1291013 觀察近年來之量測技術發展，共焦顯微術為生醫領域中獲得蓬勃發展 與廣泛之應用，乃因此項技術具備良好之光學切片能力以及空間解析度等 優點，其量測基本原理乃以光學掃描方式，獲得待測物不同高度之平面影 ^像，藉由過濾失焦訊號，保留聚焦訊號資訊方式，再經由電腦軟體進行重 - 建以形成三度空間立體影像。 >習用的光學三維量測技術係採用雷射光源，然而雷射光源容易因待測 物表面形狀不連續的部份而產生散射的問題。 數位从鏡組裝置(Digital Micromirror Devices，DMD)多用於影像或影片 投影上；F· Bitte等人於2001年提出以數位微鏡組裝置(DMD)作為主動光源 才又射的共焦顯微系統(F. Bitte et al” 3D micro-inspection goes DMD，Optics and Lasers in Engineering，2001，36 : 155_67·)，將數位微鏡組裝置(dmd^ 用於光學檢測上，產生特定結構光圖案進行量測，以dmd取代傳統共焦顯 微鏡的針孔（Pinhole)裝置，DMD的每一個微小鏡組（micr〇mirr〇r)即代 •表著影像中的每一個像素（Pixel)，當移動待測物通過聚焦面時，意即 所代表的像素對應到該待測物與物鏡之聚焦面時，將可以獲得最大的影像 強度；藉由評估深度曲線響應圖中的峰值（Peak)，可找出待_表面相對 應點之實際高度值，進而重建待測物之三維表面輪廓。 使用數位彳政鏡組裝置(DMD)作為主動光源投射具備以下的優點：維護 容易、處理參數可由軟體控制、大量使用可降低成本，以及影像高對比、 局解析、高亮度等優點；在量測精度方面，F· Bitte等人以内插法提高量測 精度，並降低雜訊的影響，避免誤判峰值；以DMD為基礎所發展共焦掃描 1291013 一光學投影單元，包括一使用DMD為基底的DLP (Digital Light Processing)投影機以及一光學投影鏡組； -影像擷取單元，包括—顯微系統以及—影像感測元件； -精松平移台單元，由—χ、γ、ζ三軸精密平移台所組成；以及 一主控制單元； 該精密平移台單元係設置機影像擷取單元的顯⑽統之下，用以放 置待測物；魅控鮮元係連接該光學投影單元，經_主㈣單元編寫 並由該DLP投影機產生一任意圖案之數位結構光源，該數位結構光源經過 該光予投〜鏡組進行該圖案的縮影、減光、濾波與聚焦後，投射進入顯微 系、、先内’藉#销微纟細£合待測物之尺寸與高度，關整該光源之景深 與可量測範圍，並將驗位結構統投射至該制物上，再由該影像感測 几件_該待測物表面反射訊號，該影像感測元件係連接該主控制單元， 可將擷取之峨傳輸至齡控鮮元，進行分析·點之聚焦函數，以推 异出待測物之三維高度，進而獲得待測物表面精確之三維輪廊資訊； 其中該光學投影鏡組係為光學準直鏡、線性偏光板、半鏡片與光 學聚焦透鏡之一種以上所組成之構造者； 於-較佳實施例中’該光學投影鏡組係由一光學準直鏡、一線性偏光 板’以及一半圓柱鏡片所組成； 其中該顯微系統係為聚焦物鏡、分光鏡、光學聚焦透鏡與針孔之一種 以上所組成之構造者； 於-較佳實施财，該顯微系祕由—聚焦物鏡、—分光鏡、二光學 1291013 聚焦透鏡，以及一針孔所組成； 於一較佳實施例中，該數位結構光源係由該主控制單元編寫並由該 DLP投影機產生，並依序通過該半圓柱鏡片、該線性偏光板、與該光學準 直鏡，再投射進入該顯微系統之分光鏡上，由該分光鏡與該聚焦物鏡調整 該光源之景深與可量測範圍，並將該數位結構光源投射至該待測物上。 其中該影像感測元件係為CCD顯像感測元件； 其中該光學投影單元，更進一步包含一光強度控制裝置，用以調整該 DLP投影機之光源強度，以獲得良好空間解析度之數位結構光源； 其中該影像線取單元，更進一步包含一高连影像處理器褒置，用以提 供影像之擷取與處理； 其中該主控制單元係為-計算機裝置，包含一以Rs_232為介面的奶 位移裝置，以及一影像擷取卡； 其中該數位結構光源之圖案係為一沿水平與垂直方向正弦變化之數位 、、。構光圖案’用以投影至待測物表面，以提高量測之空間解析，該圖案如 下列所示：1291013 Observing the development of measurement technology in recent years, confocal microscopy has been flourishing and widely used in the field of biomedicine. Therefore, the technology has good optical slicing ability and spatial resolution. The basic principle of measurement is The optical scanning method obtains a planar image of different heights of the object to be tested, and filters the out-of-focus signal, retains the focused signal information mode, and then re-establishes through the computer software to form a three-dimensional spatial stereoscopic image. > Conventional optical three-dimensional measurement technology uses a laser light source, however, the laser light source is liable to cause scattering due to a discontinuous portion of the surface of the object to be tested. The digital micromirror device (DMD) is mostly used for image or film projection; F. Bitte et al. proposed a confocal microscope system with a digital micromirror device (DMD) as the active light source in 2001. Bitte et al" 3D micro-inspection goes DMD, Optics and Lasers in Engineering, 2001, 36: 155_67·), using a digital micromirror device (dmd^ for optical detection to produce a specific structured light pattern for measurement, Replace the pinhole device of the traditional confocal microscope with dmd. Each micromirror of the DMD (micr〇mirr〇r) is used to represent each pixel (Pixel) in the image. When moving the object to be tested When focusing the surface, it means that the pixel represented by the object corresponds to the focal plane of the object to be tested and the objective lens, and the maximum image intensity can be obtained; by evaluating the peak value (Peak) in the depth curve response map, it can be found _ The actual height value of the surface corresponding to the point, and then reconstruct the three-dimensional surface contour of the object to be tested. The use of the digital Mirror Group Device (DMD) as the active light source projection has the following advantages: easy maintenance, processing parameters can be soft Control, large-scale use can reduce costs, as well as high contrast, local analysis, high brightness, etc. In terms of measurement accuracy, F· Bitte et al. use interpolation to improve measurement accuracy and reduce the impact of