Analysis method of encoder measurement data

Abstract

The present application provides an analysis method for code disk measurement data, which relates to the field of information technology, including: obtaining surface roughness measurement data of the grating area and non-grating area of ​​the code disk, segmenting the measurement data for the surface roughness mutation area by digital image processing, and obtaining the roughness distribution map of the grating area and the non-grating area; constructing a three-dimensional geometric model of the roughness mutation area according to the roughness distribution map of the grating area and the non-grating area, calculating the stress distribution state of the protective film in the roughness mutation area by the finite element analysis method, and obtaining the stress distribution cloud map of the protective film; for the transition zone between gratings of different depths, using a scanning electron microscope to obtain the microscopic morphological features of the transition zone, extracting the width parameter of the transition zone by an image processing algorithm, and establishing a correlation model between the width of the transition zone and the grating depth. It provides a scientific basis for optimizing the lamination process of the code disk protective film, and has important engineering application value.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to an analysis method for code disk measurement data. Background Art

[0002] In the analysis of the code disk measurement data, the surface roughness at the junction of the grating area and the non-grating area of ​​the code disk suddenly changes, resulting in a complex stress distribution when the protective film is attached. This uneven stress distribution may affect the measurement accuracy and service life of the code disk. Among them, the transition zone between gratings of different depths is the key area that affects the stress distribution of the protective film. The geometric shape, size and relative position of the transition zone with the adjacent gratings will have a significant impact on the stress gradient of the protective film. Specifically, the geometric parameters of the transition zone, such as the bevel angle, step height, and step width, determine the deformation degree and stress concentration of the protective film in this area. The larger the bevel angle, the more severe the deformation of the protective film and the more serious the stress concentration; the larger the step height, the more obvious the stress difference between adjacent gratings; the smaller the step width, the steeper the stress gradient. In addition, the relative position of the transition zone and the adjacent grating will also affect the stress distribution. If the transition zone coincides with the edge of the grating, or there is a certain overlap with the grating groove, it may lead to further increase in local stress. Therefore, in order to fully understand the stress distribution characteristics of the protective film at the junction of the grating area and the non-grating area of ​​the code disk, it is necessary to comprehensively consider factors such as the geometric parameters and relative position of the transition zone, and carry out systematic theoretical analysis and experimental research. By optimizing the design of the transition zone, it is expected to achieve uniformity and gradient control of the stress distribution of the protective film, thereby improving the measurement performance and reliability of the code disk. Summary of the invention

[0003] In order to solve the above technical problems, the present invention provides a method for analyzing code disk measurement data, which mainly includes:

[0004] The surface roughness measurement data of the grating area and the non-grating area of ​​the code disk are obtained. For the surface roughness mutation area, the measurement data is segmented by digital image processing to obtain the roughness distribution map of the grating area and the non-grating area; according to the roughness distribution map of the grating area and the non-grating area, a three-dimensional geometric model of the roughness mutation area is constructed, and the stress distribution state of the protective film in the roughness mutation area is calculated by the finite element analysis method to obtain the stress distribution cloud map of the protective film; for the transition zone between gratings of different depths, a scanning electron microscope is used to obtain the microscopic morphological characteristics of the transition zone, and the width parameter of the transition zone is extracted by the image processing algorithm, and a correlation model between the transition zone width and the grating depth is established; according to the correlation model between the transition zone width and the grating depth, combined with the protective film material characteristic parameters and the stress distribution cloud map, the stress distribution gradient of the protective film in the transition zone area is analyzed by the numerical simulation method, and the change law of the stress gradient with the grating depth and the transition zone width is obtained; combined with the protective film in the transition zone The stress distribution gradient in the belt area screens out key variables, including the surface roughness difference between the grating area and the non-grating area, the transition zone width and the grating depth, to form data containing variables and the resulting stress distribution of the protective film. The regression analysis model is used to fit the data and determine the weight or coefficient of each variable to obtain the main controlling factors. The model diagnosis is checked by the determination coefficient R square and the residual diagram to evaluate the goodness of fit of the model; according to the changing law of the main controlling factors and the stress gradient, the orthogonal experimental design method is used to optimize the surface roughness of the grating area and the non-grating area, the transition zone width and the grating depth process parameters to obtain the optimal process parameter combination for the uniformity of stress distribution of the protective film, including a film thickness of 0.3 to 0.5 mm; according to the optimized process parameter combination, the front and back sides of the code disc are film-laminated, and the protective film lamination stress distribution test is carried out on the code disc after the film-laminated process to verify the effectiveness of the optimized film lamination process and ensure the reliability and stability of the code disc during use.

[0005] The technical solution provided by the embodiment of the present invention may have the following beneficial effects:

[0006] The present invention discloses a method for analyzing the measurement data of a code disk. The method measures the surface roughness of the grating area and the non-grating area of ​​the code disk, analyzes the microscopic morphology of the transition zone, establishes a correlation model between the width of the transition zone and the grating depth, and calculates the stress distribution of the protective film in combination with finite element analysis. The main control factors are determined by regression analysis, and the process parameters are optimized by orthogonal experiments to achieve optimal uniformity of stress distribution of the protective film. The present invention solves the problem that stress concentration is easily generated at the junction of the grating area and the non-grating area of ​​the code disk protective film. By optimizing parameters such as surface roughness, transition zone width and grating depth, the stress distribution of the protective film is effectively improved, and the reliability and stability of the code disk are improved. The method provides a scientific basis for optimizing the lamination process of the code disk protective film and has important engineering application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] Figure 1 The figure is a flow chart of the method for analyzing the code disk measurement data of the present invention. DETAILED DESCRIPTION

[0008] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0009] like Figure 1 The analysis method of the code disk measurement data in this embodiment may specifically include:

[0010] Step S101, obtaining surface roughness measurement data of the grating area and the non-grating area of ​​the code disk, segmenting the measurement data by digital image processing for the surface roughness mutation area, and obtaining roughness distribution maps of the grating area and the non-grating area.

[0011] A contact roughness meter is used to obtain horizontal displacement coordinate values ​​and vertical height values ​​on the surface of the code disk to obtain a two-dimensional data sequence of surface morphology; Gaussian filtering and baseline correction are performed on the two-dimensional data sequence of surface morphology, and the processed data is subjected to multi-scale decomposition using wavelet transform, and the mutation point position is obtained by judging the local area variance threshold; the original data sequence is segmented according to the mutation point position, and the position coordinates, local mean and standard deviation of each data point in the data sequence are calculated using mean shift clustering to obtain the regional division results of the grating area and the non-grating area; the arithmetic mean roughness value, maximum roughness value and ten-point average roughness value of the grating area and the non-grating area are calculated according to the regional division results to obtain the roughness parameters of each area; the roughness parameters of each area are processed by open and closed operations, the missing data points are supplemented by cubic spline interpolation, and the boundary between the grating area and the non-grating area is optimized using the regional growing algorithm to obtain the regional roughness distribution map.

