TW200938804A - Method of implementing pattern reflection to measure contour of object surface - Google Patents

Abstract

Disclosed is a method of implementing pattern reflection to measure contour of object surface, comprising the steps of: generating pattern images to be displayed on a monitor by a computer using phase shifting means; projecting the pattern images displayed on the monitor towards a plane mirror; capturing pattern images reflected from the plane mirror by image capturing means; and placing a to-be tested object on the plane mirror; projecting the pattern images displayed on the monitor towards the to-be-tested object; capturing pattern images reflected from the to-be-tested object by the image capturing means; analyzing the phase deformation of these patterns by a computer using phase shifting method and phase unwrapping means; computing contours of the to-be-tested object by geometrical or vector analysis means based the slopes inferred from the phase deformation. As such, this invention provides an automatic optical measuring method to be implemented in accurate measurement of contours of object surfaces.

Description

200938804 IX. INSTRUCTIONS: [Technical Field] The present invention relates to an optical measuring method for surface topography of an object, and more particularly to a method for automatically measuring the surface size of an object by using a stripe reflection method. [Prior Art] Due to the rapid development of the semiconductor and micro-electromechanical industries today, the measurement requirements for tiny-sized components are expanding. Take the semiconductor process as an example.

❹ It is mainly used to grow several layers of different thicknesses and different materials on each die unit of the wafer. Due to the influence of the process environment, each layer of the film on the wafer is subjected to stress changes. For example, in the process of growing the thin films on the wafer, it is often necessary to deposit a thin film by a step of deuteration, physical or chemical deposition, etc., because the deposited film and the thermal mechanical properties of the wafer are different 'in the process of depositing the thin film' The influence of the increase or decrease of temperature leads to the residual stress between the heterogeneous materials due to the difference in thermal expansion coefficient, which causes the sheet-shaped wafer to easily generate a koji, and even the film-wafer interface is peeled off or excessive voids and cracks are formed. And a defect. Therefore, in the wafer manufacturing process, in order to ensure the quality level of the product, non-destructive inspection methods are usually used to screen out defective units to avoid unnecessary loss of subsequent processes. Regarding the aforementioned wafer process _ '丨~咏~" 'the film on the day, the deformation due to high temperature is between tens of micrometers (micr〇n), and the currently known detection methods such as ··· Projection embossing method, stripe feeding, shadow stacking: :Equivalent optical measuring method, although the measured out-of-plane deformation range is above tens of micrometers, it can meet the basic requirements of the above wafer surface measurement. However, the surface of the wafer is a highly reflective surface. When the above-mentioned known optical sensing method is applied to the surface topography measurement, the optical measurement may be disturbed by the high reflectivity of the surface. By achieving the precise measurement results, the semiconductor technology is becoming more and more sophisticated, and the size of the wafer product is more miniaturized. The above measurement method is even more inapplicable to the use of wafer surface topography measurement, so it needs further improvement. SUMMARY OF THE INVENTION The main object of the invention is to provide a method for measuring the surface topography of an object by using a stripe reflection method, and to improve the accuracy and correctness of the surface measurement of the object. Prior to the achievement of the present invention, the method for measuring the surface topography of an object by using the stripe reflection method is designed, and the steps thereof include: displaying a stripe image by a computer by means of phase shifting on the screen; The stripe image is clearly projected onto the plane mirror Ji., ❹ w like the stripe image reflected from the plane mirror in the county, and recorded in the computer as a reference for the measurement of the object to be tested; Z with a reflective surface The object to be tested is placed on the plane mirror; the image of the stripe displayed on the far screen is projected onto the object to be tested; and the image of the stripe taken from the object to be tested is taken by the image capture object (4) and recorded in the computer; And using the computer to analyze the phase deformation of the stripe by phase shift method and phase spread means, and calculate the slope and then calculate the surface topography of the object to be tested by analytical means. 5 200938804 The present invention is characterized by the foregoing method part _ _ _ 忐 又 又 又 又 , , , , , , , , , 本 本 本 : : : : : : : : : : : : : : : : : : : : : : : : : : The method and the higher accuracy of accurately measuring the surface of the object &#& . ^ ^ appearance, for the industry to provide a more practical measurement method for non-destructive testing of its products. [Embodiment] Referring to the flow chart shown in the first figure and the configuration shown in the second figure, the method for measuring the surface topography of the object by the stripe reflection method includes the following steps: The computer (5 〇) produces a stripe image by phase shifting on the screen (2 〇). It mainly uses a computer program H ( 2 〇) preset in the computer (50) to sequentially display a plurality of intervals. Stripe image of the stripe arranged in parallel; 7 The stripe image displayed on the screen (20) is clearly projected onto the plane mirror (丄〇) through the optical path, which mainly places a plane mirror (丄0) on the β-screen (2〇) is located on one side below, and the stripe image of the screen (20) can be projected onto the plane mirror (on the work' and then reflected in the opposite direction; the image is taken from the plane mirror 〇) Reflected fringe image, recorded in the f brain (5 〇) t, as a reference for the measurement of the object to be tested, mainly using a set of image pickers (4 〇) such as CCD, by the plane mirror ( 10) The upper side is taken from the plane mirror 1〇) The reflected fringe image is then transmitted back to the computer (50) for recording; a reflective surface of the object to be tested is placed on the plane mirror (Workshop 6 200938804) where the surface of the object to be tested is round. . etc.); it has a mirror-like reflection effect (such as crystals that the screen (2 0) is shot on the object to be tested, and the piece is better; the image of the stripe displayed on the image is transmitted through the light path. Projection to the object to be tested

