1352188 · 九、發明說明： 【發明所屬之技術領域】 本發明係有關於一種雙光探針檢測透明物體之厚度與 間隙的裝置及方法，尤指一種採用多元檢測器鎖定共軛焦 面的位置，配合利用像散法、刀緣法及臨界角法或其他可 實訑之方法，而可達到精準的定位及測量出透明物體其厚 度與間隙之目的者。 • 【先前技術】 製造液晶顯示器產業的晶圓廠商為了因應TFT_ LCD面 板輕薄平大的要求，物在平面起伏、表面粗度、密度、比 重等品質有一定的標準，因此對於基板的耐熱度、化學、 機械、電氣等性質均有參數上的限定，使得物之平坦度檢 測佔整個液晶螢幕顯示效能其中的重要一環。 有關玻璃厚度檢測的相關研究相當的多，大致上的擺 ® "又方式主要可以分為穿透式與反射式兩種；就量測行為主 要可以刀為接觸式與非接觸式兩種，隨著量测技術的進 步，目前非接觸式的光學量測技術，已經慢慢成為主流。 而目前非接觸式的光學量測技術有利用雷射超音波來 決疋玻璃厚度的方法，這種方法也是屬於非接觸式量测， 罝測物體除了物外，玻璃瓶也屬量测範圍，由於玻璃厚度 夏測在生產製造時之檢測提供非常重要之資訊，將這些資 訊加以即時監控以得到高品質之線上產品。後來，為了改 1352188 · 善玻璃檢驗時之速度及精度，有一種新型非接觸式獲取玻 璃表面粗糙度及厚度之方法被研究，其量測原理主要是利 用Sne 11疋律去f測，藉由描述光線行進於不同介質間的 關係而得玻璃表面粗之糙度及厚度。上述諸項之既有技 術，雖共同可以達到一定的量測精度，但於實際使用上卻 仍存在著諸多的缺點。 【發明内容】 # 本發月之第一目的，在於提供一種減少測量的時間， 降低因環境漂移而產生的誤差，提高測量精度的檢測透明 物體之厚度與間隙的裝置。 本發明之第二目的，在於提供一種採用兩顆光學讀寫 頭做為里測之用，目的在比對與校正等效厚度與實際厚度 的關係’以達透明物體其厚度與間隙的檢測方法。 “本發明之第二目的’在於提供—種兩個共_焦探頭於 修平口移動後，在待測物下表面所得到兩個s曲線之中心點 距離即传等效厚度，因而達到透明物體其厚度檢測的目 的。 達成上述發明目的之雙共軛焦雷射檢測透明物體其厚 度與間隙的裝置，其包括： 二光學讀取裝置’分別提供產生光束，其—顆作為檢 測源光學讀取奘罢 口 裝置’另一顆作為參考源光學讀取裝置； 聚“、、物鏡’提供該光束聚焦與改變該光束距離；及 1352188 · • -基準面，提供該光學讀取裝置的聚焦點掃描； 測量前，將檢測源光學讀取裝置與參考源光學讀取襄 置先做-次無待測物的掃描，可測得檢測源光學讀取農置 的聚焦點掃描到基準距離⑻），而參考源光學讀取農置的 聚焦點掃到基準面的距離（R1)，⑽減去⑼ 差△。再將透明之待測物放置在檢測源光學讀取裝置：: 準面上，再測量-次後，檢測源光學讀取裝置可得檢測二 φ讀取裝置的聚焦點掃到待測物上表面的距離（S2)與空氣間 隔（gap)，經由下列式子計算： S1 = R1-A t = S1 — S2 — gap 可獲得待測物實際量測之更準確的厚度（t)。 【實施方式】 I ·本發明之基本特徵及原理 •請參看圖五、六（a)及六（b)，本發明所提供之雙光探 針檢測透明物體其厚度與間隙的裝置與方法，主要包括有· 二光學讀取裝置（31)(310)，分別提供產生光束，其一 顆作為檢測源光學讀取裝置（31)，另一顆作為參考源光學 讀取裝置（310); 二聚焦物鏡（311)(312)，提供該光束聚焦而形成光探 針（313)(314)與改變該光束距離；及 一基準面（33)，供置放待測物’並提供該光學讀取裝 1352188· . 置的聚焦點掃描。 如圖至四（a)、四⑻所示，市面上之共輕焦雷射光 碟機中的光學讀取裝置（Gptieal piekupHead)即有共輕焦 檢測的架構；其共軛焦量測系統包括：雷射二極體（ιι)、 分光鏡（13)、準直透鏡（14)、聚焦物鏡（15)等。當光學讀 寫頭的雷射一極體（11)發射光束後，再經分光鏡（13) (Be⑽ Splitter)反射至準直透鏡（14)，再經由聚焦物鏡（15)聚焦 _到待測物（32)上，反射光束則循原路經由準直透鏡（14)、 分光鏡（13)後投射在位置檢測器（本發明以四象限檢測器 (16)為例）上，之後再利用像散法進行鎖焦的工作。 由於採用像散法鎖焦及四象限檢測器（丨6)(圖一中a、 B、C、D)，雖然玻璃的反射率很低，但只要能鎖定共軛焦 面的位置就可精準的定位及測量，因此厚度與間隙只要小 於聚焦物鏡（15 )的焦距都可精準的測量。其雷射二極體（11) 籲之波長λ =650 nm ’ NA=0.67，則依據繞射理論，可得聚焦 光點半徑（r)如下式： 尸=0.61 去= 0.592//m 故雷射二極體（11)，以物鏡（objective)進行聚焦，即 是利用光學成像時所產生的像散像差（aberration)來判斷 是否聚焦。像散是指垂直方向與水平方向的光學倍率不 同，因此各有一個焦點，如圖二所示，其子午光線的成像 13521.88 .位置（F〇與弧矢光線的成像位置⑹不^子午光線的成像 位f (FT)與弧矢光線的成像位置（Fs)不重合。因此對一物點 而言，成像便不再會是一個點，有可能如子午光線的成像 位置（Ft)為一水平線，弧矢光線的成像位置（Fs)為一垂直 線，在兩者之間可能是橢圓或圓。雷射二極體（U)即是利 用此方式經由聚焦物鏡〇5)，聚焦在四象限檢測器（⑻ 上，根據光點的像散現象，判讀聚焦的情形，如圖三。當 _垂直與水平焦距不同，則物體偏離透鏡前焦點或後焦點位 置時，成像光點是一個橢圓形，若物體在合焦位置時，成 像光點為一圓形，藉由四象限檢測器（16)及相關的訊號處 理，由聚焦誤差訊號（Focus Error Signal): FES=(A+C)-(B+D) (34) 可得一 S曲線，基本上s曲線的線性區為共軛焦系統可 量測的範圍，曲線的中心點對應於合焦的位置，其餘皆為 鲁離焦。 共軛焦雷射整體的量測系統，是透過光學讀取裝置所產 生之光束’經由聚焦物鏡（15)聚焦至待測物（32)上，藉由 移動平台掃描的方式進行厚度與間隙的量測，光路形式如 圖四（a)所示，其中待測物（32)為一透明材質的物體。當光 束經過聚焦物鏡（15)聚焦於焦點上形成一個光探針，而待 測物（32)往上移動時’當表面達到聚焦物鏡（15)焦點位置 時’此時為合焦的狀態’正好對應於共軛焦測量曲線的中 1352188 Γ點。當待測物（32)繼續往上移動待測物⑽等效厚产 (t )的距離’則焦點位於待測物（32)底面， ^ 到測量曲線的中心點位置，因4 4 7 U马九在介質中的折射作用， 面 層 只需移動待測物（32)等效厚度（t，）的距離，即鎖 面，而並非移動實際厚度⑴的距離才鎖焦到待測物⑽底 這種的現象，可經由光學原理的分析得知。 由圖四（b)及Snell定律可知 l*sin(9,. =nr sin^ (3-2) 且由上圖之幾何關係可得 w = ftm.Oi - = ttan^r (3-3) 卜产r ! t’Jnr2 - sin2 $ tan0r -Jl-sin2 Θ. (3-4) 再T t為将測物（32)的實際厚度，t，$密介質中 物(32)受折射影響之等效量測厚度’ &為光束之入射角： 心為光束之折射角’ nr為待測物（32)之折射率，從如） 式可得實際厚度⑴與密介質中的等效厚度（t，）、待測物 (32)之折射率⑴、人射光的角度有關，即t為t，、〜及 的函數。此推導公式的主要用意在暸解測量的厚度為— 等效厚度（t，），與實際的厚度有差。 Π ·本發明之實施例 咕配合參看圖五及圖十所示，本發明的量測方法在實施 時可有兩種方式，第-種為步階式的掃插，平移台每移動 1352188 • 一步抓取該點信號後，再繼續移動下一步，直到走完所要 移動的總仃程’測量時間較長。第二種為連續掃描的量測， 即平移台一次走完整個行程並不停頓，且在移動平台運動 的同時抓取信號’測量時間較短。 (i)步階式的掃描方式 本發明所採用之步階式量測方法，其系統的架構如圖 五所示首先藉由指令控制移動平台（18)上下做垂直移 鲁動，當光束經聚焦物鏡（311)聚焦至待測物（32)的上表面 後，反射回四象限檢測器，透過資料擷取卡（19)將訊號擷 取至電腦（20)處理，可得一個s曲線。移動平台（18)繼續 移動可在待測物（32)下表面得到另一個s曲線，藉由兩個s 曲線之中心點距離，可得一等效厚度（t，），達到檢測的目 的。 本發明的移動平台（18)可以一光柵的讀值作為移動距 _離的依據，最咼解析度為1 OOnm。在光學系統上，利用共扼 焦光學讀取裝置（31)，其na=0.67，；l=650nm;為避免其聚 焦物鏡（311)(312)微小的震動所造成的量測誤差，共軛焦 光學讀取裝置（31)的音圈（17)部份必須固定（音圈請配合 參看圖一所示）。在移動平台（18)方面，採用三軸移動平台 (18)，垂直方向（Z轴）精度為nm，最大行程1〇咖，其餘 兩軸的精度為20nm，最大行程50mm ;在軟體的部分，採用 VB與Labv i ew為開發移動平台（1 §)，用來系統控制與訊號 11 1352188· _ 處理。 根據（3-4)式’若m與0i是未知數，單顆探頭無法檢 測玻璃的厚度。因此，本發明採用兩顆光學讀取裝置 (31)(310)做為量測之用，目的在比對與校正等效厚度（t， 與實際厚度（t)的關係，等效確認參數0 i及m之後，只需 一顆光學讀取裝置（31)即可精準檢測待測物（32)。 量測時’以反射鏡當基準面（33)，以一顆光學讀取襄 φ置（31)作為檢測源光學讀取裝置（31)(test)，另一顆光學 讀取裝置（310)作為參考源光學讀取裝置（31〇) (Ref erence)，如圖六所示。