TW200923837A - Method for constructing the image of structures - Google Patents

Method for constructing the image of structures Download PDF

Info

Publication number
TW200923837A
TW200923837A TW96145080A TW96145080A TW200923837A TW 200923837 A TW200923837 A TW 200923837A TW 96145080 A TW96145080 A TW 96145080A TW 96145080 A TW96145080 A TW 96145080A TW 200923837 A TW200923837 A TW 200923837A
Authority
TW
Taiwan
Prior art keywords
image
steps
reconstruction method
following
tested
Prior art date
Application number
TW96145080A
Other languages
Chinese (zh)
Other versions
TWI342524B (en
Inventor
Yan-Jen Su
Shih-Hsuan Kuo
Original Assignee
Ind Tech Res Inst
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Ind Tech Res Inst filed Critical Ind Tech Res Inst
Priority to TW96145080A priority Critical patent/TWI342524B/en
Publication of TW200923837A publication Critical patent/TW200923837A/en
Application granted granted Critical
Publication of TWI342524B publication Critical patent/TWI342524B/en

Links

Landscapes

  • Image Analysis (AREA)
  • Length Measuring Devices By Optical Means (AREA)

Abstract

A method for constructing the image of structures is provided in the present invention. The method is to obtain a plurality of images corresponding to the different image-acquiring angle of a sample through an angle-adjustable profilometer. Then, each image is processed through an aligning procedure for compensating the error caused by the angle-adjustable profilometer while changing image-acquiring angle. Finally, an overlapping step is processed for obtaining a surface profile image corresponding to the sample. Since the present invention utilizes a simple algorithm to process the images, it is capable of reducing the calculating time cost and enhancing the precision for understanding characteristics of the structure of the sample.

