A Prediction Method for Expansion Rate in High-Purity Metals Based on Image Processing

Abstract

The present invention relates to the field of measurement technology, and specifically relates to a method for predicting the expansion rate in high-purity metals based on image processing. The method includes: obtaining the grayscale image of the grating image on the metal surface at different temperatures, and obtaining the spectrogram of the grayscale image corresponding to any temperature; determining the optimal filtering radius of the filter of the spectrogram; filtering the spectrogram according to the optimal filtering radius to obtain the optimal spectrogram; obtaining an equidistant fringe image based on the optimal spectrogram; obtaining the linear expansion coefficient of the metal at any temperature based on the equidistant fringe image, and further obtaining the linear expansion coefficients at different temperatures, so as to realize the prediction of the linear expansion coefficient of the metal at the temperature to be measured. The solution of the present invention can improve the accuracy of predicting the linear expansion coefficient of the metal.

Description

Technical Field

[0001] The present invention relates to the field of measurement technology. More specifically, the present invention relates to a method for predicting the expansion rate in high-purity metals based on image processing. Background Art

[0002] In modern industrial production, high-purity metals are widely used in various high-tech fields such as aerospace and semiconductor manufacturing due to their excellent physical and chemical properties. In the manufacturing process of high-purity metals, accurate prediction of their expansion rate is the key to ensuring product quality and performance stability.

[0003] The prior art discloses a method for measuring the linear expansion coefficient of metal materials based on image analysis. This method is based on the Moiré fringe theory, derives the functional relationship among the linear expansion coefficient of the metal material, the slope of the equidistant fringes of the Moiré fringe, and the temperature change amount of the metal material due to thermal expansion, and with the help of image analysis technology, through three-layer operation processing of image optimization, image heterodyne information processing, and image data calculation on the collected grating change images, and using the slope of the equidistant fringes of the Moiré fringe and the temperature change amount of the metal material due to thermal expansion as basic parameters, finally obtains the strain and the magnitude of the linear expansion coefficient of the metal copper material within a certain temperature change range.

[0004] For example, in the patent application document with the application publication number CN102879418A and the invention name "A Method for Measuring the Linear Expansion Coefficient of Metal Materials", it discloses performing Fourier transform, filtering processing, and thinning processing by the extreme value method on the obtained Moiré fringes to obtain the thinned equidistant fringes, and then calculating the slope and inclination angle of the equidistant fringes by the least squares method to calculate the expansion coefficient.

[0005] The above method has advantages such as non-contact, high precision, and real-time performance. However, it should be noted that when selecting low-frequency information in the above measurement method, the selection of low-frequency information is often achieved through a low-pass filter. When the low-pass filter is set improperly, there may be problems such as more interference information in the restored low-frequency information or excessive loss of low-frequency information, resulting in a lower accuracy rate of the final detection result of the expansion coefficient, and further making the subsequent prediction of the expansion coefficient inaccurate. Summary of the Invention

[0006] The object of the present invention is to propose a method for predicting the expansion rate in high-purity metals based on image processing to solve the problem that the detection of the expansion coefficient of metals in the prior art has a low accuracy rate, resulting in inaccurate subsequent predicted expansion coefficients; for this purpose, the present invention provides a solution in the following aspect.

[0007] The present invention provides a method for predicting the expansion rate in high-purity metals based on image processing, including the following steps:

[0008] Obtain the grayscale image of the grating image on the metal surface at different temperatures, and obtain the spectrogram of the grayscale image corresponding to any temperature;

[0009] Determine the optimal filtering radius of the filter for the spectrogram; filter the spectrogram according to the optimal filtering radius to obtain the optimal spectrogram; based on the optimal spectrogram, obtain the isochromatic fringe image;

[0010] Based on the isochromatic fringe image, obtain the linear expansion coefficient of the metal at any temperature, and based on the obtained linear expansion coefficients at different temperatures, realize the prediction of the linear expansion coefficient of the metal at the temperature to be measured;

[0011] Among them, the optimal filtering radius is:

[0012] Obtain the peak point sequence in the spectrogram; based on the position coordinates of each peak point in the peak point sequence, construct a triangulation model; the triangulation model includes a plurality of triangulations;

[0013] Calculate the priority value of each filtering radius, and the priority value is positively correlated with the fitness value and importance of each triangulation; the fitness value represents the possibility of belonging to the isochromatic fringe within the corresponding triangulation, and the importance is the coincidence situation of any triangulation with the circular region; the center of the circular region is the center point of the spectrogram, and the radius is any filtering radius;

[0014] Based on each filtering radius and the corresponding priority value, generate a corresponding fitting function, and take the filtering radius corresponding to the maximum value of the derivative of the fitting function as the optimal filtering radius.

