A method for calculating the accuracy of the cross-section profile of a diamond anvil
The three-dimensional point cloud data of the diamond anvil is obtained through a microscope and processed, and the accuracy of its cross-axis cross-sectional contour is calculated, which solves the problem of insufficient accuracy of traditional detection methods and realizes high-precision contour detection, which is suitable for the field of ultra-precision processing.
Patent Information
- Application Number
- CN202210526615.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Priority Date
- 2022-04-11
- Filing Date
- 2022-05-16
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2042-05-16
AI Technical Summary
Traditional methods cannot accurately detect the contour accuracy of diamond anvils, especially in the field of ultra-precision processing. The prior art is difficult to achieve the detection accuracy of nm level, and there are errors and inconsistencies in manual detection.
The three-dimensional point cloud data of the diamond anvil was obtained through a microscope, denoising and coordinate transformation were performed, and its cross-axis section profile accuracy was calculated. The best fit contour calculation method was used to achieve high-precision contour detection.
It realizes high-precision detection of diamond anvil profile accuracy, reaching the nm order, and is suitable for anvil of complex slewing symmetric microstructures, improving detection efficiency and reducing the risk of damage to the surface of the part.
Smart Images

Figure CN114972636B_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field of ultra-precision machining, and in particular relates to a method for calculating the accuracy of a cross-axis cross-section profile of a diamond anvil. Background Art
[0002] In recent years, anvils containing complex rotationally symmetrical microstructures have been widely used in the field of ultra-precision machining due to their superior performance. The anvils are accompanied by wear during use, and this wear will inevitably cause changes in the surface accuracy of the machined parts. Therefore, regular contour accuracy testing of the anvils is not only beneficial for determining the service life of the anvils, but also has important significance for controlling the surface accuracy of ultra-precision machined parts.
[0003] The anvils used in the field of ultra-precision machining have complex structures and extremely high quality control requirements. The contour detection accuracy is generally required to reach the nm level. The traditional manual detection method mainly completes the detection by manually operating the profilometer to read the numerical value. The measuring probe of the traditional profilometer causes certain damage to the surface of the part during detection, and the manual detection method is prone to false detection, missed detection and inconsistent standards. The detection accuracy of the profilometer is low and it is difficult to achieve nm-level detection accuracy, so it cannot meet the detection needs of the anvil.
[0004] In recent years, with the development of three-dimensional imaging technology, the detection accuracy of three-dimensional imaging detection equipment such as atomic force microscopes and confocal microscopes can reach the nm level. However, such equipment can often only obtain high-precision point cloud information of the anvil. Due to the lack of calculation methods for the cross-sectional profile of the point cloud, it is currently impossible to accurately obtain the contour accuracy of the anvil used in actual cutting processing.
[0005] Therefore, there is an urgent need for a method for calculating the accuracy of the cross-section profile of a diamond anvil through the axis, which can obtain the profile accuracy of the cross-section of a diamond anvil through the axis for use in the field of ultra-precision machining. Summary of the invention
[0006] In view of this, the present invention provides a method for calculating the accuracy of the cross-section profile of a diamond anvil through an axis, and the method processes the three-dimensional point cloud data of the diamond anvil obtained by a microscope to obtain the accuracy of the cross-section profile of the diamond anvil through an axis. Compared with the traditional manual detection method, the method disclosed by the present invention has high detection accuracy, and the accuracy detection of the diamond anvil profile can reach the nm level, and the method is also applicable to the profile accuracy detection of other diamond anvils with rotating microstructures.
[0007] To achieve this purpose, the present invention adopts the following technical scheme: a method for calculating the accuracy of the cross-section profile of a diamond anvil, the method being applicable to a diamond anvil with a complex rotationally symmetrical microstructure, the diamond anvil comprising a circular cutting plane, an annular base surface, and a rotational curved surface connecting the machining cutting plane and the base surface; the method comprising the following steps:
[0008] 1.1. Use a microscope to measure the diamond anvil and obtain multiple sets of three-dimensional point cloud data of the diamond anvil in the microscope coordinate system;
[0009] 1.2. After splicing and fusing multiple sets of 3D point cloud data, the 3D point cloud data P0 of the anvil is obtained;
[0010] 1.3. De-noise the three-dimensional point cloud data P0 to obtain the effective three-dimensional point cloud data P of the anvil f ;
[0011] 1.4. Calculation of the effective 3D point cloud data P of the anvil f The rotary axis l;
[0012] 1.5. Effective 3D point cloud data P for anvil f Coordinate transformation is performed to transform the effective three-dimensional point cloud data P of the anvil under the microscope coordinates. f Transform to the new coordinate system O-x'y'z' in which the z-axis and the rotation axis l direction vector are consistent, and obtain the transformed three-dimensional effective point cloud data P T ;
[0013] 1.6. Obtaining 3D effective point cloud data P T The centroid coordinates of the shape are used to further obtain the contour line f of any section P passing through the rotation axis l;
[0014] 1.7. Solve for the best fitting contour line f corresponding to any section P b , and calculate the fitted contour line f point by point b The height difference between the anvil and the contour line f can be used to obtain the contour accuracy of any cross section P of the anvil.
