1352189 . 九、發明說明： 【發明所屬之技術領域】 本發明係有關一種掃描式三角測量系統及方法，尤指一種 包括一掃瞄模組、一第一光檢測器、一第二光檢測器、一分光 裝置及一訊號處理單元，俾能藉由該第一光檢測器及該第二光 檢測器之檢測信號進行運算處理及比對求得該待測物的形貌 尺寸參數者。 • 【先前技術】 由於目前國内的經濟正面臨轉型期，各種高新科技產業崛 • 起，使得許多關鍵零組件需要更高要求的檢測技術，使廠商於 - 近年致力於精密量測系統的發展。尤其現今廠商多致力於單件 -計價方式，故需做到百分之百的全檢，不論是1C、LCD、BGA、 模具、精密零組件…等，產品的外觀尺寸、形貌或表面瑕疵… 等，均為重要檢測参數。 • 再者，尺寸或位移量測是幾何量測或是物理量測的重要目 標，而傳統的位移和表面形貌的測量方法多為接觸式的探針檢 測，其確實存在有如下所述諸多的缺點： 1. 接觸使工件變形，特別是軟質或易損傷的表面。 2. 探針容易磨損。 3. 存在接觸式的遽波效應。 據此，近代發展出了各種非接觸式的測量技術，其中最為 成功的技術有干涉儀量測系統與雷射式三角探頭等技術。其中 1352189 .二角探頭技術具有抗環境干擾能力強，精度高及使用方便等優 點’廣泛為業界使用。 雷射式三角探頭是一種單點式的高精度探頭，依對象及測 量範圍區分，精度從lmm(散射式）到2nm(反射式）不等。若只針 對散射式的三角探頭而言，是一種成像式的檢測系統，技術的 七展己十分成熟’較著名的廠商有Microtrak，Keyence， Hamamatsu，Panasonic...等，最高精度可達〇 1/zm。此系統 φ中多採用電荷耦合式感測器（CCD)或位置靈敏感測器（psd)〇 — $又產απ具有微米等級的精度及5〇ΚΗζ抽樣率的能力。特別是原 •子力顯微鏡與近場光學顯微鏡都採用位置靈敏感測器為檢測元 --件，在小位移量測時具有次奈米的精度，可測量原子的表面形 狀。 請參看第一圖所示，係習用之非接觸式的三角量測之結 構，其係以一雷射單元（6〇)(Light Source)經投射之透鏡（61) 鲁後投射在待測物dXObject)上，相對於成像透鏡（62)(Image Lens)而5，此一光點為物光點，根據Lambert散射定理形成一 個能量分佈的光錐，經透鏡成像於一光檢測器（63)psD上。若 待測物（1)位於光軸與雷射光軸的交點上，則成像於該光檢測器 (63)PSD的幾何中心，若待測物⑴移動-段距離d到位置〇〇 時，則光斑將在光檢測器（63)PSD上位移一段距離5，這是— 個量測值疋一個電子心號，由第一圖所示的幾何關係中得知： ^ __ di sin^ + ^cos^ (1) 其中，，為準直透鏡（30)CL的焦距。 值由於控制馬達轉速的不穩定、準直透鏡（64)6Z因幾何轉 換產生的非線性與旋轉反射鏡（65)的離焦誤差等因素的影響， 將造成掃描速度的不均勻。在此利用數學式進行分析與表達。 令準直透鏡（64)的焦距為/，控制馬達的轉速為ω，而射入第 —個準直透鏡（CL) (64)的入射角為0。由於反射鏡的改變角度 使·入射光束與反射光束的夾角關係為 (4) ^ = 2ωί 則掃描速度可表示如下： (5) 2ft%^ (2(〇t)~+tangent)—(η-ΐ)Δΐ dt dt 實際的掃描速度「(t)是的非線性的。在此將（5)式的各項分析 如下： 2/i»sec2(2 如） (6) 為透鏡幾何轉換的非線性誤差’一般的解決方案多採用/0透 鏡，使y與掃描角度呈現尸/θ的關係。請參看第五圖，高度 y，轉角（9與時間的關係。 2ft sec (2ωή (7) dco dt 由於控制馬達轉速的不穩定，一般的解決方案多採用控制馬速 加裝編碼器及回饋控制，或加大控制馬達的負載能力。 (8) tan( 2 ω t) dt 1352189 這是多面鏡的離焦誤差等因素的影響，一般的解決方案多採 用銳角入射及增大焦距。 不論是那一項的解決方案都是昂貴及複雜的，因此，本發 明設計一套簡易的雙感測器的測量方案來解決（5)式的問題。 請參看第四圖（a)所示’其中，該光檢測器（63)PSD是利 用P-I-N接合橫方向上所產生光電效應，以檢出入射光點照射 位置’該光檢測器（63)(PSD)的結構及等效電路如第四圖（b) 所示。其在高阻抗的基板表面上形成P層，在底層形成N層。A， β是P層上面的兩個端電極。位置檢測元件表面被光束照射後 則會產生電子及電洞對；如果ρ型是均勻性良好的電阻層，則 這些電子及電洞對到達電極接合處的數量與光點入射位置和兩 電極距離位置有關。因此取得相對應的電流值就可估算入射光 點的位置。 再請參看第四圖（b)所示，則是光檢測器（63)的等效電路， 再請參看第四圖（a)所示，其光點照射位置為q，ρ侧電極以 A，B兩者表示之，N側電極以逆偏接地。由於光點在光檢測器 (63)上的不同位置會產生不同的信號大小，經過後級的信號處 理可精準的測量出光點的坐標位置。再以一維位置光檢測器 (ID-PSD)為例，當光點位於光檢測器（63)的幾何中心時，由於 兩邊對稱，因此兩個信號的大小相等。當光點偏離中心點時， 則光點到兩個電極的幾何長度不同，因此等效電阻的大小不 同。而光束的照射點相當於一個光電流源，根據分流定律而分 1352189 . 配能量。 若以電極A與電極B的中點為中心坐標，則到光點的距離 為J。電極A與電極B間的距離為Z，其間的阻抗為必。假使外 部負載電阻為必。以等效電路模型如第四圖（b)所示，尚有等 效電阻β、β外的存在。若有入射的成像光束，則光電流信號 所產生的就是一個電流源（51,以人/幻表示）。在不影響正確結 果下進行分析，可以標準電阻的模型近似，使成為與尤有關的 |相依電阻與相依電容。The invention relates to a scanning triangulation system and method, in particular to a scanning module, a first photodetector, a second photodetector, A light splitting device and a signal processing unit are configured to perform processing and comparison on the detection signals of the first photodetector and the second photodetector to obtain a shape size parameter of the object to be tested. • [Previous Technology] As the domestic economy is facing a transition period, the rise of various high-tech industries has made many key components require more demanding detection technologies, enabling manufacturers to work on the development of precision measurement systems in recent years. . In particular, today's manufacturers are committed to a single-piece pricing method, so 100% full inspection is required, whether it is 1C, LCD, BGA, mold, precision components, etc., the appearance size, shape or surface flaw of the product, etc. All are important test parameters. • Furthermore, size or displacement measurement is an important target for geometric measurement or physical measurement, while traditional displacement and surface topography measurement methods are mostly contact probe detection, which does exist as described below. Disadvantages: 1. Contact deforms the workpiece, especially soft or vulnerable surfaces. 