測量前先以二光學讀取裝置 (31)(310)對無置放待測物之基準面（33)做一次掃描，可測 得檢測源光學讀取裝置（31)的聚焦點掃到基準面（33)的距 離（s〗），並測得參考源光學讀取裝置（31〇)的聚焦點掃到基 準面（33)的距離（R,)，（L)減去（Sl)可得一高度差（△)。 鼸確認檢測源光學讀取裝置（31)與參考源光學讀取裝置 (310)的高度差（△)後，再將待測物（32)放置在檢測源光學 讀取裝置（31)之基準面（33)之上，再測量一次後，檢測源 光學讀取裝置（31)可得S2與gap而如圖六（b)。檢測源光學 讀取裝置（31)的聚焦點掃到待測物（32)上表面的距離 (S2)，參考源光學讀取裝置（31〇)的聚焦點掃到基準面（33) 上表面的距離（R!)，經由下列各式的計算： (3-5) *5| = — Δ 1352188 ， t = Si~S2~gap ^ (3-6) 可得待測物（32) 一個實際測量的精確厚度（t)值，其中 即p為待測物（32)與基準面（33)間存在的空氣間隙。 依據圖五的系統架構與測量方式而得到的結果，在量測 前必須先檢測出基準面（33)，透過移動平台（18)掃描基準 面⑽可得如圖七（a)、圖七⑻的結果，分別為檢測源光 學讀取裝置（31)與參考源光學讀取裝置（31〇)測量之s曲 籲線，圖七（a)為參考源，圖七（b)為測試源。測試時以每步 lem逐步移動lmm，透過讀取s曲線的中心點座標，可找 出兩顆光學讀取裝置（31)(310)的高度差（△)與兩顆光學 項取裝置（31)(310)的基準點。檢測的結果可得參考源的 L(〇.9266mni)及檢測源的Sl(0.9388mm)，計算後可得△ =12. 3 ν m。 本發明採用多項式曲線擬合（p〇lyn⑽ial cUrve • Fitting)結合内插法（Intenx)lati〇n method)做為分析 s 曲線中心點的方法，由於圖六（a)、圖六（b)及圖七（a)、圖 七（b)操作中的移動平台（18)之移動步階為1//m，而檢測源 S曲線中心點位於0.938〜0.939 v οι之間，無法直接判斷 FES=0的座標，因此以多項式曲線的方式處理資料結果便 如圖八所不。接下來，本發明將大小3〇mmx3〇mm，厚度為 0.7mm的玻璃基板當做待測物（32)，檢測結果則如圖九 (a)、圖九（b)。從圖九（b)令可看到三個s曲線，分別為玻 13 丄叹188 .，表面、底面與基準面（33)面，經由圖六（b)的方式，可計 算得出厚度0. 6999 mm。 (i i )連續掃描的測量方式 旦本發明採用連續掃描的測量方式，主要的目的在減少 測量的時間，降低因環境漂移而產生的誤差，希望提高測 里的精度，其整個系統的組成架構如圖十，主要是將控制 器與光栅之間的回饋切斷，另行處理。為了移動的同:同 籲步擷取位置座標與檢測的信號，必須直接讀取並處理光柵 的信號。本發明㈣資㈣取卡⑽（竭）直接讀取光拇信 號，透過軟體的方式來辨別移動的距離，同時再以多通道 抓取檢測源光學讀取裝置⑻與參考源光學讀取裝置=) 内的四象限檢測器（16)訊號。由於光栅的位置訊號血四象 限檢測器（16)的訊號是同步處理，因此可減少誤差了本發 明仍以VB與Labview為開發移動平台（18)來建構自動 _的系統。 “ 本發明以連續掃描的測量方式縮短檢測時間，並探 對系統精度的影響針對厚度G 7mm的玻璃掃描檢:， 以0.3mm/SeC的速度移動，在3 67秒内完成測量， 結果如圖十-⑷、圖十一 (b)e圖十一(a)為參考源光學里讀 取裝置⑽）掃描基準面（33)的結果，圖十―⑻為檢測源 光學讀取裝置⑻量測待測物（32)的結果。由 遠大於移㈣度’本發明利料域平均时核料號， U^2188 透過每一位置取信號平均的方式，可以消減雜訊的影響， 如圖十二⑷、®十二（b)，亦可採用移動平均法（R晒ing Average Method)進行信號處理。此一方法的基本構相是對 原始資料取前後共2KH筆信號平均，其數學式表達如下： 1 κ1352188 · IX. DESCRIPTION OF THE INVENTION: 1. Field of the Invention The present invention relates to a device and method for detecting the thickness and gap of a transparent object by a dual optical probe, and more particularly to a position of using a multi-element detector to lock a conjugate focal plane. With the use of astigmatism, knife edge method and critical angle method or other practical methods, it can achieve accurate positioning and measurement of the thickness and clearance of transparent objects. • [Prior Art] In order to meet the requirements of TFT_LCD panel, the wafer manufacturer in the liquid crystal display industry has certain standards for the quality of planar undulation, surface roughness, density, specific gravity, etc. Chemical, mechanical, electrical and other properties have limitations on the parameters, making the flatness detection of the material account for an important part of the overall LCD display performance. There are quite a lot of related researches on glass thickness detection. In general, the pendulum® " method can be mainly divided into two types: transmissive and reflective; in terms of measurement behavior, the knife can be contactless or non-contact. With the advancement of measurement technology, the current non-contact optical measurement technology has gradually become the mainstream. At present, the non-contact optical measurement technology has a method of using laser ultrasonic waves to determine the thickness of the glass. This method is also a non-contact measurement. In addition to the object, the glass bottle is also a measurement range. Since the glass thickness of the summer test provides very important information during the manufacturing process, this information is monitored in real time to obtain high quality online products. Later, in order to change the speed and accuracy of the 1352188 · good glass inspection, a new non-contact method for obtaining the surface roughness and thickness of the glass was studied. The measurement principle is mainly to use the Sne 11 law to measure the f Describe the relationship between the light traveling through different media and the roughness and thickness of the glass surface. The above-mentioned various technologies have a certain measurement accuracy, but