noise to avoid false positives. Development of confocal scanning 1291013 based on DMD An optical projection unit comprising a DLP (Digital Light Processing) projector using DMD as a substrate and an optical projection mirror set; - an image capturing unit comprising: a microscopic system and - image sensing element; - a loose translation stage unit consisting of a three-axis precision translation stage of - χ, γ, ζ; and a main control unit; the precision translation stage unit is provided with a display (10) of the image capturing unit Next, for placing the object to be tested; the charm control fresh element is connected to the optical projection unit, and is written by the _ main (four) unit and generates an arbitrary pattern digital structure light source by the DLP projector, and the digital structure light source passes through the light ~The lens group performs the microcosm, dimming, filtering and focusing of the pattern, and then projects into the microscopy system, and the first 'borrowing' pin-sells the size and height of the object to be tested. The depth of field and the measurable range of the light source, and the inspection structure is projected onto the workpiece, and then the image is sensed by the image _ the surface of the object to be tested is reflected, the image sensing component is connected to the main control The unit can transmit the captured enthalpy to the age-control fresh element, and perform an analysis and point focusing function to push out the three-dimensional height of the object to be tested, thereby obtaining accurate three-dimensional wheel corridor information of the object to be tested; wherein the optical The projection mirror assembly is composed of one or more of an optical collimating mirror, a linear polarizing plate, a half lens and an optical focusing lens; in the preferred embodiment, the optical projection mirror assembly is composed of an optical collimating mirror, a linear polarizing plate and a semi-cylindrical lens; wherein the microscopic system is a construct composed of a focusing objective lens, a beam splitter, an optical focusing lens and a pinhole; The secret consists of a focusing objective, a beam splitter, a two-optical 1291013 focusing lens, and a pinhole; in a preferred embodiment, the digital structure light source is written by the main control unit and projected by the DLP. Generating, and sequentially passing the semi-cylindrical lens, the linear polarizing plate, and the optical collimating mirror, and then projecting into the beam splitter of the microscopic system, and adjusting the depth of field of the light source by the beam splitter and the focusing objective lens The range can be measured and the digital structure light source is projected onto the object to be tested. The image sensing component is a CCD imaging sensing component; wherein the optical projection unit further comprises a light intensity control device for adjusting the light source intensity of the DLP projector to obtain a digital structure with good spatial resolution. The image line taking unit further includes a high-connected image processor device for providing image capture and processing; wherein the main control unit is a computer device including a milk with an interface of Rs_232 a displacement device, and an image capture card; wherein the pattern of the digital structure light source is a digit that varies sinusoidally in the horizontal and vertical directions. The constitutive pattern 'is projected onto the surface of the object to be tested to improve the spatial resolution of the measurement, as shown below:
Kh j) = A sm(— ι + φ{)·Β sin(~ ] + φ2) 1 该巧為水平方向的條紋間距，:Γ2為垂直方向的條紋間距為水平方向 的in紋相位’ A為垂直方向的條紋相位，^為水平與垂直方向的光強振幅。 本發明更進-步提供一種微三維輪廓量測之方法，係包含下列步驟： 步驟辟投影步驟，藉由本發明所提供之量測祕，該主控制單元 編寫並由該DLP投影機產生—任意酵之數位結構光源，該數位結構光源 1291013 經過該光學投影鏡組進行該圖案的縮影、減光、濾波與聚焦後，投射進入 該顯微系統内，藉由該顯微系統配合該待測物之尺寸與高度，以調整該光 源之景深與可量測範圍，並將該數位結構光源投射至待測物上； 步驟二影像擷取步驟，該待測物表面之反射訊號經由該顯微系統與該 影像感測元件所擷取，該顯微系統可根據不同待測物之尺寸大小，調整量 測景深與可ϊ測範圍，以適應不同待測物之量測需求，擷取待測物於不同 掃麟度下之影像局影像感測元個，該影像感測元件可將影像資訊輸 出至該主控制單元，以進行分析； 步驟二演异法計算步驟，在固定該顯微系統之聚焦物鏡的情況下，設 定該待測物之初始掃描位置，並進行該待測物之深度掃描，可由該影像感 測元件擷取該待測物在不同深度掃描位置下，因距離該聚焦物鏡位置之不 :而產生不同聚焦程度之影像，將不職餘度之影像經由賴函數進行 y估求取最大聚焦函數評估值處之位置，即為該待測物到達該聚焦物鏡 聚焦面之位置，亦為該制物之真實表面高度值，根據此原理，即可完成 待測物之三維輪廓掃描。 【實施方式】 實施例-數絲構賴三維共絲面輪廓量測系統 請參本發明所提供之數位結構賴三料絲面輪專量測系 物置 一精密平移台單元4， 於該精密平移台單元上 由X、γ、z二軸精密平移台所組成，可將待測 ，於三度空間進行完整之掃描； 1291013 一光學投影單元2,由一使用DMD為基底wDLP投影機21以及一光 學投影鏡組22所組成，該光學投影鏡組22包含一光學準直鏡221、一線性 偏光板222，以及一半圓柱鏡片223 ; 一影像擷取單元3，由一顯微系統31與一 CCD顯像感測元件32所組 成，該顯微系統31包含一聚焦物鏡311、一分光鏡312、二光學聚焦透鏡 313與315 ’以及一針孔314 ;Kh j) = A sm(— ι + φ{)·Β sin(~ ] + φ2) 1 This is the horizontal stripe spacing, Γ2 is the vertical stripe spacing is horizontal in the in-phase phase 'A is The fringe phase in the vertical direction, ^ is the intensity amplitude of the horizontal and vertical directions. The present invention further provides a method for microscopic three-dimensional contour measurement, comprising the steps of: projecting a projection step, by which the main control unit is programmed and generated by the DLP projector - any a digital structure light source of the yeast, the digital structure light source 1291013 is subjected to microscopic, dimming, filtering and focusing of the pattern through the optical projection mirror group, and then projected into the microscopic system, and the microscopic system cooperates with the object to be tested Dimensions and heights for adjusting the depth of field and the measurable range of the light source, and projecting the digital structure light source onto the object to be tested; Step 2: an image capturing step, the reflection signal of the surface of the object to be tested is transmitted through the microscope system And the image sensing component is taken, the microscopic system can adjust the measured depth of field and the detectable range according to the size of different objects to be tested, to adapt to the measurement requirements of different objects to be tested, and to extract the object to be tested The image sensing component of the image bureau under different scanning degrees, the image sensing component can output the image information to the main control unit for analysis; and the second step performs the different calculation step. In the case of fixing the focusing objective of the microscope system, setting an initial scanning position of the object to be tested, and performing a depth scanning of the object to be tested, and the image sensing element can capture the object to be tested at different depth scanning positions. Then, the image of the different focus degree is generated due to the position of the focusing objective lens, and the image of the inoperative margin is estimated by the y function to obtain the position of the maximum focus function evaluation value, that is, the object to be tested arrives. The position of the focusing surface of the focusing objective is also the true surface height value of the workpiece. According to this principle, the three-dimensional contour scanning of the object to be tested can be completed. [Embodiment] Embodiment-Digital-wire three-dimensional co-filament surface profile measurement system, please refer to the digital structure of the three-wire wire surface wheel special measurement system provided by the invention, and a precision translation stage unit 4 is arranged for the precision translation The unit is composed of X, γ, z two-axis precision translation stage, which can be tested and scanned in three dimensions. 1291013 An optical projection unit 2, which uses a DMD as the base wDLP projector 21 and an optical The optical lens assembly 22 comprises an optical collimating mirror 221, a linear polarizing plate 222, and a semi-cylindrical lens 223. An image capturing unit 3 is displayed by a microscopic system 31 and a CCD. The imaging system 32 comprises a focusing objective 311, a beam splitter 312, two optical focusing lenses 313 and 315 ' and a pinhole 314;