The image capture device (4〇) recorded by the image capture means, recorded in the computer (50), the striped image reflected from the surface of the object, and the striped image reflected on the object to be tested is the same Using a set of objects such as CCD, the upper side of the object to be tested is retrieved from the computer (50), and the computer (50) is used to analyze the strips and images by means of reflective fringe image analysis. The phase deformation is calculated and the surface morphology of the object to be tested is obtained by the analytical means after the slope is calculated. In the step of displaying, by the computer (50), the stripe image is displayed on the screen (20) by the phase shifting means, the phase shifting means sequentially generates a plurality of images of the stripe through the computer material, the stripe material The gray scale distribution is - cosine listening. In the preferred embodiment, the 'system' uses the step phase shift method to set the computer program to generate four stripe images (Ιι, ", ", and "), the initial phase of the four stripe images. Set to 〇, verb/2, 疋, 3 /2, respectively, and the simultaneous equation is as follows: 7 200938804

/〇+ cos I h = I0 + A C0S(^ - £.) '3 = '〇+ j COS(沴-π) _.·(3· 1~ 1) '〇+ j COS(必-l£ _) In the above-mentioned simultaneous equation, "the intensity of the background" is the phase angle of the point in the vibrating towel and the pattern image.

The front phase of the cubic program is transformed by a trigonometric function and the phase value of each point between the algebraic stripe and the stripe is 0, as shown in the following equation. , can get ^,3^) = tan'1 W3) ............................... (3. 4) It should be further explained that since the stripe offset in this direction can be obtained by extending in a certain direction, the stripe 'only T to W at least two different directions of the side surface (ie, the X direction, the y direction) stripes = to be used As shown in the fourth figure, t can have enough information to reconstruct the map and, in the preferred embodiment of the present invention, 'must open 7 appearances, ^ white produces four X-direction stripe image maps by phase shift method, and four sheets y Directional stripe image: (4): The computer analyzes the image by reflection fringe image analysis: Stripe cool = summer: processing 'through the phase expansion method and the determination of the measurement area method to Yuancheng' or, in the phase expansion method and determination The quantity_domain first performs median filtering processing, wherein: the use of the median filtering processing means is caused by electronic noise generated during image capturing and conversion, and is caused by the trapping, . ^ ^ ^ The fringe image map generally has random m, so the present invention advances through the computer (1) (4) wave transport 8 200938804 calculation method to reduce noise, and its operation formula is : Λ /(^ y) = median{g(s, t)} (")e~ ........................... ..(3.4-1) Where? The median filtered image grayscale value; \ is the coordinates of all pixel points in the original image; the guest CM) is the coordinate value of all the grayscale values in the set region at coordinates (9)). The processing step of the phase unwrapping method is because the phase shift method is in the form of an inverse tangent function (tarrl).