[0015] Preferably, in step 1.1, the plurality of sets of three-dimensional point cloud data include three-dimensional point cloud data P obtained by scanning the diamond anvil by microscope. fac The 3D point cloud data P obtained by reverse scanning the diamond anvil ant .
[0016] Preferably, in step 1.2, the splicing and fusion of multiple sets of three-dimensional point cloud data is to merge the three-dimensional point cloud data P fac and 3D point cloud data P ant Take the first half separately for splicing.
[0017] Preferably, the step 1.3 comprises:
[0018] 1.3.1: Convert the 3D point cloud data P0 into a binary image. The z coordinate value of the 3D point cloud data P0 is proportionally converted into the pixel value of the binary image. The x and y values are converted into the horizontal and vertical index values of the binary image. Then, the area T between the inner contour ring line O3 of the base surface and the edge of the image is precisely extracted in the binary image. e ;
[0019] 1.3.2: Extracting region T e The index values x and y of all points in T are found according to the index values. e The three-dimensional point cloud data points corresponding to all the points in the anvil are deleted from the three-dimensional point cloud data P0, and the retained point cloud data constitute the effective three-dimensional point cloud data P0. f .
[0020] Preferably, in step 1.3.1, the area T between the inner contour O3 of the base surface and the edge of the image is precisely extracted. e The process is as follows:
[0021] 1.3.1.1: Solve the gradient values of all data points in the three-dimensional point cloud data P0;
[0022] 1.3.1.2: Convert the 3D point cloud data P0 into a binary image, and convert the z coordinate value of the 3D point cloud data P0 into a binary image pixel value, where the pixel value of all points in P0 whose gradient value is greater than 1 / 10 of the microscope error value is 1, and the pixel value of all other points is 0; the x and y values are converted into the horizontal and vertical index values of the binary image, and the binary image I1 of the 3D point cloud data P0 is obtained;
[0023] 1.3.1.3: After multiple dilation and erosion processing, thinning and clipping operations are performed on the binary image I1, the processed binary image I2 is obtained;
[0024] 1.3.1.4: Search outward from the center of the binary image I2 to extract loop lines O1 and O2, which are the inner and outer contours of the rotational surface on the plane of the binary image I2 respectively;
[0025] 1.3.1.5: Let the area enclosed by contour lines O1 and O2 be T r , and let T r The pixel value of all points in the region is 1;
[0026] 1.3.1.6: Extract the three-dimensional point cloud data P of the base surface from the three-dimensional point cloud data P0 using the random sampling consensus method B , the P B With holes and error points;
[0027] 1.3.1.7: 3D point cloud data P B Convert to binary image I B , and set the 3D point cloud data P B The pixel values of all projected points projected into the binary image plane are 1, and the remaining points in the binary image plane are 0;
[0028] 1.3.1.8: Eliminate binary images I B and region T r The intersection of , extracts the area T1 where the pixel value is 1;
[0029] 1.3.1.9: Perform “hole filling” processing on region T1 to remove the small holes caused by the error points and obtain region T h , and the area T h The pixel values of all points are set to 1, and the pixel values of other positions are set to 0, and the binary image I is obtained. h ;
[0030] 1.3.10: For binary image I h Perform dilation processing to obtain a binary image I n ;
[0031] 1.3.11: Using binary image I n Subtract I h , and obtain a binary image I containing the inner and outer contours of the base surface e ;
[0032] 1.3.12: From binary image I e Search outward from the center, extract the ring line O3 in the center, and extract O3 and the binary image I e The area enclosed by the edge is T e .
[0033] Preferably, the random sampling consensus method in step 1.3.1.6 adopts the RANCAS algorithm, the number of iterations of the algorithm is set to 5000 times, and the maximum value of the machining tolerance of the diamond anvil plane and the maximum error of the microscope height measurement is used as the segmentation threshold of the algorithm.