[0012] Exemplarily, a contact roughness meter is used to scan and sample the surface of the code disk in the radial direction, and the surface height data is collected at a sampling interval of 0.1 micron horizontal displacement and a vertical resolution of 0.01 micron. The horizontal position coordinates and vertical height values ​​of each measuring point are recorded at the same time to obtain a two-dimensional data sequence of surface morphology. The collected surface morphology data sequence is subjected to Gaussian filtering denoising and baseline correction processing, and the processed data is subjected to multi-scale decomposition using wavelet transform. The local area variance threshold is set to 1.5 times the variance of the previous level, and the gradient change rate in the local area of ​​each measuring point is calculated. When the gradient change rate exceeds the preset threshold and the duration exceeds 5% of the sampling period, the point is marked as a mutation point. Starting from the marked mutation point position, the original data sequence is segmented, and mean shift clustering is used to divide different areas. The cluster feature vector contains the position coordinates, local mean and standard deviation of the data point. The number of cluster centers is set to 2, and the iterative calculation is performed until the change in the cluster center position is less than 0.1%, and the grating area and the non-grating area are distinguished. For the data of the divided grating area and non-grating area, the arithmetic mean roughness value, the maximum roughness value and the ten-point average roughness value are calculated respectively. The local noise is removed by opening and closing operations. The missing points are supplemented by cubic spline interpolation. The regional growth algorithm is used to optimize the boundary between the grating area and the non-grating area to generate the roughness distribution results of each area. In the roughness measurement of the grating code disk, the surface profile measurement range is set to 10 mm × 10 mm, the range is ±50 μm, the probe stylus radius is 2 μm, the contact force is set to 0.2 mN, and the probe is scanned in the radial direction at a speed of 0.5 mm per second. The horizontal displacement data and vertical height values ​​collected by the probe stylus constitute the original measurement data sequence, the vertical resolution reaches 0.01 μm, and the horizontal sampling interval is 0.1 μm. The raw data is filtered through a Gaussian filter with a cutoff wavelength of 0.8 mm to remove high-frequency noise, and the least squares method is used to fit the baseline and perform data correction. The db4 wavelet was used to perform a three-layer decomposition of the data sequence, and the variance in a local window of 256 data points was calculated. If the local variance exceeded 1.5 times the upper layer variance and lasted for more than 5% of the sampling period, i.e., 12 data points, the region was marked as a mutation region. Mean shift clustering was used to classify the marked mutation region data. The feature vector consisted of the position coordinates of the data point, the mean and standard deviation in a window of 50 data points, and the bandwidth parameter was set to 0.8. The iteration was stopped when the position change of the cluster center was less than 0.1%.The classified data of the grating area and non-grating area were statistically analyzed separately, and the arithmetic mean roughness value Ra of the grating area was calculated to be 0.8 microns, the maximum roughness value Rz was 4.2 microns, and the ten-point average roughness value Rq was 1.2 microns. The Ra of the non-grating area was 0.2 microns, Rz was 1.5 microns, and Rq was 0.4 microns. Opening and closing operations with a structure element size of 5×5 were used to eliminate local noise. For missing areas with an interval of less than 5 data points, cubic spline interpolation was used to supplement the data. Finally, the 8-neighborhood regional growing algorithm was used to optimize the boundary between the grating area and the non-grating area, and the growth threshold was set to twice the standard deviation of the local area.

[0013] Step S102, constructing a three-dimensional geometric model of the roughness mutation area according to the roughness distribution diagrams of the grating area and the non-grating area, calculating the stress distribution state of the protective film in the roughness mutation area by the finite element analysis method, and obtaining the stress distribution cloud diagram of the protective film.

[0014] The local curvature value is calculated according to the two-dimensional coordinates and height values ​​of the measuring points in the roughness distribution map, the first surface function is constructed by cubic Hermite interpolation for the area where the curvature value is greater than the preset threshold, and the second surface function is constructed by cubic spline interpolation for the remaining areas, and the calculation domain geometric model is obtained by the first surface function and the second surface function; the calculation domain geometric model is meshed by twenty-node hexahedral units, if a mutation area is detected, a preset grid size is set at the mutation area, and the grid topological relationship is obtained by transitioning outward according to the preset size ratio; a fixed constraint is applied to the bottom of the calculation domain according to the grid topological relationship, a pressure load is applied to the top of the calculation domain, a displacement symmetric boundary condition is applied to the side of the calculation domain, and the calculation domain boundary condition is obtained by Coulomb friction contact at the interface between the protective film and the substrate; the displacement field under the calculation domain boundary condition is solved by the Newton iteration method, and the Gaussian point stress value is calculated by the displacement field; the node stress value is restored according to the Gaussian point stress value, and the Mises stress distribution cloud map, the maximum principal stress distribution cloud map and the maximum shear stress distribution cloud map are obtained, and the stress gradient value is calculated by the stress distribution cloud map.