there are still many shortcomings in practical use. SUMMARY OF THE INVENTION The first object of the present invention is to provide a device for reducing the time of measurement, reducing the error caused by environmental drift, and improving the thickness and gap of a transparent object with improved measurement accuracy. A second object of the present invention is to provide a method for detecting the thickness and gap of a transparent object by using two optical reading heads for the purpose of measuring, in order to compare and correct the relationship between the equivalent thickness and the actual thickness. . "The second object of the present invention is to provide a two-co-focus probe that moves at the flattening opening, and the distance between the center points of the two s-curves obtained on the lower surface of the object to be tested is the equivalent thickness, thereby achieving a transparent object. The purpose of thickness detection. The apparatus for detecting the thickness and the gap of a transparent object by the double conjugate focal laser which achieves the above object includes: two optical reading devices respectively providing a generated light beam, which is optically read as a detection source. a stop device 'another optical reading device as a reference source; a poly ", an objective lens" provides the beam to focus and change the beam distance; and a 1352188 · - reference plane to provide a focus point scan of the optical reading device; Before the measurement, the detection source optical reading device and the reference source optical reading device are first-timed to scan without the object to be tested, and the detection source optically reads the focus of the agricultural spot to the reference distance (8), and The reference source optically reads the distance from the focus point of the farm to the reference plane (R1), and (10) subtracts the difference (9) from Δ. Then, the transparent object to be tested is placed on the detecting source optical reading device: on the quasi surface, and after measuring again, the detecting source optical reading device can detect that the focus point of the two φ reading device is swept onto the object to be tested. The distance between the surface (S2) and the air gap (gap) is calculated by the following equation: S1 = R1 - A t = S1 - S2 - gap A more accurate thickness (t) of the actual measurement of the object to be tested can be obtained. [Embodiment] I. Basic features and principles of the present invention. Referring to Figures 5, 6(a) and 6(b), the apparatus and method for detecting the thickness and gap of a transparent object by the dual optical probe provided by the present invention, Mainly comprising two optical reading devices (31) (310), respectively providing a light beam, one of which serves as a detection source optical reading device (31) and the other as a reference source optical reading device (310); Focusing the objective lens (311) (312), providing the beam to focus to form the optical probe (313) (314) and changing the beam distance; and a reference surface (33) for placing the object to be tested 'and providing the optical reading Pick up the focus point scan of 1352188·. As shown in Figures 4(a) and 4(8), the optical reading device (Gptieal piekupHead) in the common light-focus laser disc player on the market has a common light focus detection architecture; the conjugate focal length measurement system includes: Laser diode (ι), beam splitter (13), collimating lens (14), focusing objective (15), etc. When the laser body (11) of the optical pickup emits a light beam, it is reflected by a beam splitter (13) (Be(10) Splitter) to the collimating lens (14), and then focused by the focusing objective lens (15) to be tested. On the object (32), the reflected beam is then projected through the collimating lens (14) and the beam splitter (13) and then projected on the position detector (the four-quadrant detector (16) of the present invention is taken as an example), and