[0015] In the above solution, by means of the triangulation model, the relationship between the region enclosed by each peak point in the spectrogram corresponding to any temperature and the isochromatic fringe is analyzed, and different filtering radii can be adaptively analyzed, so as to determine the optimal filtering radius of the spectrogram. Filter the spectrogram with the optimal filtering radius to obtain the filtered optimal spectrogram, and then obtain the isochromatic fringe image, and finally realize the calculation of the expansion coefficient; that is, the solution of the present invention has a corresponding optimal filtering radius for the spectrograms at different temperatures. When filtering different spectrograms, it has a stable filtering effect, preventing the problem that the corresponding isochromatic fringes obtained have too much interference information or too little information missing, making the detection of the expansion coefficient of pure metal stable when using isochromatic fringes, improving the accuracy of the detection of the expansion coefficient of pure metal, and being able to more accurately predict the expansion coefficient of the metal at the temperature to be measured.

[0016] In one embodiment, the process of obtaining the grating image is:

[0017] Use the metal to be measured as the screen, and conduct a Moiré fringe experiment using two gratings with the same grating parameters; during the Moiré fringe experiment, the included angle between the two gratings is the set angle, and a Moiré fringe image is formed on the surface of the metal to be measured, and the Moiré fringe image is used as the grating image.

[0018] In one embodiment, the moderation value is:

[0019] ;

[0020] wherein, is the ratio of the relative included angle of the j-th triangular mesh to 90°, where the relative included angle is the included angle between the straight line formed by the center of gravity of the j-th triangular mesh and the center point of the spectrogram and the straight line formed along the angle β direction, is the ratio of the Euclidean distance value from the center of gravity of the j-th triangular mesh to the center point of the spectrogram to the diagonal length of the grayscale image, is the ratio of the average spectral intensity of all internal points within the j-th triangular mesh to the spectral intensity of the center point of the spectrogram, is the exponential function with the natural constant e as the base, , and θ is the set angle.

[0021] The above solution can determine the possibility that the triangular mesh belongs to equidistant fringes by obtaining the moderation value of each triangular mesh.

[0022] In one embodiment, the process of obtaining the importance is as follows:

[0023] Randomly select any filtering radius, and with the center point of the spectrogram as the center, obtain the circular region corresponding to any filtering radius;

[0024] Count the number of overlapping data between the circular region and the j-th triangular mesh, and take the ratio of the number of overlapping data to the number of all internal points within the j-th triangular mesh as the importance of any filtering radius.

[0025] The above solution can measure the overlapping situation between the circular region under any filtering radius and each triangular mesh by obtaining the importance.

[0026] In one embodiment, the priority value is:

[0027] ; where J is the total number of triangular meshes, is the importance of the s-th filtering radius, is the moderation value of the j-th triangular mesh.

[0028] The above solution can adaptively analyze different filtering radii by obtaining the priority value.

[0029] In one embodiment, before constructing the triangular mesh model, it further includes the step of screening the peak point sequence to obtain a new peak point sequence, specifically:

[0030] Obtain the spectral intensity of each peak point in the peak point sequence;

[0031] Based on the position coordinates and corresponding spectral intensities of each peak point, divide the peak point sequence to obtain an interference sequence, a weak feature sequence, and a strong feature sequence; among them, the interference sequence is the category with the smallest average spectral intensity, the weak feature sequence is the category with the average spectral intensity in the middle, and the strong feature sequence is the category with the largest average spectral intensity;

[0032] Remove other remaining peak points in the interference sequence and the strong feature sequence except the central peak point to obtain the weak feature sequence and the central peak point;

[0033] Combine the weak feature sequence and the central peak point to form a new peak point sequence.

[0034] The above solution reduces the interference problem of other peak points by removing all other peak points in the strong feature sequence, improving the accuracy.

[0035] In one embodiment, the specific process of predicting the linear expansion coefficient of the metal at the temperature to be measured is as follows: perform data fitting on the temperature and linear expansion coefficient of the metal at different temperatures to obtain a fitting curve, and substitute the temperature to be measured into the fitting curve to obtain the predicted linear expansion coefficient.

[0036] In one embodiment, after obtaining the grayscale image of each grating image, the grayscale image is further denoised to obtain a denoised grayscale image.

