A crack deformation measurement method based on reliable sub-region DIC

Abstract

The present application relates to the field of solid experimental mechanics, and discloses a crack deformation measurement method based on reliable sub-region DIC, comprising: extracting feature points in the region of interest on the target image and the reference image by using the SIFT algorithm and matching; taking the statistical region of the feature vector corresponding to each feature point matched on the reference image as a reliable sub-region to perform correlation operation, and obtaining the displacement of the center point of the reliable sub-region; selecting a fitting region with each measuring point in the reference image as the center, and judging the crack in the fitting region based on the Delaunay triangulation algorithm and the coefficient of variation; in the case that there is a crack in the fitting region, using the 3D Hough Transform algorithm to divide the center points of the reliable sub-regions in the fitting region into two categories; selecting one category of the center points of the reliable sub-regions according to the position of the measuring point to perform displacement fitting, obtaining the fitting coefficient and determining the deformation parameters of the measuring point. The method can be used for crack deformation measurement under complex conditions, and improves the applicability of the digital image correlation method.

Description

Technical Field

[0001] This invention relates to the field of experimental mechanics of solids, and relates to, but is not limited to, a method for measuring crack deformation based on reliable subregion DIC. Background Technology

[0002] In the field of experimental mechanics of solids, digital image correlation (DIC) has become an important method for measuring object deformation. Its advantages include simple operation, low requirements for the measurement environment, high accuracy, non-contact measurement, and full-field deformation measurement. The standard DIC method can only guarantee measurement accuracy when the object's deformation is continuous, mainly because the correlation calculation uses shape functions based on continuous deformation theory. In many cases, it is necessary to measure discontinuous deformation of objects, especially in the study of crack formation and evolution. Therefore, there is a need to develop DIC methods suitable for measuring discontinuous deformation.

[0003] The existing discontinuous deformation DIC methods have the following drawbacks: (1) Discontinuous deformation is difficult to identify. Some methods generally use a correlation coefficient threshold to judge discontinuous deformation, that is, it is believed that discontinuous deformation has occurred when the correlation coefficient is lower than a certain value. However, in reality, there are other factors that cause the correlation coefficient to drop, such as the destruction of the speckle pattern. (2) It can only handle simple discontinuous deformation. Some methods idealize discontinuous deformation, believing that the formation of discontinuity will not cause damage to the speckle field around it. However, in reality, the speckle field around the formation of discontinuity will inevitably be destroyed, especially for some granular materials. Once the speckle pattern around the discontinuous position is damaged, the measurement accuracy is difficult to guarantee. Summary of the Invention

[0004] In view of this, embodiments of the present invention provide a crack deformation measurement method based on reliable sub-region DIC, which at least solves the problems of difficulty in identifying discontinuous deformation and low accuracy in identifying complex deformation.

[0005] The specific technical solutions of this invention are as follows:

[0006] This invention provides a crack deformation measurement method based on reliable subregion DIC, the method comprising:

[0007] In a series of deformed object images, a reference image and a target image are selected. The SIFT algorithm is used to extract and match feature points within regions of interest (ROIs) in both images. The statistical region of the feature vector corresponding to each matched feature point in the reference image is used as a reliable sub-region for correlation calculations to obtain the displacement of the reliable sub-region's center point. During the correlation calculations, the displacement of the matched feature points is used as the initial displacement value for iteration. A fitting region is selected centered on each measurement point in the reference image, and crack detection is performed on the fitting region based on the Deloni triangulation algorithm and the coefficient of variation. If cracks exist in the fitting region, the 3D Hough Transform algorithm is used to divide the reliable sub-region center points within the fitting region into two categories. Based on the location of the measurement point, one category of reliable sub-region center points is selected for displacement fitting, and the fitting coefficients are obtained to determine the deformation parameters of the measurement point.

[0008] Furthermore, the crack detection of the fitted region based on the Deloni triangulation algorithm and the coefficient of variation includes: connecting the center points of N reliable sub-regions within the fitted region into q triangles using the Deloni triangulation algorithm in the XOY plane; and calculating the plane equation A of each triangle in the O-xyu and O-xyv spaces. i x+B i y+C i +Z=0; where Z represents the displacement component u in the x-direction or v in the y-direction at the center point of the reliable sub-region, A i B i C i Let C be the coefficients of the plane equation, i = 1, 2, 3, ..., q; and let C be the coefficients of all triangle plane equations. i Form a matrix Q = {C1, C2, C3, ..., C} q}, calculate the coefficient of variation C of Q. v ; in C v If the value is greater than 1, it is determined that the fitted region contains cracks or mismatched points.