[0015] Exemplarily, the local curvature value is calculated based on the two-dimensional coordinates and height values ​​of the measurement points in the roughness distribution map. The surface function is constructed by using cubic Hermite interpolation in the area where the curvature is greater than 0.5 per micron, and cubic spline interpolation is used in the remaining areas. The boundary of the mutation area is expanded by 5 times the width of the mutation area as the calculation domain range, and the position continuity deviation of the surface is checked to be less than 0.01 microns, and the tangential continuity deviation is less than 0.1 degrees. For the constructed geometric surface, the calculation domain is meshed using twenty-node hexahedral units. The mesh size is set to 0.1 microns in the mutation area, and it transitions layer by layer outward according to the size ratio of 1.2. The size ratio of adjacent units is controlled to be less than 1.5, the minimum unit angle is greater than 30 degrees, and the unit distortion rate is less than 0.8, and the mesh topology relationship with the number of nodes generated is within 500,000. For the linear elastic constitutive relation of the protective film, the elastic modulus is set to 2000 MPa, the Poisson's ratio is set to 0.3, a fixed constraint is applied at the bottom of the calculation domain, a 0.1 MPa normal pressure and a 0.02 MPa tangential force are applied to the top, and a displacement symmetric boundary condition is set on the side. Coulomb friction contact is used at the interface between the protective film and the substrate, and the friction coefficient is 0.3. Newton iteration is used to solve the displacement field, the convergence residual is set to be less than 0.001, and the number of iterations does not exceed 100 steps. The maximum stress difference under the three mesh densities is less than 5%, the node displacement field data is extracted, the Gaussian point stress value is calculated and restored to the node, the Mises stress, maximum principal stress and maximum shear stress distribution cloud map are output, and the stress gradient value of the stress concentration area is calculated. In the three-dimensional modeling and stress analysis of the grating code disk surface, the height data obtained in the 10 mm × 10 mm measurement area is firstly calculated for curvature, and the area with a curvature radius less than 2 microns is selected as the mutation area, and the area with a curvature radius greater than 2 microns is selected as the transition area. The measurement area is divided into 20 × 20 grid units, and 4 × 4 control points are used in each unit to construct the cubic Hermite basis function. Position and tangential continuity constraints are used at the boundaries of adjacent units. The position deviation at the control boundary is within 0.01 microns, and the tangential angle deviation is within 0.1 degrees. Cubic spline interpolation is used to construct a smooth transition surface in the expansion area. Adaptive meshing is used for the constructed geometric surface. A base grid size of 0.1 microns is set in the mutation area, and it is expanded layer by layer at a ratio of 1.2 outward. A total of 480,000 twenty-node hexahedral units are generated. The unit quality control indicators include the minimum angle greater than 35 degrees and the distortion rate less than 0.75. A linear elastic material model is selected for the protective film. The material parameters include elastic modulus of 2000 MPa, Poisson's ratio of 0.3, and density of 1.2 grams per cubic centimeter. Displacement constraints are applied to the six surfaces at the bottom of the calculation domain, and a normal pressure of 0.1 MPa and a tangential force of 0.02 MPa are applied to the top. The Coulomb friction model is used for the contact surface with the substrate, and the friction coefficient is 0.3.The incremental iterative solver was used to calculate the stress field. The displacement convergence residual was set to 0.001, the stress convergence residual was set to 0.01 MPa, and the maximum number of iterations was limited to 100 steps. The grid independence was verified by comparing the stress calculation results under three different grid densities. The final output stress cloud map showed a maximum Mises stress of 225 MPa, a maximum principal stress of 180 MPa, and a maximum shear stress of 120 MPa in the mutation area, and the stress gradient reached 50 MPa per micron.

[0016] Step S103, for the transition zone between gratings of different depths, a scanning electron microscope is used to obtain the microscopic morphological features of the transition zone, the width parameter of the transition zone is extracted through an image processing algorithm, and a correlation model between the width of the transition zone and the grating depth is established.

[0017] An electron microscope is used to scan the grating transition zone area, and an original scanned image is obtained based on the scan. The original scanned image is subjected to median filtering and grayscale histogram matching processing to obtain an enhanced image; wavelet transform decomposition is performed on the enhanced image, and a feature layer with the largest local variance is selected based on the decomposition result. The edge detection result is obtained by calculating the local variance on the feature layer and setting an adaptive threshold; a morphological operation is performed based on the edge detection result, and a transition zone contour binary image is obtained through the morphological operation. If the transition zone contour binary image has edge breaks, a closed operation connection is performed; a skeleton extraction algorithm is applied to the transition zone contour binary image, and the edge point curvature is calculated based on the skeleton extraction result. The edge point pairs are screened by the curvature to obtain the transition zone width data; a support vector regression model is constructed for the transition zone width data, and a mapping relationship between the transition zone width and the grating depth is obtained based on the model parameter optimization result, and a corresponding model with a determination coefficient greater than a preset threshold is established through the mapping relationship.

[0018] Exemplarily, an electron scanning microscope is used to scan and image the grating transition zone area under the conditions of an acceleration voltage of 20 kilovolts, a working distance of 10 millimeters, an electron beam current of 50 picoamperes, and secondary electron signal collection. The scanning magnification is set to 5000 times, the resolution is 2048×1536 pixels, and the original scanned image is subjected to a 5×5 window median filter to remove noise. The contrast is increased to 0.8 and the signal-to-noise ratio is increased to 35 decibels through grayscale histogram matching. For the enhanced image, a four-layer decomposition is performed using a wavelet transform, and the third layer with the largest local variance of the image is selected as the feature layer. The local variance is calculated in a sliding window of 16×16 pixels, and the adaptive threshold is set to 1.8 times the local mean to extract the edge. The extracted edge image is subjected to a closing operation to connect the broken edges, and an opening operation is performed to remove scattered noise, so as to obtain a continuous and complete transition zone contour binary image. Apply the thin skeleton extraction algorithm to the binary image to obtain the edge line with a single pixel width, calculate the local curvature of the edge point, remove the abnormal points with a curvature greater than 0.5 per pixel, set the scan line at an interval of 5% of the image height in the vertical direction, find the left and right edge point pairs on each scan line, match the edge point pairs by the minimum distance criterion, and calculate the Euclidean distance between each pair of points to obtain the transition zone width data. According to the grating depth measurement data and transition zone width data, the abnormal values ​​exceeding the range of 3 times the standard deviation are removed, and the mapping relationship is established by support vector regression. The kernel function parameters are optimized by five-fold cross validation to take values ​​in the range of 0.005 to 0.05, and the penalty coefficient is taken in the range of 50 to 200. The parameter combination with the smallest root mean square error is selected, and the corresponding relationship between the transition zone width and the grating depth with a determination coefficient greater than 0.95 is fitted. In the grating transition band feature extraction and correlation analysis, for the grating structure with a depth range of 0.5 to 5 microns, an electron microscope was used to scan the 100×100 micron field of view at an accelerating voltage of 20 kilovolts. The electron beam current was controlled at 50 picoamperes, and the working distance was maintained at 10±0.1 mm. The secondary electron signals were collected to form an original image of 2048×1536 pixels. A 5×5 window median filter was used to eliminate salt and pepper noise from the original image. The image contrast was increased from 0.4 to 0.8 and the signal-to-noise ratio was increased from 25 decibels to 35 decibels through cumulative grayscale histogram matching. After image enhancement, four-layer discrete wavelet decomposition is used. By calculating the standard deviation of the wavelet coefficients of each layer, it is found that the standard deviation of the third layer coefficients reaches 0.45, which is significantly higher than other layers. A 16×16 pixel sliding window is used to calculate the local variance on this layer. The local variance mean is 0.2, and the edge extraction threshold is set to 0.36. The extracted edge image is closed by a structure element with a radius of 3 pixels, and then opened by a structure element with a radius of 2 pixels to obtain a continuous transition zone contour.The Zhang thin skeleton algorithm is used to extract the edge line with single pixel width from the contour image, calculate the local curvature of the edge point, remove the noise points with curvature greater than 0.5 per pixel, set a scan line every 80 pixels in the vertical direction, match the left and right edge point pairs with the closest distance on the scan line, and the median of the transition zone width data obtained by statistics is 2.8 microns, the standard deviation is 0.4 microns, and the outliers that deviate from the median by 3 times the standard deviation are removed. The mapping relationship is established by support vector regression, and the optimal kernel function parameter is determined to be 0.02 and the penalty coefficient is 150 through five-fold cross validation. The determination coefficient obtained by fitting reaches 0.97, and the root mean square error is 0.15 microns.