[0037] The beneficial effects of the present invention are:

[0038] The solution of the present invention has a corresponding optimal filtering radius for spectrograms at different temperatures. When filtering different spectrograms, it has a stable filtering effect, making it stable when obtaining the expansion coefficient of pure metal using equal-interval fringes, that is, improving the accuracy of detecting the expansion coefficient of pure metal, and thus being able to more accurately predict the expansion coefficient of the metal at the temperature to be measured. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] By reading the following detailed description with reference to the accompanying drawings, the above and other objects, features, and advantages of the exemplary embodiments of the present invention will become readily understood. In the drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein:

[0040] Figure 1 The flowchart of steps of a method for predicting the expansion rate in high-purity metal based on image processing in this embodiment;

[0041] Figure 2 Schematic diagram of local Moiré fringes;

[0042] Figure 3 Schematic diagram of Moiré fringes with an arrangement direction;

[0043] Figure 4 is Figure 3 the spectrogram of. Detailed implementation manners

[0044] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative efforts belong to the scope of protection of the present invention.

[0045] Next, the specific implementation manners of the present invention will be described in detail with reference to the accompanying drawings.

[0046] The scenario targeted by the present invention is as follows: In the process of measuring the linear expansion coefficient of a metal material based on the characteristics of a grating and Moiré fringes, when filtering the spectrogram of the grating image during the detection of the linear expansion coefficient of the metal at a certain temperature, the selection of the low-pass filter is not considered. Therefore, when filtering the spectrogram, there may be errors, resulting in inaccurate linear expansion coefficients calculated subsequently, thereby affecting the prediction of the linear expansion coefficient at the temperature to be measured.

[0047] Based on the above problems, the present invention provides a method for predicting the expansion rate in high-purity metal based on image processing. This method mainly focuses on the selection of the low-pass filter and does not improve other processes in the grating image (such as Fourier transform of the grating image, inverse Fourier transform of the filtered spectrogram, thinning process of the isochromatic fringe image, and least squares method process).

[0048] Among them, the expansion rate in the present invention is actually the linear expansion coefficient of the metal, which is used to represent the change in the length value caused by temperature changes. Among them, the linear expansion coefficient is a type of thermal expansion coefficient; the thermal expansion coefficient also includes the area expansion coefficient and the volume expansion coefficient. Different metals have different expansion rates, that is, the expansion amount per unit length at different temperatures.

[0049] Specifically, as Figure 1 shown, a method for predicting the expansion rate in high-purity metal based on image processing in this embodiment includes the following steps:

[0050] At step S1, a grayscale image of the grating image on the metal surface at different temperatures is obtained, and a spectrogram of the grayscale image corresponding to any temperature is obtained.

[0051] In this embodiment, by collecting grating images on the surface of the metal to be measured at different temperatures, each grating image is grayscale processed to obtain grayscale images corresponding to each temperature; the different temperatures are at least two temperatures set according to a set temperature interval.

[0052] Since noise is often high-brightness points, in this embodiment, each grayscale image is also denoised, and specifically, minimum filtering can be used for denoising to obtain a denoised grayscale image.

[0053] In this embodiment, the specific process of obtaining the grating image is as follows: taking the metal to be measured as a screen, and using two gratings with the same grating parameters to perform a Moiré fringe experiment; during the Moiré fringe experiment, the included angle between the two gratings with the same grating parameters is a set angle, which is θ, and a Moiré fringe image is formed on the surface of the metal to be measured, and the Moiré fringe image is used as the grating image.

[0054] In the above, when performing the Moiré fringe experiment, an existing grating device is used. The included angle between the grating lines of the two gratings with the same grating parameters in the grating device is θ (the included angle θ is generally very small). Taking the metal to be measured as a screen, Moiré fringes are formed on the metal to be measured. At the same time, a high-definition camera is installed near the grating device to collect the Moiré fringes on the surface of the metal to be measured to obtain the grating image. Among them, since the Moiré fringe is a black-and-white striped pattern formed by the superposition of two cosine gratings, the grating image is a black-and-white striped pattern.

[0055] Exemplarily, when the included angle between the grating lines of the two gratings with the same grating parameters is θ, light and dark striped patterns appear, as Figure 2 shown, where I is the dark fringe, the position of the I line is opaque, II is the bright fringe, and the light transmittance at the position of the II line is the largest. This kind of light and dark striped pattern is the Moiré fringe. At this time, the light energy distribution after the light passes through the two gratings is an ideal triangular wave; however, in actual applications, there is a diffraction phenomenon of light, and the illumination light source has a width, and the grating pitches, slit widths, and grating line widths of the two gratings are not exactly equal. Therefore, the actual light energy distribution is an approximate sine wave.