[0009] Furthermore, the 3D Hough Transform algorithm is used to classify the reliable sub-region center points within the fitting region into two categories, including: classifying the N reliable sub-region center points (x...) within the fitting region... j ,y j ,u j ) and (x j ,y j ,v j Transform to spherical coordinates according to the following correspondence: θ-φ-r u and θ-φ-r v In two parameter spaces, where j∈[1,N]:

[0010]

[0011] In the formula, u j v j Let θ and φ be the components of the displacement at the center point of the reliable sub-region in the x and y directions, respectively, and let θ and φ be the spherical coordinate representations of the plane normal vector, where θ∈(0,π) and φ∈(-π,π). u It is any point (x) j ,y j ,u j The distance from the origin of the polar coordinate system. rv is any point (x) j ,y j ,v j The distance from the origin of the polar coordinate system.

[0012] θ-φ-r u Space and θ-φ-r v The space is divided into multiple small grids according to a preset size, and the calculated r u r v The count of the small square where the value falls is incremented by 1; where each small square represents a possible solution in the parameter space; θ-φ-r u Space and θ-φ-r v The cell with the largest cumulative count in the space is designated as the target cell; the center point of the reliable sub-region that falls within the target cell is designated as the first category, and the center point of the reliable sub-region that does not fall within the target cell is designated as the second category.

[0013] Furthermore, the preset dimensions are as follows: the length of each small grid in both the θ and φ dimensions is π / 180; r u The length of each small cell in the dimension is taken as |max(u1,u2,…,u N )-min(u1,u2,…,u N )| / 10;r v The length of each small cell in the dimension is taken as |max(v1,v2,…,v N )-min(v1,v2,…,v N )| / 10.

[0014] Furthermore, the method further includes: when there are at least two target cells, comparing the fitting residuals for each target cell's corresponding classification; and taking the classification corresponding to the target cell with the smallest fitting residual as the best classification result.

[0015] Further, the step of selecting one type of reliable sub-region center point for displacement fitting based on the location of the measuring point includes: when the difference in the number of center points of the two types of reliable sub-regions is greater than 3, selecting one type of reliable sub-region center point where the measuring point is located for displacement fitting; when the difference in the number of center points of the two types of reliable sub-regions is less than or equal to 3, using the SVM algorithm to determine the boundary of the two types of reliable sub-region center points, and selecting one type of reliable sub-region center point for displacement fitting based on the positional relationship between the measuring point and the boundary.

[0016] Furthermore, the fitting process uses a first-order polynomial function, which can be expressed as:

[0017]

[0018] In the formula, , respectively, are the fitted values ​​of the horizontal displacement and the vertical displacement at the center point of the reliable sub-region; a0, a1, a2, b0, b1, and b2 are the fitting coefficients;

[0019] The fitting coefficients are obtained by using the least squares method. The horizontal displacement u(x,y) of the measuring point is a0x+a1y+a2, the vertical displacement v(x,y) is b0x+b1y+b2, and the deformation parameters of the measuring point are [u(x,y),v(x,y),a0,a1,b0,b1].

[0020] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0021] In this embodiment of the invention, firstly, SIFT feature matching is performed on the region of interest in the reference image. Then, a reliable sub-region is selected for correlation calculation. Next, a fitting region is determined with the measurement point as the center. Crack judgment and reliable sub-region center point classification are performed on the fitting region. Finally, a suitable class of reliable sub-region center points is selected for displacement fitting, the fitting coefficients are obtained, and the deformation parameters of the measurement point are calculated. Thus, a crack processing flow is added to the reliable sub-region digital image correlation method, improving the accuracy of crack deformation measurement. It can be used for crack deformation measurement in complex situations, enhancing the applicability of digital image correlation methods. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort, wherein:

[0023] Figure 1A schematic flowchart of the crack deformation measurement method based on reliable sub-region DIC provided in an embodiment of the present invention;

[0024] Figure 2 A schematic diagram of the reliable sub-region and the fitting region of the measurement point provided in an embodiment of the present invention;