then used. The astigmatism method works for locking the focus. Due to the astigmatism lock focus and four-quadrant detector (丨6) (a, B, C, D in Figure 1), although the reflectivity of the glass is very low, as long as the position of the conjugate focal plane can be locked, it can be accurate. The positioning and measurement, so the thickness and clearance can be accurately measured as long as the focal length of the focusing objective (15) is smaller. The laser diode (11) calls the wavelength λ = 650 nm ' NA=0.67, according to the diffraction theory, the focal spot radius (r) can be obtained as follows: corp.=0.61 to = 0.592//m The emitter diode (11) is focused by an objective lens, that is, the astigmatism aberration generated by optical imaging is used to judge whether or not focusing is performed. Astigmatism means that the optical magnifications in the vertical direction and the horizontal direction are different, so each has a focus, as shown in Fig. 2, the imaging of the meridional light is 13521.88. The position (the imaging position of the F〇 and the sagittal rays (6) is not the meridian light. The imaging position f (FT) does not coincide with the imaging position (Fs) of the sagittal ray. Therefore, for an object point, the imaging is no longer a point, and it is possible that the imaging position (Ft) of the meridional ray is a horizontal line. The imaging position (Fs) of the sagittal ray is a vertical line, which may be an ellipse or a circle between the two. The laser diode (U) is focused on the four quadrants by focusing the objective lens 〇 5) On the detector ((8), according to the astigmatism phenomenon of the light spot, the case of focusing is judged, as shown in Fig. 3. When the _ vertical and horizontal focal length are different, the imaging spot is an ellipse when the object deviates from the front focus or the back focus position of the lens. If the object is in the focus position, the imaging spot is a circle. The four-quadrant detector (16) and the associated signal processing are used by the Focus Error Signal: FES=(A+C)- (B+D) (34) can get an S curve, basically s The linear region of the line is the range that can be measured by the conjugate focal system, the center point of the curve corresponds to the position of the focus, and the rest is the defocus. The overall measurement system of the conjugated focus laser is transmitted through the optical reading device. The generated beam is focused onto the object to be tested (32) via the focusing objective (15), and the thickness and the gap are measured by scanning on the moving platform. The optical path is as shown in FIG. 4(a), wherein the object to be tested is (32) is a transparent material object. When the beam is focused by the focusing objective lens (15) to form a light probe, and the object to be tested (32) moves upwards, when the surface reaches the focus of the focusing objective (15) When 'the state of the focus is now' corresponds to the middle 1352188 Γ point of the conjugate focal measurement curve. When the object to be tested (32) continues to move up the distance of the object (10) equivalent thick yield (t)' The focus is on the bottom surface of the object to be tested (32), ^ to the center point of the measurement curve. Due to the refraction of the 4 4 7 Ma Ma in the medium, the surface layer only needs to move the equivalent thickness of the object to be tested (32) (t, The distance, ie the lock surface, is not the distance to move the actual thickness (1) The phenomenon of locking the focus to the bottom of the object to be tested (10) can be known through optical analysis. From Figure 4(b) and Snell's law, l*sin(9,. =nr sin^ (3-2) is known and The geometric relationship of the above figure can be obtained as w = ftm.Oi - = ttan^r (3-3) 卜 production r ! t'Jnr2 - sin2 $ tan0r -Jl-sin2 Θ. (3-4) Then T t is measured The actual thickness of the object (32), the equivalent thickness of the object (32) affected by the refraction in the dense medium '& is the angle