Description

200923837 九、發明說明: 【發明所屬之技術領域】 尤其疋指一種影像 本發明係有關一種影像處理方法 重建方法。 【先前技術】 在微小結構物的量測上,光干涉技術因為具有快速、 =度高、以及非破壞性檢_優點,而f遍地被應用。 =對於傾斜度過高的表面形貌,則會因為反射光的角户 =法=備:確地偵測到,而發生量測資“ ’係為f知光學顯微鏡以同轴 二=㈣’反射絲法反射回光學顯微鏡之結構示咅 地表面斜率較大時,尤其是當光學顯微; 法進入會因為待測物12表面反射光訊號13益 彳 頰微知,而無法取得待測物表面形貌資料。 動至,是將待測物或量測儀器位移、並擺 位移行測量,而得到多張量測影像;然後由 據彼此的:連張量測影㈣空間關係;再根 境如:旋轉中、、=取後的結果。前述技術之量測環 三維重建運^ ^u+徑、旋轉角度等,都必需提供給 所揭4如.美國專利US.Pat.No.5,940,181號, = 於非球面待測物的量測技術就⑽ 作供的環境。採料種對位方式的需求是:量測儀 200923837 …須要能精確地提供諸如位移量及擺㈣度等各項環境 . 變數,否則,一點微小的誤差就會讓影像在疊合時,、因為 空間座標的轉換而造成極大的偏移 。不過前述之方式,二 • #㈣㈣裝置以將量測時儀器的誤差限^在—定的範圍 . 二内::,有_環境參數,但是如此-來儀器的成 本和複雜度就會有所增加。 另外’在美國專利US. Pat. No. 6, 449, 048號所揭露的 〇 料,則是將干涉儀傾斜—肖度,使其與制物橫移方向 不成垂直’然後直接以傳統上垂直掃描干涉儀(VSI)、或 相移干涉儀(PSI)的硬體,連續掃描制物表面。此方法 不需要使用影像疊合的技術即可取得表面形貌,但是沒有 - 解決當形貌的傾斜度過大時,資料取得的問題。 ' θ另外,在影像疊合的演算上,Besl andMcKay於1992年 提的 Iterative Closest point(Icp)是一種具有普遍性 Ϊ演算方式’它以疊代(iteratiQn)的方式不斷地重覆比對 〇 貞料,直賴誤差㈣至—個相接受的敎值為止。但 是根據Rusinkiewicz所做的分析,貯仍有其極限存在, 例如沒有㈣特徵的形貌就會造成它運算時的不穩定性, 使得演算發散’誤差無法收斂。 【發明内容】 ' 本發明之影像重建方法係擷取關於待測物之不同角卢 之影像資訊,再利用軟體影像對位與疊合技術,重建出^ 測物原始形貌,取得待測物實際形貌尺寸及其粗縫度。 200923837 本發明之影像重建方法利用簡單之演 表面形貌,不但可以縮短演算時門、,# 、制重建影像 之結構特徵的精確度。 、並且提尚掌握待測物 本發明之影像重建方法可以斜 — 大型待測物,取的不同方位之影傻^热法傾斜或旋轉的 技術解決檢測表面形貌的問題。 错由衫像對位與疊合 有下法,包括 正衫像,(C)重複步驟(a)至⑹至少—次 、=凋 整影像,·以及⑷疊合該複數個 =後數個調 重建影像。 正〜像以仵到該待測物之 在另-實施财,本發明提供— =二n,(a;r影像·角度以對-待=取- _該:=::==之複數個微結構; 你 τ衫像调整程序以得刭一哺敕與 像’該調整程序包括有:(b = 5周正〜 該初始影# $& , 、 車疋轉角度並對 轴旋轉= 之旋轉調整;以及⑽決定-第二 轉^以二ΐ對步驟(bl)所得到之影像進行第二軸之旋 ㈣影像;⑹錢㈣⑷至⑹至少一 【實施方式】 貝番查委員能對本發明之特徵、目的及功能有 200923837 更進-步的認知與瞭解,下文特將本發明 部結構以及設計的理念原由進行說明,以使得J = 以了解本發明之特點,詳細說明陳述如下:—委貝可200923837 IX. Description of the invention: [Technical field to which the invention pertains] In particular, the invention relates to an image processing method reconstruction method. [Prior Art] In the measurement of minute structures, the optical interference technique is applied because of its fast, high degree, and non-destructive inspection advantages. = For the surface topography with too high inclination, it will be because the angle of the reflected light = method = ready to be detected, and the amount of measurement is "following the optical microscope with the coaxial two = (four)' The structure of the reflection wire method reflected back to the optical microscope shows that the slope of the surface of the ground surface is large, especially when the optical microscope is used; the method enters because the surface of the object to be tested 12 reflects the light signal 13 and can not obtain the object to be tested. The surface topography data is moved to measure the displacement of the object to be tested or the measuring instrument, and obtain a plurality of measurement images; and then according to each other: the tensor measurement (4) spatial relationship; The environment is as follows: the results of the rotation, the = after the removal. The three-dimensional reconstruction of the measurement ring of the above-mentioned technology, ^ ^ + diameter, rotation angle, etc., must be provided to the disclosure of 4 US Patent. Pat. No. 5,940,181 , = The measurement technique of the aspherical object to be tested is (10) The environment for the supply. The demand for the alignment method of the seed is: the measuring instrument 200923837 ... needs to accurately provide various environments such as displacement and pendulum (four) degrees. Variable, otherwise, a slight error will make the image superimposed, The conversion of the space coordinates causes a great offset. However, in the foregoing way, the two #(4)(4) devices are used to limit the error of the instrument during the measurement to a certain range. Two::, there are _ environmental parameters, but so - The cost and complexity of the instrument will increase. In addition, the material disclosed in U.S. Patent No. 6,449,048 is the tilting of the interferometer. The object traverse direction is not vertical' and then directly scans the surface of the workpiece with the hardware of a conventional vertical scanning interferometer (VSI) or phase shifting interferometer (PSI). This method does not require the use of image overlay technology. Get the surface topography, but not - solve the problem of data acquisition when the slope of the topography is too large. ' θ In addition, in the calculation of image overlay, the Iterative Closest point (Icp) proposed by Besl and McKay in 1992 is a kind It has a universal Ϊ calculus method. It continually repeats the comparison data in an iteratiQn manner, depending on the error (4) to an acceptable threshold. However, according to the analysis made by Rusinkiewicz, the storage still Have their poles Existence, for example, without the feature of (4) features will cause instability in its operation, so that the calculation divergence 'error can not converge. 【 SUMMARY OF THE INVENTION The image reconstruction method of the present invention draws different angles about the object to be tested. The image information is then reconstructed by the software image alignment and superposition technique to reconstruct the original shape of the object and obtain the actual shape and the roughness of the object to be tested. 200923837 The image reconstruction method of the present invention utilizes a simple surface. The shape can not only shorten the calculus time, the accuracy of the structural features of the reconstructed image, but also the master of the object to be tested. The image reconstruction method of the present invention can be oblique-large object to be tested, taking different directions. Shadow silo ^ Thermal tilt or rotation technology solves the problem of detecting surface topography. Wrong by the shirt like the alignment and superposition, including the shirt image, (C) repeat steps (a) to (6) at least - times, = full image, · and (4) superimpose the plural = the last few Rebuild the image. The positive ~ like to 仵 to the object to be tested in another - implementation of the money, the present invention provides - = two n, (a; r image · angle to - to = take - _ the: =:: = = a plurality of Microstructure; your tau shirt is like a tweaking program to get a squat with like the 'adjustment program included: (b = 5 weeks positive ~ the initial shadow # $&, rut angle and axis rotation = rotation And (10) the decision - the second turn is to perform the second axis rotation (four) image on the image obtained by the step (bl); (6) the money (four) (4) to (6) at least one [embodiment] Features, purposes, and functions have a further step-by-step understanding and understanding of 200923837. The following is a description of the structure of the invention and the concept of the design, so that J = to understand the characteristics of the present invention, the detailed statement is as follows: can

請參閱圖二所示,該圖係為本發明影像 =示意圖:該影像重建方法2包括有下列步驟^ 公:乂::0 ’5周整影像擷取角度以對一待測物擷取-初妒 f象。在本步財,歸勒上财微結構,該微社射 t週期性微結構,但不以此為限。麻該待測物之影像所 三二可以調整影像擷取角度之機構,其係可為圖 置31右:,β之機構。以圖三八為例,該影像擷取裝 -Η、有—影賴取單元3G,其上具有—導姉31以及 «軸32。在該影像擷取單元3〇之一側設置 ^ 33,其係可進行-旋轉運動。在該旋轉減33上疋= 相^個滑軌^與35,其中之一滑軌34係與該導執槽31 π ’而另一滑執35係與該樞轴32相樞接。藉由改變Referring to FIG. 2, the figure is the image of the present invention=schematic diagram: The image reconstruction method 2 includes the following steps: 乂::0 '5 weeks of the entire image capturing angle to capture a sample to be tested - The first time f icon. In this step, the financial micro-structure is attributed to the micro-structure, but it is not limited to this. The image of the object to be tested can adjust the angle of the image capturing angle, and the system can be the right of the figure: 31, the mechanism of β. Taking Figure 38 as an example, the image capture device has a -3, which has a guide 31 and a «axis 32. On one side of the image capturing unit 3 is set to ^33, which is capable of -rotating motion. In the rotation minus 33, 相 = phase slides ^ and 35, one of the slide rails 34 is coupled to the guide groove 31 π ' and the other slide 35 is pivotally coupled to the pivot shaft 32. By changing

