A method, system, electronic device and medium for reconstructing a non-continuous deformation displacement field

Abstract

The application provides a non-continuous deformation displacement field reconstruction method, system, electronic equipment and medium, and belongs to the field of material dynamics analysis. The method comprises the following steps: collecting a reference image and a non-continuous deformation image; arranging control points in the reference image; performing digital speckle correlation calculation according to the control points in the reference image and the non-continuous deformation image to obtain an initial displacement field; drawing a ZNCC distribution map according to the reference image, a ZNCC value and a deformation subset, and determining a non-continuous boundary track and a non-continuous deformation region; removing data in the non-continuous deformation region in the initial displacement field, and filling a blank region to obtain a predicted displacement field; determining a non-continuous control point subset according to the non-continuous boundary track, the non-continuous deformation region, the predicted displacement field and the reference image; performing digital speckle correlation calculation according to the non-continuous control point subset to update the predicted displacement field, and obtaining a non-continuous deformation displacement field. The application improves the reconstruction accuracy of the non-continuous deformation displacement field.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of material kinetics analysis, and particularly to a non-continuous deformation displacement field reconstruction method based on adaptive subset correlation matching. BACKGROUND

[0002] With the continuous development of aerospace, mechanical manufacturing and special equipment industries, the service performance requirements of various materials and structures are becoming more and more demanding. In the structural design stage, measuring the deformation of materials or structures under loading and reconstructing the deformation displacement field are crucial for analyzing the dynamic mechanical behavior of materials and structures.

[0003] Digital speckle correlation is an optical measurement technology based on image processing and numerical calculation, and is an important test method for measuring material and structure deformation. It has the advantages of non-contact, full-field measurement and simple equipment. However, digital speckle correlation requires that the deformation of the measured surface be continuous, i.e., the deformation displacement field surface be continuous. In actual applications, non-continuous deformation often occurs in areas such as cracks, dislocations, slips and component connections. In non-continuous deformation areas, the deformation displacement field is complex, and the subset and shape function are difficult to adapt, which will result in a significant decrease in the measurement accuracy of digital speckle correlation in non-continuous deformation areas, or even the inability to complete the measurement. SUMMARY

[0004] The purpose of the present application is to provide a non-continuous deformation displacement field reconstruction method, system, electronic device and medium, which can improve the reconstruction accuracy of non-continuous deformation displacement field.

[0005] To achieve the above purpose, the present application provides the following solutions:

[0006] A non-continuous deformation displacement field reconstruction method, comprising:

[0007] acquiring a reference image and a non-continuous deformation image; the reference image is a speckle image before deformation of an object; the non-continuous deformation image is a speckle image after deformation of the object;

[0008] arranging control points in the reference image according to a pre-set region of interest and subset spacing;

[0009] performing digital speckle correlation calculation according to the control points in the reference image and the non-continuous deformation image to obtain a plurality of reference subsets, a plurality of deformation subsets, ZNCC values of each reference subset and an initial displacement field;

[0010] drawing a ZNCC distribution map according to the reference image, the ZNCC values of each reference subset and the deformation subsets, and determining a non-continuous boundary trajectory by using a peak searching algorithm according to the ZNCC distribution map;

[0011] Based on the reference image and the discontinuous boundary trajectory, the discontinuous deformation region is determined;

[0012] Data in discontinuous deformation regions in the initial displacement field are removed, and blank regions are filled with data in continuous regions to obtain the predicted displacement field; the continuous regions are the regions in the initial displacement field excluding the discontinuous deformation regions.

[0013] Based on the discontinuous boundary trajectory, the discontinuous deformation region, the predicted displacement field, and the reference image, a subset of discontinuous control points is determined;

[0014] Based on the subset of discontinuous control points, digital speckle correlation calculations are performed to update the predicted displacement field, thereby obtaining the discontinuous deformation displacement field.

[0015] Optionally, the number of reference subsets and the number of deformable subsets are the same as the number of control points;

[0016] Based on the control points in the reference image and the discontinuous deformation image, digital speckle correlation calculation is performed to obtain multiple reference subsets, multiple deformation subsets, ZNCC values ​​for each reference subset, and the initial displacement field, specifically including:

[0017] For any control point in the reference image, determine the reference subset corresponding to the control point and multiple preliminary deformation subsets corresponding to the control point in the discontinuous deformation image;

[0018] Based on the reference subsets and multiple preliminary deformed subsets corresponding to the control points, the ZNCC values ​​of the deformed subsets and each reference subset corresponding to the control points are determined using the zero-mean normalized cross-correlation criterion and the first-order shape function.

[0019] The initial displacement field is determined based on the deformation subset corresponding to each control point.