of incidence of the beam: the center is the angle of refraction of the beam' nr is the object to be tested (32 The refractive index, from the equation, can be obtained from the actual thickness (1) and the equivalent thickness (t,) in the dense medium, the refractive index of the object (32) (1), the angle of the human light, that is, t is t, ~ and the function. The main purpose of this derivation formula is to understand that the measured thickness is the equivalent thickness (t,), which is different from the actual thickness. Π · Embodiments of the present invention 咕 Referring to FIG. 5 and FIG. 10 , the measuring method of the present invention can be implemented in two ways, the first type is a stepped sweeping, and the translation stage is moved 1352188 per movement. After grabbing the signal in one step, continue to move the next step until the total process to be moved is 'measured longer. The second type is the measurement of continuous scanning, that is, the translation stage does not pause once in a complete stroke, and the signal is captured while the mobile platform is moving. (i) Step-by-step scanning method The step-by-step measuring method adopted by the present invention has a system architecture as shown in FIG. 5, which firstly controls the moving platform (18) to vertically move the rudder by the command, when the beam passes through After focusing the objective lens (311) to the upper surface of the object to be tested (32), it is reflected back to the four-quadrant detector, and the signal is captured by the data capture card (19) to the computer (20) to obtain an s curve. The moving platform (18) continues to move to obtain another s curve on the lower surface of the object to be tested (32). By the distance between the center points of the two s-curves, an equivalent thickness (t,) can be obtained for the purpose of detection. The mobile platform (18) of the present invention can use the reading value of a grating as the basis for the moving distance, and the final resolution is 100 nm. In the optical system, a co-focus optical reading device (31) is used, which has a value of 0.7.67; l=650 nm; in order to avoid measurement errors caused by the slight vibration of the focusing objective lens (311) (312), conjugate The voice coil (17) part of the focal optical reading device (31) must be fixed (please refer to Figure 1 for the voice coil). In the mobile platform (18), the three-axis mobile platform (18) is adopted, the vertical direction (Z-axis) precision is nm, the maximum stroke is 1 〇, the other two axes have an accuracy of 20 nm, and the maximum stroke is 50 mm; in the software part, VB and Labv ew are used to develop the mobile platform (1 §) for system control and signal 11 1352188· _ processing. According to the formula (3-4), if m and 0i are unknown, a single probe cannot detect the thickness of the glass. Therefore, the present invention uses two optical reading devices (31) (310) for measurement purposes, in order to compare and correct the equivalent thickness (t, the relationship with the actual thickness (t), equivalent confirmation parameter 0 After i and m, only one optical reading device (31) can accurately detect the object to be tested (32). When measuring, use the mirror as the reference surface (33), and take an optical reading 襄φ (31) As the detection source optical reading device (31), another optical reading device (310) is used as the reference source optical reading device (31 〇), as shown in Fig. 6. The first reference surface (33) of the object to be tested is scanned by the two optical reading devices (31) (310), and the focus point of the optical reading device (31) of the detecting source is detected to be swept to the reference surface. (33) the distance (s), and the distance (R,) of the focus point of the reference source optical reading device (31〇) to the reference plane (33) is measured, and (L) minus (S1) is obtained. A height difference (Δ) 鼸 After confirming the height difference (Δ) between the