像執34與35之滑執桿間的相對位置可以調整該影 像擷取早凡30之傾斜角度。 凊參閱圖三Β所示,在本實施例之影像擷取裝置3,具 操取單元3〇,其兩侧分別具有至少—導引塊%、。 單元3〇之一側具有一旋轉車由體37,該旋轉 传盘兮旦之兩側分別連接有一滑執槽38。每一個滑執槽38 二、彡像擷取單元3〇兩側之導引塊36相滑接,使得該 ^取單元30可以改變傾斜角度。步驟20影像擷取之 又如圖三C所示,其所得到之影像如圖四所示。 再回到圖二所示,得到該初始影像之後,接著進行步 200923837 驟21,對該初始影像進 里 像。本步驟之用咅在此^ μ正耘序以得到一調整影 a或者是圖三21之演算法彌補利㈣^ 差。等差所4 榻取裝置在操取影像時所造成之举 左决差所造成之原因, 丨k取芝决 影像的角度時,機構M = H〜像擷取單元之揭取 的干擾,如费動ΪΓΓ身所造成之誤差外,外部環境 搌勁4’也會影響擷取影 由本發明之步驟21調整所擷取之影像。'又。因此’藉The relative position between the sliders of the 34 and 35 can be adjusted to capture an angle of 30. As shown in FIG. 3A, the image capturing device 3 of the present embodiment has a processing unit 3, which has at least one guiding block % on both sides. One side of the unit 3 has a rotating vehicle body 37, and a sliding groove 38 is connected to each side of the rotating disk. Each of the sliding grooves 38 and the guiding blocks 36 on both sides of the image capturing unit 3 are slidably connected, so that the drawing unit 30 can change the inclination angle. Step 20 image capture is also shown in Figure 3C, and the resulting image is shown in Figure 4. Returning to Figure 2, after obtaining the initial image, proceed to step 200923837, step 21, to enter the image into the initial image. This step is used to obtain an adjustment shadow a or the algorithm of Figure 3 21 to make up the difference. The cause of the left-handed difference caused by the 4th reclining device in the operation of the image, 丨k takes the angle of the image of the image, the mechanism M = H~ the interference extracted by the capturing unit, such as In addition to the errors caused by the slamming of the body, the external environment 4 4' will also affect the image captured by step 21 of the present invention. 'also. So borrow

C 性微膜在該待測物係為具有週期 待測物Π 為了方便說明對位校正之方式,在兮 寺川物上建立一座標系以利說明。接著, 一μ 該圖係為本發明影像敗 > 〗圖所示, :,影像調整程 Π圖角度^該初始影像進行第-軸之旋轉調 利用-轉換方式,尋找出該初始影像 該待測物係為增亮膜,因此太竇二:特徵衫像。由於 Ο 增亮膜内微結構之邊界。至於特:影像 :=:有所改變’並不以本發明所述Si: 二所示’該尋找該初始影像内之特徵影 微υί'Γ中’每—個斜線9G代表著增亮膜内 步驟2100所述之轉換方式,係為擷取影像知己和特徵 的演算法’在本實施㈣湘霍夫㈣的演算法(h〇_ Transform),其係為影像處理中從影像令 基本方法之-。至於霍夫轉換尋找幾何特徵之 200923837 f技術’在此不作贅述。再回到圖七a所示,步 後,再進行步驟21〇1,決定竽特 .2100之 中,主要A刹田士 & Λ寺徵衫像的斜率。在本步驟 率。最後ΓϋΛ 彡像龍之邊界的斜 取後再進订步驟21〇2根據該斜率決 ^ °〜出斜率之後,可將 轉 = 圖四之影像為例,旋轉五中之2軸,以 示。 轉㈣减影像調整成如圖八所 再回到圖六所示,隨後,進 旋轉角度,並對步驟2Jf) 11,決定一第二軸 刳敫⑴曰心 所得到之影像進行第二軸之旋棘 發明決定該第二軸旋轉角度流圖係為本 合之方法更具有;=像:=r擬合-平面。擬 值對圖八之影1 ,办像處理方法先以閥 所示,其中出來,其· 來,如圖九C所示,相像有代物之特徵影像。再 對於同一個區塊91而if線92通過各個區㈣, 有-第-交點_以二;條個咖會 九C中對應複數個特徵影像:影像區 面方程式,在本步驟令,首先對圖九口出後數個子平 進行複數次的取樣,如圖九D所示,所影像特徵 之區塊想像有複數條橫戴線92料,^ 該影像 92與區塊相交之第—交點之 橫戴線 後再以最小平方法擬合 〃 Z軸向度關係,然 口出代表该魏個第一交點920所形 200923837 平面93,如圖九E所示。至於取樣之第-交 占920數置則可更具需求而定,並無特定之限制。乂 然後,同樣地,再對圖九c中 921進行取檨,诒田r τ丄 κ币一 乂點 之-第-子平面Q4平方法擬合出該於該第二交點奶 ^弟一子+面94,如圖九以斤示。然後,擬合 ==與5亥第二子平面94之法向量’也就是對該苐-子 千面93以及該第二子平 乐十 該擬人平而q7^办上干 之法向置作平均計算,得到The C-type microfilm in the object to be tested has a periodic object to be tested. In order to facilitate the explanation of the alignment correction, a standard system is established on the 寺寺川物 to illustrate. Then, a μ is the image of the present invention, as shown in the figure, the image is adjusted by the angle of the image, and the initial image is rotated by the first axis to find the initial image. The measuring system is a brightness enhancement film, so too sinus two: characteristic shirt image. Because Ο brightens the boundaries of the microstructure inside the film. As for the special: image: =: some changes 'not the Si described in the present invention: 'shown' 'Looking for the feature in the initial image υ ί Γ ' Γ 'every slash 9G represents the brightness enhancement film The conversion method described in step 2100 is an algorithm for capturing image confession and features 'in this implementation (4) Xiang Hoff (four) algorithm (h〇_Transform), which is the basic method for image processing from image processing. -. As for the Hoff transform to find the geometric features of the 200923837 f technology ' will not be described here. Returning to Figure 7a, after step, proceed to step 21〇1 to determine the slope of the main A-Shaws & At this step rate. Finally, 斜 斜 彡 边界 边界 边界 边界 边界 边界 边界 边界 边界 边界 边界 边界 边界 边界 边界 边界 边界 边界 边界 边界 边界 龙 边界 龙 龙 龙 龙 龙 龙 根据 根据 根据 根据 根据 根据 根据 根据 根据 根据 根据 根据 根据 根据 根据 根据 根据 根据 根据 根据Turn (4) the reduced image is adjusted as shown in Figure 8 and then back to Figure 6. Then, enter the rotation angle, and step 2Jf) 11, determine a second axis 刳敫 (1) to obtain the image of the second axis The spine invention determines that the second axis rotation angle flow diagram is more suitable for the method of this combination; = like: = r fit - plane. The value of the pseudo-value is shown in Fig. 1. The image processing method is first shown by the valve, and the one coming out, as shown in Fig. 9C, is similar to the characteristic image of the surrogate. For the same block 91 and the if line 92 passes through the respective areas (4), there is a - first-intersection point _ two; a café will nine corresponding to a plurality of characteristic images: image area equation, in this step, first Figure 9 shows the sampling of a plurality of sub-levels. As shown in Figure 9D, the block of image features imagines a plurality of horizontal lines 92, ^ the intersection of the image 92 and the block. After the horizontal line is worn, the 〃Z axial relationship is fitted by the least square method, and the mouth represents the 200923837 plane 93 of the first intersection 920, as shown in FIG. 9E. As for the number of the first-to-pay 920 of the sampling, it can be more demanded, and there is no specific limit.乂 Then, in the same way, take the 921 in Fig. 9c, and the 第田r 丄 丄 币 乂 - 第 第 第 第 第 第 第 第 第 第 第 第 第 第 第 第 第+ face 94, as shown in Figure IX. Then, fitting == and the normal vector of the second sub-plane 94 of 5 hai, that is, the normalization of the 苐-子千面 93 and the second child Average calculation, get