[0025] Figure 3 This is a schematic diagram of the crack judgment process provided in an embodiment of the present invention, wherein (a) represents the distribution of the center points of the reliable sub-region, (b) represents the division of several triangles, and (c) represents the spatial position of the triangles;

[0026] Figure 4 This is a schematic diagram showing the relative positional relationship between the measuring point and the center points of two types of reliable sub-regions provided in an embodiment of the present invention; wherein (a) represents the side with a large difference in the number of the two types of points and the measuring point is on the side with more reliable sub-region center points; (b) represents the side with a small difference in the number of the two types of points and the measuring point is on the side with more reliable sub-region center points; and (c) represents the side with a small difference in the number of the two types of points and the measuring point is on the side with fewer points.

[0027] Figure 5 A schematic diagram of the boundary line determined by the SVM algorithm provided in this embodiment of the invention;

[0028] Figure 6 A schematic diagram of deformation measurement results of a similar material specimen provided in an embodiment of the present invention; (a) partially shows the image before deformation; (b) the image after deformation and the position of the measuring point; (c) magnified view of region 1; (d) magnified view of region 2;

[0029] Figure 7 The cloud diagram shows the deformation measurement results of similar material specimens provided in the embodiments of the present invention; (a) represents the horizontal displacement field; (b) represents the vertical displacement field. Detailed Implementation

[0030] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. The following embodiments are used to illustrate the present invention, but are not intended to limit the scope of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0031] In the following description, references are made to “some embodiments,” which describe a subset of all possible embodiments. However, it is understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0032] It should be noted that the terms "first, second, and third" used in the embodiments of the present invention are only used to distinguish similar objects and do not represent a specific ordering of objects. It is understood that "first, second, and third" can be interchanged in a specific order or sequence where permitted, so that the embodiments of the present invention described herein can be implemented in an order other than that illustrated or described herein.

[0033] It will be understood by those skilled in the art that, unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which these embodiments of the invention pertain. It should also be understood that terms such as those defined in general dictionaries should be understood to have the same meaning as in the context of the prior art and should not be interpreted in an idealized or overly formal sense unless specifically defined as herein.

[0034] When performing deformation measurements using the standard DIC method, it is necessary to select a reference image and a target image. Typically, the image before deformation is used as the reference image, and the image after deformation is used as the target image. The key to deformation measurement using the DIC method is finding the correspondence between the reference image and the target image. Considering the non-uniformity of object deformation, a small region is usually selected on the reference image to find the correspondence between the reference and target images.

[0035] The Reliable Subregion DIC method is a novel extension of the standard DIC method. It adjusts the subregion's position to avoid discontinuous areas, making it simpler to implement than subregion segmentation methods and extended finite element methods. The core steps are: first, selecting a reference subregion as the reliable subregion by feature matching of the images before and after deformation; then, obtaining the deformation parameters of the reliable subregion through correlation calculations; finally, obtaining the deformation parameters of the measurement point by least-squares fitting of the displacement of the center point of the reliable subregion within a certain area around the measurement point. It should be noted that the reliable subregion is centered on a feature point, and the position of the feature point depends on the image's grayscale distribution, which means that the feature point and the manually placed measurement point may not completely coincide. Therefore, it is necessary to calculate the deformation of the measurement point based on the deformation of the reliable subregion surrounding it.

[0036] Figure 1 A flowchart illustrating a crack deformation measurement method based on reliable sub-region DIC provided in this embodiment of the invention is shown below. Figure 1 As shown, the method includes at least the following steps:

[0037] Step S110: Select a reference image and a target image from a series of deformed object images, and use the SIFT algorithm to extract feature points in the regions of interest on the target image and the reference image and perform matching.

[0038] Here, the image before deformation is typically used as the reference image, and the image after deformation is used as the target image. The key to deformation measurement using the DIC method is finding the correspondence between the reference image and the target image. Considering the non-uniformity of object deformation, a small region is usually selected on the reference image to find the correspondence between the reference image and the target image.

[0039] Although SIFT is a feature point matching algorithm, it is also, to some extent, a region matching algorithm. This is because feature point matching in SIFT is achieved by finding the shortest Euclidean distance between the feature vectors of the feature points, and the feature vectors are obtained by statistically analyzing the pixel gray-level gradients within a certain region (statistical region) around the feature point. If the feature points on the images before and after deformation can be matched, then the regions surrounding the feature points in the images have similarity.