detection source optical reading device (31) and the reference source optical reading device (310), the object to be tested (32) is placed on the detection source for optical reading. Above the reference surface (33) of the device (31), after another measurement, the detection source optical reading device (31) can obtain S2 and gap as shown in Fig. 6(b). The detection source optical reading device (31) The distance from the focus point to the upper surface of the object to be tested (32) (S2), the distance from the focus point of the reference optical reading device (31〇) to the upper surface of the reference surface (33) (R!), via the following Calculation of various formulas: (3-5) *5| = - Δ 1352188 , t = Si~S2~gap ^ (3-6) Available object (32) An actual measured exact thickness (t) value, Where p is the air gap existing between the object to be tested (32) and the reference plane (33). According to the system architecture and measurement method of Figure 5, the reference plane (33) must be detected before the measurement. Scanning the reference plane (10) through the mobile platform (18) can be obtained as shown in Fig. 7(a) and Fig. 7(8), which are measured by the detection source optical reading device (31) and the reference source optical reading device (31〇), respectively. Qu Yu line, Figure 7 (a) is the reference source, Figure 7 (b) is the test source. During the test, each step lem moves lmm gradually, by reading the center point coordinates of the s curve, you can find two The height difference (Δ) of the reading device (31) (310) and the reference point of the two optical item taking devices (31) (310). The result of the detection can be obtained from the reference source L (〇.9266mni) and the detection source. Sl (0.9388mm), after calculation, can obtain △ = 12. 3 ν m. The present invention uses polynomial curve fitting (p〇lyn (10) ial cUrve • Fitting) combined with interpolation (Intenx) lati〇n method) as analysis s The method of the center point of the curve, because the moving step of the moving platform (18) in the operations of FIG. 6(a), FIG. 6(b) and FIG. 7(a), and FIG. 7(b) is 1//m, and the detection is The center point of the source S curve is located between 0.938 and 0.939 v οι, and the coordinates of FES=0 cannot be directly judged. Therefore, the data is processed in a polynomial curve as shown in Fig. 8. Next, in the present invention, a glass substrate having a size of 3 mm x 3 mm and a thickness of 0.7 mm is used as the object to be tested (32), and the detection results are shown in Fig. 9 (a) and Fig. 9 (b). From Figure 9(b), we can see three s curves, which are glass 13 丄 188. The surface, the bottom surface and the reference surface (33) are calculated by the way of Figure 6(b). . 6999 mm. (ii) Measurement method of continuous scanning The present invention adopts a continuous scanning measurement method, and the main purpose is to reduce the measurement time, reduce the error caused by the environmental drift, and hope to improve the accuracy of the measurement, and the composition of the whole system is as follows: Figure 10, mainly to cut off the feedback between the controller and the grating, and deal with it separately. In order to move the same: to capture the position coordinates and the detected signal, the signal of the grating must be directly read and processed. The invention (4) capital (4) card acquisition (10) (exhaustion) directly reads the optical thumb signal, and discriminates the moving distance through the soft body, and simultaneously captures the detection source optical reading device (8) and the reference source optical reading device with multiple channels= The four quadrant detector (16) signal inside. Since the signal of the raster position signal four-quadrant detector (16) is synchronized, the error can be reduced. The present invention still uses VB and Labview to develop a mobile platform (18) to construct an automatic _ system. "The invention shortens the detection time by the continuous scanning