[0056] Among them, in this embodiment, the direction where the I line or the II line is located is the arrangement direction of the Moiré fringe, and the arrangement direction of the Moiré fringe is perpendicular to the bisector of the included angle between the two grating lines; Figure 2 The included angle between the arrangement direction of the Moiré fringe in and the horizontal direction is , and the angle of the included angle is half of the set angle θ. It should be noted that since the included angle θ is generally very small, therefore, the included angle Generally, it is also very small.

[0057] In this embodiment, the two gratings can be attached to the metal to be measured, or there can be a gap between them and the metal to be measured, and the gap can be set according to the actual situation.

[0058] Among them, the different temperatures above can be multiple set temperatures, and the multiple set temperatures are based on the set initial temperature and set at a set temperature interval; for example, when the initial temperature is 70 °C, the set temperature interval is 10 °C, and the number of multiple set temperatures is 7, the multiple set temperatures are specifically 70 °C, 80 °C, 90 °C, 100 °C, 110 °C, 120 °C, 130 °C. Of course, the initial temperature and the set temperature interval above can also be set according to the actual situation, and they are not limited to the specific values set above. For example, the initial temperature can also be 25 °C, and the set temperature interval can also be 5 °C or 15 °C; and the setting of the number is not limited to the specific number above, and it can also be 8, 9 or more.

[0059] In one embodiment, a Fourier transform is also performed on the grayscale image corresponding to any temperature to obtain a corresponding spectrogram. It should be noted that the spectrogram in this embodiment is a centered spectrogram; and the change of the grayscale image has a perpendicular relationship with the change of the spectrogram, that is, the direction in the spectrogram and the spatial domain direction are perpendicular to each other. Therefore, the arrangement direction of the equidistant fringe image in the grayscale image is perpendicular to the direction in the spectrogram. At this time, with the center point as the origin on the spectral image, the straight line parallel to the straight line formed in the direction of angle β (β is the difference between 90 degrees and the included angle between the arrangement direction of the Moiré fringes and the horizontal direction) has a greater possibility of belonging to the equidistant fringes.

[0060] At step S2, determine the optimal filtering radius of the filter of the spectrogram; filter the spectrogram according to the optimal filtering radius to obtain the optimal spectrogram; based on the optimal spectrogram, obtain the equidistant fringe image.

[0061] At the same time, since the equidistant fringes are the main features in the Moiré fringe image, the equidistant fringes belong to low-frequency information. And in the spectrogram, the data closer to the center belongs more to low-frequency information; and the gray level of the equidistant fringes fluctuates greatly in the Moiré fringe image, so the corresponding spectral intensity in the spectrogram is greater. Therefore, it is necessary to filter the spectrogram to remove high-frequency information and only retain low-frequency information to obtain the equidistant fringe image.

[0062] Specifically, in this embodiment, a circular low-pass filter is selected for filtering. The center of the spectrogram is used as the center of the circular low-pass filter, and the spectrogram is filtered by setting the radius of the circular low-pass filter to eliminate high-frequency information and only retain low-frequency information. However, when setting the radius of the low-pass filter, if the filtering radius is set improperly, there will be a problem of poor filtering effect, which in turn results in a low accuracy rate for detecting the expansion rate of the metal to be measured.

[0063] Therefore, it is necessary to screen out the optimal filtering radius from multiple different filtering radii set. Specifically, the process of obtaining the optimal filtering radius is as follows:

[0064] Step S21, obtain the peak point sequence in the spectrogram; based on the position coordinates of each peak point in the peak point sequence, construct a triangulation model; the triangulation model includes multiple triangulations; calculate the fitness value of any triangulation; for any filtering radius, calculate the importance of any triangulation and this any filtering radius; take the product of the fitness value and the importance as the priority of this any filtering radius, and take the sum of the priorities of all triangulations as the priority value of this any filtering radius.

[0065] Specifically, obtain the position coordinates of all peak points in the peak point sequence and construct a triangulation model. The triangulation model includes multiple triangulations, and each triangulation includes vertices, edges, and internal points. Since the triangulation model is a prior art, it will not be elaborated here too much.

[0066] The triangulation model in this embodiment is a Delaunay triangulation model.