[0019] Step S104, based on the correlation model between the transition zone width and the grating depth, combined with the protective film material characteristic parameters and the stress distribution cloud map, a numerical simulation method is used to analyze the stress distribution gradient of the protective film in the transition zone area, and the change law of the stress gradient with the grating depth and the transition zone width is obtained.

[0020] A cubic spline interpolation function is constructed according to the grating depth data and the transition zone width data, and the cubic spline interpolation function is used to generate the grid point coordinates of the continuous curved surface contour of the transition zone; the grid point coordinates are used to divide the tetrahedral unit grid, and the tetrahedral unit grid has a preset central area unit size value and an edge area unit size value; fixed constraints and load conditions are applied to the tetrahedral unit grid, and the displacement field data and stress field data of the transition zone area are obtained through an explicit integral solver; the principal stress data and equivalent stress data of the transition zone area are extracted according to the stress field data, and the principal stress data and equivalent stress data are distributed along the depth direction and width direction of the transition zone; the stress gradient value is calculated for the principal stress data and the equivalent stress data, and the radial basis function is used to interpolate the stress gradient value. If the residual root mean square error of the interpolation operation is less than a preset threshold, the stress gradient change data with depth and width are obtained.

[0021] Exemplarily, according to the grating depth and transition zone width data in the associated database, cubic spline interpolation is used to construct the continuous surface profile of the transition zone. For the grating depth range of 0.5 to 5 microns, the vertical resolution is set to 2% of the depth value, and for the transition zone width range of 1 to 10 microns, the horizontal resolution is set to 5% of the width value. The grid points are encrypted at the curvature greater than 0.5 per micron, and the position continuity error is less than 0.01 micron, and the curvature continuity error is less than 0.1 per micron. For the constructed transition zone structure, the protective film material parameters are imported, including elastic modulus 2000 MPa, Poisson's ratio 0.3, yield strength 100 MPa, fracture toughness 0.8 MPa meter square root, and tetrahedral units are used for meshing. The unit size in the central area of ​​the transition zone is 0.05 micron, the unit size in the edge area is 0.5 micron, the adjacent unit size ratio is less than 1.2, the grid node Jacobian matrix condition number is less than 5, the unit distortion rate is less than 0.8, and the minimum dihedral angle is greater than 30 degrees. A fixed constraint was set at the bottom, a normal pressure of 0.1 MPa and a tangential force of 0.02 MPa were applied at the top, periodic boundary conditions were applied at the lateral boundaries, and the displacement field and stress field were calculated using an explicit integral solver. The time step was set to the minimum unit size divided by 1 / 10 of the material sound velocity, and the iterative residual was less than 0.001. The maximum stress difference under the three mesh densities was controlled within 5%, and the principal stress and equivalent stress distribution in the transition zone were extracted. Based on the stress field data, the stress gradient values ​​were extracted along the depth and width of the transition zone, and the outliers exceeding 3 times the standard deviation were removed. The radial basis function was used to interpolate the stress gradient data, and a Gaussian kernel function containing 50 uniformly distributed sampling points was selected. The kernel parameter was set to 0.01, and the interpolation results were cross-validated by five folds. The residual root mean square error was less than 0.1 MPa per micron, and the stress gradient data with depth and width were output. In the numerical simulation of stress distribution in the grating transition zone, a three-dimensional model of the grating structure with a depth range of 0.5 to 5 microns was established based on the measured depth and width data. The vertical resolution at a depth of 2 microns was set to 0.04 microns, and the horizontal resolution at a width of 4 microns was 0.2 microns. The grid spacing was reduced to 0.02 microns at the inflection point where the curvature value reached 0.6 per micron. The position continuity error of the transition zone contour curve constructed by cubic spline interpolation was controlled within 0.008 microns, and the curvature continuity error was less than 0.08 per micron. For the established geometric model, the material parameters of the polyimide protective film were imported, with an elastic modulus of 2000 MPa, Poisson's ratio of 0.3, a yield strength of 100 MPa, and a fracture toughness of 0.8 MPa m square root. Tetrahedral units were used for meshing, and a baseline unit size of 0.05 microns was set in the central area of ​​the transition zone, which was increased layer by layer to 0.5 microns outward. The ratio of adjacent unit sizes was controlled within 1.15. The mesh quality test showed that the maximum value of the node Jacobian matrix condition number was 4.2, the maximum value of the unit distortion rate was 0.75, and the minimum dihedral angle was 32 degrees.In the calculation boundary condition setting, a fixed constraint is applied at the bottom, a 0.1 MPa normal pressure and a 0.02 MPa tangential shear stress are applied at the top, a periodic displacement constraint is used for the lateral boundary, an explicit integration algorithm is selected to solve the stress field, the time step is set to 0.01 nanoseconds, and the calculation is stopped after the iterative residual drops to 0.0008. The comparison and verification of the three grid sizes of 0.05, 0.1 and 0.2 microns show that the maximum stress difference is 4.2%. The final extracted stress field data shows that the maximum Mises stress of 225 MPa appears in the central area of ​​the transition zone, the principal stress gradient reaches 50 MPa per micron along the depth direction and 35 MPa per micron along the width direction. The radial basis function interpolation model is constructed using 50 uniformly distributed sampling points, and the root mean square error of cross-validation is 0.08 MPa per micron.

[0022] Step S105, combining the stress distribution gradient of the protective film in the transition zone area, screen out key variables, including the surface roughness difference between the grating area and the non-grating area, the transition zone width and the grating depth, to form a model containing variables and the resulting stress distribution data of the protective film, use a regression analysis model to fit the data and determine the weight or coefficient of each variable to obtain the main controlling factors, and check the model diagnosis through the determination coefficient R square and residual plot to evaluate the goodness of fit of the model.