After the judgment "phase cutoff" is in the interval of seven, the obtained phase map is not continuous. Therefore, the phase that has been truncated is restored to the original continuous phase map by the Macy phase unwrapping method. The processing step of the determination measurement area method is to further distinguish the measured image range from the background by the phase map restored by the phase expansion method, wherein the phase map obtained by using the phase mask method is separated. The gray scale distribution of the measured object and the background image pixel, and then the image threshold value method is used to divide the gray scale distribution of the measured object and the background image to obtain a measurement range. In the phase mask method just described, the equation shown below is used to make the difference between the phase change region and the phase change region without the phase change region, and the measured object is separated from the background gray scale. /('Lee w suppress "I...........................(3. 3-1) In the above, I Guang I4 is the aforementioned The phase map obtained in phase shift. The image threshold method is a simple method for image segmentation. The selection of the threshold value of the δ» boundary value is selected. In the present invention, the setting is ·· ' 9 Τ = 2μ ο ❹ dAB = 2 h-tana ····· where: the distance between the point A and the point B when the image picker (2 〇) is to be mirrored ( 3 〇), force 200938804 '(3. 3-2 ) where #• the overall gray level of the image. For the aforementioned @ 4 &

The ice analysis means, in the preferred embodiment of the present invention, V uses a geometric analysis method 推 s to derive any one of the measurement area methods, so that the offset of the striate... integral, in order to correctly calculate the strip Regarding the mentioned handles, the soil #靼- "analysis method, the basic principle is to use the geometric relationship and the angle function first, by the shadow] (four) in the relative relationship between the image processing point, the mirror surface and the stripe image, The mirror height is the phase of the oblique μ 1 . In the relationship between the geometric solution and the crepe offset, the object is integrated based on the object table and the pixel is added to the true length. As shown in the fifth figure, the image is taken in the vertical (y-direction) to set the image capture point (4 〇) to offset, and the twill image (ie, the screen (2 0) ) and the reference h' strip L, the image fetching g (40) from the reference mirror 〇) is the Λ point in the striated image, and the 镜 point in the strip can be observed from the mirror surface (2) to be measured, from the reference mirror surface and Waiting for the image of the fringe, the stripe offset between the gauge faces is the height h The difference between the m and the difference is (2. 200938804 face (3〇) i-slope change θ, from the law of reflection, this slope change will cause a total 2θ angular offset when reflected The angle of the fringe image observed by the mirror surface (30) is shifted to point C, so that the offset distance Ac is: dBc = (L - h)[x^n(a + 2Θ) - tan α ]..................( 2 2 - 2 ) From the (formula 2.2]) and (formula 2.2_2), the change in height difference and slope is obtained in the y direction. The resulting offset is: ❹

^ = (L-/z)[tan(a + 2^)-tan a ]-2Atan or ............(2· 2-3) 畲 fringe image and reference mirror (1 Distance between 〇) (1) is far: the height difference (h)N· between the reference mirror (10) and the mirror to be tested (3 〇), the former formula (2.2-3) can be further simplified as: ^ = Z[ Tan(a + 26>)-tana]........................(2 2-4) Afterwards, the re-use and angle formulas are further refined. The formula can be obtained as follows: dy ^tan 2Θ -(l + tan2 g ) (2.2-5) 1 - tan or tan 2Θ When 0 is less than 0.50, and tan α tan, 26> "丨, therefore, the above formula can be further Simplified to: do = 2 recognize (lWa) ..................... (2.2-6) can be known from the above formula, Under the condition of the same slope change value and the fixed image capture angle (α), increasing the distance between the object and the stripe can increase the stripe offset distance and improve the measurement sensitivity of the measurement system. Assuming that the surface of the object to be tested is a perfect plane, the gray-scale distribution of the fringe image reflected through the surface is a cosine function that changes along the y direction, and the light intensity distribution of the captured fringe image is: I(y) = a( y) + b(y)ο〇5[(2π/py)y]................ (2. 2-7) π 200938804 In the above formula, a: background Light intensity; b. Amplitude, A: y-direction stripe spacing. When there is a slope change on the surface of the object to be tested, the slope change will cause the stripe to deform. The intensity of the stripe after deformation is: / (y) = a(y) + b(y)cos^/py)y + φ{γ )].........(2.2-8) where the river is less) The phase change value produced by the stripe deformation in the y direction ❹ 上 In the above formula, the unit of the phase change value due to the stripe deformation is the diameter ( Rad), the relationship between the phase change value and the length unit is. Two and 9{y) = ^±.................... '

