V發明說明： 【發明所屬之技術領域】 本發明侧於-種沿著-表面形麵上之練形成截 面輪廓（line profile)的方法，尤其是一種沿著—正交 表面形貌圖上之截線形成截面輪廓的方法。 【先前技術】 按’利用白光干涉（white light interferometry) 之特性，以非接觸方式對物件表面進行測量之形貌測量裝 置，係廣泛應用於對於精密度有高度要求之產品，諸如： 半導體晶圓、液晶顯示器之玻璃面板等。 請參照第一圖所示，係一典型非接觸式形貌測量裝置 之示意圖。如圖中所示，此形貌測量裝置具有一寬頻光源 、一準直透鏡（collimation j^ns) 20、一 45 度分光鏡 (Splitter) 30、一成像目鏡40、一影像感測裝置50、一 干涉顯微物鏡組60、一掃描平台70與一電腦系統8〇。寬 頻光源10所提供之光線，透過準直透鏡20形成平行光， 投射至45度分光鏡30。此平行光係受到45度分光鏡之反 射30，照射至干涉顯微物鏡組60。 干涉顯微物鏡組60係位於掃描平台70之上方，並且， 對準掃描平台70所承載之待測物體90的表面。此干涉顯 微物鏡組60具有一顯微物鏡62、一反射鏡64與一分光鏡 (beamsplitter) 66。其中，顯微物鏡62係位於反射鏡 64之上方，而反射鏡64係位於分光鏡66之上方。來自45 度分光鏡30之光線，透過顯微物鏡62，照射至分光鏡66， 而被分光鏡66分成兩道光線。其中一道光線係照射至反射 1282407 : 鏡64，而另一道光線則是照射至待測物體9〇表面。 二 此二道光線分別經由反射鏡64與待測物體9〇表面反 射後，投射回分光鏡66再度重合。值得注意的是，由於此 二道光線行經距離之差異（即此二道光線之光程差 (Optical Path Difference, 0PD))，將在重合後之光線 内產生干涉效果。而此光線再向上投射，經過顯微物鏡62 與45度分光鏡30，最後，透過成像目鏡4〇聚焦於影像感 測裝置50 〇 Ψ 前述二道光線之光程差的大小，係受到干涉顯微物鏡 組60與掃描平台70之距離的影響。因此，藉由改變干涉 顯微物鏡組60或是掃描平台70之垂直位置，即可改變光 程差之大小，而使影像感測裝置50接收到一系列具有不同 · _ 光程差之干涉影像。透過電罈系终8〇分析，計算此一系列 干涉影像中，各個晝素位置在不同之干涉影像中的光強 度，即可形成如第二圖之干涉圖譜。此干涉圖譜係一理想 之干涉圖譜，透過計算此干涉圖譜中之波包的峰值，即可 • 推導出零光程差之位置，以確認此畫素位置所對應之高 ， 度。以同樣的方式，計算其他晝素位置所對應之高度，即 可得到此待測物體9〇之表面形貌。 值得注意的是，由於影像感測裝置5〇之感測元件（未 圖不）係成陣列排列，透過感測元件所攫取之干涉影像的 資料型態，必然係屬二維正交分布之資料類型。也因此， 如第三圖所示，依據此二維正交分布之資料所建構的表面 形貌圖’係由正交排列之方格（晝素）所構成，而各個方 格以不同之色彩或是圖案，表現其所對應之高度值。 责二;而，、在實際之應用上’例如：光柵、玻璃基板上之 畫素疋件等，通常都能以截面輪#呈現沿著特定截線 方向之高低起伏變化。由於前述影像感職置5α_致之 資料型態，僅能直接推算出正交方向之截面輪廓^是所 欲觀察之截線（如圖中IS2線段）的方向與正交方向（即 X軸與Y輛之方向）有所偏離，則必織轉待測物體，使 截線之方向與正交方向—致，方能獲得截面輪廓。 為了克服此問題，如第三圖所示，對於一給定截線 S1 -S2’一習知方法係依據此截線之起點幻與 表面形_上的鋪錄，得出—直線絲式代表此截線 si-S2。然後’將不同之χ座標值分別代入此直線方程式 中’以得出相對應之γ座標值。此γ座標值，係用以判斷 在不同之X座標處，依據表面形貌圖上哪一個方格所對應 之咼度值，表示截面之高度值Ί _ 值知注意的是，依據此直線方程式所計算出之γ座標 值往往並非整數，而造成方格判定上之困難。同時，由於 截線S1 -S2之長度係大於其投射於χ座標上之長度。然而， 透過代入不同之X座標值所獲致之資料點數（即圖中之粗 框方格數），鱗同於此齡投射於χ紙之線段長度所對 應之資料點數。此資料點數與截線S1—S2之長度無法匹 配，而難以表達真實之截面輪廓。 爰疋，就一已知之正交表面形貌資料圖中，如何迅速 且準確地獲彳于任意方向之截線輪廊，對於非接觸式形貌量 測裝置之應用，將會產生顯著之影響。 【發明内容】 1282407 辉/月yi日細j正替換3 不僅Γ _提供—獅賴面麵之方法， 上的需求二截面輪廊’同時還可以兼顧運算速度 面!^7知之表面形細而言’本翻所提供之形成截 分別形成-平行於此截線之虛擬線。然後，Γ表面:’ 長為單位，於此二虛= 「虛擬區段___彳_ ^ = 個方格所對庫之古疮姑、，, X我度乘上各 加權f㈣亚虛_段域線之距離 加推5十异後，以估算此待測點之高度值。 所附^ ί㈣之伽觸神可以㈣町的發明詳述及 所附圖式得到進一步的瞭解。 4及 【實施方式】 方法二^四Α至D圖所示’係本發明形成截面輪摩之 貌圖，tr之示意圖。第四A圖係一已知之表面形 a a線段則是欲形成截面輪摩（line profile) 形貌圖係劃分為多個正交分布之晝素方 。"各個旦素方格分別有其對應之數值以表示高度值。 $ ’如第四Β圖所示，沿著㈣截線之走向，建立 貫穿1==(㈣中以虛線表示之方格）’而a_a戴線係 於义查7方格之中央位置。待測方格的邊長L係等同 素方格（即圖中以實線表示之方格，即原始方格 ^#邊長S。* a__a截線上之係以制方格之邊長^ ’、、、早立’劃分出多個待测資料點PI，P2, P3, P4。此a-a戴 8 1282407V invention description: [Technical field to which the invention pertains] The present invention is directed to a method of forming a line profile along a surface profile, in particular, along a quadrature surface topography A method of forming a cross-sectional profile by a stub. [Prior Art] A topography measuring device that measures the surface of an object in a non-contact manner by using the characteristics of white light interferometry is widely used for products with high precision, such as: semiconductor wafer , glass panel of liquid crystal display, etc. Please refer to the first figure for a schematic diagram of a typical non-contact topography measuring device. As shown in the figure, the topography measuring device has a broadband source, a collimation lens 20, a 45 degree splitter 30, an imaging eyepiece 40, an image sensing device 50, An interference microscope objective set 60, a scanning platform 70 and a computer system 8A. The light supplied from the broadband source 10 is formed into parallel light through the collimator lens 20 and projected onto the 45-degree beam splitter 30. This parallel