I 一主控制單元1 ’係連接該光學投影單元2與該影像擷取單元3，用以 調控該絲郷單元2之結構細譜輸$，以及處賴影像擷取單元3所 得之影像資訊； 該光學投景彡單元2係設置於該影像擷取單元3之側邊，與該分光鏡312 位於同一水平線上，其排列由右至左依序為該〇1^投影機21、該半圓柱鏡 片223、該線性偏光板222、該光學準直鏡221、該分光鏡312 ;該影像擷 取單元3之各元制位於同―垂直線上，其湖由下社依序為該聚焦物 鏡311，該分光鏡312，該光學聚焦透鏡313，該針孔314，該光學聚焦透 鏡315，以及該CCD顯像感測元件32 ;該精密平移台單元4則設置於該聚 焦物鏡311之下；該主控制單元i係連接該DLp投影機21，以及該ccd 顯像感測元件32。 將待測物該精密平移台單元4上，由該主控鮮元！編寫並由該 DLP投影機21產生-任意_之數健構統，練位轉絲依序經過 該半圓柱鏡片223、該線性偏光板222，以及該光學準直鏡221，以進行該 數位結構光源圖案的縮影、減光、濾波與聚焦後，該數位結構光源投射至 11 1291013 該分光鏡312 ’並配合該顯微系統31根據該待測物之尺寸與高度，以調整 該光源之景深與可量測範圍，並將該數位結構光源投射至該待測物上，該 待測物表面反射訊號依序經過該聚焦物鏡3U、該分光鏡312、該光學聚焦 • 透鏡313、該針孔314,以及該光學聚焦透鏡315，藉由該針孔314達到空 間濾波之效果’再由該CCD顯像感測元件32擷取該待測物表面反射訊號， 該CCD顯像感測元件32係連接該主控制單元1，可將擷取之訊號傳輸至該 主控制單元1，進行分析影像點之聚焦函數，以推算出該待測物之三維高 ®度，進而獲得該待測物表面精確之三維輪廓資訊。 當該CCD顯像感測元件32擷取該待測物表面反射訊號時，在固定該 顯微系統之聚焦物鏡祀的情況下，需設定該待測物之初始掃描位置，並 進行該待測物之深鱗描，可_⑽顯像_元件32棘該待測物在 不Π木度掃描位置下’因距離該聚焦物鏡311位置之不一而產生不同聚焦 程度之影像，將不同聚_度之影像經由M、函數進行評估，求取最大聚 鲁焦函數評估值處之位置，即為該待測物到達該聚焦物鏡3ιι聚焦面之位置， 亦為該待測物之真實表面高度值，根據此原理，即可完成待測物之三維輪 廊掃描。 /在該主控制單元1中，可利用程式編寫任意之數位結構光圖案投射至 待測物i由於考慮到當待測物為—表面平滑之物體時，制量測技術會產 Γ7頻貝Λ不足的問題，因此本發明提供一沿水平與垂直方向正弦變化之 該正弦變化 數位結構光ϋ案進行翻彳物之量測，以提高量測之空間解析， 結構光可表示成： 12 1291013 公式1 公式1中，7；為水平方向的條紋間距，G為垂直方向的條紋間距，^為水平 方向的條紋相位，A為垂直方向的條紋相位，為水平與垂直方向的光強 振幅，所投影出之圖案如附件一所示。 實施例二結構光之正弦波週期對量測結果的影響 在本實施例中，比較投影以主動式正弦結構光時，不同正弦波週期對 量測結果的影響，該正弦波職脑在該結構光巾，單_個完整正弦波所 佔的像素(Pixel)總數，以一階高為2·37,的校正塊進行量測結果的比較， 該正弦波週期分別為5、1G、3〇、5G、7G、9G與削，量測時之垂直掃描 間距為Ο.ίμιη其誤差分析如表—所示，圖二為不同正弦波週期結構光與量 測誤差之關係圖。 表一 不同正弦波週期之校正塊階高誤差表 期 5 10 30 50 70 90 110 含域平均誤差(μηχ) 0.251 0.071 1.55 6.461 8.578 10.367 17 9Q8 bb(%) 主域平均誤差標準差(μπι) 10.6 2.99 65.4 273 362 "—----- 438 759 0.565 0.183 3.681 14.385 20.778 24.755 45.199 附件m皮週麟1G，附件二6之正弦波為％;味其結果可 觀察出胃才又影正弦波週期為10之結構光時，其量測結果最佳，量測誤差 田j右將正弦波週期縮小至5，嘗試提高空間解析時，則因受光學投影系 、、先之限制，造成結構光品f不佳，反而使量測誤差提高，而當結構光之正 13 1291013 弦波週期繼續提高達到30以上時，因結構光所產生的量測空間解析與高頻 資訊變化量也隨正弦波週期的提高而降低，而形成較大的量測誤差。 實施例三主動光強對量測結果的影響 由於不同主動光強會影響聚焦深度反應曲線的品質，其原因為不同待 測物的表面粗縫度、反射率甚至顏色可能皆不相同，因此若在同一量測條 件下進行不同待測物的量測，將產生量測上的誤差，而其中一個鉍響量測 _結果的重要因素即為絲光強，只有在適#的光強下，不同待職表面的 組織才可做最佳的呈現；表二為在不社動光強的情形下的誤差分析，其 影像相對平均灰階分別為22、33、44、52、61、74、科、％、ιΐ4、⑽與 145，圖二為不同主動光強與量測誤差之關係圖。 表二不同主動光強之校正塊階高誤差表 影像相 均灰階 22 33 誤差 44 52 61 74 84 98 114 130 145 平均誤差 (μπι) 9.04 6.545 0.498 0.471 0.261 0.179 0.073 0.636 1.863 3.177 3.365 平均誤差 百分比(％) 381 276 21.01 19.87 11.01 7.53 3.06 26.84 78.61 134 142 全域平均 誤差標準差| 22.47 18.76 1.692 1.172 」 0,634 0.451 0.182 2.22 7.256 14:46 15.32 附件三為在不同主動光強下，所得之校正塊全域誤差圖，附件三A之 =像相對平均灰職度為8彳，附件三b之影像相對平均灰階強度為a ;比 較其結果可觀Μ ’ #影像姆平均灰㈣度為U時，因光料足，造成 待測物表Φ!職無法進行良好的呈現，因此造成量耻有較大的誤差’而 14 1291013 當影像相對平均灰階強度為84時，此時的主動光強可使投影結構光至校正 塊階高上時，其影像對比度最佳，因此能獲得最佳的量測結果。 實施例四垂直掃描間距對量測結果的影樂 於本實施例中，探討在不同垂直掃描間距下對量測結果的影響，以一 階南為2.37μιη的校正塊進行量測結果的比較，垂直掃描間距分別為〇·6、 〇·5、〇·4、〇·3、0.2及αίμπι，表三為在不同垂直掃描間距的誤差分析，圖 四為不同垂直掃描間距與量測誤差之關係圖。 表三不同垂直掃描間距之校正塊階高誤差表 ^直掃描間距 0.1 0.2 0·3 0.4 0.5 0.6 平均誤差 0.067 0.116 0.175 0.234 0.291 0.38 平均誤差百分比 2.83 4.88 7.36 9.86 12.29 16.03 全域平均誤差標準差(μιη) 0.164 0.296 I 0.467 0.634 0.799 L053The main control unit 1 is connected to the optical projection unit 2 and the image capturing unit 3 for regulating the structure of the silk unit 2 and the image information obtained by the image capturing unit 3; The optical projection unit 2 is disposed on the side of the image capturing unit 3, and is located on the same horizontal line as the beam splitter 312. The arrangement is from right to left, and the projector 21 and the semi-cylindrical are sequentially arranged. The lens 223, the linear polarizing plate 222, the optical collimating mirror 221, and the beam splitter 312; the elements of the image capturing unit 3 are located on the same vertical line, and the lake is ordered by the lower body as the focusing objective lens 311. The beam splitter 312, the optical focusing lens 313, the pinhole 314, the optical focusing lens 315, and the CCD imaging sensing element 32; the precision translation stage unit 4 is disposed under the focusing objective lens 311; The control unit i is connected to the DLp projector 21, and the ccd imaging sensing element 32. The object to be tested is placed on the precision translation stage unit 4, and the main control unit is fresh! The DLP projector 21 is