史 私式’如圖九G所示。再回到圖九A 擬人平面m旦步根據圖九〇所得到之 的法向置得到該第二軸之旋轉角度。因此,口要 :該第:軸係為圖五軸。請參閲圖十The history of private is shown in Figure 9G. Returning to Fig. 9A, the anthropomorphic plane m denier is obtained according to the normal direction obtained in Fig. 9 to obtain the rotation angle of the second axis. Therefore, the mouth should: The first: the shaft is the five axes of the figure. Please refer to Figure 10.

=為進行完影像調整程序之結果說明示S A'之景完步驟211所得到之影像,並對該圖十= for the result of the image adjustment procedure, the image obtained by step S211 is shown in step S211, and the figure is ten.

軸位置分析置取一橫截線’所得到的Z軸高度與X 以及第-船 十B所示。經過旋轉第-軸(Z軸) 苐一軸(X轴)之後,可以發現圖十β中之 =各斜線所對應之高度⑺已經顯示 該_ 之—側面81 n 待測物8 再回到圖二所示’步,驟21之後 就是判斷是否所有的角度已經操取完畢=仃二驟22 ’也 :為_楊":處::先為, 再回到需要在取得另—側面之f彡像村。因此, 乂驟20,调整影像擷取裝置(圖 之嶋取角度,形成如圖三D之位置並^ = 200923837 再度進行步驟 另一侧影像之位置。影像調整程序調整該待測物 例,該待測物C與圖三D之截取影像角度為 因此需重複進行步驟2二=:81與82鄉 之影像。如果像圖三E的=兩田人,完成類取所有表面 84與85則需要會;;的、、、。構’因為有三個結構斜面83、 步驟至行步驟2〇與21三次。因此,重複 以前述之次數為1數需根據待測物表面之結構而定,並不 複數張對應;同:;擷理初始影像之後可以得到 23,豐合該複數個調 驟 行步驟挪,===!!更包括有下列步驟:首先進 m ^西影像進行對位並結合成—組 〜象由料將該複數個輕 必須透過對位的方式,也就是$敕70正心像, 每-張,敫… 轴與”由之位移,使 n衫像與相鄰之調整影像可以準確結 =中可以透過在初始影像上形成對位記號的方式^ 订、至於對位5己號形成方法’一般而言可以該待測物 上以-對位光源投射至少一對位記號至該待測物上,使每 一個初始影像上具有該對位記號。例如圖十二中,利用雷 ,光投射於待測物上形成兩對位記號95肖96,以利將^ 3:合影像時可以對位。該投射光源一般而言可以選擇為雷 射,但不以此為限。由於每一張調整影像上具有該對位記 號,因此可以利用該對位記號作為基準將影像結合在一起 即可。 200923837 對該步驟230之後,接著進行步驟231, 驟中,由程序以形成該重建影像。在本步 ^ 由於精微量測的結果為相對高度。如圖十三所一 對疊合後之影像於同-位置之橫截線所通過:區 域其X軸位置與2軸之高度關係圖。由圖十三中,The axial position analysis takes the Z-axis height obtained by a cross-sectional line' as shown by X and the first ship XB. After rotating the first axis (Z axis) and the first axis (X axis), it can be found that the height corresponding to each oblique line in Fig. 10 (7) has already displayed the _ side 81 n the object to be tested 8 and then return to Fig. 2 The step shown, after step 21 is to determine whether all the angles have been manipulated = 仃二骤22 'also: _杨": at:: first, then back to the need to obtain another side of the f彡Like the village. Therefore, in step 20, the image capturing device is adjusted (the angle of the image is taken to form a position as shown in FIG. 3D and ^=200923837. The position of the other side image is further performed. The image adjusting program adjusts the object to be tested, and the image adjustment program adjusts the object to be tested. The interception image angle of the object to be tested C and Fig. 3D is such that the image of step 2 ==81 and 82 is required to be repeated. If it is like the figure of Fig. 3E = two fields, the completion class takes all the surfaces 84 and 85. Because there are three structural slopes 83, steps to steps 2〇 and 21 three times. Therefore, repeating the number of times mentioned above depends on the structure of the surface of the object to be tested, and The plural number corresponds to; the same:; after the initial image is processed, 23 can be obtained, and the plural number of steps is moved, ===!! More includes the following steps: first, the m ^ west image is aligned and combined into - group ~ like the material to be the light must pass through the alignment, that is, $ 敕 70 positive heart image, each - Zhang, 敫 ... axis and "displaced by, so that the n-shirt image and adjacent adjustment image can Accurate knot = can be achieved by forming a registration mark on the initial image ^ 订定,对对5己号 forming method 'Generally, at least one pair of bit marks can be projected onto the object to be tested on the object to be tested, so that each of the initial images has the alignment mark For example, in Figure 12, using lightning, light is projected onto the object to be tested to form two pairs of marks 95 Xiao 96, so that the image can be aligned when the image is combined. The projection light source can generally be selected as a laser. However, it is not limited to this. Since each of the adjusted images has the alignment mark, the image can be combined by using the alignment mark as a reference. 200923837 After the step 230, the process proceeds to step 231. In the step, the reconstructed image is formed by the program. In this step, the result of the micro-measurement is relative height. As shown in Fig. 13, the image of the superimposed image passes through the cross-section of the same-position: the region The relationship between the position of the X-axis and the height of the 2-axis. From Figure 13,