Py .............................. Substituting the formula (2.2-6) into the formula (2.2.2-9): 2,) to give n,)............... (2<2_1〇) In the above formula, ♦ η:%, n is the optical configuration magnification (10) amplification of the 〇Dtical rnnf,, · optical configuration) ^ η ^ represents a large phase change under the same surface slope change, π thousand changes will produce a large phase change, so the larger the η value, the sensitivity of the measurement system can be improved. Further finishing the above formula can be obtained: .............................................. ............... 2-11) Because the slope of the surface of the object is λ j, VL VII is the differential of the degree, and the slope of the surface of the object is integrated in the y direction. Get v ancient a & morale %. The surface height in the y direction is good ω, its integral 八·H(y)= jedy.......................... ............................2-12) From the above derivation process, J is in the fringe reflection method, and draws 12 200938804 Phase shift changes after the fringe image 1 (four) value and slope change, and then into a phase change to the phase change I progress through the computer integral outside the relationship between Pixel (Pixel), pixel and true ^ slope and pixel y = Ry -S{pixel)............The length relationship of two lengths is: Image resolution in the V y direction (2.2 13)

Rp(6)0: In the above derivation process, the height relationship between the tjy direction (longitudinal direction) and the X direction (lateral direction) cannot be obtained from the two-degree ❹-strip offset of the line. Therefore, it is necessary to The step-by-step relationship can obtain the correct and complete three-dimensional height. The height of the two X directions is related to the lateral (X-direction) stripe offset, such as the quasi-mirror (10) and the mirror to be tested (3 〇). The base stripe offset is ', but d has no component in the x direction due to the 曰 and the south. Therefore, the height h does not have the stripe offset in the X direction. It-person, when the image capturing point observes the mirror to be tested (3 〇), if the mirror surface (30) ay axis is rotated, a slope change point will be generated, which will cause the mirror to be tested ( 3〇) The observed B point is shifted to C, and the resulting stripe offset is seven c. Therefore, the intersection point of the image manipulation point and the stripe image plane is D. In the state of Dj, square, c, five, and five · common plane, it is known from the law of reflection that Li = far, and must be proved , and ADBC' is a similar triangle, so ^γϊ = 2@. Also Ζ(:Ό'Β = 2Ζ£Ό5=2 eve, so the slope shift caused by the change of slope is: = —/2) tan2p ............(2.2-17) 13 200938804 Therefore, you can get The difference between the height difference and the slope is the total offset of the square ^: °^^ {L — h)tB.n 2β ......... wood....... .........................(2.2-18) Although the distance L between the fringe images is much larger than the reference mirror (丄〇) to be tested ( 3 Ο ) The height difference (8), i can enter _step: ❹ dx = Ltax\ 2β ........ When the stone is within 0. 50........... ................................. The above can be further simplified as: dx ~ 2pL... .............. The light intensity distribution of the fringe image taken along the X direction I(x) = a(x) + K^)cos[(2^ / ρχ)χ] ........................ where a: the intensity of the background light; b · amplitude, • (2.2-20) % i • (2.2-21) Where: χ direction stripe spacing. The stripe strength after deformation is: Ο I(x) = a{x) + ft(x)cos[(2^· I ρχ)χ + φ{χ)\ (2.2-22) where 〆χ): because The phase change value produced by the directional stripe deformation offset. The relationship between the phase change value and the length unit is: ¢)( χ)= 2π, dx Px (2.2-23) Substituting (2.2-20) into (2.2-23) gives: (2.2-24) φ(χ) = (2π / px)dx = .............

Px further refines the above formula: 200938804 e_<p(x).pr ......(2. 2 - 2 5 ) 4πΣ The slope of the surface of the object due to the inclination of the surface of the object and the differential of its southness along

Integral in the X direction, the surface height of the media $丨丨Q X to the X direction /f(x), the integral formula is: .......(2.2-26) .....C 2. 2 ~ 2 7) mx)= \βάχ............................. The relationship between the length of the pixel and the true length is: x = K-^{pixel)................