[0040] Step S120: The statistical region of the feature vector corresponding to each matched feature point on the reference image is used as a reliable sub-region for correlation calculation to obtain the displacement of the center point of the reliable sub-region.

[0041] Here, the displacement of the matched feature point is used as the initial displacement value for the iteration during the relevant calculation process.

[0042] When cracks or invalid regions appear around feature points, the grayscale gradient of the image changes, the feature vector changes, and the feature points may even disappear, which can significantly affect feature point matching. Therefore, a reference sub-region is selected using the statistical region of the feature vectors of the matched feature points. This avoids the reference sub-region from containing cracks or invalid regions after deformation. To distinguish it from the reference sub-region of the traditional DIC method, the reference sub-region selected by matching image features before and after deformation is called the reliable sub-region.

[0043] The reliable subregions are randomly distributed, which is not conducive to calculating the deformation field. To obtain the deformation parameters at the measuring point, it is necessary to fit the displacement of the center point of the reliable subregion within a certain area around the measuring point, such as... Figure 2 As shown, the left side is the reference image, the right side is the target image, and the dashed box centered on the measurement point is the fitting region.

[0044] The statistical region of the eigenvector is related to the scale of the feature points, which leads to the selection of less uniform sizes for reliable subregions. Optionally, the size of the reliable subregion is 12[σ]. oct ]×12[σ oct ](σ oct (Scale of feature points).

[0045] Step S130: Select a fitting region centered on each measurement point in the reference image, and determine the crack in the fitting region based on the Deloni triangulation algorithm and the coefficient of variation.

[0046] Here, the measurement point is a selected, equally spaced pixel point on the reference image or the center point of a predefined region of interest (ROI). The fitting region is a suitably sized area surrounding the measurement point; this area can be circular, rectangular, or other shapes, depending on the crack morphology and image resolution. Preferably, a rectangular area of ​​50 pixels × 50 pixels to 60 pixels × 60 pixels is selected.

[0047] When a crack is present within the fitted region, the gradient of the displacement of adjacent matched feature points on both sides of the crack changes significantly compared to other locations. Therefore, the gradient of the displacement of adjacent matched feature points can be used for crack detection. Considering that the selection of the gradient threshold for crack detection is highly empirical, this embodiment of the invention uses a Deloni triangulation algorithm and the coefficient of variation for crack detection.

[0048] When processing crack detection tasks in reference images, combining the Delaunay triangulation algorithm and the coefficient of variation (CV) is an effective method to assess the likelihood of cracks within the fitted region. The Delaunay triangulation algorithm is applied to the center points of reliable sub-regions within the fitted region to generate a triangular mesh. The generated triangular mesh is then examined, especially those near the edges of the fitted region or containing outliers. Cracks can cause significant changes in the shape of the triangular mesh, such as uneven side lengths, excessively large or small angles, etc. The coefficient of variation, the ratio of the standard deviation to the mean, measures the dispersion of the data. In crack detection, cracked regions may exhibit high grayscale variation because there can be significant grayscale differences between the crack edges and the interior. If the triangular mesh has an anomalous shape and a high coefficient of variation, the region is likely to contain a crack.

[0049] It is worth noting that before crack detection, the reference image may need to be preprocessed, such as denoising and contrast enhancement, to improve the accuracy of crack detection.

[0050] Step S140: In the case of cracks in the fitting region, the 3D Hough Transform algorithm is used to divide the center points of reliable sub-regions in the fitting region into two categories.

[0051] Here, the 3D Hough Transform (HT) algorithm is a powerful tool, particularly suitable for detecting and analyzing geometry from noisy data, especially when multiple shapes or models exist. The 3D Hough Transform algorithm determines the geometry present in the data by transforming reliable sub-region center points from the original space to the parameter space and utilizing counting and fitting residuals. It is particularly suitable for processing datasets containing multiple models or shapes.

[0052] First, the parameter space of the 3D Hough Transform needs to be defined. For crack detection, the parameter space may include the crack's orientation, location, and some form of scale (such as crack width or length). The specific parameters are chosen based on the crack's characteristics and the nature of the image. The center points of each reliable sub-region are transformed from the original image space to the 3D Hough Transform's parameter space. In the parameter space, an accumulator array is used to record the vote count for each parameter combination. Each center point increments its count in the accumulator array based on its corresponding parameter. Then, by detecting the peaks in the accumulator array, the most likely crack parameters can be identified.