programmed and generated by the DLP projector 21, and the training rotary wire sequentially passes through the semi-cylindrical lens 223, the linear polarizing plate 222, and the optical collimating mirror 221 to perform the digital structure. After the miniature, dimming, filtering and focusing of the light source pattern, the digital structure light source is projected to the 11 1291013 the beam splitter 312 ′ and cooperates with the microscope system 31 according to the size and height of the object to be measured to adjust the depth of field of the light source and Measure the range and project the digital structure light source onto the object to be tested. The surface reflection signal of the object to be tested sequentially passes through the focusing objective lens 3U, the beam splitter 312, the optical focusing lens 313, and the pinhole 314. And the optical focusing lens 315, the spatial filtering effect is achieved by the pinhole 314. The CCD imaging sensing component 32 captures the surface reflection signal of the object to be tested, and the CCD imaging sensing component 32 is connected. The main control unit 1 can transmit the captured signal to the main control unit 1 to analyze the focus function of the image point to calculate the three-dimensional height of the object to be tested, thereby obtaining the surface of the object to be tested. Three-dimensional contour News. When the CCD imaging sensing component 32 captures the surface reflection signal of the object to be tested, in the case of fixing the focusing objective lens of the microscope system, the initial scanning position of the object to be tested needs to be set, and the test is performed. The deep scale drawing of the object can be _(10) imaging_component 32. The object to be tested is not in the scanning position of the wood. The image with different focusing degrees due to the difference in the position of the focusing objective lens 311 will be different. The image of degree is evaluated by M and function, and the position of the evaluation value of the maximum poly-rug function is obtained, that is, the position of the object to be measured reaches the focal plane of the focusing objective 3 ιι, which is also the true surface height value of the object to be tested. According to this principle, the three-dimensional corridor scan of the object to be tested can be completed. / In the main control unit 1, the program can be used to write an arbitrary digital structure light pattern to be projected onto the object to be tested i. Since the object to be tested is an object having a smooth surface, the measurement technique will produce a 7-frequency measurement. Insufficient problem, therefore, the present invention provides a sinusoidal digital structure optical sinusoidal variation in horizontal and vertical sinusoidal measurement to improve the spatial resolution of the measurement, and the structured light can be expressed as: 12 1291013 Formula 1 In Equation 1, 7 is the stripe pitch in the horizontal direction, G is the stripe pitch in the vertical direction, ^ is the fringe phase in the horizontal direction, and A is the fringe phase in the vertical direction, which is the horizontal and vertical intensity of the light intensity, projected The pattern is shown in Annex 1. Embodiment 2 Effect of the sinusoidal period of the structured light on the measurement result In the present embodiment, when the active sinusoidal structured light is projected, the influence of different sinusoidal periods on the measurement result is compared, and the sine wave is in the structure. Light towel, the total number of pixels (Pixel) occupied by a single sine wave, compared with the calibration block of the first-order height of 2.37, the sine wave period is 5, 1G, 3〇, 5G, 7G, 9G and cutting, the vertical scanning spacing when measuring is Ο. ίμιη its error analysis is shown in the table - Figure 2 is the relationship between the different sinusoidal periodic structure light and measurement error. Table 