[0023] The surface morphology data of the grating area and the non-grating area are obtained, and the arithmetic mean roughness value and the maximum roughness value are calculated according to the surface morphology data, and the arithmetic mean roughness value and the maximum roughness value are weighted and summed using weight coefficients to obtain surface roughness characteristic values; the minimum and maximum normalization processing is performed on the grating depth data, the transition band width data and the stress gradient data to obtain a standardized data matrix, and the Mahalanobis distance is used to perform outlier detection on the standardized data matrix, and data points outside the range of three times the standard deviation are eliminated to obtain a preprocessed data matrix; the covariance matrix is ​​calculated according to the preprocessed data matrix, and the eigenvalue decomposition of the covariance matrix is ​​performed to obtain the characteristic A vector matrix is ​​prepared, wherein the cumulative variance contribution rate is calculated through the eigenvector matrix, and the number of principal components is determined according to the set variance contribution rate threshold; the principal components are subjected to maximum variance orthogonal rotation to obtain a rotated load matrix, and variables whose absolute values ​​of load coefficients exceed the load threshold and have significant loads only on a single principal component are screened according to the rotated load matrix to obtain a characteristic variable set; the characteristic variable set is used to construct a regression equation of variables and stress distribution, quadratic terms and interaction terms are introduced as candidate variables, variables are screened according to the Akaike information criterion, standardized partial regression coefficients are calculated, significant variables are determined by t-test, and variable weights are obtained by calculating partial correlation coefficients according to the significant variables.

[0024] Exemplarily, based on the surface morphology data of the grating area and the non-grating area, the weighted sum of the arithmetic mean roughness and the maximum roughness is calculated, with the weights being 0.6 and 0.4 respectively, the minimum and maximum standardization of the grating depth, transition band width and stress gradient data is performed, the Mahalanobis distance is calculated to detect outliers, the data points exceeding 3 times the standard deviation are removed, the Pearson correlation coefficient matrix is ​​used to calculate the correlation between variables, and the variance inflation factor is calculated to identify collinear variables. For the preprocessed data matrix, principal component analysis is used to extract eigenvectors, the number of principal components is determined by the cumulative variance contribution rate, a variance contribution rate threshold of 85% is set, the extracted principal components are subjected to maximum variance orthogonal rotation, the rotation step is set to 1 degree, the load coefficient of each variable on the principal component is calculated, and the variables with an absolute value of the load coefficient greater than 0.6 and only having a significant load on a single principal component are screened. The mapping relationship between variables and stress distribution was constructed by stepwise regression based on Akaike information criterion, and quadratic terms and interaction terms were introduced as candidate variables. The standardized partial regression coefficient was calculated, and variables with a significant level less than 0.01 were screened by t-test. The partial correlation coefficient was calculated to determine the variable weight, and the Cook distance was used to identify the influencing points. The regression equation after removing the influencing points was refitted. The adjusted determination coefficient and the predicted determination coefficient were calculated according to the regression model. The normality and homogeneity of variance of the residuals were analyzed by Shapiro-Wilk test and Levene test, and the skewness coefficient and kurtosis coefficient of the residuals were calculated. The autocorrelation and heteroscedasticity of the residuals were determined by Durbin-Watson test and Bruce-Bacon test, and the prediction interval width of the 95% confidence interval was calculated. In the multivariate regression analysis of stress distribution of grating structure protective film, the experimental data obtained were first preprocessed. The arithmetic mean roughness of the grating area was 0.8 microns, the maximum roughness was 4.2 microns, the arithmetic mean roughness of the non-grating area was 0.2 microns, the maximum roughness was 1.5 microns, and the weighted roughness difference was 2.1 microns. After data standardization, the maximum Mahalanobis distance was 2.8, and 5 abnormal points were eliminated. The Pearson correlation coefficient matrix showed that the correlation coefficient between the grating depth and the transition band width was 0.72, the correlation coefficient with the stress distribution was 0.85, and the maximum variance expansion factor was 2.3. In the principal component analysis, the first principal component explained 65% of the variance, the second principal component explained 22% of the variance, and cumulatively explained 87% of the total variance. After the maximum variance rotation, the load matrix showed that the loads of the grating depth and stress distribution on the first principal component were 0.85 and 0.82 respectively, and the load of the transition band width on the second principal component was 0.78. The quadratic term of depth and the interaction term of depth and width were introduced in the stepwise regression process. Four significant variables were screened out based on the Akaike information criterion, and their standardized regression coefficients were 0.75, 0.42, 0.38 and 0.25, respectively. Three points with Cook distance greater than 0.5 were identified as influencing points.The adjusted determination coefficient of the final fitted regression model reached 0.92, the predicted determination coefficient was 0.89, the p-value of the Shapiro-Wilk test was 0.23, the p-value of the Levene test was 0.18, the skewness coefficient of the residual was 0.15, the kurtosis coefficient was 2.8, the Durbin-Watson statistic was 2.1, the p-value of the Bruce-Bacon test was 0.21, and the average width of the prediction interval was 15% of the stress value.

[0025] Step S106, according to the changing law of the main controlling factors and stress gradient, the orthogonal experimental design method is used to optimize the surface roughness, transition band width and grating depth process parameters of the grating area and the non-grating area, and obtain the process parameter combination with the best uniformity of stress distribution of the protective film, including a film thickness of 0.3 to 0.5 mm.

[0026] The parameters of the arithmetic mean roughness of the grating area, the arithmetic mean roughness of the non-grating area, the width of the transition zone, the grating depth and the film thickness are obtained, and a five-factor four-level orthogonal test table is constructed according to the parameters, wherein the adjacent level spacing in the orthogonal test table is one-fourth of the parameter variation range; for the parameter combination in the orthogonal test table, the stress distribution is calculated by using the finite element method, and the stress distribution is obtained by setting the bottom fixed constraint, the top pressure and the tangential force boundary conditions; the mean square error, the peak-to-valley ratio and the distribution entropy of the stress gradient in the radial and circumferential directions are calculated according to the stress distribution, and the weight coefficient of the mean square error, the peak-to-valley ratio and the distribution entropy is determined by the entropy weight method; for the weight coefficient, the significance level of each parameter is calculated by the F test, and the parameters with the significance level less than 0.01 are subjected to interaction effect analysis, and the main effect and the interaction effect contribution rate are obtained by variance decomposition; according to the main effect and the interaction effect contribution rate, a local encrypted grid is constructed at the significant level of the main effect parameter, and the response surface is constructed by ordinary Kriging interpolation, and the interpolation error is calculated by cross-validation on the response surface, and the confidence interval of the optimal parameter combination is obtained based on the Monte Carlo method.