[0053] Based on the results of the 3D Hough Transform, reliable sub-region center points are divided into two categories: one category of reliable sub-region center points has parameters that correspond to peak values ​​in the accumulator array, indicating that they are located on or very close to the crack. The other category of reliable sub-region center points has parameters that do not match any significant peak values, possibly located outside the crack, or misidentified as reliable sub-regions due to noise, artifacts, or other reasons.

[0054] Step S150: Select one type of reliable sub-region center point based on the location of the measuring point for displacement fitting, obtain the fitting coefficient, and determine the deformation parameters of the measuring point.

[0055] Here, in order to obtain the deformation parameters at the measuring point, the displacement of the center point of a reliable sub-region around the measuring point is fitted. Considering that the size of the fitting region is relatively small, a first-order polynomial function is selected as the fitting function.

[0056] In this embodiment of the invention, firstly, SIFT feature matching is performed on the region of interest in the reference image. Then, a reliable sub-region is selected for correlation calculation. Next, a fitting region is determined with the measurement point as the center. Crack judgment and reliable sub-region center point classification are performed on the fitting region. Finally, a suitable class of reliable sub-region center points is selected for displacement fitting, the fitting coefficients are obtained, and the deformation parameters of the measurement point are calculated. Thus, a crack processing flow is added to the reliable sub-region digital image correlation method, improving the accuracy of crack deformation measurement. It can be used for crack deformation measurement in complex situations, enhancing the applicability of digital image correlation methods.

[0057] In some embodiments, the step S130 above, "judging the crack in the fitted region based on the Deloni triangulation algorithm and the coefficient of variation," is specifically implemented through some processes:

[0058] S131, in the XOY plane, the Deloni triangulation algorithm is used to connect the center points of N reliable sub-regions in the fitting region into q triangles.

[0059] Here, as Figure 3 As shown, (a) represents the distribution of reliable sub-region center points, (b) represents the division into several triangles, and (c) represents the spatial location of the triangles.

[0060] S132, calculate the plane equation A of each triangle in the O-xyu and O-xyv spaces. i x+B i y+C i +Z=0.

[0061] Here, Z represents the displacement component u in the x-direction or v in the y-direction at the center point of the reliable sub-region, and A i B i C i The coefficients are the plane equations, i = 1, 2, 3, ..., q; the parameters of the plane equations for each triangle need to be determined.

[0062] S133, the coefficients C in the equations of all triangular planes i Form a matrix Q = {C1, C2, C3, ... ….. C q}, calculate the coefficient of variation C of Q. v .

[0063] Here, the coefficient of variation is the ratio of the standard deviation to the mean, used to measure the dispersion of the data. A threshold is set based on the distribution of the coefficient of variation. If the coefficient of variation of the fitted region exceeds this threshold, the region is considered to potentially contain cracks. v It can be represented as in Let Q be the standard deviation; Let Q be the mean of Q.

[0064] S134, in C v If the value is greater than 1, it is determined that the fitted region contains cracks or mismatched points.

[0065] Here, if the coefficient of variation C of the fitted region v If the value exceeds a set threshold of 1, the region is considered to potentially contain cracks. If the fitted region contains cracks or mismatched points, further classification of the reliable sub-region center points is needed to reduce fitting error. If C v If the threshold value of 1 is not exceeded, displacement fitting of the center point of the reliable sub-region can be performed directly.

[0066] In some embodiments, step S140 above, "using the 3D Hough Transform algorithm to divide the reliable sub-region center points within the fitting region into two categories," is specifically implemented through the following steps:

[0067] S141, the center points (x) of N reliable sub-regions within the fitting region j ,y j ,u j ) and (x j ,y j ,v j Transform to spherical coordinates according to the following correspondence: θ-φ-r u and θ-φ-r v In two parameter spaces, where j∈[1,N]:

[0068]

[0069] In the formula, u j v j Let θ and φ be the components of the displacement at the center point of the reliable sub-region in the x and y directions, respectively, and let θ and φ be the spherical coordinate representations of the plane normal vector, where θ∈(0,π) and φ∈(-π,π). u It is any point (x) j ,y j ,u j The distance from the origin of the polar coordinate system. rv is any point (x) j ,y j ,v j The distance from the origin of the polar coordinate system.