2. The probe is prone to wear. 3. There is a contact chopping effect. Accordingly, various non-contact measurement techniques have been developed in recent times, among which the most successful technologies include interferometer measurement systems and laser-type triangular probes. Among them, 1352189. The two-corner probe technology has the advantages of strong resistance to environmental interference, high precision and ease of use, and is widely used in the industry. The laser-type triangular probe is a single-point, high-precision probe that is distinguished by object and measurement range, with accuracy ranging from lmm (scattering) to 2nm (reflective). If it is only an imaging type detection system for the scattering type triangular probe, the seven exhibitions of the technology are very mature. The more famous manufacturers include Microtrak, Keyence, Hamamatsu, Panasonic, etc., and the highest precision is up to 〇1. /zm. This system φ uses a charge-coupled sensor (CCD) or a position sensitive sensor (psd) 〇 — which produces απ with micron-level accuracy and a 5-inch sampling rate. In particular, both the original force microscope and the near-field optical microscope use a position sensitive sensor as the detection element, which has sub-nano precision in small displacement measurement and can measure the surface shape of the atom. Please refer to the first figure, which is a non-contact triangulation structure that is projected by a projecting lens (61) with a laser unit (6 〇) (Light Source) and projected onto the object to be tested. dXObject), relative to the imaging lens (62) (Image Lens) 5, this spot is the object spot, forming a light cone of energy distribution according to the Lambert scattering theorem, and imaging through a lens to a photodetector (63) On psD. If the object to be tested (1) is located at the intersection of the optical axis and the optical axis of the laser, it is imaged at the geometric center of the PSD of the photodetector (63). If the object to be tested (1) moves by a distance d to the position 〇〇, then The spot will be displaced by a distance of 5 on the photodetector (63) PSD, which is a measured value 疋 an electronic heart number, known from the geometric relationship shown in the first figure: ^ __ di sin^ + ^cos ^ (1) where, is the focal length of the collimating lens (30) CL. The value is due to the instability of the control motor speed, the nonlinearity of the collimating lens (64) 6Z due to geometric transformation, and the defocusing error of the rotating mirror (65), which causes uneven scanning speed. Here, mathematical analysis is used for analysis and expression. The focal length of the collimating lens (64) is /, the rotational speed of the control motor is ω, and the incident angle of the first collimating lens (CL) (64) is zero. Since the angle of change of the mirror makes the angle between the incident beam and the reflected beam (4) ^ = 2ωί, the scanning speed can be expressed as follows: (5) 2ft%^ (2(〇t)~+tangent)—(η- ΐ) Δΐ dt dt The actual scanning speed "(t) is