light system is reflected by a 45 degree beam splitter 30 and is incident on the interference microscope objective set 60. The interference microscope objective set 60 is located above the scanning platform 70 and is aligned with the surface of the object 90 to be tested carried by the scanning platform 70. The interference microscope lens set 60 has a microscope objective 62, a mirror 64 and a beam splitter 66. The microscope objective 62 is located above the mirror 64, and the mirror 64 is located above the beam splitter 66. The light from the 45-degree beam splitter 30 is transmitted through the microscope objective 62 to the beam splitter 66, and is split into two beams by the beam splitter 66. One of the rays illuminates to reflect 1282407: mirror 64, and the other illuminates the surface of the object to be tested. The two rays are reflected by the mirror 64 and the surface of the object to be tested 9 respectively, and then projected back to the beam splitter 66 to overlap again. It is worth noting that due to the difference in the distance between the two rays (i.e., the optical path difference (OPD) of the two rays), interference will occur in the overlapping rays. The light is then projected upwards, through the microscope objective 62 and the 45-degree beam splitter 30, and finally, through the imaging eyepiece 4, focusing on the image sensing device 50 〇Ψ the optical path difference of the two rays is subjected to interference display. The effect of the distance between the micro-objective lens set 60 and the scanning platform 70. Therefore, by changing the vertical position of the interference microscope objective group 60 or the scanning platform 70, the optical path difference can be changed, and the image sensing device 50 receives a series of interference images having different _ optical path differences. . Through the analysis of the final 8 电 of the electric circuit, the interference intensity of each element in the interference image is calculated in the series of interference images, and the interference spectrum as shown in the second figure can be formed. The interferogram is an ideal interferogram. By calculating the peak value of the wave packet in the interferogram, the position of the zero optical path difference can be derived to confirm the height and degree of the pixel position. In the same way, the height corresponding to the position of the other elements is calculated, and the surface topography of the object to be tested is obtained. It is worth noting that since the sensing elements (not shown) of the image sensing device 5 are arranged in an array, the data type of the interference image captured by the sensing element is necessarily a two-dimensional orthogonal distribution data. Types of. Therefore, as shown in the third figure, the surface topography map constructed according to the data of the two-dimensional orthogonal distribution is composed of orthogonally arranged squares (different elements), and each square has a different color. Or a pattern that expresses its corresponding height value. Secondly, in practical applications, such as gratings, pixel elements on glass substrates, etc., the cross-section wheel # can usually exhibit fluctuations along the height of a particular section. Due to the above-mentioned image sensory 5α_ data type, it is only possible to directly calculate the cross-sectional profile of the orthogonal direction ^ is the direction of the intercept line (the IS2 line segment in the