[0027] For example, according to the 95% confidence interval of historical data, the range of the arithmetic mean roughness of the grating area is 0.6 to 1.0 microns, the range of the arithmetic mean roughness of the non-grating area is 0.1 to 0.3 microns, the range of the transition zone width is 2 to 6 microns, the range of the grating depth is 1 to 5 microns, and the range of the film thickness is 0.3 to 0.5 mm. A five-factor four-level orthogonal table is constructed, and the adjacent level spacing is set to 25% of the parameter variation range, and the physical feasibility index of each parameter combination is calculated. For the parameter combination in the orthogonal test table, the finite element simulation is used to calculate the stress distribution under each group of parameters, and the bottom fixed constraint, the top pressure of 0.1 MPa, the tangential force of 0.02 MPa, and the lateral periodic boundary conditions are set. The mean square error, peak-to-valley ratio and distribution entropy of the stress gradient in the radial and circumferential directions are calculated, and the entropy weight method is used to determine the weight coefficients of the three indicators to construct a comprehensive evaluation function for stress uniformity. Based on the simulation results, the 95% confidence interval of the range value is calculated by bootstrap sampling, the significance level of each parameter is calculated by F test, the parameters with significance probability less than 0.01 are analyzed for pairwise interaction effects, and the contribution rate of main effect and interaction effect is calculated by variance decomposition to generate parameter sensitivity ranking results. According to the parameter sensitivity results, a local encrypted grid is constructed at the significant level of the main effect parameter, and the grid spacing is set to 1 / 5 of the original parameter spacing. The response surface is constructed by ordinary Kriging interpolation, and the interpolation error is less than 5% through cross-validation calculation. The 95% confidence interval of the optimal parameter combination is calculated based on the Monte Carlo method, and the parameter value range that meets the process stability requirements is output. In the research on process parameter optimization of grating protective film, based on statistical analysis of 500 sets of historical data, the 95% confidence interval of the arithmetic mean roughness of the grating area was determined to be 0.68 to 0.92 microns, and that of the non-grating area was 0.15 to 0.25 microns. The transition band width was 2.5 to 5.5 microns, the grating depth was 1.2 to 4.8 microns, and the film thickness was 0.32 to 0.48 mm. A five-factor four-level orthogonal table was constructed, and the horizontal spacing of each parameter was set to 0.08 microns, 0.03 microns, 1 micron, 1.2 microns and 0.05 mm, respectively. Four groups of unreasonable parameter combinations were eliminated through physical feasibility calculations. Finite element calculations were performed for the remaining 76 parameter combinations, with a normal pressure of 0.1 MPa and a tangential force of 0.02 MPa. The calculated stress gradient mean square error ranged from 15 to 45 MPa per micron, the peak-to-valley ratio was between 1.8 and 3.2, and the distribution entropy was between 2.1 and 4.5. The weights of the three indicators calculated by the entropy weight method were 0.42, 0.35, and 0.23, respectively. The range values ​​were calculated by 1000 self-help resamplings. The range confidence interval of the grating depth was 28 to 35 MPa per micron, with a significance probability of 0.002, and the range confidence interval of the transition zone width was 20 to 26 MPa per micron, with a significance probability of 0.008. The contribution rate of the interaction effect of the two parameters reached 15%.A 50×50 local encrypted grid was constructed in the neighborhood of the main effect parameters, and the response surface was established through ordinary Kriging interpolation. The root mean square error of cross-validation was 4.2%. The optimal grating depth of 3.2±0.2 microns and the transition zone width of 4.1±0.3 microns were determined through 10,000 Monte Carlo samplings. The corresponding comprehensive evaluation index of stress gradient reached the optimal value of 0.85.

[0028] Step S107, according to the optimized process parameter combination, the front and back sides of the code disc are film-laminated, and the protective film lamination stress distribution test is performed on the code disc after the film lamination to verify the effectiveness of the optimized film lamination process and ensure the reliability and stability of the code disc during use.

[0029] A laser confocal microscope is used to scan the surface of the protective film, and coordinate alignment is performed according to the reference mark points to obtain the film thickness distribution surface data; for the film thickness distribution surface data, high-frequency noise is removed by low-pass filtering, and baseline drift is removed by high-pass filtering to obtain the filtered film thickness distribution surface; the root mean square deviation and peak-to-valley value are calculated according to the filtered film thickness distribution surface, and the area beyond the standard deviation range is marked to obtain the data of the abnormal film thickness area; a grating interferometer is used to measure the stress distribution of the protective film, and the displacement field is reconstructed by the mass conservation constrained iterative phase unwrapping algorithm to obtain the stress distribution cloud map; multi-band wavelet decomposition is performed on the stress distribution cloud map, the position and intensity of the stress mutation area are calculated, the statistical characteristic data of the stress field are obtained, and the corresponding relationship between the stress mutation and the film thickness distribution is established.

[0030] Exemplarily, a laser confocal microscope is used to set 4 reference mark points on both sides for coordinate alignment, with a scanning range of 100 mm × 100 mm, a sampling interval of 0.1 mm, a detection depth range of 0.2 to 0.6 mm, and a vertical resolution of 0.01 mm. The film thickness distribution surface is obtained by morphological reconstruction, and the measurement data is temperature compensated, with the compensation coefficient set to 0.01 mm per degree Celsius. For the film thickness distribution surface obtained by scanning, a low-pass filter with a cutoff frequency of 0.5 per mm is used to remove high-frequency noise, and a high-pass filter with a cutoff frequency of 0.02 per mm is used to remove baseline drift, and the root mean square deviation and peak-to-valley value of the surface are calculated. The abnormal area beyond the range of 3 times the standard deviation is marked, and the position coordinates and area proportion of the abnormal film thickness area are recorded to establish a spatial mapping relationship between the film thickness anomaly and the stress distribution. The stress distribution state of the protective film after lamination was measured by grating interferometer. The spatial resolution of the interferometer was set to 0.1 mm, the phase resolution was π / 100, the ambient temperature was controlled at 23±0.5 degrees Celsius, and the humidity was controlled at 45±5%. The iterative phase unwrapping algorithm with mass conservation constraint was used to reconstruct the displacement field distribution, calculate the stress distribution cloud map, and obtain the main direction and anisotropy index of the stress field. According to the stress distribution cloud map, the stress gradient data in the transition zone area was extracted, and wavelet decomposition was performed on 4 frequency bands, with the frequency band ranges of 0.1-0.2, 0.2-0.4, 0.4-0.8, and 0.8-1.6 per millimeter, respectively. The position and intensity of the stress mutation area were calculated, and the statistical characteristics of the stress field were analyzed, including the mean square error, skewness, kurtosis, and autocorrelation length, and the corresponding relationship between the stress mutation and the film thickness distribution was established. In the grating code disk protective film lamination performance evaluation experiment, four metal reference mark points with an interval of 25 mm were set on the front and back sides by laser confocal microscopy, and the least squares registration algorithm was used to achieve the correspondence of the front and back coordinates. The registration accuracy was better than 0.02 mm. The scan was performed at an interval of 0.1 mm within the range of 100 mm × 100 mm. The original data obtained after temperature compensation showed that the average film thickness was 0.42 mm and the standard deviation was 0.03 mm. After data processing using 0.5 per mm and 0.02 per mm bandpass filters, 5 abnormal film thickness areas were identified. The largest abnormal area was 0.8 square millimeters, located at the edge of the grating transition zone, and the stress gradient at the corresponding position reached 45 MPa per millimeter. The grating interferometry measurement was carried out under the conditions of constant temperature of 23.2 degrees Celsius and relative humidity of 43%. The stress distribution was obtained by phase unwrapping with mass conservation constraints. The calculation showed that the angle between the main direction of the stress field and the grating direction was 12 degrees, and the anisotropy index was 1.4. Wavelet multi-scale decomposition was performed in the transition zone. The stress mutation intensity in the 0.1-0.2 per millimeter frequency band was 28 MPa, 35 MPa in the 0.2-0.4 band, 42 MPa in the 0.4-0.8 band, and 31 MPa in the 0.8-1.6 band. The mean square error of the stress field was 15 MPa, the skewness coefficient was 0.8, the kurtosis coefficient was 3.2, the autocorrelation length was 2.5 mm, and the correlation coefficient with the film thickness distribution reached 0.82.