步,驟230之後,相同位置的高度值並不—定會相 、,:|如在圖十三中’在疊合之後會有不同之高▲差 1。因此在步,驟231中,透過一等化(c〇取㈣)的 匕’’ f相同位置的高度值調成相等以形成如圖十四A所 Γί結果。在本步驟231中,可利用最小平方法,找出誤 2小的等化參數,經過等化後,就完成4合影像的動作: 取後經過等化步驟所形成之影像如圖十四W示。。After step 230, the height value of the same position is not fixed, and :|| as in Fig. 13 'has a different height ▲ difference 1 after the superposition. Therefore, in step 231, the height values of the same position of the 匕'' of the equalization (c) are equalized to form a result as shown in Fig. 14A. In this step 231, the least square method can be used to find the equalization parameter with a small error of 2, and after equalization, the action of the 4-fold image is completed: the image formed by the equalization step is as shown in FIG. Show. .

惟以上所述者,僅為本發明之實施例,當不能以之限 制本發明範圍。即大凡依本發明申料㈣圍所做之 ^化及修Ιΐ,仍將不失本發明之要義所在,亦不脫離本發 明之精神和範圍,故都應視為本發明的進一步實施狀況。 綜合上述,本發明提供之影像重建方法,不但可以解 決習用技術中因為影像擷取裝置動作誤差所產生 準卻之問題,更可以簡短演算相,提升影像重建^不 因此可以滿足業界之需求,進而提高該產業之競爭力以及 帶動週遭產業之發展,誠已符合發明專利法所狀申請發 明所需具備之要件,故爰依法呈提發明專利之申請,謹請 貴審查委員允撥時間惠予審視,並賜准專利為禱。 200923837 【圖式簡單說明】 圖一為習知光學顯微鏡以同軸光檢測斜面時,反射光無法 反射回光學顯微鏡之結構示意圖。 圖二係為本發明影像重建方法實施例流程示意圖。 圖三A與圖三B係為可調整傾角之影像截取裝置示意圖。 圖三C與圖三D係為截取待測物影像角度示意圖。 圖三E係為另一種待測物表面輪廓樣態示意圖。 圖四係為利用圖三A或圖三B所截取之待測物影像。 圖五係為本發明之待測物立體結構示意圖。 圖六係為本發明影像調整程序流程示意圖。 圖七A係為決定第一轴旋轉角度流程示意圖。 圖七B係為尋找該初始影像内之特徵影像結果示意圖。 圖八係為旋轉Z轴調整結果之影像。 圖九A係為本發明決定該第二軸旋轉角度流程示意圖。 圖九B至圖九G係為擬合平面流程示意圖。 圖十A與圖十B係為進行完影像調整程序之結果說明示意 圖。 圖十一係為本發明疊合影像流程示意圖。 圖十二係為具有對位記號之影像示意圖。 圖十三係為影像疊合後具有高度差之結構示意圖。 圖十四A係為將疊合後影像進行等化步驟所得到之截面輪 廓圖。 圖十四B係為將疊合後影像進行等化步驟所形成之最終影 200923837 像。 【主要元件符號說明】 11- 光學顯微鏡 12- 待測物 13- 反射光訊號 2- 影像重建方法 20〜23-步驟 210〜211-步驟 2100〜2102-步驟 2110〜2111-步驟 230〜231_步驟 3- 影像擷取裝置 30- 影像擷取單元 31- 導軌槽 3 2 -枢轴3 2 33-旋轉軸體 34、35-滑軌 36-導引塊36。在該影像擷取單元30之一側具有一 3 7-旋轉軸體 38-滑執槽 8-待測物 8卜82、83、84、85-結構斜面 15 200923837 9 0 -斜線 91 -區塊 92- 橫戴線 920- 第一交點 921- 第二交點 93- 第一子平面 94- 第二子平面 95、96-對位記號 97-擬合平面However, the above is only an embodiment of the present invention, and the scope of the present invention is not limited thereto. It is to be understood that the present invention is not limited to the spirit and scope of the present invention, and should be considered as further implementation of the present invention. In summary, the image reconstruction method provided by the present invention not only solves the problem of the accuracy caused by the motion error of the image capturing device in the conventional technology, but also can shorten the calculation phase and improve the image reconstruction, so that the requirements of the industry can not be met. To improve the competitiveness of the industry and to promote the development of the surrounding industries, Chengcheng has met the requirements for applying for inventions in the invention patent law. Therefore, the application for invention patents is submitted according to law. Please ask the review committee to allow time for review. And grant a patent as a prayer. 200923837 [Simple description of the drawing] Figure 1 is a schematic view showing the structure of a conventional optical microscope when the inclined surface is detected by coaxial light, and the reflected light cannot be reflected back to the optical microscope. FIG. 2 is a schematic flow chart of an embodiment of an image reconstruction method according to the present invention. FIG. 3A and FIG. 3B are schematic diagrams of an image intercepting device with adjustable tilt angle. Figure 3C and Figure 3D are schematic diagrams for intercepting the image angle of the object to be tested. Figure 3E is a schematic diagram of another surface profile of the object to be tested. Figure 4 is an image of the object to be tested taken by using Figure 3A or Figure 3B. Figure 5 is a schematic view showing the three-dimensional structure of the object to be tested of the present invention. Figure 6 is a schematic flow chart of the image adjustment program of the present invention. Figure 7A is a schematic diagram of the process of determining the angle of rotation of the first axis. Figure 7B is a schematic diagram of the result of finding a feature image in the initial image. Figure 8 is an image of the rotation Z-axis adjustment result. Figure 9A is a schematic flow chart of determining the rotation angle of the second shaft according to the present invention. Figure 9B to Figure 9G are schematic diagrams of the fitting plane flow. Fig. 10A and Fig. 10B are schematic diagrams showing the results of the image adjustment procedure. Figure 11 is a schematic flow chart of the superimposed image of the present invention. Figure 12 is a schematic diagram of an image with a registration mark. Figure 13 is a schematic diagram showing the structure of the height difference after the image is superimposed. Fig. 14A is a sectional profile obtained by performing an equalization step on the superimposed image. Figure 14B is the final shadow 200923837 image formed by the equalization step of the superimposed image. [Main component symbol description] 11- Optical microscope 12-Test object 13- Reflected light signal 2 - Image reconstruction method 20~23-Step 210~211-Step 2100~2102-Step 2110~2111-Step 230~231_Step 3- Image capture device 30 - Image capture unit 31 - Track slot 3 2 - Pivot 3 2 33 - Rotary shaft body 34, 35 - Slide rail 36 - Guide block 36. On one side of the image capturing unit 30, there is a 3 7-rotating shaft body 38 - sliding groove 8 - object to be tested 8 82, 83, 84, 85 - structural inclined surface 15 200923837 9 0 - oblique line 91 - block 92- Transverse line 920 - First intersection 921 - Second intersection 93 - First sub-plane 94 - Second sub-plane 95, 96 - Alignment 97 - Fit plane

Claims (1)