figure) and the orthogonal direction (ie the X-axis) If there is a deviation from the direction of the Y vehicle, the object to be measured will be woven, so that the direction of the intercept line and the orthogonal direction will be obtained to obtain the cross-sectional contour. In order to overcome this problem, as shown in the third figure, a conventional method for a given cut line S1 - S2' is based on the starting point of the cut line and the surface shape _, which is obtained - a straight line representation This section line si-S2. Then 'substituting different χ coordinate values into this linear equation respectively' to obtain the corresponding γ coordinate value. The gamma coordinate value is used to determine the height value of the cross section according to the square value corresponding to the square topography on the X coordinate, and the value of the cross section is _ _ value, according to the linear equation The calculated gamma coordinate value is often not an integer, which makes it difficult to determine the square. At the same time, the length of the cut lines S1 - S2 is greater than the length projected on the stile. However, the number of data points obtained by substituting different X coordinate values (i.e., the number of square boxes in the figure) is the number of data points corresponding to the length of the line segment projected on the crepe paper at this age. The number of points of this data cannot match the length of the cut lines S1 - S2, and it is difficult to express the true cross-sectional profile.爰疋, in a known orthogonal surface topography data map, how to quickly and accurately obtain the intercepting wheel gallery in any direction will have a significant impact on the application of the non-contact topography measuring device. . [Summary of the Invention] 1282407 Hui / Yue yi day fine j is replacing 3 Not only Γ _ Provided - the method of the lion's face, the demand for the two-section porch 'at the same time can also take into account the speed of the surface! ^7 know the surface is thin The formation of the section provided by the present section is formed separately - a virtual line parallel to this section line. Then, Γ surface: 'long as unit, here two virtual = "virtual segment ___彳_ ^ = squares of the library of the ancient sore,, X, I multiply each weighted f (four) sub-virtual _ After the distance of the segment field line is increased by 5, the height value of the point to be measured is estimated. The attached ^ ί (4) can be further understood by the details of the invention and the drawings. Embodiments] Method 2^4Α to DFig. is a schematic diagram of the cross-sectional wheel of the present invention, and a schematic diagram of tr. The fourth A-picture is a known surface-shaped aa line segment to form a cross-section wheel (line) Profile) The topographic map is divided into a number of orthogonal distributions of the prime side. " Each denier square has its corresponding value to represent the height value. $ 'As shown in the fourth diagram, along the (four) section The direction of the line is established through the 1== (the square indicated by the dotted line in (4)) and the a_a line is in the central position of the square of the 7th square. The side length L of the square to be tested is equivalent to the square (ie In the figure, the square is indicated by the solid line, that is, the original square ^# side length S. * The a__a line is divided into a plurality of sides of the square ^ ', ,, and early ' Measurement data points PI, P2, P3, P4. This a-a wear 81282407