nonlinear. Here, the equations of (5) are analyzed as follows: 2/i»sec2(2) (6) is the non-transformation of the lens geometry Linear error 'general solution mostly uses /0 lens, so that y and scan angle show corpse / θ relationship. Please refer to the fifth picture, height y, corner (9 vs. time. 2ft sec (2ωή (7) dco Dt Due to the instability of the control motor speed, the general solution is to control the horse speed to install the encoder and feedback control, or to increase the load capacity of the control motor. (8) tan( 2 ω t) dt 1352189 This is a polygon mirror The influence of factors such as defocus error, the general solution mostly adopts acute angle incidence and increase the focal length. No matter which solution is expensive and complicated, the present invention designs a simple dual sensor. The measurement scheme to solve the problem of (5). Please refer to the figure 4 (a) The photodetector (63) PSD is a photoelectric effect generated by the PIN junction in the lateral direction to detect the incident spot illumination position. The structure and equivalent circuit of the photodetector (63) (PSD) are as shown in the fourth figure. (b) shows a P layer on the surface of a high-impedance substrate and an N layer on the bottom layer. A, β are the two end electrodes on the P layer. The surface of the position detecting element is irradiated with a beam to generate electrons. If the p-type is a uniformity of the resistive layer, the number of these electrons and holes to the junction of the electrodes is related to the position of the spot and the distance between the two electrodes. Therefore, the corresponding current value can be estimated. The position of the incident light spot. Please refer to the fourth figure (b), which is the equivalent circuit of the photodetector (63). Please refer to the fourth figure (a), the spot illumination position is q. The ρ side electrode is represented by both A and B, and the N side electrode is reversely grounded. Since the light spot produces different signal sizes at different positions on the photodetector (63), the signal processing through the subsequent stage can be accurately performed. Measuring the coordinate position of the spot. Then one-dimensional position photo inspection For example, when the spot is located at the geometric center of the photodetector (63), the two signals are equal in size because of the bilateral symmetry. When the spot is off the center point, the spot is two The geometric length of the electrodes is different, so the equivalent resistance is different. The irradiation point of the beam is equivalent to a photocurrent source, which is divided into 1352189 according to the law of shunt. If the center point of the electrode A and the electrode B is the center coordinate, Then, the distance to the spot is J. The distance between the electrode A and the electrode B is Z, and the impedance between them is necessary. If the external load resistance is necessary, the equivalent circuit model is as shown in the fourth figure (b). The existence of the equivalent resistance β, β. If there is an incident imaging beam, the photocurrent signal produces a current source (51, expressed in human/phantom). The analysis can be performed without affecting the correct result, and the model of the standard resistance can be approximated to make the dependent resistance and the dependent capacitance.