[0034] Preferably, in step 1.5, the effective three-dimensional point cloud data P of the anvil f The process of coordinate transformation is as follows:
[0035] 1.5.1: Rotate the z-axis of the microscope coordinate system O-xyz clockwise around the x-axis by an angle θ so that the z-axis rotates to Direction, rotation angle The rotation matrix at this time is Then rotate the rotated z-axis around the y-axis by an angle of β to Direction, rotation angle The rotation matrix is now Get the rotation matrix R T =R x R y , in the above formula represent the average gradient respectively; and They represent the effective three-dimensional point cloud data of the anvil P f The gradient value of any point (x, y, z) in the x and y directions, d x With d y The effective three-dimensional point cloud data P of the anvil are shown respectively f The spacing between sampling points in the x and y directions; n represents the effective three-dimensional point cloud data P of the anvil f The total number of midpoints; z represents the coordinate value of the anvil effective three-dimensional point cloud data point in the microscope coordinate system O-xyz;
[0036] 1.5.2: Obtain the rotated 3D point cloud data P e , P e =P f R T ;
[0037] 1.5.3: The rotated 3D point cloud data P e The effective three-dimensional point cloud data P of the anvil after translation is obtained. T , assuming P e The coordinates of any point in the e ,y e ,z e ), the point is at P T The coordinates of the corresponding point in are (x', y', z'), then the two satisfy: x' = x e -x1,y'=y e -y1,z'=z e , where x1=∑ t x t / n,y1=∑ t y t / n,x t and t is the 3D point cloud data P e The horizontal and vertical coordinates of any point in the anvil, t represents the tth point cloud data, and t∈[1,n], n represents the effective three-dimensional point cloud data P of the anvil f The total number of midpoints.
[0038] Preferably, in step 1.6, the contour line f of the anvil on any cross section P is obtained by data interpolation, and the process is as follows:
[0039] 1.6.1: Determine the number of points and spacing for data interpolation;
[0040] Determine the three-dimensional effective point cloud data P T The row and column closest to the x'axis and y'axis of the new coordinate system, and determine the number of data points n on the row and column r and n c , take n r and n c The larger value in is used as the number of contour interpolation points n m , with the microscope measurement sampling spacing l r is the interpolation point spacing;
[0041] 1.6.2: Calculate the interpolation point coordinates of the cross-sectional profile: x = k l r ·cosγ,y=k·l r sinγ; where Z * is a positive integer; γ is the angle between any section through the rotation axis l and the x'Oz' plane;
[0042] 1.6.3: Perform bicubic interpolation at the interpolation point coordinates to obtain the coordinate value z' at the interpolation point k , and then obtain the point sequence P of the three-dimensional contour of any interpolated section P n ;
[0043] 1.6.4: The three-dimensional contour point sequence P n Convert to a two-dimensional contour sequence;
[0044] 1.6.5: Use cubic spline interpolation to connect P n The missing positions in the point sequence on the two-dimensional contour line sequence are used to obtain the anvil cross-section contour line f.
[0045] Preferably, the best fitting contour line f of the contour line of any cross section P of the anvil is solved in step 1.7 b The process is: the ideal contour line f0 corresponding to any section P of the anvil is translated up and down along the contour line f of any section P of the anvil, so that the sum of square errors between f0 and the ordinates of the corresponding points on f is minimized. At this time, the ideal contour line is the best fitting contour line f b .
[0046] The beneficial effects of the present invention are as follows: the accuracy calculation method of the through-axis cross-sectional profile of the diamond anvil disclosed in the present invention proposes a specific high-precision profile calculation method for the anvil data measured by a high-precision measuring instrument, thereby realizing high-precision non-contact automated calculation of the anvil profile accuracy. Compared with the traditional manual detection method, the calculation method has high detection accuracy and can achieve calculation accuracy of the order of nm. At the same time, it does not damage the surface of the part, does not require manual participation throughout the process, and improves the calculation efficiency. The method is also suitable for the profile accuracy detection of other diamond anvils with rotating microstructures. BRIEF DESCRIPTION OF THE DRAWINGS
[0047] Figure 1 It is a flow chart of the method for calculating the accuracy of the cross-axis cross-sectional profile of the diamond anvil of the present invention;
[0048] Figure 2 is the rotation contour line of the diamond anvil structure in the embodiment of the present invention;
[0049] Figure 3 (a) is the binary image I1 of the three-dimensional point cloud data P0; (b) is the image after three dilations of the binary image I1; (c) is the image after three erosions of the dilated image; (d) the skeleton structure is extracted from the eroded image; (e) the image is imaged after debranching; (f) the binary image I2 is obtained after the stray points of the image are deleted;
[0050] Figure 4 (a) is a schematic diagram of the inner and outer contours of the surface of revolution 2 on the binary image I2 plane; (b) the area T enclosed by the loops O1 and O2 on the binary image I2 r Schematic diagram; (c) binary image I B (d) Schematic diagram of region T1; (e) binary image I h Schematic diagram of the base surface in the binary image I e Schematic diagram of inner and outer contours on;
[0051] Figure 5 The region T in the embodiment e Schematic diagram of
[0052] Figure 6 (a) is the effective three-dimensional point cloud data P of the anvil in the embodiment f (b) is the effective three-dimensional point cloud data P of the anvil f The cross-axis section contour line;
[0053] Figure 7 is the best fitting contour line f in the embodiment b Schematic diagram after superimposition with the through-axis section contour line f. DETAILED DESCRIPTION
[0054] Those skilled in the art will appreciate that the embodiments described herein are intended to help readers understand the principles of the present invention, and should be understood that the protection scope of the present invention is not limited to such specific statements and embodiments. Those skilled in the art can make various other specific variations and combinations that do not deviate from the essence of the present invention based on the technical revelations disclosed by the present invention, and these variations and combinations are still within the protection scope of the present invention.
