optimization process, by image center (Cx, Cy), correction The point and its corresponding image are obtained by Γ, /, r. We use the correction parameter and the evaluation function/(cost function) to map the correction point γ back to the image plane, and define it as nonlinear optimization. The main purpose is to minimize j and obtain the best image center (Cx, Cy) through a well-established numerical analysis method. The second step is the projection device (including the digital reflection device 15 and the 97302.doc -11 - 1249019 The correction of the projection lens 16) is optimized to be similar to the CCD camera calibration method described above, the only difference being the generation of the correction point. First, the calibration plate 20 is removed (i.e., the calibration plate 20 is not required at this step). Then, an optical pattern 30 (shown in FIG. 4 ) is projected by the projection device to the platform 11 , wherein the optical pattern 30 includes a plurality of dots 3 1 , a center distance of the dots 3 1 and the calibration plate 20 . The center distances of the circular holes 2 1 are the same. Next, the image of the dots 31 is captured by the CCD camera 17 which has been corrected, and the X-Y coordinates are calculated by the corrected CCD camera 17 mode. Next, the platform 11 is moved in a vertical direction so that the platform 11 is at different heights, for example: Z=2 mm, Z=0 mm, Z=_2 mm, Z=-3 mm, etc., to generate a group of non-coplanar corrections. Point, as shown in Figure 5. Then, the projection device, the coordinates conversion relationship between the CCD camera 17 and the platform 11 are obtained by the least square method. In an embodiment of the present invention, the calibration result of the CCD camera 17 and the projection device is as follows: Table 1 shows the calibration result of the CCD camera and the projection device CCD camera rotation matrix Γ-0.99696 -0.01998 0.075301 0.01537 -0.97544 -0.03913 [0.07603 -0.03877 0.99635J Γ-0.88017 -0.00834 -0.474591 0.00653 -0.97862 -0.00880 [-0.47450 -0.01107 0.88019J Translation vector Γ 0.00-0.00237 Ί 0.00362 0.05947" "0.00592 1 0.00183 -0.13754" Effective focal length / ( m) 0.09567 0.19280 Distortion rate κ 3.76e-8 1.20e-8 Image center (pixel) (28.93, 479.42) (831.0, 420.2) The measurement method using the calibration method of the present invention is to first calculate the matching 97302.doc - 12- J249〇j9 Confrontation and cQd, and the system's edge out; only positive parameters, and substituted into (4) Yin, solve the "three-dimensional information K of the surface contour of the workpiece"). Because of this active type: measurement system! It can project the entire optical pattern at a time, not only can project different optical patterns according to the geometry of the part to be measured, but also without the need of a moving mechanism. The whole process is measured by the workpiece to be measured. Therefore, the structure of the measurement system is simple and has a relatively accurate measurement accuracy. The above embodiments are merely illustrative of the principles of the invention and its advantages, and are not intended to limit the invention. Therefore, those skilled in the art can modify and change the above embodiments in light of the spirit of the present invention. The scope of the invention should be as set forth in the appended claims. BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a schematic view showing a conventional active 3D measuring system; FIG. 2 is a schematic view showing the mathematical model of the active 3〇 measuring system established by the present invention. 4 shows an optical pattern projected to the platform in the correction step of the projection device of the present invention; and FIG. 5 shows a schematic diagram of parameter correction of the projection device in the present invention. [Main component symbol description] 1 Conventional active 3D measurement system 4 Reference plane 5 Projection plane 6 Image plane 97302.doc -13- CCD coordinate system Projection coordinate system World coordinate system Platform Light source generator Lens processing unit Digital reflection device projection Lens CCD camera calibration flat hole optical pattern dot digital light source processor digital micro lens device-14-