200923837 十、申請專利範圍: 1. 一種影像重建方法,包括有下列步驟: ⑷:整影像擷取角度以對—待測物擷取—初始影 ㈦對該初始影像進行—影像罐程序以得到一調 整影像;200923837 X. Patent application scope: 1. An image reconstruction method, comprising the following steps: (4): The entire image capture angle is used to capture the object to be tested - the initial shadow (7) is performed on the initial image - the image canister program to obtain a Adjust the image; (C);複步驟⑷至⑹至少-次,以得到複數個調 整影像;以及 ⑷$合该複數個調整影像以得到該待測物 影像。 咬 ’其中該待 如申明專利範圍第1項所述之影像重建方法 測物係具有一微結構。 3·如申料利範㈣2項所狀影像錢方法, 、、’°構係為具有週期性排列之微結構。 n 4.如申請專利範圍第丨項所述之影像重建方法,其中料 像調整程序更具有下列步驟: … (bl)決定一第一軸旋轉角度並對該初始影像進行第 一軸之旋轉調整;以及 (b2)決定一第二軸旋轉角度,並對步驟所得到 之影像進行第二軸之旋轉調整以得到該調^影 5.如申請專利範圍第4項所述之影像重建方法,其中決^ 该第一軸旋轉角度之方式更具有下列步驟: ’、疋 經由一轉換決定該初始影像内之特徵影像; 17 200923837 決定該特徵影像之斜率;以及 根據該斜率決定該第一軸旋轉角度。 6. 如申請專利範圍第5項所述之影像重建方法,其中該特 徵影像係為待側物之—特徵結構之邊界。 7. 如申請專利範圍第5項所述之影像重建方法,其中該轉 換係為擷取幾何特徵的演算法。 8. 如申明專利範圍帛7項所述之影像重建方法,其中該幾 何特徵係為直線,且触該幾何特徵之演算法係為霍 轉換(Hough Transform)。 如申請專利範圍第4項所述之影像重建方法,其中 該第二軸旋轉角度之方式更具有下列步驟: 根據步驟(bl)所得到的影像洲數值分析擬合一平 面;以及 據該平面決定該第二軸旋轉角度。(C); repeating steps (4) to (6) at least once to obtain a plurality of adjusted images; and (4) combining the plurality of adjusted images to obtain the image of the object to be tested. The image reconstruction method described in item 1 of the patent application scope has a microstructure. 3. For example, the method of image money in the two items of the application of Lifan (4) is a microstructure with periodic arrangement. n. The image reconstruction method according to the scope of claim 2, wherein the image adjustment program further has the following steps: (bl) determining a first axis rotation angle and performing a first axis rotation adjustment on the initial image. And (b2) determining a second axis rotation angle, and performing a second axis rotation adjustment on the image obtained in the step to obtain the image. 5. The image reconstruction method according to claim 4, wherein The first axis rotation angle has the following steps: ', 决定 determines a feature image in the initial image through a conversion; 17 200923837 determines the slope of the feature image; and determines the first axis rotation angle according to the slope . 6. The method of image reconstruction according to claim 5, wherein the feature image is a boundary of a feature to be a side object. 7. The image reconstruction method of claim 5, wherein the conversion is an algorithm for capturing geometric features. 8. The image reconstruction method according to claim 7, wherein the geometric feature is a straight line, and the algorithm touching the geometric feature is a Hough Transform. The image reconstruction method according to claim 4, wherein the second axis rotation angle has the following steps: fitting a plane according to the image continent numerical analysis obtained in the step (bl); and determining according to the plane The second axis is rotated by an angle. 10·如申請專利範圍第9項所述之影像重建方法,並中決 疋^面之方程式方法更包括有下列步驟: 、 尋找该初始影像内之複數個特徵影像; 分:對f複數個特徵影像相互對應之影像位置擬合 出禝數個子平面方程式;以及 對=數個子平面方程式進行擬合以得到該平面方 核式。 U·如申請專利範圍第丨項所述 步驟⑷更具有下列步驟:⑥像重建方法,其中該 組合影 將該複數個調整料進行對位並結合成 18 200923837 如申过專;;進订一耗程序以形成該重建影像。 位,、第11項所述之影像重建方法,立t對 ^是數個調整影像之方式更包括有下列步驟·/、 利用-對位光源投射—對位記號至該待測物上 母一個初始影像上具㈣對位記號;以及 利用邊對位記號結合該複數個調整影像。 13.—種影像重建方法,包括有下列步驟: (a)调整料擷取角度以對—待測物擷取一初始影 ^該待測物具有呈週期性排列之複數個‘ (b)對該初始影像進行一影像調整程序以得到一調 整影像,該調整程序包括有·· ^ (bl)決定-第一軸旋轉角度並對該初始影像進 行第一軸之旋轉調整;以及 (b2)決定一第二軸旋轉角度,並對步驟(bi)所 得到之影像進行第二轴之旋轉調整以 該調整影像; (〇重複步驟⑷至⑻至少—次,以得到複數個調 整影像;以及 (d)疊合該複數個調整影像以得到該待測物之重建 影像。 14· &如申請專利範圍第13項所述之影像重建方法,其中決 定該第一軸旋轉角度之方式更具有下列步驟: 19 200923837 -二由一轉換決定該初始影像内之特徵影像; 決定該特徵影像之斜率;以及 根據該斜率決定該第一軸旋轉角度。 15‘^^專利範圍第14項所述之影像重建方法,其中該 寺试衫像係為該徵結構之邊界。 l6.2請專利朗第Η項所述之影像錢方法,其中該 奐係為擷取幾何特徵之演算法。 1 # m圍第16項所述之影像重射法,其中該 夬鍊始徵係為直線,且擷取該幾何特徵之演算法係為霍 夫轉換(H〇ugh Transf〇rm)。 巧隹 定今申第月專利视圍第13項所述之影像重建方法,其中決 μ弟—軸旋轉角度之方式更具有下列步驟: 根據步驟(bl)所得到的影像利用數值分析擬合 面;以及 據該平面決定該第二軸旋轉角度。 18項所述之影像重建方法,其中決 亥/面^方程式方法更包括有下列步驟·· 1 4初始影像内之對應該複數微結構之特徵影 分:對:複數個特徵影像相互對應 出稷數個子平面方程式;以及 罝擬口 對=數個子平面方程式進行擬合以得到該平面方 20. 如申請專圍第13項所述之影像重建方法,其中該 20 200923837 乂驟⑷更具有下列步驟: 將=複數個調整影像進料位並結合成—組 像,以及 對該組合影像進行—等化程序以形成該重建影像。 .如申請專職SI第2〇項所述之影像重射法,並 位該複數個調整影像之方式更包括有下列步驟:、 利用一對位光源投射一對位記號至該待測物上,使 每一個初始影像上具有該對位記號;以及 利用該對位記號結合該複數個調整影像。