[0070] Here, we first transform the center points of the N reliable sub-regions from the Cartesian coordinate system to the parameter space. The purpose of this transformation is to change the problem from finding a point set of a specific shape in the Cartesian coordinate system to finding the peak value in the parameter space.

[0071] S142, θ-φ-r u Space and θ-φ-r v The space is divided into multiple small grids according to a preset size, and the calculated r u r v The count of the small square where the value falls is incremented by 1; where each small square represents a possible solution in the parameter space.

[0072] Here, the parameter space is divided into multiple small grids (or "bins"), each corresponding to a region within the parameter space. Each small grid represents a possible crack characteristic solution. For the center point of each reliable sub-region, its corresponding position in the parameter space is calculated using the transformation formula described above, based on its coordinates in the Cartesian coordinate system. If the calculated result falls within a certain small grid, the cumulative count of that small grid is incremented by 1.

[0073] S143, θ-φ-r u Space and θ-φ-rv The cell with the largest cumulative count in the space is selected as the target cell.

[0074] Here, after transforming the center points of all reliable sub-regions, the algorithm searches for the target cell with the largest cumulative count in the parameter space. These peaks represent possible model parameters.

[0075] S144, the reliable sub-region center points that fall within the target small grid are classified as the first category, and the reliable sub-region center points that do not fall within the target small grid are classified as the second category.

[0076] In this way, the complexity of classification and recognition is simplified by transforming the problem of classifying reliable subregion center points into a peak detection problem in the parameter space.

[0077] In some embodiments, the preset dimensions are as follows: the length of each small grid in both the θ and φ dimensions is π / 180; r u The length of each small cell in the dimension is taken as |max(u1,u2,…,u N )-min(u1,u2,…,u N )| / 10;r v The length of each small cell in the dimension is taken as |max(v1,v2,…,v N )-min(v1,v2,…,v N )| / 10.

[0078] Here, in this embodiment of the invention, the segmentation dimensions of each dimension of the parameter space are preset according to the expected crack characteristics or detection requirements. These dimensions can be adjusted as needed in practical applications.

[0079] In some embodiments, the method further includes: when there are at least two target cells, comparing the fitting residuals for each target cell's corresponding classification; and taking the classification corresponding to the target cell with the smallest fitting residual as the best classification result.

[0080] Here, there may be more than one cell with the largest count, meaning there could be multiple classifications. The presence of at least two target cells indicates the existence of multiple peaks, implying the possible existence of multiple models. In this case, the algorithm needs to fit each possible model and calculate the fitting residual for each model. Finally, the algorithm selects the model with the smallest fitting residual as the optimal classification.

[0081] Calculate the fitting residuals: For each classification (i.e., each maximum count cell), fit the corresponding reliable subregion center point to the geometry represented by that classification and calculate the fitting residuals. The fitting residuals are the deviations between the actual reliable subregion center points and the fitted geometry (model predictions). The smaller the residuals, the better the model fit. By comparing the fitting residuals of different models, the algorithm can select the optimal model, working effectively even when multiple models exist in the data.

[0082] In some embodiments, the above step S150, "selecting one type of reliable sub-region center point for displacement fitting based on the location of the measuring point," is specifically implemented through the following steps:

[0083] S151, when the difference in the number of center points of the two types of reliable sub-regions is greater than 3, the center point of one type of reliable sub-region where the measuring point is located is selected for displacement fitting.

[0084] S152, when the difference in the number of center points of the two reliable sub-regions is less than or equal to 3, the SVM algorithm is used to determine the boundary of the center points of the two reliable sub-regions, and the center point of one reliable sub-region is selected for displacement fitting based on the positional relationship between the measuring point and the boundary.