u係位於其中一待測方格 個待測資料點pi P2 Pq W在八σ, 51 、/ ’Ρ3，Ρ4係分別位於所對應之待測方格 的中央位置。並且，待丨資# 格之數目 行列貝枓點之數目，係等同於待測方 然後’如第四c圖所示，於待測方格内，形成至少一 虛擬線（圖中係於a—A7 ㈣Γ 截線兩侧，分別形成一平行於此 a-a截線之虛擬線V1與V2 )。虛擬㈣與v2係平行於㈣ 截線。。此二虛擬線V1，V2亦係貫穿關細方格，並且，The u system is located in one of the squares to be tested. The data points to be tested are pi P2 Pq W at eight sigma, 51 , / ’ Ρ 3 , and the Ρ 4 lines are respectively located at the center of the corresponding square to be tested. And, the number of the number of rows of the waiting grid is equal to the number of the beauties, and then, as shown in the fourth c diagram, at least one virtual line is formed in the square to be tested (the figure is tied to a —A7 (4) 两侧 On both sides of the line, a virtual line V1 and V2 parallel to the aa line are formed respectively. The virtual (four) and v2 systems are parallel to the (iv) cut line. . The two virtual lines V1 and V2 also run through the closed squares, and
分別被廷些待測方格齡為多個虛擬線區段R1，R2。這些 劃分出來之虛擬線區射的長度係等同於待測方狀 邊長s，亦係等同於表面形貌圖上之晝素方格的邊長l。各 ^職料點P1，P2, P3, P4係位於相對應之二虛擬線區段 Kl, R2所構成之四邊形的中央位置 隨後，如細D ®所示’由待财格糊分出之虛擬 線區段Rl，R2，又可區分為多個段落Αβ，Βς⑶，EF分別落 於不同之畫素方格（原始方格）内。將這些段落之長度（即 • 纟虛舰段落於各俩始方格㈣長度），乘JL各個晝素方 t (原始方格）所對應之高紐，加權即可估算各個待侧 方才。所對應之回度值。就一較佳實施例而言，以此虛擬線 區段R1，R2與a-a截線之距離加權後，即可估算各個待測 方格（即各個待測資料點P1，P2, P3, P4)所對應之高度值。 舉例而言，假定虛擬線VI，V2與a-a截線之距離d均 為待測方格之邊Μ _分之—。在估算待測方格所對應 之j度值的過程中，虛擬線VI與V2係具有相同之權重。 如第四D圖所示，虛擬線區段R1係被劃分為段落 1282407 朴年丨月0日修(更，正替換j AB，BC，CD，虛擬線區段R2係被劃分為段落ef ;其中，段 落AB係位於晝素方格B1内，段落BC係位於晝素方格β2 内，段落CD係位於晝素方格B3内，段落EF係位於畫素方 格B2内。各個段落AB，BC，CD, EF之長度，決定了各個晝素 方格Bl，B2, B3所對應之高度值，對於此待測方格之高度值 的貢獻程度。 因此，待測方格（待測資料點）所對應之高度值Η : H = [(h(Bl)x l(ABh h(B2)x (l(BC^l(EF))+h(B3)x l(CD))] ^ (l{Rl)+l(R2)) 其中，h(Bl)，h(B2)，h(B3)分別為方格B1，B2，B3m 對應之高度值，·1(ΑΒ)，1(BC)，1(CD)，1(EF)分別是段落 AB，BC，CD，EF之長度；而1(R1)，1(R2)分別是虛擬線區段 R1與R2之長度。 換吕之，各個待測方格（待測點）所對應之高度值， 係將相對應之一戎虛擬區段Rl，R2，落於各個原始方格 B1，B2，B3内之長度，乘上各個原始方格βι，Β2 Β3所對應 之高度值，加總後除以二虛擬區段R1，R2之總長度。 雖然上述實施例係利用二虛擬線V1與V2，以獲得截 面輪廓資料。然亦不限於此。本發明之主要特徵在二，利 用虛擬線之建立，以決定各個畫素方格對於此截線上各個 資料點的貢獻度，藉以獲得截面輪_料。因此，不限於 使用二虛擬線。惟，截線之兩側最好分別具有至少一虛擬 線。其次’各個虛擬線與截線之距離不宜太大。就一較佳 實施例而言，虛擬線與截線之距離最好是小於待測方格之 邊長的-半，以確健躲區段触紅晝素雜，係與 截線最為接近之晝素方格。 ^ 1282407 進一步來說，同時請參照第四八至1)囝 以在a-a截線上，以原始方格之邊長為單可 複數個待測點P1，P2，P3，P4，而非如第四B 釗分出 列待測方格。本實施例可以於此a〜a截線^不建立— 成至少一平行於截線之虛擬線VI，V2。接下來，貝’、’別形 擬線VI，V2上分別劃分出複數個虛擬區則於^些虛 應於截線上各個待測點P1，P2, P3, P4。最’刀別對 區段Rl，R2落於各個原始方格Β1，β2, B3 e 虛擬The squares to be tested are respectively a plurality of virtual line segments R1, R2. The length of the divided virtual line region is equivalent to the square length s of the square to be tested, which is equivalent to the side length l of the elementary square on the surface topography. The respective material points P1, P2, P3, and P4 are located at the central position of the quadrilateral formed by the corresponding two virtual line segments K1, R2, and then, as shown by the thin D ® 'virtual The line segments R1, R2 can be further divided into a plurality of paragraphs Αβ, Βς(3), and EF respectively fall in different pixel squares (original squares). The length of these paragraphs (ie, the length of the virtual ship's paragraph in the first square (four)), multiplied by the high-intensity corresponding to each of the JL's squares (original squares), can be weighted to estimate each side. The corresponding return value. In a preferred embodiment, after the distance between the virtual line segments R1, R2 and the aa line is weighted, each square to be tested can be estimated (ie, each of the data points to be tested