[0055] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments.
[0056] A method for calculating the accuracy of a cross-section profile of a diamond anvil. As an embodiment, the method is applicable to a diamond anvil having a complex rotationally symmetrical microstructure. Figure 2 The contour line shown in the figure is obtained by rotating, including a circular cutting plane 1 with a smaller radius, a circular base surface 3 with a larger radius, and a rotational surface 2 connecting the machining cutting plane 1 and the base surface 3. The process of calculating the contour accuracy of any cross-axis section of the diamond anvil by this method is as follows: Figure 1 As shown, specifically including:
[0057] Step 1: Use a microscope to perform forward and reverse scanning on the diamond anvil. The forward scanning obtains the three-dimensional point cloud data P fac , reverse scanning to obtain the three-dimensional point cloud data P ant ,When the microscope scans, the height measurement direction is approximately consistent with the anvil rotation axis direction;
[0058] The atomic force microscope is a contact-type measuring device, which adopts a line-by-line scanning method to acquire the three-dimensional point cloud data of the diamond anvil. However, due to the inertia of the contact end of the atomic force microscope, when encountering a position with a sudden change in height difference, the measuring end is no longer close to the object, which is prone to measurement errors. Therefore, it is necessary to scan from left to right to obtain the forward data and from right to left to obtain the reverse data. In this way, the error caused by the rapid change in object height can be removed through data fusion.
[0059] Step 2: Take positive scans to obtain 3D point cloud data P fac And reverse scanning to obtain the three-dimensional point cloud data P ant The three-dimensional point cloud data of the first half are spliced and fused to obtain the three-dimensional point cloud data P0 of the anvil;
[0060] Step 3: De-noise the 3D point cloud data P0 to obtain the effective 3D point cloud data P of the anvil f , the specific denoising process is as follows:
[0061] 1. Convert the 3D point cloud data P0 into a binary image. The z coordinate value of the 3D point cloud data P0 is converted into the pixel value of the binary image in proportion. The x and y values are converted into the horizontal and vertical index values of the binary image. Then, the area T between the inner contour ring line O3 of the base surface and the edge of the image is precisely extracted in the binary image. e , as follows:
[0062] (1) Solve the gradient values of all data points in the three-dimensional point cloud data P0. The process is as follows:
[0063]
[0064]
[0065]
[0066] Among them, x, y, and z are the coordinates of the midpoint of the three-dimensional point cloud data in the microscope coordinate system O-xyz, respectively. x With d y They represent the spacing between the sampling points of the three-dimensional point cloud data P0 in the x and y directions respectively; n represents the effective three-dimensional point cloud data P of the anvil f The total number of midpoints; z represents the coordinate value of the anvil effective three-dimensional point cloud data point in the microscope coordinate system O-xyz, and M represents the gradient value;
[0067] (2) The three-dimensional point cloud data P0 is converted into a binary image. The z coordinate value of the three-dimensional point cloud data P0 is converted into a binary image pixel value. The pixel values of all points in P0 whose gradient values are greater than 1 / 10 of the microscope error value are replaced with 1, and the pixel values of all other points are replaced with 0. The x and y values are converted into the horizontal and vertical index values of the binary image. The binary image I1 of the three-dimensional point cloud data P0 is obtained, as shown in Figure 3 (a)
[0068] (3) After performing multiple dilation, erosion, thinning and clipping operations on the binary image I1, a processed binary image I2 is obtained; as an embodiment, the structural operator used in the dilation and erosion processing is The number of dilation and erosion is 3 times, and the thinning and clipping are mainly to remove the data branches and stray points in the two-dimensional image. The process is as follows Figure 3 As shown in (b) to (f);
[0069] (4) Search outward from the center of the binary image I2 to extract loop lines O1 and O2, which are the inner and outer contours of the rotational surface on the plane of the binary image I2, respectively. Figure 4 (a)
[0070] (5) Let the area enclosed by contour lines O1 and O2 be T r , and let Tr The pixel value of all points in the region is 1, such as Figure 4 (b)
[0071] The above steps (1) to (5) roughly extract the inner and outer contours of the surface of revolution 2 on the binary image I2 plane, and the area between the two contours. Due to the limitations of the above operation process, the data points currently extracted are all rough extraction results. The next step is to perform a series of operations to accurately extract the effective point cloud.