10. The image reconstruction method according to claim 9 of the patent application, and the equation method of the method further includes the following steps:: searching for a plurality of feature images in the initial image; The image positions corresponding to the images are fitted with a plurality of sub-plane equations; and a plurality of sub-plane equations are fitted to obtain the plane square nucleus. U. The step (4) as described in the scope of the patent application scope has the following steps: 6 image reconstruction method, wherein the combination shadows the plurality of adjustment materials to be aligned and combined into 18 200923837, for example; The program is consumed to form the reconstructed image. Bit, the image reconstruction method according to item 11, the method of adjusting the image is further including the following steps: /, using the -position light source to project - the alignment mark to the parent of the object to be tested The initial image has a (4) alignment mark; and the edge alignment mark is used to combine the plurality of adjustment images. 13. The image reconstruction method comprises the following steps: (a) adjusting the material extraction angle to obtain an initial image of the object to be tested, and the object to be tested has a plurality of periodic intervals (b) The initial image is subjected to an image adjustment program to obtain an adjusted image, the adjustment program includes: determining a first axis rotation angle and performing a first axis rotation adjustment on the initial image; and (b2) determining a second axis rotation angle, and the second axis rotation adjustment is performed on the image obtained in step (bi) to adjust the image; (〇 repeat steps (4) to (8) at least once to obtain a plurality of adjustment images; and (d And superimposing the plurality of adjustment images to obtain a reconstructed image of the object to be tested. The image reconstruction method according to claim 13, wherein the method of determining the rotation angle of the first axis further comprises the following steps. : 19 200923837 - two determines a feature image in the initial image by a conversion; determines a slope of the feature image; and determines the first axis rotation angle according to the slope. 15 '^^ Patent scope item 14 The method of image reconstruction, wherein the temple test shirt image is the boundary of the structure of the sign. l6.2 The method of image money described in the patent Langdi item, wherein the line is an algorithm for extracting geometric features. The image re-shooting method described in Item 16 of the present invention, wherein the 始 chain eigensystem is a straight line, and the algorithm for extracting the geometric feature is H霍夫ugh Transf〇rm. The image reconstruction method according to Item 13 of the patent application of the first month of the present invention, wherein the method of determining the angle of rotation of the shaft further comprises the following steps: using the numerical analysis to fit the surface according to the image obtained in the step (bl); The plane determines the rotation angle of the second axis. The image reconstruction method according to Item 18, wherein the method of determining the resolution/face equation further comprises the following steps: 1 4 Characterization of the corresponding complex microstructure in the initial image: : a plurality of feature images corresponding to each other with a plurality of sub-plane equations; and a virtual port pair = a plurality of sub-plane equations to obtain the plane square. 20, wherein the image reconstruction method described in claim 13 is 20 2009 23837 Step (4) further has the following steps: Having = a plurality of adjusted image feeding positions and combining them into a group image, and performing an equalization process on the combined image to form the reconstructed image. If applying for full-time SI item 2 The image re-shooting method further includes the following steps: and using a pair of bit light sources to project a pair of bit marks onto the object to be tested, so that each of the initial images has the pair a bit mark; and combining the plurality of adjusted images with the alignment mark.
TW96145080A 2007-11-28 2007-11-28 Method for constructing the image of structures TWI342524B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
TW96145080A TWI342524B (en) 2007-11-28 2007-11-28 Method for constructing the image of structures