[0067] The above internal points are actually the data constituting the cross-section of the triangulation; since the cross-section of the triangulation is a region, the internal points are the data other than the vertices and edges in the triangulation. All internal points constitute a cross-section data sequence. For example, the cross-section data sequence of the jth triangulation is .

[0068] It should be noted that the reason for not considering the vertex and edge information of the triangulation is that the vertices and edges may be the shared information of multiple different triangulations, and thus it is difficult to determine the attribution of the vertices and edges. Therefore, in order to avoid the problem of difficult attribution determination, only the cross-section data sequence composed of the internal points of the triangulation is considered.

[0069] Furthermore, before constructing the triangulation model, it also includes the step of screening the peak point sequence to obtain a new peak point sequence. Specifically:

[0070] First, obtain the spectral intensities of the peak points in the peak point sequence. Specifically, in this embodiment, a peak point detection algorithm is used to detect peak points in the spectrogram, obtaining a peak point sequence, where the information of each peak point in the peak point sequence includes position coordinates and spectral intensity. Since the peak point detection algorithm is a prior art, it will not be elaborated here.

[0071] Secondly, based on the position coordinates and corresponding spectral intensities of the peak points, divide the peak point sequence to obtain an interference sequence, a weak feature sequence, and a strong feature sequence; among them, the interference sequence is the category with the smallest average spectral intensity, the weak feature sequence is the category with the average spectral intensity in the middle, and the strong feature sequence is the category with the largest average spectral intensity.

[0072] In this embodiment, to classify the peak point sequence, the k-means clustering algorithm can be specifically used, where k is set to 3; during the classification process, the clustering distance is obtained by calculating the Euclidean distance of the information of any two peak points; of course, a hyperparameter f can also be set to adjust the calculated Euclidean distance, specifically f = 0.3.

[0073] It should be noted that since there may be some obvious features, some weakly obvious features, and interference noise data in the spectral intensity information of each peak point in the peak point sequence, it is necessary to divide them into three categories.

[0074] Furthermore, since the information of each peak point in the peak point sequence is three-dimensional data (abscissa, ordinate, and spectral intensity) with inconsistent dimensions, before classification, it is also necessary to perform normalization processing on the peak point sequence.

[0075] Then, remove other remaining peak points except the central peak point in the interference sequence and the strong feature sequence, obtaining the weak feature sequence and the central peak point. Since the purpose of this embodiment is to obtain low-frequency information, using the above classification idea, the interference sequence can be removed, and at the same time, other remaining data except the central peak point in the strong feature sequence can be removed, finally obtaining the weak feature sequence and the central peak point.

[0076] The reason for removing the remaining data other than the central peak point in the strong feature sequence in the above embodiments is as follows: The central peak point in the strong feature sequence is the highest spectral intensity, that is, the center point of the spectrogram; the texture of the equidistant fringe image is an obvious stripe, and the arrangement direction of the stripe is perpendicular to the direction of the corresponding spectral information in the spectrogram. Also, since the equidistant fringe image is obvious stripe information and has a high spectral intensity, the spectral information of the obvious stripe information passes through the center point of the spectrogram. Therefore, the center point of the spectrogram needs to be retained; while all other peak points except the central peak point in the strong feature sequence may have interference problems, resulting in inaccuracies. Therefore, all other peak points in the strong feature sequence are removed.

[0077] Exemplarily, as Figure 3 and Figure 4 shown, Figure 4 is Figure 3 's spectrogram. Assuming Figure 3 the arrangement direction of the equidistant fringe image in Figure 3 is along the direction of the black line in Figure 3 (the angle between the black line in ) and the horizontal direction is ), then in the spectrogram, the direction of the spectral information corresponding to the equidistant fringe image is along the direction of the black line in Figure 4 ; at this time, Figure 4 the angle between the black line in Figure 4 and the horizontal direction is β. It should be noted that Figure 3 and Figure 4 the direction of the black line in Figure 4 can have two types, which can be specifically determined according to the actual situation.

[0078] Finally, the weak feature sequence and the central peak point are combined to form a new peak point sequence. In this embodiment, the position coordinates of each peak point in the new peak point sequence can be used to construct a triangular mesh model, which can improve the accuracy of the constructed triangular mesh model and reduce the interference of noise.