P1, P2, P3, P4). The corresponding height value. For example, suppose the distance d of the virtual line VI, V2 and the a-a cut line is the edge of the square to be tested. In the process of estimating the j-degree value corresponding to the square to be tested, the virtual line VI and the V2 system have the same weight. As shown in the fourth D diagram, the virtual line segment R1 is divided into paragraphs 1282407, and the replacement of j AB, BC, CD, and the virtual line segment R2 is divided into paragraphs ef; Among them, the paragraph AB is located in the alizarin square B1, the paragraph BC is located in the alizarin square β2, the paragraph CD is located in the alizarin square B3, and the paragraph EF is located in the pixel square B2. The length of BC, CD, and EF determines the height value corresponding to each of the squares B, B2, and B3, and the contribution to the height value of the square to be tested. Therefore, the square to be tested (the data point to be tested) The corresponding height value Η : H = [(h(Bl)xl(ABh h(B2)x (l(BC^l(EF))+h(B3)xl(CD))] ^ (l{Rl ) +l(R2)) where h(Bl), h(B2), h(B3) are the height values corresponding to squares B1, B2, and B3m, respectively, ·1(ΑΒ), 1(BC), 1( CD), 1 (EF) is the length of paragraphs AB, BC, CD, EF, respectively; and 1 (R1), 1 (R2) are the lengths of the virtual line segments R1 and R2, respectively. The height value corresponding to the grid (to be measured) is the length of one of the corresponding virtual segments R1, R2, which falls within each of the original squares B1, B2, and B3. Multiply the height values corresponding to the respective original squares βι, Β2 Β3, and add up to the total length of the two virtual segments R1, R2. Although the above embodiment utilizes the two virtual lines V1 and V2 to obtain the cross-sectional profile. The data is not limited to this. The main feature of the present invention is that the creation of virtual lines is used to determine the contribution of each pixel square to each data point on the cut line to obtain a cross-section wheel. Therefore, no It is limited to use two virtual lines. However, it is preferable that both sides of the cut line have at least one virtual line respectively. Secondly, the distance between each virtual line and the cut line is not too large. In a preferred embodiment, the virtual line and the cut line The distance is preferably - half of the length of the side of the square to be tested, so as to ensure that the hiding section touches the red sputum, and the line is the closest to the cut line. ^ 1282407 Further, please refer to The fourth eight to one) 囝 in the aa section, the length of the original square can be a single number of points to be measured P1, P2, P3, P4, instead of the fourth square 钊 points out the square to be tested This embodiment can be used for the a~a cut line ^ not established - at least one parallel The virtual line VI, V2 of the cut line. Next, the multiple virtual areas are divided into multiple virtual areas on the 'B', 'Dimensional line VI, V2, and then some of the points to be measured P1, P2, P3 on the cut line. P4. The most 'cutting pair' section Rl, R2 falls on each original square Β1, β2, B3 e virtual