[0072] (6) Using the random sampling consensus method, extract the three-dimensional point cloud data P of the base surface 3 from the three-dimensional point cloud data P0 B , the extracted P B With holes and error points. The random sampling consensus method uses the RANCAS algorithm, sets the number of iterations of the algorithm to 5000 times, and takes the maximum value of the machining tolerance of the diamond anvil plane and the maximum error of the microscope height measurement as the segmentation threshold of the algorithm;
[0073] (7) The three-dimensional point cloud data P B Convert to binary image I B , and set the 3D point cloud data P B The pixel values of all projected points on the binary image plane are 1, and the remaining points on the binary image plane are 0, so the binary image I is obtained. B ,like Figure 4 (c) shows that, obviously, Figure 4 (c) shows that the binary image I B Including the inner contour of the surface of revolution 2;
[0074] (8) Eliminate binary image I B and region T r The intersection of the two images is used to extract the area T1 with a pixel value of 1, that is, the binary image I is removed. B The inner contour of the upper revolution surface 2 is as follows: Figure 4 (d)
[0075] (9) Since there are holes in area T1, it is necessary to perform “hole filling” processing on image T1 to remove the small holes caused by the error points. The area after hole filling is T h , and the area T h The pixel values of all points are set to 1, and the pixel values of other positions are set to 0, and the binary image I is obtained. h ,like Figure 4 (f)
[0076] (10) For binary image I h Perform dilation processing to obtain a binary image I n , and then use the binary image I n Subtract Ih , and obtain a binary image I containing the inner and outer contours of the base surface 3 e , from image I e Search outward from the center and extract the ring line O3 in the center (i.e., the base plane 3 in the binary image I h The inner contour line on Figure 4 As shown in (f), O4 is the outer contour of the base surface 3; extract O3 and the binary image I e The area enclosed by the edge is T e ,like Figure 5 Shown
[0077] 2. Extract area T e The index values x and y of all points in T are found according to the index values. e The three-dimensional point cloud data points corresponding to all the points in the anvil are deleted from the three-dimensional point cloud data P0, and the retained point cloud data constitute the effective three-dimensional point cloud data P0. f The effective three-dimensional point cloud data of the anvil P f It includes all point cloud data within the revolution surface 2 and the cutting plane 1. The reason why the point cloud data outside the revolution surface 2 is eliminated is that the base surface 3 of the anvil itself is a rough surface, which has little effect on the actual cutting. Therefore, when calculating the accuracy of the anvil profile section, the base surface 3 and the invalid data outside its area must be eliminated. The valid three-dimensional point cloud data P f The three-dimensional data point cloud and contour lines are as follows Figure 6 As shown in (a) and (b).