Where ': X-direction image resolution: image pixel degree, so the combination can be obtained correctly. The above-mentioned derivation can be obtained in the x-direction (ie, horizontal). The height of each line has previously derived the relationship between the heights of the lines in the @ y direction, the complete three-dimensional height. Morphology. The vector analytic method matrix and the reflection vector conversion matrix offset conversion slope formula are completed, and the main method is to use vector rotation conversion ‘recombining the stripe shift formula and the stripe, wherein:

The rotation to the ^ transformation matrix, which uses the -. Ding π π π double-macro matrix type table

Not vector' and further refine the matrix transformation of A derivation vector rotation. In the case, when the vector is rotated along a set axis (^ ^ y, ζ any), the setting of the octet /, 顼 considers the variation of the scalar amount on the plane of the remaining two axes, the vector is on the plane The projection has a constant vector size, but only changes direction after rotation. j 燹 As shown in the seventh and eighth figures, the p vector (X, y, ζ) is timed by a mathematical moment. Because of the movement conversion and other conversion ports I ’, the formula is:

15 200938804 Since the z-axis component is constant when the vector is rotated along the z-axis, the projection of the vector on the x, y plane, the vector size is unchanged, and only the direction after the rotation changes, so the vector size of the z component is not considered: r = "X2 + y2 The x and y directions before and after rotation are: x = r cos <f\

^ = r sin φχ x = rcos φ2 y — -rsin φ2 and 戎 = 0 — for, substituting the above formula X = rcos = Γ〇〇3(φ — φι) = jccos φ +_ysin φ γ=-Γ5ΐηφ2 =-rsin ((p-</>x) = -xsin φ +jcos φ Therefore, based on the aforementioned geometric relationship, after derivation and collation, it is known that a vector rotates along the X axis (0 angle) and rotates along the y axis (0 The angles of rotation and the transformation matrix along the Z axis (彡 angle) are: II 1 X yz

Cos φ -sin φ 0 0 sin φ cos φ 0 0 0 0 1 0 0 0 0 1 (2.3-10) 1 0 0 0 0 cos -sin φ 0 0 sin φ cos φ 0 0 0 0 1 (2.3-11 ) 16 200938804 Ο _χ' y ζ' l] y cos φ 0 · -sin φ O' 0 1 0 0 sin φ 0 cos φ 0 0 0 0 1 Reference, as shown in the ninth figure... (2.3-12) ^ = [4 反射 The reflection vector generated in a plane E = = [tomb < 'the unit normal vector of plane E and = [meaning ' %], let the projection of 2 on and above be %, then: N0

N

N —· /V = (2.3-13) and ′ 乂 and its vector in the unit normal vector and the projected vector watt are c: <^=] + ΛΓ. =:5 + ((-])· and) and..............................(2. 3-14) 1 · —► 5 is the reflection vector of J' so 2 is symmetric with respect to the cloud with respect to %, so 5 = ζ ; reflection vector B = iV0 + C2 = 3 + 2((-]). ❹ Αχ-2{α·Ν )Νχ Ay-2^N)Ny Αζ-2{Α·Ν)Νζ~ 0' ο ο 1 ο ο