[0079] In this embodiment, the fitness value of the j-th triangular mesh is: ; where is the ratio of the relative angle of the j-th triangular mesh to 90°, where the relative angle is the angle formed by the line connecting the centroid of the j-th triangular mesh and the center point of the spectrogram and the line formed along the angle β direction. Among them, the smaller it is, the more the frequency domain distribution in the frequency domain conforms to the characteristic distribution of the equidistant fringes, and the smaller it is, the better. is the ratio of the Euclidean distance value from the centroid of the j-th triangular mesh to the center point of the spectrogram to the diagonal length of the grayscale image. The smaller it is, the closer it is to the center of the frequency domain, and the more the internal points of the j-th triangular mesh are low-frequency information, and the equidistant fringe image is low-frequency information. is the ratio of the average spectral intensity of all internal points within the j-th triangular network to the spectral intensity of the center point of the spectrogram, where The lower it is, the higher the possibility that the internal points within the j-th triangular network belong to interference data, which is not conducive to generating equidistant fringes; the moderate value characterizes the possibility of belonging to equidistant fringes within the corresponding triangular network.

[0080] Among them, is an exponential function with the natural constant e as the base, where the angle β is the difference between 90° and the included angle That is, , θ is a set angle, and the included angle is half of the set angle.

[0081] Among them, the diagonal length of the grayscale image can be obtained according to the size of the grayscale image; for example, when the size of the grayscale image is n×n, the diagonal length is ; when the size of the grayscale image is n×m, the diagonal length is .

[0082] Furthermore, using the function to perform a negative correlation mapping on and , the larger is, the more likely the determined filtering radius can be preferentially selected.

[0083] In this embodiment, since the constructed triangular network model uses the position coordinates of the peak points, and the peak points are themselves in the spectrogram, the triangular network corresponding to the peak points is also in the spectrogram. Therefore, the Euclidean distance value from the centroid of the triangular network to the center point of the spectrogram can be directly calculated.

[0084] It should be noted that the spectral intensity of the internal points within the triangular network determines the characteristic situation of the internal points. Therefore, it is possible to use the spectral intensity to determine the possibility of interference, that is, if the spectral intensities of all internal points within the profile data sequence are generally low, then all internal points within the profile data sequence are features that are not obvious. For restoring the equidistant fringe image, all internal points within the profile data sequence may be interference factors.

[0085] In this embodiment, the process of obtaining the importance is as follows:

[0086] ​Randomly select any filtering radius, and with the center point of the spectrogram as the center, obtain the circular area corresponding to any filtering radius; count the number of overlapping data between the circular area and the j-th triangular mesh, calculate the ratio of the number of overlapping data to the number of all internal points of the j-th triangular mesh, and take this ratio as the importance between the filtering radius and the triangular mesh.

[0087] The number of overlapping data mentioned above is obtained by statistically analyzing the position coordinates of the data within the s-th filtering radius and the position coordinates of the internal points within the j-th triangular mesh. The data with repeated position coordinates is the overlapping data. The larger the number of overlapping data, the larger the overlapping part between the circular area under the s-th filtering radius and the j-th triangular mesh.

[0088] In this embodiment, the priority values corresponding to each filtering radius are specifically as follows: for any filtering radius, take the product of the moderation value of any triangular mesh and the corresponding importance as the priority of this filtering radius, and take the sum of the priorities of all triangular meshes as the priority value of this filtering radius.

[0089] Specifically, taking the s-th filtering radius as an example, the priority value corresponding to the s-th filtering radius : ; where J is the total number of triangular meshes, is the importance of the s-th filtering radius, is the moderation value of the j-th triangular mesh.

[0090] In this embodiment, since there is only partial repetition between the circular area under the s-th filtering radius and the j-th triangular mesh, so weight according to the repeated ratio, and accumulate the moderation values of all triangular meshes covered under the s-th filtering radius. Take the weighted accumulated sum as the priority value corresponding to the s-th filtering radius.

[0091] Step S22, based on each filtering radius and the corresponding priority value, generate a corresponding fitting function, and take the filtering radius corresponding to the maximum value of the derivative of the fitting function as the optimal filtering radius.

[0092] After obtaining the priority values under different filtering radii, use the least squares method to fit different filtering radii and their corresponding priority values to obtain a corresponding fitting function. Take the derivative of the fitting function to obtain the filtering radius corresponding to the maximum derivative as the optimal filtering radius. However, since only integers and no decimals are available when selecting pixel points during filtering, the optimal filtering radius needs to be rounded to obtain the final optimal filtering radius.

[0093] Other fitting methods can also be used for the fitting in the above embodiment, such as the gradient descent method.