[0085] Here, based on the relative positions of the measuring points and the cracks, the possible scenarios after classification are as follows: Figure 4 As shown, (a) represents the side with a large difference in the number of points between the two types, where the measuring point is located on the side with more reliable sub-region center points; (b) represents the side with a small difference in the number of points between the two types, where the measuring point is located on the side with more reliable sub-region center points; and (c) represents the side with a small difference in the number of points between the two types, where the measuring point is located on the side with fewer points. After classifying the reliable sub-region center points, it is necessary to select one type of point for displacement fitting based on the location of the measuring point. For Figure 4 In case (a), the type of points with a larger number are directly identified as the center points of the reliable fitting sub-region. This is because there is a significant difference in the number of points of the two types when the measuring point is far from the crack surface. However, for Figure 4 In cases (b) and (c), the number of reliable sub-region center points cannot be used to determine the fitted reliable sub-region center point. This is because when the measuring point is close to the crack surface, the difference in the number of points of the two types is small, making it impossible to determine the class to which the measuring point belongs. To address this situation, SVM-supported data selection is employed. Figure 5 As shown, the SVM algorithm is used to determine the boundary w of the two classes of points. T When X+n=0, the class of the reliable sub-region center point can be determined based on the positional relationship between the measurement point and the boundary. w is the boundary normal vector, n is the intercept, and X is the coordinates of the reliable sub-region center point. Through extensive testing, the authors found that when the difference in the number of points in the two classes is less than 3, [a problem occurs]. Figure 4 Cases (b) and (c) are as follows.

[0086] In some embodiments, the fitting process employs a first-order polynomial function, which can be expressed as:

[0087]

[0088] In the formula, , respectively, are the fitted values ​​of the horizontal displacement and the vertical displacement at the center point of the reliable sub-region; a0, a1, a2, b0, b1, and b2 are the fitting coefficients;

[0089] The fitting coefficients are obtained by using the least squares method. The horizontal displacement u(x,y) of the measuring point is a0x+a1y+a2, the vertical displacement v(x,y) is b0x+b1y+b2, and the deformation parameters of the measuring point are [u(x,y),v(x,y),a0,a1,b0,b1].

[0090] This invention utilizes the proposed crack deformation measurement method based on reliable sub-region DIC to conduct uniaxial compression experiments on similar materials, such as... Figure 6 The deformation measurement results of similar material specimens are shown. The calculation region (i.e., the fitting region) is as follows: Figure 6 As shown in section (a), the measurement point interval is 5 pixels, the number of measurement points is 6710 (122 rows × 55 columns), and the fitting region size is 51 pixels × 51 pixels. To more intuitively reflect the deformation of the specimen, only the positions of the measurement points on the deformed image are given in the measurement results. Figure 6 As shown in (b), (c), and (d), the morphology and boundary of the crack can be accurately observed from the location of the measuring point after deformation, with no obvious abnormal locations. Figure 7 As can be seen in sections (a) and (b), the positions of the measuring points after deformation are basically close to the boundary of the actual crack. This indicates that the method proposed in this invention has good stability for displacement measurement of actual cracks.

[0091] This invention incorporates a crack processing procedure into the reliable sub-region digital image correlation method, improving the accuracy of crack deformation measurement. The key points are as follows: (1) The Deloni triangulation algorithm and coefficient of variation are used to determine if cracks are present within the fitted region. (2) 3D-Hough transform is used to classify reliable sub-region center points. (3) After classification, when the number of reliable sub-region center points in both classes is small, the SVM algorithm is used to select the reliable sub-region center point class. This invention can be used for crack deformation measurement in complex situations, improving the applicability of digital image correlation methods.

[0092] It should be understood that the phrase "one embodiment" or "an embodiment" throughout the specification means that a specific feature, structure, or characteristic related to the embodiment is included in at least one embodiment of the invention. Therefore, "in one embodiment" or "in an embodiment" appearing throughout the specification does not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. It should be understood that in the various embodiments of the invention, the sequence numbers of the above-described processes do not imply a sequential order of execution; the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the invention. The sequence numbers of the above-described embodiments of the invention are merely descriptive and do not represent the superiority or inferiority of the embodiments.

[0093] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0094] In the several embodiments provided by this invention, it should be understood that the disclosed methods can be implemented in other ways. The methods disclosed in the several method embodiments provided by this invention can be arbitrarily combined without conflict to obtain new method embodiments. The features disclosed in the several method embodiments provided by this invention can be arbitrarily combined without conflict to obtain new method embodiments.

[0095] The above description is merely an embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A crack deformation measurement method based on reliable sub-region DIC, characterized in that, include: In a series of deformed object images, a reference image and a target image are selected. The SIFT algorithm is used to extract feature points in the regions of interest on the target image and the reference image and then match them. The statistical region of the feature vector corresponding to each matched feature point on the reference image is taken as a reliable sub-region for correlation operation to obtain the displacement of the center point of the reliable sub-region; wherein, the displacement of the matched feature point is used as the initial displacement value of the iteration during the correlation operation. A fitting region is selected centered on each measurement point in the reference image, and cracks are judged in the fitting region based on the Deloni triangulation algorithm and the coefficient of variation. In the case of cracks in the fitted region, the 3D Hough Transform algorithm is used to classify the center points of reliable sub-regions within the fitted region into two categories. Based on the location of the measuring point, select one type of reliable sub-region center point for displacement fitting, obtain the fitting coefficient, and determine the deformation parameters of the measuring point.