1 Correction block step height error period for different sine wave periods 5 10 30 50 70 90 110 Field average error (μηχ) 0.251 0.071 1.55 6.461 8.578 10.367 17 9Q8 bb(%) Main field mean error standard deviation (μπι) 10.6 2.99 65.4 273 362 "------ 438 759 0.565 0.183 3.681 14.385 20.778 24.755 45.199 Annex m Pi Zhoulin 1G, Annex II 6 sine wave is %; taste results can observe the stomach before the sine wave cycle When the structure light is 10, the measurement result is the best, and the measurement error field J right reduces the sine wave period to 5, and when trying to improve the spatial analysis, the optical projection system is restricted by the optical projection system. f is not good, but the measurement error is improved, and when the sinusoidal period of the structured light is 13 1391013 continues to increase to more than 30, the measurement space analysis and the high-frequency information change due to the structured light also follow the sine wave period. The increase is reduced and a large measurement error is formed. Example 3 Effect of Active Light Intensity on Measurement Results Since different active light intensities affect the quality of the depth of focus response curve, the reason is that the surface roughness, reflectance and even the color of different objects to be tested may be different, so Measurement of different analytes under the same measurement conditions will produce errors in the measurement, and one of the important factors of the measurement is the intensity of the filament, which is different only under the light intensity of # The organization of the on-the-job surface can be optimally presented; Table 2 shows the error analysis in the case of non-communication, and the relative grayscale of the image is 22, 33, 44, 52, 61, 74, and , %, ιΐ4, (10) and 145. Figure 2 shows the relationship between different active light intensities and measurement errors. Table 2 Calibration of different active light intensity block height error table image phase-average gray scale 22 33 Error 44 52 61 74 84 98 114 130 145 Average error (μπι) 9.04 6.545 0.498 0.471 0.261 0.179 0.073 0.636 1.863 3.177 3.365 Average error percentage ( %) 381 276 21.01 19.87 11.01 7.53 3.06 26.84 78.61 134 142 Global standard deviation error standard | 22.47 18.76 1.692 1.172 ” 0,634 0.451 0.182 2.22 7.256 14:46 15.32 Annex 3 is the global error map of the obtained correction block under different active light intensities. Annex A A = like the relative average gray position is 8 彳, the relative average gray level intensity of the image of Annex III b is a; comparing the results is considerable # '# image um average gray (four) degree is U, due to the light material , causing the object to be tested Φ! The job can not be performed well, so the shame has a large error' and 14 1291013 When the image relative average gray level intensity is 84, the active light intensity at this time can make the projected structure light When the step height is corrected, the image contrast is optimal, so the best measurement results can be obtained. Embodiment 4: The vertical scanning pitch affects the measurement results. In this embodiment, the influence of the vertical scanning interval on the measurement results is discussed. The comparison results of the first-order South 2.37μηη correction block are compared, and the vertical measurement is performed. The scanning pitches are 〇·6, 〇·5, 〇·4, 〇·3, 0.2, and αίμπι, Table 3 is the error analysis at different vertical scanning intervals, and Figure 4 is the relationship between different vertical scanning spacing and measurement error. . Table 3 Correction block height error table of different vertical scanning intervals ^ Straight scanning pitch 0.1 0.2 0·3 0.4 0.5 0.6 Average error 0.067 0.116 0.175 0.234 0.291 0.38 Average error percentage 2.83 4.88 7.36 9.86 12.29 16.03 Global average error standard deviation (μιη) 0.164 0.296 I 0.467 0.634 0.799 L053