Ra = P 丄^— (9)Ra = P 丄^— (9)
. X ^ = p -^― (10) ' 在此令橫向的總電阻為R，則 ' R=Rs-hRb (11) 根據分流定律： (12) (13) (14)X ^ = p -^― (10) ' Here, the total resistance in the lateral direction is R, then ' R = Rs - hRb (11) according to the law of shunting: (12) (13) (14)
I〇(Rb+RL] (Ra + Rb + 2.RLj J〇{Ra+RL) (jR.a + Rb + ΊΕΐ^ 令橫向的總電流為I。，則 /〇 = /*-/- la (15) 1352189 將入射能量做正規化處理：I〇(Rb+RL) (Ra + Rb + 2.RLj J〇{Ra+RL) (jR.a + Rb + ΊΕΐ^ Let the total current in the lateral direction be I., then /〇= /*-/- la (15) 1352189 Normalize the incident energy:
( (/；〇+/?/)___(R b + R 〇!>- Ια) _ [ {R 〇 + Ji b + 2 R L') (βα + Rb + 2RL) (/6 + /e) 一 ~Γ~ (R 〇 + RL) R b + R L \( (/;〇+/?/)___(R b + R 〇!>- Ια) _ [ {R 〇+ Ji b + 2 R L') (βα + Rb + 2RL) (/6 + /e )~~Γ~ (R 〇+ RL) R b + RL \
\ Λ~〇 + R b + 2 R L R a + R b + 2 R L J CRa - R b} Τβ 〇 + R b + 2 R L )\ Λ~〇 + R b + 2 R L R a + R b + 2 R L J CRa - R b} Τβ 〇 + R b + 2 R L )
(β + 2 RL ) 2 pX ^Xfi + 2 Λ x( + 、2f> Xfb + Ja， (16) 由上式可以看出等號右式與位置坐標J成正比，而與其它 參數無關，且光檢測器（63)所檢測出來的位置為入射光能量的 重心位置。只要根據數學式，適當的設計信號處理電路即可測 出光點的重心的坐標位置。 然而在檢測表面形貌或瑕疵時，通常採用平台掃描的方 式，使得檢測速度被降低，因此習用結構因設計不佳所造成上 述諸多的缺失，確實有再改良的必要。本發明人等承蒙國科會 汁晝補助研究，經不斷地努力研發下，終於研發出一種掃描式 的一角量測系統及方法的本發明，不僅可以改善習用結構缺 失遊具有南精度及南測速等之特點。 【發明内容】 本發明之主要目的在於提供—種可崎決三角量測系統逐 點測量的速㈣題’故而具有快速檢驗物體形貌的能力，不僅 1352189 •可以改善系統的架構與量測方法，而且在不需要精密词服控制 馬=及透鏡的情況下進行測量，可以獲得在±1_範圍内對線性 "^準差小於〇 〇145的精度，因而具有高精度、高測速以及成 本相對低廉等特點的掃描式三角測量系統及方法。 本毛明達成上述二角量測系統之功效所採用的技術手段 二/、包括一掃瞄模組、一第一光檢測器、一第二光檢測器、 一t光裝置及-訊號處理單元’該掃猫模組用以產生一平行光 鲁軸掃描的掃描光束，並以該分光裝置將該掃描光束分成至少兩 束，其-該掃描光朿由該第-光檢測器接收，用以做為位置檢 ‘測與錢補償另-該掃描光㈣作為三肖量測的測試光 束，該測試光束經過-待測物後散射，再經由一成像透鏡而成 -像於該第二光檢測器上，且該訊號處理單元分別與該第一光檢 測器、該第二光檢測器及該掃描模組電性連接以控制該掃描模 “運=光檢測器及該第二光檢測器之檢測信號進 鲁行運算處理及比對，進而求得該待測物的形貌尺寸參數者。 本發明達成上述三角量測方法所採用的技術手段在於，其 係以掃瞄模組產生-平行光軸掃描的掃描光束，再以分光裝置 將該掃描光束分成二束，使其-該掃描光束由一第一光檢測器 接收而為-參考光信號’再使另一該掃描光束作為三角量測的 測試光束而投射至-待測物後散射，使由該待測物散射後的該 測試光束經由-成像透鏡成像於-第二光檢測器上而為一檢測 光信號，並將該參考光信號及該檢測光信號進行運算處理及比 12(β + 2 RL ) 2 pX ^Xfi + 2 Λ x( + , 2f> Xfb + Ja, (16) It can be seen from the above equation that the equal sign right is proportional to the position coordinate J, and is independent of other parameters, and The position detected by the photodetector (63) is the position of the center of gravity of the incident light energy. As long as the mathematical design is appropriate, the signal processing circuit can appropriately measure the coordinate position of the center of gravity of the spot. However, in detecting the surface topography or flaws At the time, the scanning method of the platform is usually used, so that the detection speed is lowered. Therefore, the conventional structure is often lacked due to poor design, and there is indeed a need for further improvement. The present inventors have been subjected to the research of the National Science Council's juice subsidy. With continuous