[0094] It should be noted that the reason for the above fitting is that since both the moderation value and the importance are values that are always greater than or equal to 0, and the moderation value of each triangular mesh in the triangular mesh model is unchanged, for any triangular mesh, the product of its moderation value and importance increases as the filtering radius increases (the importance increases as the filtering radius increases). Therefore, the priority values of all triangular meshes must be a data that increases as the filtering radius increases. Therefore, in this embodiment, it is necessary to fit the priority values to obtain the position of the maximum slope of the fitting function (where the maximum slope is obtained by differentiating the fitting function). Since the position of the maximum slope represents the point where the priority value changes the most, the filtering radius corresponding to the point where the priority value changes the most at this time has a better priority value.

[0095] The purpose of using the Delaunay triangular mesh model in this embodiment is that there may still be a lot of interference information in the peak point sequence, and not all peak points in the peak point sequence are pixel points strongly associated with the equidistant fringes. If the filtering radius of the low-pass filter is too small, too much information will be lost and relatively obvious equidistant fringes cannot be obtained. Therefore, it is necessary to screen the filtering radius by virtue of the relationship between the Delaunay triangular mesh model and the filtering radius to determine whether the filtering radius is appropriate.

[0096] After obtaining the optimal filtering radius, a circular filter is created with the optimal filtering radius as the low-pass filter of the spectrogram, so that only the data covered by the filter is retained after filtering, and the filtered spectrogram data is obtained. Then, the inverse Fourier transform is performed on the filtered spectrogram to restore the spectrogram to the spatial domain to obtain the equidistant fringe image.

[0097] So far, the optimal filtering radius of the spectrogram corresponding to each temperature can be obtained in this embodiment to filter the spectrogram at each temperature. Since each spectrogram corresponds to an optimal filtering radius, each optimal spectrogram after filtering can be obtained, improving the accuracy of the equidistant fringe image of each optimal spectrogram obtained subsequently.

[0098] At step S3, the linear expansion coefficient of the metal at any temperature is obtained based on the equidistant fringe image, and then the linear expansion coefficients at different temperatures are obtained, thereby realizing the prediction of the linear expansion coefficient of the metal at the temperature to be measured.

[0099] In this embodiment, the equidistant fringe image is subjected to thinning processing and least squares processing to obtain the linear expansion coefficient of the metal to be measured at any temperature.

[0100] Specifically, first binarize the grayscale image corresponding to the equidistant fringes, and then thin the binary image using a thinning algorithm. Since there are multiple fringe lines in the equidistant fringes, obtain the fringe lines corresponding to all the fringes. For each fringe line, use the least squares method for fitting to obtain the angle value between the fitted curve and the X-axis, and obtain the difference between this angle value and the included angle θ between the two grating lines. Thus, the slope and difference of the fringes can be obtained, and the strain of the metal material can be calculated. Furthermore, obtain the maximum value of the strain among all the fringes, and based on the maximum value of the strain, the linear expansion coefficient of the metal material can be calculated. It should be noted that since the process of measuring the linear expansion coefficient of the metal material based on the characteristics of the grating and the Moiré fringes is prior art, it will not be elaborated here too much.

[0101] Thus, the linear expansion coefficients of the metal at multiple different temperatures can be obtained.

[0102] In this embodiment, use the obtained linear expansion coefficients of the metal at multiple different temperatures, and based on the existing prediction model, predict the linear expansion coefficient at the temperature to be measured, so as to complete the prediction of the expansion rate in high-purity metal.

[0103] Among them, the prediction model can be data fitting, the arima algorithm, the random forest algorithm, etc. Since the specific prediction process of the above prediction models is prior art, it will not be elaborated here too much.

[0104] Exemplarily, perform data fitting on the temperature and the linear expansion coefficient of the metal at different temperatures to obtain a fitting curve, and substitute the temperature to be measured into the fitting curve to obtain the predicted linear expansion coefficient.

[0105] In the description of this specification, the meaning of "a plurality" is at least two, such as two, three or more, etc., unless otherwise specifically defined.

[0106] Although this specification has shown and described multiple embodiments of the present invention, it is obvious to those skilled in the art that such embodiments are provided only by way of example. Those skilled in the art will think of many changes, alterations and alternative ways without departing from the spirit and idea of the present invention. It should be understood that various alternative solutions to the embodiments of the present invention described herein can be adopted in the process of practicing the present invention.