2. The method according to claim 1, characterized in that, The crack detection of the fitted region based on the Deloni triangulation algorithm and the coefficient of variation includes: In the XOY plane, the Deloni triangulation algorithm is used to connect the center points of N reliable sub-regions within the fitting region into q triangles. Calculate the plane equation A for each triangle in the O-xyu and O-xyv spaces. i x+B i y+C i +Z=0; where Z represents the displacement component u in the x-direction or v in the y-direction at the center point of the reliable sub-region, A i B i C i Let be the coefficients of the plane equation, i = 1, 2, 3, ..., q; The coefficients C in the equations of all triangular planes i Form a matrix Q = {C1, C2, C3, ..., C} q }, calculate the coefficient of variation C of Q. v ; In C v If the value is greater than 1, it is determined that the fitted region contains cracks or mismatched points.

3. The method according to claim 1, characterized in that, The 3D Hough Transform algorithm is used to classify the reliable sub-region center points within the fitted region into two categories, including: The center points (x) of N reliable sub-regions within the fitting region are defined. j ,y j ,u j ) and (x j ,y j ,v j Transform to spherical coordinates according to the following correspondence: θ-φ-r u and θ-φ-r v In two parameter spaces, where j∈[1,N]: In the formula, u j v j Let θ and φ be the components of the displacement at the center point of the reliable sub-region in the x and y directions, respectively, and let θ and φ be the spherical coordinate representations of the plane normal vector, where θ∈(0,π) and φ∈(-π,π). u It is any point (x) j ,y j ,u j The distance from the origin of the polar coordinate system. rv is any point (x) j ,y j ,v j The distance from the origin of the polar coordinate system. θ-φ-r u Space and θ-φ-r v The space is divided into multiple small grids according to a preset size, and the calculated r u r v The count of the small square where the value falls is incremented by 1; where each small square represents a possible solution in the parameter space; θ-φ-r u Space and θ-φ-r v The cell with the largest cumulative count in the space is selected as the target cell. The reliable sub-region center points that fall within the target grid are classified as the first category, and the reliable sub-region center points that do not fall within the target grid are classified as the second category.

4. The method according to claim 3, characterized in that, The preset dimensions are as follows: The length of each small cell in both the θ and φ dimensions is π / 180. r u The length of each small cell in the dimension is taken as |max(u1,u2,…,u N )-min(u1,u2,…,u N )| / 10; r v The length of each small cell in the dimension is taken as |max(v1,v2,…,v N )-min(v1,v2,…,v N )| / 10.

5. The method according to claim 3 or 4, characterized in that, The method further includes: When there are at least two target cells, compare the fitting residuals for each target cell under the corresponding category. The classification corresponding to the target small grid with the smallest fitting residual is taken as the best classification result.

6. The method according to any one of claims 1 to 4, characterized in that, The step of selecting a reliable sub-region center point for displacement fitting based on the location of the measuring point includes: If the difference in the number of center points between the two reliable sub-regions is greater than 3, the center point of the reliable sub-region where the measuring point is located is selected for displacement fitting. When the difference in the number of center points of the two reliable sub-regions is less than or equal to 3, the SVM algorithm is used to determine the boundary of the center points of the two reliable sub-regions, and the center point of one reliable sub-region is selected for displacement fitting based on the positional relationship between the measuring point and the boundary.

7. The method according to any one of claims 1 to 4, characterized in that, The fitting process uses a first-order polynomial function, which can be expressed as: In the formula, , respectively, are the fitted values ​​of the horizontal displacement and the vertical displacement at the center point of the reliable sub-region; a0, a1, a2, b0, b1, and b2 are the fitting coefficients; The fitting coefficients are obtained by using the least squares method. The horizontal displacement u(x,y) of the measuring point is a0x+a1y+a2, the vertical displacement v(x,y) is b0x+b1y+b2, and the deformation parameters of the measuring point are [u(x,y),v(x,y),a0,a1,b0,b1].

Images