附件四A 之垂直掃描間距為〇·1μηι，附件四B之垂直掃描間距為G._ :比較其結果 可觀察出’當深度掃描間距加大時’ _直解析度降低，使聚焦函數在聚 焦深度反應鱗之軸定位上鼓較大的偏差量，因此造成制誤差之增 加，反之亦然 實施例五校正階高之量測結果與分析 為了對本毛明所提供之系統的性能進行驗證與校正，使該量測系統 15 1291013 更具有可信度，本實施例採用TaylorHobson所製造之標準校正塊進行量 /則系統的校正工作’使用50倍物鏡進行量測，F.o.v.為373χ280μιη2，校 正塊階咼為2.37μιη ;附件五Α為該校正塊的實體圖，附件五Β為該校正塊 之篁測剖面圖’附件五C為三維輪廓重建圖，附件五ρ為三維輪廓量測俯 - 視圖，附件六為校正塊階高之全域誤差圖；利用本量測實例進行量測精確 度之分析，得到校正塊階高之全域量測平均誤差為〇 〇7〇3μιη，全域量測誤 差百分比在2.97 %範圍以内。 實施例六半導逋微金凸塊(Micr〇 G〇lden Bump)之量測實例與結果分析 因微金凸塊其表面組織屬較粗糙之待測物，因此不需投影主動式結構 光即可進行量測，且因Micro Golden Bump底部受一金屬平板所承載，因此 不須對其不同區域間進行光強的調整即可完整的觀察其組織變化進行掃 瞒’使用50倍物鏡進行量測，R〇.v·為373χ28〇μπι2 ;附件七a為Micr〇 Golden Bump的實體圖，附件七3為Micro Golden Bump的量測剖面圖，附 •件七C為三維輪廓重建圖，附件七D為三維輪廓量測俯視圖，附件七E為 濾除非聚焦訊號點後的三維輪廓重建圖，附件七F為濾除非聚焦訊號點後 的俯視圖；經實際量測後，評估單一 Micro Golden Bump的長寬約為175μπι χ7〇μηι，兩約為 〇 實施例七系統垂直量測解析度之實測結果與分析 經由貫驗發現，當以〇·〇5μπι為間距進行垂直掃描時，因系統聚焦函數 之解析能力已達極限，造成聚焦深度反應曲線產生大量之高頻雜訊，如附 16 1291013 並非用以限制本發明之專利範圍，凡未脫離本發明技藝精神所為之等饮實 施或變更’例如：光學投影鏡組與顯微系統使用不同的鏡片與排列等變化 之等效性實施例，均應包含於本案之專利範圍中。 上所述本案不但在光學投影系統上確屬創新，並能較習用物品增 進上述夕項功效’應已充分符合新職及進步性之法定發明翻要件，差 依法提出申請，懇請貴局核准本f發明專利申請案，以勵發明，至感德 便。 I【圖式簡單說明】 圖—為本發贼位結構光微三維共絲面輪廓制系統裝置示意圖； 圖二為不同正弦波週期之校正塊階高誤差圖； 圖二為不同主動光強之校正塊階高誤差圖； 圖四為不同垂直掃描間距之校正塊階高誤差圖；以及 圖五為重覆性之量測結果。 .附件一 A為正弦波週期為1〇之校正塊階高全域誤差圖； 附件一B為正弦波週期為7〇之校正塊階高全域誤差圖； 附件二為本發明正弦變化的結構光圖案； 附件二A為影像相對平均灰階強度為84時之校正猶高全域誤差圖； 附件二B為影像相對平均灰階強度為22時之校正塊階高全域誤差圖； 附件四A為垂直掃描間距為〇1哗時之校正塊階高全域誤差圖·， 附件四B為垂直掃描間距為〇 0陣時之校正塊階高全域誤差圖； 附件五A為標準校正塊之實體圖； 19 1291013 附件五B為標準校正塊之量測剖面圖； 附件五C為校正塊階高之三維輪廓重建圖； 附件五D為校正塊階高之三維輪廓量測俯視圖； 附件六為校正塊階高之全域誤差圖； 附件七A為半導體微金凸塊的實體圖； 附件七B為半導體微金凸塊的量測剖面圖； 附件七C為半導體微金凸塊的三維輪廓重建圖； 附件七D為半導體微金凸塊的三維輪廓量測俯視圖； 附件七E為濾除非聚焦訊號點後的三維輪廓重建圖； 附件七F為濾除非聚焦訊號點後的俯視圖； 附件八為垂直掃描間距為0.05μιη時之聚焦深度反應曲線；以及 附件九為垂直掃描間距為〇·1μπι時之聚焦深度反應曲線。 【主要元件符號說明】 1 主控制單元 2 光學投影單元 21 DLP投影機 22 光學投影鏡組 221 光學準直鏡 222 線性偏光板 223 半0柱鏡 3 影像擷取單元 31 顯微系統 311 聚焦物鏡 312 分光鏡 313 光學聚焦透鏡 20 1291013 314針孔 315光學聚焦透鏡 32 CCD顯像感測元件 4精密平移台單元The vertical scanning pitch of Annex A is 〇·1μηι, and the vertical scanning pitch of Annex IV B is G._ : Comparing the results, it can be observed that 'when the depth scanning interval is increased' _ straight resolution is reduced, so that the focusing function is in focus The axis of the deep reaction scale locates the larger deviation of the drum, thus causing an increase in the error, and vice versa. The measurement result and analysis of the fifth step of the correction step are to verify and correct the performance of the system provided by Ben Maoming. In this embodiment, the measurement system 15 1291013 is more credible. In this embodiment, the standard correction block manufactured by Taylor Hobson is used to perform the calibration work of the system. The measurement is performed using a 50-fold objective lens, and the Fov is 373 χ 280 μιη 2, and the block order is corrected. 2.37μιη; Annex 5Α is the solid figure of the correction block, Annex 5Β is the survey profile of the correction block' Annex 5 C is the 3D contour reconstruction diagram, Annex 5 ρ is the 3D contour measurement depth-view, attachment The sixth is to correct the global error map of the block height; using the measurement example to analyze the accuracy of the measurement, the global measurement average error of the corrected block height is 〇〇7〇3μι , Global percentage measurement error is within the range of 2.97%. Example 6 Measurement example and result analysis of semi-conducting micro-gold bumps (Micr〇G〇lden Bump) Due to the rough surface of the micro-gold bumps, the active structure light is not required to be projected. It can be measured, and because the bottom of the Micro Golden Bump is carried by a metal plate, it is not necessary to adjust the light intensity between different areas to completely observe the tissue changes for the broom 'Measure with a 50x objective lens , R〇.v· is 373χ28〇μπι2; Annex VII is the physical diagram of Micr〇Golden Bump, Annex VII is the measurement profile of Micro Golden Bump, attached item 7 C is the 3D contour reconstruction diagram, Annex 7D For the 3D contour measurement top view, Annex VII E is the 3D contour reconstruction map after filtering the focus signal point, and Annex VII F is the top view after filtering the