rRm’B3所對應之高度值，並^虛= 幻’^2與a-a截線之距離加權後，估算待測點朽，段，⑺ 此外，同時請參照第四AB圖，本發明亦可 P2，物咖。本實施例 #==； ，分卿成—传於a~a截線之 f擬區段R1，R2。然後’將二虛擬區段R1，R2落 :3B1f長度，乘上各姆始方格 子應之阿度值，亚以該二虛擬區段R1，R2與& 距離加權後，估算此待測點P2之高度。 、 值得注意的是，本發明以虛擬線區段R1，R2上各段落 之長度，決定各個晝素方格，對於制資料點p2之高度= 所佔之權重’只涉及-維（即長度）之計算。相較之下， ，過計算待測方格落於各個晝素方格内之面積大小，以決 疋各個畫素方格所佔之權重的方法，賴可能有較佳之精 確度。但是，由於此方法必須進行二維（即面積）之計算， 將會造成計算4之大幅增加，而導錢算速度明顯下降。 1282407 日修;I止替誕g 區“ 示之f知方法’本發明以虛擬線 声°，又尤合士上各段落之長度，決定各個晝素方格之貢獻 ^明^有如習知方法中，方格判定上之困難。同時，本 ί方㈣以兼顧待測資料點Ρ1，Ρ2,Ρ3,Ρ4周圍各個晝 者之1面輪廊估算上之貢獻度，因此，較能求得直 p"l Ρ2 pt p7f 5 位’於伽數目’係制独晝素方格之邊長為單rRm'B3 corresponds to the height value, and ^ imaginary = illusion ^ ^ 2 and aa section distance weighted, estimate the point to be measured, segment, (7) In addition, please refer to the fourth AB diagram, the invention may also P2 , coffee. In this embodiment, #==; , 分卿成- is transmitted in the a~a section of the f-segmented section R1, R2. Then 'the two virtual segments R1, R2 fall: 3B1f length, multiplied by the Adu value of each m-square, and the second virtual segment R1, R2 and & distance weighted, the estimated point to be measured The height of P2. It should be noted that, in the present invention, each pixel square is determined by the length of each paragraph on the imaginary line segments R1, R2, and the height of the data point p2 = the weight occupied by only the - dimension (ie length) Calculation. In contrast, the method of calculating the size of the squares to be tested in each element square to determine the weight of each pixel square may have better precision. However, since this method must be calculated in two dimensions (ie, area), it will cause a large increase in the calculation 4, and the speed of the money guiding calculation is significantly reduced. 1282407 日修; I stop for the birth of the g area "show the f know method" The invention uses the virtual line sound °, and the length of each paragraph of the special paragraph, determines the contribution of each element square ^ ^ ^ as the conventional method In the middle, the square is difficult to judge. At the same time, this 395 (4) takes into account the contribution of the data points of the data to be tested Ρ1, Ρ2, Ρ3, Ρ4, and the evaluation of the one-sided porch. Therefore, it is more straightforward. p"l Ρ2 pt p7f 5 digits 'the number of gamma' system is the side of the square
中，來成截面ΐ斤产劃分之區段數目，因而不會有如習知方法 $成截面輪廓之資料點數與截線長度無法匹配之問題。 制太’施例詳細說明本發明’而非限 卷月之_，而且熟知此類技 ，的改變及調整，仍將不失本發明之要義:在； 不脫離本發明之精神和範圍。 受我所在’亦 【圖式簡單說明】 圖係-典型非接觸式形貌測量裝置之示意圖。In the middle, the number of sections of the section is divided into sections, so that there is no problem that the number of data points and the length of the section cannot be matched as in the conventional method. The present invention is intended to be illustrative of the present invention, and is not to be construed as limiting the scope of the invention. According to my location, the figure is a schematic diagram of a typical non-contact topography measuring device.
第二圖係一理想上之干涉圖譜。 第三圖係一典型表面形貌圖， 計算截面輪腐之方法。 ΜΙΠ時顯不一傳統 第=A至D圖係本發明形成戴面輪廓方法—較佳實於 例之不意圖。 貝也 【主要元件符號說明】 寬頻光源10 準直透鏡20 45度分光鏡30 12 1282407 了蛑f月4曰修(更)正替換頁 成像目鏡40 干涉顯微物鏡組60 掃描平台70 電腦系統80 影像感測模組50 待測資料點P1，P2，P3，P4 虛擬線V1，V2 區段R1，R2 畫素方格B1，B2，B3The second picture is an ideal interference spectrum. The third figure is a typical surface topography diagram, and the method of calculating the cross-section rot. ΜΙΠ 显 显 第 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 Bei also [main component symbol description] Broadband light source 10 Collimating lens 20 45 degree beam splitter 30 12 1282407 蛑f月4曰修(more) replacement page imaging eyepiece 40 interference microscope objective group 60 scanning platform 70 computer system 80 Image sensing module 50 data points to be tested P1, P2, P3, P4 virtual line V1, V2 section R1, R2 pixel squares B1, B2, B3
1313