B lA Ay A, Ο 1 ο Ο Ο 1 '2{ν·α)Νχ -2(NA)Ny -2(ϊν·1)ΛΓ; ................ ............(2.3-16) For the stripe offset method, as shown in the ninth and tenth figures, the 〇 point is the coordinate origin and the xy plane is a mirror surface (i 〇 ), the redundant representative image picker (40) observes the vector of the upper point of the mirror surface (1〇), α is the image building device (40) observation angle 'then: must = [〇sina _cosa]; mirror normal 17 200938804 The quantity iV = [0 0 1], let i? be the reflection vector, and ^ and and enter the previous method (2.3-16), then: R l] = [i?x Ry Rz l] [AOx AOy AOz l] 1 0 0 0 1 0 0 0 1 0 0 0 -2 (N^A〇)Nx -2(N.Ad)Ny -.2(nA〇)Nz Ο =[0 sin a a cos, a = [0 sin a cos a The cloud is formed by the defect X = « Less = Z = where 0 0 1 0 0 0 0 1 0 0 0 0 2cos a 1 '· (2.3-17) ........ .............(2.3-18) The constant to be found represents the plane equation of the screen. ❹ 6 is the L - h solution (2.3-18), (2.3-19) [ Γ _ ...........··(2.3-20) ("", the intersection of vector and screen: V {L~h)tBna (L~h))... .......................(2.3-21) The position of the image picker screen before the slope change has not yet been generated. The following i隹一牛, the money mirror observes the slope and further changes the slope of the vertical and horizontal directions, and the vertical slope is added first, as shown in the eleventh figure, 0. 0) Generate a slope change along the y-axis, that is, around (9)/mirror (1) When the mirror method is used, the normal vector of the human face is ', and the 法 is substituted. (2 3_u 7 }, the normal vector L — h ]—-- cos a ...... a substitution (2·3_18) can be obtained (2.3-19) 18 200938804 $ is X ι]=[^

Ny Nz [0 0 11] 1 0 0 0 1 0 0 0 0 cos^ -sin0 0 0 sin^ COS0 0 0 0 0 1 0 cos <9 sin0 0 0 0 -sin Θ 0 cos0 0 0 1 (2.3- 22) =[〇sin^ cos0 1]

Then add the lateral slope change. As shown in Fig. 12, the mirror surface (10) produces a slope change along the X-axis, that is, after the y-axis is rotated by /3 degrees, so that the mirror normal vector is $, /3 substitution type (2.3-12), then normal vector

Nxy is 'X l' = [{Ny)x (Ny)y (ΝΑ =[0 sin(^) cos(^) 1]

Cos;0 0 - sin P -sin 夕0 0 0 0 COSy0 0 0 0 0 cos;0 0 -siny^ o 1 siny9 0 0 0 0 COSy0 0 0 0 0 =[cos ^ sin/? sin^ cos^cosy ^ l]....................(2.3-23) After the slope is changed, the mirror normal vector is $, so that the redundant reflection vector is $, which will be ^ With % substitution (2.3-16), then:

=[ΑΟχ ΑΟγ AOz l]· 19 200938804 1 0 0 0 0 1 0 0 0 0 10 -2{W^-A〇){Nxy)x -2{Ν;·Αδ){Ν^)ν -2( λΓ·Ι〇)(^)ζ 1 [0 sin(4)—cos(4) l] 1 0 ο -2 (sin ^ sin α -cos θ cos β cos a )cos Θ sin β 0 1 Ο 0 -2 (sin Θ sin a -cos Θ cos β cos cir )sin ^ 0 0 0 0 1 0 -2 (sin Θ sin a -cos Θ ^cos β cos a ) cos Θ cos β 1 [-2(sin 0 sin a -cos A.cos Θ cos> a)cos Θ sin.々sin(a) - 2(sin Θ sin a -cos Θ cos β cos a )sin Θ -cos a -2(sin Θ sin a -cos Θ cos β >cos a )cos Θ cos β ❹ .(2.3-24) Then the straight line equation for the vector formed by the defect is:

JC

(rA y = t2* ........................................... ( 2. 3. 25) ^ = h*{K)z In the above formula, t2 is a constant to be determined, and the plane equation of the screen (20) is: z = L_h ·.·... ....................................... (2. 3 — 26) 20 200938804 Solution Preform, available: 匕L-hΚΙ (2.3-27) Substituting t2 into equation (2.3-25) The intersection of the reflection vector ζ and the screen surface is: r(R) -(Lh) (R^) -(Lh) ),~~- 7~-fly——(Lh) ...............(2.3-28) The above equation is the image picker after the slope change Mirror observation of the position on the screen, subtracting the equation (2.3-21) from the equation (2.3-28), the phase shift between the X and Y directions caused by the slope change is 0X ^ , - one ( ί η \ \ 0 y · χ {φχ Φγ) tana tana (Lh) 2cos Θ sin β (cos Θ cos β cos a -sin 0 sin a ) cos Θ cos β (2cos a cos Θ cos β -2sin Θ sin a )-cos a sin a -2sin a sin2 Θ +2sin Θ cos Θ cos β cos a cos Θ cos β (2cos a cos Θ cos β -2sin Θ -sin a )-cos a /T , N sin 2β cos a cos2 Θ -sin a sin β sin 2Θ (Lh) ( (Lh) tan a (Lh) 2cos a cos2 Cos2 β -cos a -sin a sin 2Θ cos β sin a cos 20 +sin 20 cos. # cos_ a tan a 2cos丨a cos2 Θ cos2 々-cos a -sin a sin 2Θ cos The last addition of the two mirror height difference h The resulting stripe offset: 2 htana. The integrated height difference h and the phase shift in the X and y directions caused by the slope change are & U: (Z —/z) (sin 20 .cos a cos2 0 -sin a -sin sin 20 ) 2cos a cos2 Θ cos2 β -cos a -sin a isin 2Θ cos β sin a cos 2Θ +sin 2Θ cos β cos a 2cos a cos2 Θ cos2 β -cos a -sin a sin 2Θ cos β -tan a (Lh)-2 htana (2.3-30) 21 200938804 As can be seen from the above formula, when the mirror surface (1 ο) produces a height and slope change, the corresponding vertical and horizontal stripe offset, but the actual measurement time Then, the stripe offset is converted into a slope change, and then the contour of the side object is obtained. The composition variables included in the former formula are complicated, so the first step is simplified and then solved. Where: siny0 => β if both cold and 0 are extremely small (<0_ 50), and L>>h, then substituted for sin^ => 0 1

Equation (2. 3-30) Simplification 2β (cosor - esina)

L 2sina-sina + 2^cos a 2cos α-cos α-2 Θ sin a 2cos a-cos a-2 Θ sin a tan a 2β{\-θ\^ηα) 1-2 Θ tan a

=L

=L tan a+ 20 --tan a 1-2 Θ tan a 2β -2βθ\ζηα 1-20 tan a le + lOtan2 a 1-20 tan a (2.3-31) order ΧΦ

Yφ)= 2β{\-θίοηα) 1-20 tan a 20(1 + tan2 a) 1-20 tan a (2.3-33) Decomposed Cubic Formula ΧΦ (2.3-34) 2β{\ - Otana) 1-2 Θ tan a 2(9(1 + tan2 a) 1-2 Θ tana and can be obtained: 0: cos a {β θ) 2+γφ sin 2a 2 ′′ β = φν COS2 ( χφ 2 + γφ sinacosa substituted 4L + ^ Sin 2a 2L+^ysin 2a (2.3-37) 22 200938804, 'Where the offset and % are the length units, the relationship between the phase change value and the length unit is: φ(χ)~2π'^ Ρχ < .Ν2π·ώ ... 9{y) =-i- I Py Ρχ : transverse stripe spacing: vertical stripe spacing Substituting the formula (2^3-38) into the formula (2·3-37) gives Py-9 (y) C0:

Ο 2Ρχ·φ(χ) , β θ, ---_; - -__JTy 'fKJ' )^\Ja UL + Py-<P(y)^ 2a 4iZ^(3;)sin 2« J.. ..........(2.3-39) The phase deviation of the measured offset, coffee, and cloth ^^ into the equation (2·3 39) can be used to obtain the horizontal and vertical slope changes cold and 0 . Based on the slope of the surface of the object as the difference of its height', if the object is to be obtained, the slope of the surface of the object is integrated along the y and y directions, and the relationship between the length of the pixel and 2 is added as in the equation (2·2_13), (2·2·27). The surface height of the test piece in the y direction can be obtained, and its integral \βάχ H(y) = \θάγ .......................(2.3-40) The relationship between the pixel and the actual length of the substitution: y~Ry' S(pixei^ X~RX' 8{pixei^ The image resolution of the V y direction · The image resolution of the X direction Ap&o: the image pixel It is known that the height of the A^y direction is combined with the X direction and the height of the (2.2-13) (2.2-27) Y direction 23 200938804 to obtain a complete three-dimensional height profile of the object. In summary, the present invention applies the stripe. The method of measuring the surface topography of the object by the reflection method is mainly based on the optical stripe reflection measurement method combined with the geometric analysis or the vector analysis operation method, and then the calculation is correctly measured by the computer in an automatic manner and with high precision. The appearance of the surface of the object provides a more practical measurement method for the non-destructive detection of its products for the industry. BRIEF DESCRIPTION OF THE DRAWINGS The flow chart of the present invention is applied to the measurement of the appearance of an object by the stripe reflection method. The second figure is a schematic diagram of the basic apparatus for measuring the surface topography of an object by the stripe reflection method of the present invention. The third and fourth figures are respectively a schematic diagram of the stripe along the X direction and the γ direction. The fourth figure is a relationship between the mirror height and the stripe deformation plane. The sixth figure is a three-dimensional relationship between the mirror height and the stripe deformation of the present invention. The seventh diagram is a vector rotation diagram. The eighth diagram is a vector top view. The ninth diagram is a vector reflection diagram. The ten diagrams show the plane geometry configuration diagram before the slope change. The tenth diagram is the plane geometry configuration diagram along the Υ direction. A picture is a plane geometry diagram in the X direction. [Main component symbol description] (ι〇) plane mirror (reference mirror) 24 200938804 (2 0 ) screen (3 0 ) mirror to be tested (4 0 ) image capture device (5 0) Computer 〇

25

Claims (1)

200938804 X. Applying for a patent garden: Method for measuring the surface topography of an object 1 The application of the stripe reflection step includes: displaying a stripe image on the screen by a computer with a phase of the sergeant. Make the turtle on the screen + & a Appreciation, the image of the stripe on the '/, is clearly projected onto a plane mirror to pick up the shirt as a means of extraction ^ ^ ^ Β «< ^ from the plane image reflected by the plane mirror, brigade, · ' recorded In the electric bulb, make or 仗, Β! ΑΑ* ❹ as a reference for the measurement of the object to be tested; I place the object to be tested with a reflective surface on the plane mirror; project the image of the stripe displayed on the screen to On the object to be tested; X~ image captures the stripe image reflected from the object to be tested and records it in the computer; and analyzes the phase distortion of the stripe by using a computer to reflect the image of the fringe' After calculating the slope, the surface morphology of the object to be tested is obtained by the analytical means. 2. The method for measuring a surface topography of an object by using a stripe reflection method as described in claim i, wherein the step of generating a stripe image by a computer by phase shifting is displayed on a screen, using a four-step phase shift The method sets the computer program to generate four stripe image images in the X direction and the y direction, and the initial phases of the four stripe images in each direction are set to 0, π / 2, π, and 3 π /2 ° 3 respectively. The method for measuring the surface topography of an object by the stripe reflection method described in the above 2, wherein the step of analyzing the deformation amount of the stripe by using a computer to reflect the fringe image analysis means using the phase unwrapping method 26 200938804 and determining the measurement area law. 4) A method for measuring a surface topography of an object by using a stripe reflection method as described in claim 3, wherein the determined measurement area method is measured by using a phase mask method to obtain a phase map obtained by the phase mask method The gray scale distribution of the object and the background image pixels is further divided by the image threshold method to divide the gray scale distribution of the object and the background image to obtain a measurement range. 5. The method for measuring the surface topography of an object by the stripe reflection method as described in claim 4, wherein median filtering is performed to reduce noise before performing the phase unwrapping method and the determining the measurement domain method. 6. The U-strip reflection method according to the U.S. Patent No. 4 or 5, wherein the U-strip reflection method measures the shape of the surface of the object, and the analysis means uses the geometric analysis method to estimate the slope of the measurement area and integrate it. 7. A method for measuring the surface topography of an object by applying a stripe reflection method as described in Item 4, R, No. 4 of the Patent Application No. 4, wherein the analytical means uses the analytical method to calculate the measurement area. Slope and integrate. 8. The method of measuring the surface topography of the object by the stripe reflection method according to the application of the patent scope 7th Nauru Ganzi private s* main fan, the vector analytic method using the vector rotation transformation matrix and The reflection vector is dragged to the #change matrix' and combined with the stripe offset formula and the stripe offset conversion slope formula. XI, 圏: as the next page 27