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
TW96145080A TWI342524B (en) 2007-11-28 2007-11-28 Method for constructing the image of structures

Publications (2)

Publication Number Publication Date
TW200923837A true TW200923837A (en) 2009-06-01
TWI342524B TWI342524B (en) 2011-05-21

Family

ID=44728805

Family Applications (1)

Application Number Title Priority Date Filing Date
TW96145080A TWI342524B (en) 2007-11-28 2007-11-28 Method for constructing the image of structures

Country Status (1)

Country Link
TW (1) TWI342524B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
TWI578269B (en) * 2015-12-14 2017-04-11 財團法人工業技術研究院 Method for suturing 3d coordinate information and the device using the same

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
TWI618644B (en) * 2016-05-06 2018-03-21 財團法人金屬工業研究發展中心 Image overlay method

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
TWI578269B (en) * 2015-12-14 2017-04-11 財團法人工業技術研究院 Method for suturing 3d coordinate information and the device using the same

Also Published As

Publication number Publication date
TWI342524B (en) 2011-05-21

Similar Documents

Publication Publication Date Title
Sapirstein A high-precision photogrammetric recording system for small artifacts
Fontana et al. Three-dimensional modelling of statues: the Minerva of Arezzo
Sutton et al. Image correlation for shape, motion and deformation measurements: basic concepts, theory and applications
Guidi et al. High-accuracy 3D modeling of cultural heritage: the digitizing of Donatello's" Maddalena"
JP3419213B2 (en) 3D shape data processing device
JP4851239B2 (en) Image processing apparatus and processing method thereof
JP2005514606A (en) Three-dimensional three-dimensional measurement system and method
JP5571007B2 (en) Sphere shape measuring device
Nguyen et al. Accuracy assessment of fringe projection profilometry and digital image correlation techniques for three-dimensional shape measurements
CN105571514B (en) The device and method of optical element is quickly adjusted in rotation translation absolute sense method
Bergström et al. Virtual projective shape matching in targetless CAD-based close-range photogrammetry for efficient estimation of specific deviations
US20100157312A1 (en) Method of reconstructing a surface topology of an object
CN109506629B (en) Method for calibrating rotation center of underwater nuclear fuel assembly detection device
JP2006119121A (en) Shape measurement method and device for optical member
CN105627945B (en) Non-spherical element center and the measurement apparatus and measuring method of cylindrical center shift amount
JP4809134B2 (en) Image processing apparatus and processing method thereof
JP4851240B2 (en) Image processing apparatus and processing method thereof
TW200923837A (en) Method for constructing the image of structures
Chen et al. A novel mirrored binocular vision sensor based on spherical catadioptric mirrors
Shaw et al. Automatic registration method for hybrid optical coordinate measuring technology
Zheng et al. A non-coplanar high-precision calibration method for cameras based on an affine coordinate correction model
Li et al. Research on measurement method of grinding wheel profile based on image mosaic
Díaz et al. One step geometrical calibration method for optical coherence tomography
JP4813263B2 (en) Image processing apparatus and processing method thereof
JP4922905B2 (en) Method and apparatus for measuring position variation of rotation center line

Legal Events

Date Code Title Description
MM4A Annulment or lapse of patent due to non-payment of fees