Claims

1. A method for predicting expansion rate in high-purity metals based on image processing, characterized in that: The following steps are involved: Obtain grayscale images of the grating image of the metal surface at different temperatures, and obtain a frequency spectrum of the grayscale image corresponding to any temperature; Determine the optimal filter radius of the filter of the spectrum graph; filter the spectrum graph according to the optimal filter radius to obtain an optimal spectrum graph; obtain an equidistant fringe image based on the optimal spectrum graph; The linear expansion coefficient of the metal at any temperature is obtained based on the arithmetic difference fringe image, and the linear expansion coefficient of the metal at the temperature to be measured is predicted based on the obtained linear expansion coefficients at different temperatures; Wherein, the optimal filtering radius is: Acquire a peak point sequence in the spectrum graph; construct a triangulated network model based on the position coordinates of each peak point in the peak point sequence; the triangulated network model includes multiple triangulated networks; calculate the priority value of each filter radius, the priority value is positively correlated with the fitness value and importance of each triangulated network; the fitness value represents the possibility of belonging to equidistant stripes in the corresponding triangulated network, and the fitness value for: ； in, is the ratio of the relative angle of the jth triangulation to 90°, where the relative angle is the angle between the straight line formed by the center of gravity of the jth triangulation and the center point of the spectrum graph and the straight line formed along the angle β direction, is the ratio of the Euclidean distance from the centroid of the jth triangulated network to the center of the spectrum graph to the diagonal length of the grayscale image, is the ratio of the mean spectral intensity of all internal points in the jth triangulated network to the spectral intensity of the center point of the spectrum graph, is an exponential function with the natural constant e as base, , θ is the set angle; the importance is the overlap between any triangulated network and the circular area; the center of the circular area is the center point of the spectrum graph, and the radius is any filter radius; Based on each filter radius and the corresponding priority value, a corresponding fitting function is generated, and the filter radius corresponding to the maximum value of the derivative of the fitting function is taken as the optimal filter radius.

2. The method for predicting expansion rate in high-purity metal based on image processing according to claim 1, characterized in that: The acquisition process of the grating image is as follows: The metal to be tested is used as the screen, and a moiré fringe experiment is performed using two gratings with the same grating parameters. During the moiré fringe experiment, the angle between the grating lines of the two gratings is a set angle, and a moiré fringe image is formed on the surface of the metal to be tested, and the moiré fringe image is used as the grating image.

3. The method for predicting expansion rate in high-purity metal based on image processing according to claim 1, characterized in that: The process of obtaining the importance is as follows: Randomly select any filter radius, take the center point of the spectrum graph as the center of the circle, and obtain the circular area corresponding to any filter radius; The number of overlapping data between the circular area and the j-th triangulated network is counted, and the ratio of the number of overlapping data to the number of all internal points in the j-th triangulated network is used as the importance of any filtering radius.

4. The method for predicting expansion rate in high-purity metal based on image processing according to claim 2 or 3, characterized in that: The priority value for: ； Among them, J is the total number of triangulated networks, is the importance of the sth filter radius, is the appropriate value of the jth triangulation.

5. The method for predicting expansion rate in high-purity metal based on image processing according to claim 1, characterized in that: Before constructing the triangulated network model, the steps of screening the peak point sequence to obtain a new peak point sequence are also included, specifically: Obtain the spectrum intensity of each peak point in the peak point sequence; Based on the position coordinates of each peak point and the corresponding spectrum intensity, the peak point sequence is divided into an interference sequence, a weak feature sequence, and a strong feature sequence; wherein the interference sequence is a category with the smallest spectrum intensity mean, the weak feature sequence is a category with the spectrum intensity mean in the middle, and the strong feature sequence is a category with the largest spectrum intensity mean; Eliminate the remaining peak points except the central peak point in the interference sequence and the strong feature sequence to obtain the weak feature sequence and the central peak point; The weak feature sequence and the central peak point are combined into a new peak point sequence.

6. The method for predicting expansion rate in high-purity metal based on image processing according to claim 1, characterized in that: The specific process of realizing the prediction of the linear expansion coefficient of the metal at the temperature to be measured is: performing temperature and linear expansion coefficient data fitting on the linear expansion coefficient of the metal at different temperatures to obtain a fitting curve, substituting the temperature to be measured into the fitting curve to obtain the predicted linear expansion coefficient.

7. The method for predicting expansion rate in high-purity metal based on image processing according to claim 1, characterized in that: After the grayscale image of each grating image is acquired, the grayscale image is further denoised to obtain a denoised grayscale image.

Images