[0031] Based on the above embodiments of the present invention, relevant personnel can make various changes and modifications without departing from the technical concept of the present invention through the above description. The technical scope of the present invention is not limited to the content in the specification, and its technical scope must be determined according to the scope of the claims.

Claims

1. A method for analyzing code disc measurement data, characterized in that: The method comprises: The surface roughness measurement data of the grating area and the non-grating area of ​​the code disk are obtained. For the surface roughness mutation area, the measurement data is segmented by digital image processing to obtain the roughness distribution map of the grating area and the non-grating area; according to the roughness distribution map of the grating area and the non-grating area, a three-dimensional geometric model of the roughness mutation area is constructed, and the stress distribution state of the protective film in the roughness mutation area is calculated by the finite element analysis method to obtain the stress distribution cloud map of the protective film; for the transition zone between gratings of different depths, a scanning electron microscope is used to obtain the microscopic morphological characteristics of the transition zone, and the width parameter of the transition zone is extracted by the image processing algorithm, and a correlation model between the transition zone width and the grating depth is established; according to the correlation model between the transition zone width and the grating depth, combined with the protective film material characteristic parameters and the stress distribution cloud map, the stress distribution gradient of the protective film in the transition zone area is analyzed by the numerical simulation method, and the change law of the stress gradient with the grating depth and the transition zone width is obtained; combined with the protective film in the transition zone The stress distribution gradient in the belt area screens out key variables, including the surface roughness difference between the grating area and the non-grating area, the transition zone width and the grating depth, to form data containing variables and the resulting stress distribution of the protective film. The regression analysis model is used to fit the data and determine the weight or coefficient of each variable to obtain the main controlling factors. The model diagnosis is checked by the determination coefficient R square and the residual diagram to evaluate the goodness of fit of the model; according to the changing law of the main controlling factors and the stress gradient, the orthogonal experimental design method is used to optimize the surface roughness of the grating area and the non-grating area, the transition zone width and the grating depth process parameters to obtain the optimal process parameter combination for the uniformity of stress distribution of the protective film, including a film thickness of 0.3 to 0.5 mm; according to the optimized process parameter combination, the front and back sides of the code disc are film-laminated, and the protective film lamination stress distribution test is carried out on the code disc after the film-laminated process to verify the effectiveness of the optimized film lamination process and ensure the reliability and stability of the code disc during use.

2. The method according to claim 1, characterized in that: The method of obtaining the surface roughness measurement data of the grating area and the non-grating area of ​​the code disk, segmenting the measurement data by digital image processing for the surface roughness mutation area, and obtaining the roughness distribution map of the grating area and the non-grating area includes: A contact roughness meter is used to obtain the horizontal displacement coordinate value and vertical height value on the surface of the encoder disk to obtain a two-dimensional data sequence of the surface morphology; Gaussian filtering and baseline correction are performed on the surface morphology two-dimensional data sequence, wavelet transform is used to perform multi-scale decomposition on the processed data, and the mutation point position is obtained by local area variance threshold judgment; The original data sequence is segmented according to the mutation point position, and the position coordinates, local mean and standard deviation of each data point in the data sequence are calculated by mean shift clustering to obtain the regional division result of the grating area and the non-grating area; According to the area division results, the arithmetic mean roughness value, the maximum roughness value and the ten-point average roughness value of the grating area and the non-grating area are calculated respectively to obtain the roughness parameters of each area; The roughness parameters of each region are processed by opening and closing operations, missing data points are supplemented by cubic spline interpolation, and the boundary between the grating area and the non-grating area is optimized by regional growing algorithm to obtain a regional roughness distribution map.

3. The method according to claim 1, characterized in that The method of constructing a three-dimensional geometric model of a roughness mutation region according to the roughness distribution diagrams of the grating region and the non-grating region comprises: The local curvature value is calculated according to the two-dimensional coordinates and height values ​​of the measuring points in the roughness distribution map. For the areas where the curvature value is greater than the preset threshold, cubic Hermite interpolation is used to construct the first surface function, and for the remaining areas, cubic spline interpolation is used to construct the second surface function. The calculation domain geometric model is obtained through the first surface function and the second surface function.

4. The method according to claim 3, characterized in that The method of calculating the stress distribution state of the protective film in the roughness mutation area by the finite element analysis method to obtain the stress distribution cloud map of the protective film includes: The computational domain geometric model is meshed using twenty-node hexahedral units. If a mutation region is detected, a preset mesh size is set at the mutation region, and a mesh topology relationship is obtained by transitioning outward according to the preset size ratio. According to the grid topological relationship, a fixed constraint is applied to the bottom of the calculation domain, a pressure load is applied to the top of the calculation domain, a displacement symmetric boundary condition is applied to the side of the calculation domain, and a Coulomb friction contact is used at the interface between the protective film and the substrate to obtain the boundary condition of the calculation domain; Newton iteration method is used to solve the displacement field under the boundary conditions of the calculation domain, and Gauss point stress value is calculated through the displacement field; The node stress value is restored according to the Gaussian point stress value, the Mises stress distribution cloud map, the maximum principal stress distribution cloud map and the maximum shear stress distribution cloud map are obtained, and the stress gradient value is calculated through the stress distribution cloud map.