[0078] Step 4: Calculate the effective 3D point cloud data P of the anvil f The rotation axis l of the anvil is not possible to completely coincide with the z-axis of the microscope coordinate system when the anvil is placed because the microscope is used to scan the anvil. Therefore, the anvil rotation axis cannot be used as the effective three-dimensional point cloud data P of the anvil. f The rotary axis needs to be recalculated. The specific calculation method is as follows:
[0079] (1) Solving the effective three-dimensional point cloud data P of the anvil f The centroid of i x i / n,y0=∑ i y i / n,z0=∑ i z i / n, where (x i ,y i ,z i ) is the effective 3D point cloud data P fThe coordinate value of any point in the anvil, where i represents the i-th point cloud data, and i∈[1,n]; n represents the effective three-dimensional point cloud data P of the anvil f Total number of midpoints;
[0080] (2) Solve the average gradient value, in and The calculation of is shown in formulas (1) and (2);
[0081] (3) The equation of the rotation axis l is obtained:
[0082] Step 5: Valid 3D point cloud data P for the anvil f Coordinate transformation is performed to transform the effective three-dimensional point cloud data P of the anvil under the microscope coordinates. f Transform to the new coordinate system O-x'y'z' in which the z-axis and the rotation axis l direction vector are consistent, and obtain the transformed three-dimensional effective point cloud data P T ;
[0083] (1) Rotate the z-axis of the microscope coordinate system O-xyz clockwise around the x-axis by an angle θ so that the z-axis rotates to Direction, rotation angle The rotation matrix at this time is Then rotate the rotated z-axis around the y-axis by an angle of β to Direction, rotation angle The rotation matrix is now Get the rotation matrix R T =R x R y ;
[0084] (2) Obtain the rotated 3D point cloud data P e , P e =P f R T ;
[0085] (3) The rotated 3D point cloud data P e The effective three-dimensional point cloud data P of the anvil after translation is obtained. T , assuming P e The coordinates of any point in the e ,y e ,z e ), the point is at P T The coordinates of the corresponding point in are (x', y', z'), then the two satisfy: x' = x e -x1,y'=y e -y1,z'=z e , where x1=∑ t x t / n,y1=∑ t y t / n,x t and t is the 3D point cloud data P e The horizontal and vertical coordinates of any point in the anvil, t represents the tth point cloud data, and t∈[1,n], n represents the effective three-dimensional point cloud data P of the anvil f The total number of midpoints.
[0086] Step 6: Obtain 3D effective point cloud data P T The centroid coordinates of the anvil and the contour line f of any section P passing through the rotary axis are further obtained. The process is as follows:
[0087] (1) Calculate the three-dimensional effective point cloud data P T The centroid coordinates of the anvil and the arbitrary section P through the rotation axis are obtained;
[0088] (2) Obtain the contour line f of the anvil corresponding to the cross section P by numerical interpolation;
[0089] First, determine the number and spacing of data interpolation points: Determine the three-dimensional effective point cloud data P T The row and column closest to the x' and y' axes, and the number of data points n on the row and column r and n c , take n r and n c The larger value in is used as the number of contour interpolation points n m , with the microscope measurement sampling spacing l r is the interpolation point spacing;
[0090] Next, calculate the interpolation point coordinates of the cross-sectional profile P: x = k·l r ·cosγ,y=k·l r sinγ; where Z * is a set of positive integers; γ is the angle between the section P through the axis of rotation and the x'Oz' plane;
[0091] After that, bicubic interpolation is performed at the coordinates of the interpolation point to obtain z' at the interpolation point k Coordinate values, and then obtain the point sequence P of the three-dimensional contour of any interpolated section P n ;
[0092] Then, the three-dimensional contour point sequence P n (x n ,y n ,z n ) is converted into a two-dimensional contour sequence (ρ n ,h n ),in h n =z n ;
[0093] Finally, cubic spline interpolation is used to connect P n The vacant positions in the point sequence on the two-dimensional contour line sequence are used to obtain the contour line f of the anvil cross section P.
[0094] Step 7: Find the best fitting contour line f corresponding to any section P b , and calculate the fitted contour line f point by point b The difference between the contour line f of any section P and the contour accuracy of any cross-axis section P of the anvil can be obtained, and the best fitting contour line f can be solved. b The process is: the ideal contour line f0 corresponding to any section P of the anvil is translated up and down along the contour line f of any section P, so that the sum of square errors between f0 and the ordinates of the corresponding points on f is minimized. At this time, the ideal contour line is the best fitting contour line f b , the best fitting curve f can also be solved by simulated annealing method b .
[0095] The reason for solving the best fitting contour line f b The contour accuracy of any section P is calculated because it is impossible to determine the position of the ideal contour line f0 on the plane where the contour line f of any section P of the anvil is located. Therefore, the above method is used to obtain the best fitting contour line f of f0 on the contour line f. b Perform precision calculations, such as Figure 7 shown.
Claims
1. A method for calculating the accuracy of a cross-section profile of a diamond anvil, characterized in that the method is applicable to a diamond anvil with a complex rotationally symmetrical microstructure, wherein the diamond anvil comprises a circular cutting plane, an annular base surface, and a rotational curved surface connecting the cutting plane and the base surface; the method comprises the following steps: 1.
1. Use a microscope to measure the diamond anvil and obtain multiple sets of three-dimensional point cloud data of the diamond anvil in the microscope coordinate system; 1.
2. After splicing and fusing multiple sets of 3D point cloud data, the 3D point cloud data P0 of the anvil is obtained; 1.
3. De-noise the three-dimensional point cloud data P0 to obtain the effective three-dimensional point cloud data P of the anvil f ; 1.