measurement method, and detects the influence on the accuracy of the system. For the glass scanning inspection of the thickness G 7mm:, the movement is performed at a speed of 0.3 mm/SeC, and the measurement is completed within 3 67 seconds. X-(4), Fig. 11(b)e Figure 11(a) shows the result of scanning the reference plane (33) in the reference source optical reading device (10), and Fig. 10-(8) is the measurement source optical reading device (8) measurement The result of the object to be tested (32). By far more than the shift (four degrees), the average material number of the invention in the region of the invention, U^2188, by means of signal averaging at each position, can reduce the influence of noise, as shown in Fig. 10. Two (4), ® 12 (b), can also use the moving average method (R drying Average Method) for signal processing. The basic structure of this method is the average of 2KH pen signal before and after the original data, its mathematical expression As follows: 1 κ
Gi -ΪΚ + Ί/…= (3-7) 其中i為位置座標，κ為窗口參數。移動平均法可以減 少隨機變㈣影響’具有平滑數據的作用。但過度的平滑 可月έ造成訊號處理的失真，因而合理的選用窗口參數κ值， 是運用移動平均數的關鍵。圖十三⑷、圖十三（b)，為透 過移動平均法處理的結果，經過多項式内插法找出s曲線 的中。點，對於圖十二（a)、圖十三（b)所檢測的厚度為 〇·6998 以 m。 瓜.結論 本發明所提供之雙光探針檢測透明物體其厚度與間隙 的裝置’與前述引證案及其他習用技術相互比較時，更具 有下列之優點： 本發明利用共軛焦的概念設計一套檢測透明物體（如 LCD玻璃基板）的間隙與厚度m建立在以雙顆光學讀 取裝置的量測架構上以精確量測待測物的厚度與間隙。由 '快速與精確’適合用來作檢測玻璃基板 的工具之一，且誤差可在±0.29 am以内。 15 丄丄δδ .m所述，僅為本發明之—可行實施例，並非用以限 =發明之專㈣ϋ，凡舉依據·^申請專圍所述之 谷特徵以及其精神而為之其他變化的等效實施，皆應 匕含於本發明之㈣範_。本發明所具體衫於申料 利範圍之結構特徵’未見於同類物品，且具實用性盥進步 性·，已符合發明專利要件4依法具文提出申請謹請釣 局依法核予專利，以維護本申請人合法之權益。 ♦【圖式簡單說明】 圖一為習用共軛焦系統架構示意圖； 圖二為習用像散像差的原理示意圖； 圖二為習用像散法鎖焦方式示意圖； 圖四（a)為共軛焦量系統鎖焦至待測物上表面的示意圖； 圖四（b)為共軛焦量系統鎖焦至待測物下表面的示意圖； 圖五為本發明步階式之短焦距量測系統架構圖； 圖/、（a)為本發明短焦距量測系統僅測量基準面，找出△ 值； 圖六（b)為本發明短焦距量測系統放上待測物檢測，量測實 際厚度； 圖七（a)係本發明以參考源光學讀取裝置量測基準面的結 果圖； 圖七（b)係本發明以檢測源光學讀取裝置量測基準面的結 果圖； 16 . 1352188 圖八為本發明多項式曲線擬合法分析s曲線中心值； 圖九⑷係本發明轉考源光學讀取I置量職璃厚度的 結果圖； 圖九⑻係本發明以檢測源光學讀取裝置量測玻璃厚度的 測系統架構示意圖； 取裝置連續掃描檢測 參 圖十為本發明連續掃描式之小間距量 圖十一（a)係本發明以參考源光學讀 玻璃厚度之未處理信號的結果圖； 源光料取裝置連_描式檢 'J圾稱异度之未處理k號的結果圖； =二⑷係本發明以參考源料讀取裝置連續掃描檢測 玻璃厚度之經過時域平均處理的結果圖； 圖十二⑻係本發明以檢測源光學讀取裝置連續掃描檢測 玻璃厚度之經過時域平均處理的結果圖； 圖十三（a)係本發明以參考源光學讀取裝 I置連續掃描檢測 玻璃厚度之經過移動平均處理的結果圖；及 圖十三⑻係本發明以檢測源光學讀料置連續掃描檢測 玻璃厚度之經過移動平均處理的結果圖。 【主要元件符號說明】 (11)雷射二極體 （13)分光鏡 (U)準直透鏡 （15)聚焦物鏡 (16)檢測器 （17)音圈 17 1352188· (18)移動平台 (20)電腦 (311)(312)聚焦物鏡 (32)待測物 (19)資料擷取卡 (31)(310)光學讀取裝置 (313)(314)光探針 (33)基準面 (Ft )子午光線的成像位置 (Fs)弧矢光線的成像位置 (t)實際厚度 • (t’）等效厚度 (0 0入射角 (0 <·)折射角 (η〇折射率 (51) 檢測源的聚焦點掃到反射鏡時移動平台移動的距離 (52) 檢測源的聚焦點掃到待測物上表面的距離 • (△)高度差 (Rl)參考源的聚焦點掃到反射鏡時移動平台移動的距離Gi -ΪΚ + Ί/...= (3-7) where i is the position coordinate and κ is the window parameter. The moving average method can reduce the effect of random (4) influences with smooth data. However, excessive smoothing can cause distortion of the signal processing, so the reasonable selection of the window parameter κ value is the key to using the moving average. Figure 13 (4) and Figure 13 (b) show the results of the moving average method and find the middle of the s-curve by polynomial interpolation. The thickness detected for Fig. 12(a) and Fig. 13(b) is 〇·6998 and m. Melon. Conclusion The apparatus for detecting the thickness and the gap of a transparent object provided by the present invention has the following advantages when compared with the aforementioned cited documents and other conventional techniques: The present invention utilizes the concept design of conjugated coke. The gap and thickness m of the cover for detecting transparent objects (such as LCD glass substrates) are established on the measurement architecture of the two optical reading devices to accurately measure the thickness and gap of the test object. It is suitable for use as a tool for detecting glass substrates by 'quick and precise' and the error can be within ±0.29 am. 15 丄丄δδ .m is only a feasible embodiment of the present invention, and is not intended to be limited to the invention (four), which is based on the characteristics of the valley and other changes thereof. The equivalent implementation of the invention should be included in the (four) model