focus signal point; after the actual measurement, the length and width of the single Micro Golden Bump are evaluated. About 175μπι χ7〇μηι, two are about 〇. The measured results and analysis of the vertical measurement of the system of the seventh example. It is found through the static test that when the vertical scanning is performed with the spacing of 〇·〇5μπι, the system focusing function The resolution capability has reached the limit, causing the depth-of-focus response curve to generate a large amount of high-frequency noise. For example, the accompanying claims are not intended to limit the scope of the invention, and the implementation or alteration of the beverage is not departing from the spirit of the invention. The equivalent embodiments of the lens and the microscope system using different lenses and arrangements, etc., should be included in the patent scope of the present application. The above-mentioned case is not only innovative in the optical projection system, but also can improve the above-mentioned effects of the above-mentioned items compared with the conventional items. 'It should be fully in line with the new invention and progressive legal inventions, and the application is submitted according to law. f Invention patent application, in order to invent invention, to the sense of virtue. I [Simplified illustration of the diagram] Figure - Schematic diagram of the light micro-three-dimensional conjugated surface contour system of the thief-position structure; Figure 2 is the error block height error diagram of different sine wave periods; Figure 2 is the different active light intensity Correction block step height error map; Figure 4 is the correction block step height error map for different vertical scan intervals; and Figure 5 is the repetitive measurement results. Annex I A is the correction block height and total error map of the sine wave period of 1 ;; Annex I B is the correction block step height global error diagram of the sine wave period of 7 ;; Annex II is the sinusoidal structural light pattern of the present invention. Annex II A is the corrected Jupiter global error graph when the image relative average gray level intensity is 84; Annex B B is the corrected block height global error map when the image relative average gray level intensity is 22; Annex A is vertical scanning Correction block height and total error map when the spacing is 〇1哗·, Annex BB is the correction block height global error map when the vertical scanning spacing is 〇0 array; Annex 5A is the entity diagram of the standard correction block; 19 1291013 Annex V is the measurement profile of the standard calibration block; Annex 5 C is the 3D contour reconstruction of the correction block height; Annex 5D is the 3D contour measurement top view of the correction block height; Annex 6 is the correction block height The global error map; Annex VII A is the physical diagram of the semiconductor micro-gold bump; Annex VII B is the measurement profile of the semiconductor micro-gold bump; Annex VII C is the three-dimensional contour reconstruction of the semiconductor micro-gold bump; Annex VII Semi-conductive The three-dimensional contour measurement top view of the micro gold bump; Annex VII E is the three-dimensional contour reconstruction map after filtering the focus signal point; Annex VII is the top view after filtering the focus signal point; Annex VIII is the vertical scanning pitch of 0.05 μιη The depth response curve is focused; and the ninth is the depth of focus response curve when the vertical scanning pitch is 〇·1μπι. [Main component symbol description] 1 Main control unit 2 Optical projection unit 21 DLP projector 22 Optical projection mirror group 221 Optical collimation mirror 222 Linear polarizing plate 223 Half 0 cylinder lens 3 Image capturing unit 31 Microscope system 311 Focusing objective lens 312 Beam splitter 313 Optical focusing lens 20 1291013 314 pinhole 315 optical focusing lens 32 CCD imaging sensing element 4 precision translation stage unit