5. The method according to claim 1, characterized in that The method for obtaining the microscopic morphological features of the transition zone between gratings of different depths by using a scanning electron microscope, extracting the width parameter of the transition zone by an image processing algorithm, and establishing a correlation model between the width of the transition zone and the grating depth includes: Scanning the grating transition zone region with an electron microscope, obtaining an original scanned image based on the scanning, and performing median filtering and grayscale histogram matching processing on the original scanned image to obtain an enhanced image; Performing wavelet transform decomposition on the enhanced image, selecting a feature layer with the largest local variance according to the decomposition result, and obtaining an edge detection result by calculating the local variance on the feature layer and setting an adaptive threshold; Performing a morphological operation according to the edge detection result, obtaining a transition zone contour binary image through the morphological operation, and performing a closing operation to connect the transition zone contour binary image if there is edge breakage; Applying a skeleton extraction algorithm to the transition zone contour binary image, calculating the edge point curvature according to the skeleton extraction result, and obtaining the transition zone width data by screening the edge point pairs through the curvature; A support vector regression model is constructed for the transition band width data, a mapping relationship between the transition band width and the grating depth is obtained according to the model parameter optimization result, and a corresponding model with a determination coefficient greater than a preset threshold is established through the mapping relationship.

6. The method according to claim 1, characterized in that According to the correlation model between the transition zone width and the grating depth, the protective film material characteristic parameters and the stress distribution cloud map are combined, and a numerical simulation method is used to analyze the stress distribution gradient of the protective film in the transition zone area, and the change law of the stress gradient with the grating depth and the transition zone width is obtained, including: Constructing a cubic spline interpolation function according to the grating depth data and the transition zone width data, wherein the cubic spline interpolation function is used to generate grid point coordinates of a continuous curved surface profile of the transition zone; The grid point coordinates are used to divide a tetrahedral unit grid, wherein the tetrahedral unit grid has a preset center area unit size value and an edge area unit size value; Applying fixed constraint conditions and load conditions to the tetrahedral unit grid, and obtaining displacement field data and stress field data of the transition zone region through an explicit integral solver; Extracting principal stress data and equivalent stress data of the transition zone region according to the stress field data, wherein the principal stress data and equivalent stress data are distributed along the depth direction and the width direction of the transition zone; The stress gradient value is calculated based on the principal stress data and the equivalent stress data, and the stress gradient value is interpolated using a radial basis function. If the residual root mean square error of the interpolation operation is less than a preset threshold, the stress gradient variation data with depth and width are obtained.

7. The method according to claim 1, characterized in that The stress distribution gradient of the protective film in the transition zone region is combined to screen out key variables, including the surface roughness difference between the grating area and the non-grating area, the transition zone width and the grating depth, to form data containing variables and the resulting stress distribution of the protective film, including: Acquire surface topography data of the grating area and the non-grating area, calculate an arithmetic mean roughness value and a maximum roughness value according to the surface topography data, and perform weighted summation of the arithmetic mean roughness value and the maximum roughness value using weight coefficients to obtain a surface roughness characteristic value; Performing minimum and maximum normalization processing on the grating depth data, transition zone width data and stress gradient data to obtain a standardized data matrix, using Mahalanobis distance to perform outlier detection on the standardized data matrix, and eliminating data points outside the range of three times the standard deviation to obtain a preprocessed data matrix; Calculating a covariance matrix according to the preprocessed data matrix, performing eigenvalue decomposition on the covariance matrix to obtain an eigenvector matrix, calculating a cumulative variance contribution rate through the eigenvector matrix, and determining the number of principal components according to a set variance contribution rate threshold; The principal components are subjected to maximum variance orthogonal rotation to obtain a rotated load matrix, and variables having load coefficients with absolute values ​​exceeding a load threshold and having significant loads only on a single principal component are screened according to the rotated load matrix to obtain a characteristic variable set.

8. The method according to claim 7, characterized in that The regression analysis model is used to fit the data and determine the weight or coefficient of each variable to obtain the main controlling factors, and the model diagnosis is checked by the determination coefficient R square and residual diagram to evaluate the goodness of fit of the model, including: The characteristic variable set is used to construct a regression equation of variables and stress distribution, quadratic terms and interaction terms are introduced as candidate variables, variables are screened using the Akaike information criterion, standardized partial regression coefficients are calculated, significant variables are determined using the t-test, and variable weights are obtained by calculating partial correlation coefficients based on the significant variables.

9. The method according to claim 1, characterized in that: According to the changing law of the main control factors and stress gradient, the orthogonal experimental design method is used to optimize the surface roughness of the grating area and the non-grating area, the transition zone width and the grating depth process parameters, and obtain the process parameter combination with the best uniformity of stress distribution of the protective film, including a film thickness of 0.3 to 0.5 mm, including: Obtain the arithmetic mean roughness of the grating area, the arithmetic mean roughness of the non-grating area, the transition band width, the grating depth and the film thickness parameters, and construct a five-factor four-level orthogonal test table based on the parameters, wherein the adjacent level spacing in the orthogonal test table is one-fourth of the parameter variation range; For the parameter combination in the orthogonal test table, the stress distribution is calculated by using the finite element method, and the stress distribution is obtained by setting the bottom fixed constraint and the top pressure and tangential force boundary conditions; Calculating the mean square error, peak-to-valley ratio and distribution entropy of the stress gradient in radial and circumferential directions according to the stress distribution, and determining the weight coefficients of the mean square error, peak-to-valley ratio and distribution entropy by using an entropy weight method; For the weight coefficient, the significance level of each parameter is calculated by F test, and the parameters with significance level less than 0.01 are subjected to interaction effect analysis, and variance decomposition is used to obtain the main effect and interaction effect contribution rate; According to the contribution rates of the main effects and interaction effects, a local encrypted grid was constructed at the significant level of the main effect parameters, and the response surface was constructed using ordinary Kriging interpolation. The response surface calculated the interpolation error through cross-validation, and the confidence interval of the optimal parameter combination was obtained based on the Monte Carlo method.

10. The method according to claim 1, characterized in that According to the optimized process parameter combination, the front and back sides of the code disc are film-laminated, and the protective film lamination stress distribution test is performed on the code disc after the film lamination to verify the effectiveness of the optimized film lamination process and ensure the reliability and stability of the code disc during use, including: The surface of the protective film is scanned using a laser confocal microscope, and coordinates are aligned according to the reference marking points to obtain the film thickness distribution surface data; For the film thickness distribution surface data, high-frequency noise is removed by low-pass filtering, and baseline drift is removed by high-pass filtering to obtain a filtered film thickness distribution surface; Calculating the root mean square deviation and the peak-to-valley value according to the filtered film thickness distribution surface, marking the area beyond the standard deviation range, and obtaining data of the abnormal film thickness area; The stress distribution of the protective film is measured by using a grating interferometer, and the displacement field is reconstructed by using a mass conservation constrained iterative phase unwrapping algorithm to obtain a stress distribution cloud map. The stress distribution cloud map is subjected to multi-band wavelet decomposition, the position and intensity of the stress mutation area are calculated, the statistical characteristic data of the stress field are obtained, and the corresponding relationship between the stress mutation and the film thickness distribution is established.

Images