of the present invention. The structural features of the specific shirts in the scope of the invention are not found in the same kind of articles, and they are practical and progressive. They have met the requirements of the invention patents and have been submitted in accordance with the law. The legal rights of the applicant. ♦ [Simple diagram of the diagram] Figure 1 is a schematic diagram of the conventional conjugate focal system architecture; Figure 2 is a schematic diagram of the conventional astigmatic aberration; Figure 2 is a schematic diagram of the conventional astigmatism locking method; Figure 4 (a) is conjugate Schematic diagram of the focus system locking focus to the upper surface of the object to be tested; Figure 4 (b) is a schematic diagram of the conjugate focal length system locking focus to the lower surface of the object to be tested; Figure 5 is a step-type short focal length measuring system of the present invention Figure /, (a) The short focal length measurement system of the present invention only measures the reference plane and finds the Δ value; Figure 6 (b) shows the short focal length measurement system of the present invention on which the object to be tested is placed, and the measurement is actually performed. Figure 7 (a) is a result of the measurement of the reference plane by the reference source optical reading device of the present invention; Figure 7 (b) is a result of the measurement of the reference plane by the source optical reading device of the present invention; 1352188 FIG. 8 is a polynomial curve fitting method for analyzing the center value of the s curve according to the present invention; FIG. 9 (4) is a result chart of the optical reading I of the transfer source of the present invention; FIG. 9 (8) is the optical reading of the detection source of the present invention Schematic diagram of the system for measuring the thickness of the glass of the device; The continuous scanning detection reference numeral 10 is the small scanning amount of the continuous scanning type of the present invention. FIG. 11( a) is a result diagram of the unprocessed signal of the reference optical reading glass thickness of the present invention; the source light material picking device is connected with the 'J garbage is said to be the result of the unprocessed k number of the different degrees; = two (4) is the result of the time domain average processing of the continuous scanning and detecting glass thickness by the reference source material reading device of the present invention; Fig. 12 (8) is the present invention The result of the time-domain averaging process for continuously scanning and detecting the thickness of the glass by the detecting source optical reading device; FIG. 13(a) shows the moving average processing of the glass thickness of the continuous scanning detection glass by the reference source optical reading device. The result graph; and FIG. 13 (8) is a graph showing the results of the moving average processing of the glass thickness of the continuous optical scanning by the detection source optical reading material. [Main component symbol description] (11) Laser diode (13) Beam splitter (U) Collimating lens (15) Focusing objective lens (16) Detector (17) Voice coil 17 1352188· (18) Mobile platform (20 Computer (311) (312) Focusing objective lens (32) Object to be tested (19) Data capture card (31) (310) Optical reading device (313) (314) Optical probe (33) Reference plane (Ft) Imaging position of the meridian ray (Fs) Imaging position of the sagittal ray (t) Actual thickness • (t') Equivalent thickness (0 0 incident angle (0 < ·) refraction angle (η 〇 refractive index (51) Detection source The distance the moving platform moves when the focus point is swept to the mirror (52) The distance from the focus point of the detection source to the upper surface of the object to be tested • (Δ) The height difference (Rl) The focus of the reference source is swept to the mirror. The distance the platform moves