efforts and research and development, the present invention has finally developed a scanning one-angle measuring system and method, which not only can improve the characteristics of the conventional structure missing tour, but also has the characteristics of south precision and south speed measurement, etc. The main purpose of the present invention is to Provides the speed (four) questions of the point-by-point measurement of the Razaki triangulation measurement system. Therefore, it has the ability to quickly check the shape of the object, not only 1352189. • It can improve the architecture and quantity of the system. The method, and the measurement without the need of precision word control horse = and lens, can obtain the precision of linearity "^ in the range of ±1_ is less than 〇〇145, thus having high precision, high speed and Scanning triangulation system and method with relatively low cost, etc. The technical means used by the present invention to achieve the above-mentioned two-angle measurement system is to include a scanning module, a first photodetector, and a second a photodetector, a t-light device, and a signal processing unit for generating a scanning beam of a parallel optical axis scan, and dividing the scanning beam into at least two beams by the beam splitting device, the scan The optical 接收 is received by the first photodetector for use as a position detection and a compensation for the money. The scanning light (4) is used as a three-dimensional measurement test beam, and the test beam is scattered by the object to be tested, and then An imaging lens is formed on the second photodetector, and the signal processing unit is electrically connected to the first photodetector, the second photodetector, and the scan module to control the scan mode. Operation = light detection The detection signal of the second photodetector and the second photodetector are processed and compared, and the shape size parameter of the object to be tested is obtained. The technical means adopted by the present invention to achieve the above triangular measurement method is that The scanning beam is generated by the scanning module to scan the parallel optical axis, and then the scanning beam is split into two beams by a beam splitting device, so that the scanning beam is received by a first photodetector as a reference optical signal. The other of the scanned light beams is projected as a triangularly-measured test beam and projected onto the object to be tested for backscattering, so that the test beam scattered by the object to be tested is imaged onto the second photodetector via the imaging lens. Detecting an optical signal, and performing arithmetic processing on the reference optical signal and the detected optical signal and ratio 12