4. Calculation of the effective 3D point cloud data P of the anvil f The rotary axis l; 1.
5. Effective 3D point cloud data P for anvil f Coordinate transformation is performed to transform the effective three-dimensional point cloud data P of the anvil under the microscope coordinates. f Transform to the new coordinate system O-x'y'z' in which the z-axis and the rotation axis l direction vector are consistent, and obtain the transformed three-dimensional effective point cloud data P T ; 1.
6. Obtaining 3D effective point cloud data P T The centroid coordinates of the shape are used to further obtain the contour line f of any section P passing through the rotation axis l; 1.
7. Solve for the best fitting contour line f corresponding to any section P b , and calculate the fitted contour line f point by point b The height difference between the anvil and the contour line f, and then the contour accuracy of any cross section P of the anvil is obtained; The step 1.3 comprises: 1.3.1: Convert the 3D point cloud data P0 into a binary image. The z coordinate value of the 3D point cloud data P0 is proportionally converted into the pixel value of the binary image. The x and y values are converted into the horizontal and vertical index values of the binary image. Then, the area T between the inner contour ring line O3 of the base surface and the edge of the image is precisely extracted in the binary image. e ; 1.3.2: Extracting region T e The index values x and y of all points in T are found according to the index values. e The three-dimensional point cloud data corresponding to all points in the anvil are obtained, and these three-dimensional point cloud data are deleted from the three-dimensional point cloud data P0, then the retained three-dimensional point cloud data constitute the anvil effective three-dimensional point cloud data P f ; The area T between the inner contour O3 of the base surface and the edge of the image is extracted in step 1.3.
1. e The process is as follows: 1.3.1.1: Solve the gradient values of all data points in the three-dimensional point cloud data P0; 1.3.1.2: Convert the 3D point cloud data P0 into a binary image, and convert the z coordinate value of the 3D point cloud data P0 into a binary image pixel value, wherein the pixel value of all points in P0 whose gradient value is greater than 1 / 10 of the microscope error value is 1, and the pixel value of all other points is 0; the x and y values are converted into the horizontal and vertical index values of the binary image, and the binary image I1 of the 3D point cloud data P0 is obtained; 1.3.1.3: After performing multiple expansion and erosion processing, as well as thinning and clipping operations on the binary image I1, the processed binary image I2 is obtained; 1.3.1.4: Search outward from the center of the binary image I2 to extract loop lines O1 and O2, which are the inner and outer contours of the rotational surface on the plane of the binary image I2 respectively; 1.3.1.5: Let the area enclosed by contour lines O1 and O2 be T r , and let T r The pixel value of all points in the region is 1; 1.3.1.6: Extract the three-dimensional point cloud data P of the base surface from the three-dimensional point cloud data P0 using the random sampling consensus method B , the P B With holes and error points; 1.3.1.7: 3D point cloud data P B Convert to binary image I B , and set the 3D point cloud data P B The pixel values of all projected points projected into the binary image plane are 1, and the remaining points in the binary image plane are 0; 1.3.1.8: Eliminate binary images I B and region T r The intersection of , extracts the area T1 where the pixel value is 1; 1.3.1.9: Perform “hole filling” processing on region T1 to remove the small holes caused by the error points and obtain region T h , and the area T h The pixel values of all points are set to 1, and the pixel values of other positions are set to 0, and the binary image I is obtained. h ; 1.3.1.10: For binary image I h Perform dilation processing to obtain a binary image I n ; 1.3.1.11: Using binary image I n Subtract I h , and obtain a binary image I containing the inner and outer contours of the base surface e ; 1.3.1.12: From binary image Ie Search outward from the center, extract the ring line O3 in the center, and extract O3 and the binary image I e The area enclosed by the edge is T e ; In step 1.4, the rotary axis l is calculated, and the specific calculation method is as follows: 1.4.1: Determine the effective 3D point cloud data P of the anvil f The centroid of i x i / n,y0=Σ i y i / n,z0=Σ i z i / n, where (x i ,y i ,z i ) is the effective 3D point cloud data P f The coordinate value of any point in the anvil, where i represents the i-th point cloud data, and i∈[1,n]; n represents the effective three-dimensional point cloud data P of the anvil f Total number of midpoints; 1.4.2: Solve for the average gradient value, in and The calculation of is as follows: Among them, x, y, and z are the coordinates of the midpoint of the three-dimensional point cloud data in the microscope coordinate system O-xyz, respectively. x With d y Respectively represent the spacing between the sampling points of the three-dimensional point cloud data P0 in the x and y directions; and They represent the effective three-dimensional point cloud data of the anvil P f The gradient value of any point (x, y, z) in the x and y directions, n represents the effective three-dimensional point cloud data P of the anvil f The total number of midpoints; z represents the coordinate value of the anvil effective three-dimensional point cloud data point in the microscope coordinate system O-xyz; 1.4.3: Get the equation of the rotation axis l: In step 1.5, the effective three-dimensional point cloud data P of the anvil is f The process of coordinate transformation is as follows: 1.5.1: Rotate the z-axis of the microscope coordinate system O-xyz clockwise around the x-axis by an angle θ so that the z-axis rotates to Direction, rotation angle The rotation matrix at this time is Then rotate the rotated z-axis around the y-axis by an angle of β to Direction, rotation angle The rotation matrix is now Get the rotation matrix R T =R x R y , in the above formula represent the average gradient respectively; and They represent the effective three-dimensional point cloud data of the anvil P f The gradient value of any point (x, y, z) in the x and y directions, d x With d y The effective three-dimensional point cloud data P of the anvil are shown respectively f The spacing between sampling points in the x and y directions; n represents the effective three-dimensional point cloud data P of the anvil f The total number of midpoints; z represents the coordinate value of the anvil effective three-dimensional point cloud data point in the microscope coordinate system O-xyz; 1.5.2: Obtain the rotated 3D point cloud data P e , P e =P f R T ; 1.5.3: The rotated 3D point cloud data P e The effective three-dimensional point cloud data P of the anvil after translation is obtained. T , assuming P e The coordinates of any point in the e ,y e ,z e ), the point is at P T The coordinates of the corresponding point in are (x', y', z'), then the two satisfy: x' = x e -x1,y'=y e -y1,z'=z e , where x1 = Σ t x t / n,y1=Σ t y t / n,x t and t is the 3D point cloud data P e The horizontal and vertical coordinates of any point in the anvil, t represents the tth point cloud data, and t∈[1,n], n represents the effective three-dimensional point cloud data P of the anvil f Total number of midpoints; In step 1.6, the contour line f of the anvil on any cross section P is obtained by data interpolation, and the process is as follows: 1.6.1: Determine the number of points and spacing for data interpolation; Determine the three-dimensional effective point cloud data P T The row and column closest to the x'axis and y'axis of the new coordinate system, and determine the number of data points n on the row and column r and n c , take n r and n c The larger value in is used as the number of contour interpolation points n m , with the microscope measurement sampling spacing l r is the interpolation point spacing; 1.6.2: Calculate the interpolation point coordinates of the cross-sectional profile: x = k l r ·cosγ,y=k·l r sinγ; where Z * is a set of positive integers; γ is the angle between any section through the axis of rotation l and the x'Oz' plane; 1.6.3: Perform bicubic interpolation at the interpolation point coordinates to obtain the coordinate value z' at the interpolation point k , and then obtain the point sequence P of the three-dimensional contour of any interpolated section P n ; 1.6.4: The three-dimensional contour point sequence P n Convert to a two-dimensional contour sequence; 1.6.5: Use cubic spline interpolation to connect P n The missing positions in the point sequence on the two-dimensional contour line sequence are used to obtain the anvil cross-section contour line f.
2. The method for calculating the accuracy of the cross-sectional profile of a diamond anvil according to claim 1, characterized in that: In step 1.1, the plurality of sets of three-dimensional point cloud data include three-dimensional point cloud data P obtained by scanning the diamond anvil by microscope. fac The 3D point cloud data P obtained by reverse scanning the diamond anvil ant .
3. The method for calculating the accuracy of the cross-sectional profile of a diamond anvil according to claim 2, characterized in that: In step 1.2, the splicing and fusion of multiple sets of three-dimensional point cloud data is to merge the three-dimensional point cloud data P fac and 3D point cloud data P ant Take the first half separately for splicing.
4. The method for calculating the accuracy of the cross-section profile of a diamond anvil according to claim 1, characterized in that: The random sampling consensus method in step 1.3.1.6 adopts the RANCAS algorithm, sets the number of iterations of the algorithm to 5000 times, and uses the maximum value of the machining tolerance of the diamond anvil plane and the maximum error of the microscope height measurement as the segmentation threshold of the algorithm.
5. The method for calculating the accuracy of the cross-section profile of a diamond anvil according to claim 1, characterized in that: The best fitting contour line f of the contour line of any cross section P of the anvil is obtained in step 1.7 b The process is: the ideal contour line f0 corresponding to any section P of the anvil is translated up and down along the contour line f of any section P of the anvil, so that the sum of square errors between f0 and the ordinates of the corresponding points on f is minimized. At this time, the ideal contour line is the best fitting contour line f b .