[0020] Optionally, based on the reference subset corresponding to the control point and multiple preliminary deformed subsets, the ZNCC values ​​of the deformed subset corresponding to the control point and each reference subset are determined using the zero-mean normalized cross-correlation criterion and a first-order shape function, specifically including:

[0021] For any initial deformation subset corresponding to the control point in the discontinuous deformation image, the similarity between the reference subset and the initial deformation subset is determined by using the zero-mean normalized cross-correlation criterion based on the gray values ​​of each pixel in the reference subset and the gray values ​​of each pixel in the initial deformation subset.

[0022] determine a high-quality deformation subset corresponding to the control point according to the similarity between the reference subset and each preliminary deformation subset; the similarity between the reference subset and the high-quality deformation subset is a ZNCC value of the reference subset;

[0023] adjust the shape of the high-quality deformation subset by using a first-order shape function to obtain the deformation subset corresponding to the control point.

[0024] Optionally, the similarity between the reference subset and the preliminary deformation subset is determined by using the following formula:

[0025]

[0026] wherein C ZNCC is the similarity between the reference subset and the preliminary deformation subset, N is the number of pixel points in the reference subset, f(x i ,y i ) is the gray value of pixel (x i ,y i ) in the reference subset, g(x' i ,y' i ) is the gray value of pixel (x' i ,y' i ) in the preliminary deformation subset, f m is the average pixel gray value of the reference subset, and g m is the average pixel gray value of the preliminary deformation subset.

[0027] Optionally, a ZNCC distribution map is drawn according to the reference image, each reference subset, the ZNCC value of each reference subset, and the deformation subset, and specifically includes:

[0028] The size of the ZNCC distribution map is initialized according to the reference image to obtain an initial ZNCC distribution map; the size of the initial ZNCC distribution map is the same as the size of the reference image.

[0029] The gray value of each pixel in the initial ZNCC distribution map is determined according to the ZNCC value of each reference subset to obtain the ZNCC distribution map.

[0030] Optionally, a non-continuous deformation region is determined according to the reference image and the non-continuous boundary trajectory, and specifically includes:

[0031] The non-continuous boundary trajectory is set in the reference image; the non-continuous deformation region is a region in the reference image with a distance to the non-continuous boundary trajectory within a set threshold range.

[0032] Optionally, a non-continuous control point subset is determined according to the non-continuous boundary trajectory, the non-continuous deformation region, the predicted displacement field, and the reference image, and specifically includes:

[0033] In the reference image, a shortest distance of a non-continuous control point to the non-continuous boundary trajectory is determined, and a direction vector of the shortest distance is determined, so as to obtain a shortest direction vector; the non-continuous control point corresponding to the shortest direction vector is a target control point; the non-continuous control point is a control point in a non-continuous deformation region;

[0034] The predicted displacement field is fitted as a continuous curved surface;

[0035] According to the shortest direction vector and the continuous curved surface, an initial rectangle of a subset is determined; the edges of the initial rectangle of the subset are perpendicular to or parallel to the shortest direction vector;

[0036] According to the non-continuous boundary trajectory and the target control point, the initial rectangle of the subset is divided into a target rectangle and a non-target rectangle; the target rectangle is a region of the initial rectangle of the subset containing the target control point, and the non-target rectangle is a region of the initial rectangle of the subset not containing the target control point;

[0037] The sum of the square of the gray gradient of the pixels in the target rectangle is calculated;

[0038] It is judged whether the sum of the square of the gray gradient of the pixels in the target rectangle is greater than a set gray threshold value; if yes, the target rectangle is a non-continuous control point subset; otherwise, the edge length of the target rectangle is enlarged in equal proportion, and the enlarged target rectangle is taken as an initial rectangle of a subset for division until the sum of the square of the gray gradient of the pixels in the target rectangle is greater than the set gray threshold value.

[0039] To achieve the above object, the application further provides the following scheme:

[0040] A non-continuous deformation displacement field reconstruction system, comprising:

[0041] An image acquisition module, configured to acquire a reference image and a non-continuous deformation image; the reference image is a speckle image before deformation of an object; the non-continuous deformation image is a speckle image after deformation of the object;

[0042] A control point arrangement module, connected with the image acquisition module, configured to arrange control points in the reference image according to a pre-set region of interest and a subset interval;

[0043] An initial displacement field determination module, connected with the image acquisition module and the control point arrangement module respectively, configured to perform digital speckle correlation calculation according to the control points in the reference image and the non-continuous deformation image, so as to obtain a plurality of reference subsets, a plurality of deformation subsets, a ZNCC value of each reference subset and an initial displacement field;

[0044] The trajectory determination module is connected with the initial displacement field determination module, and is configured to draw a ZNCC distribution map according to the reference image, the ZNCC values of each reference subset and the deformation subset, and determine a non-continuous boundary trajectory by using a peak searching algorithm according to the ZNCC distribution map.

[0045] The deformation region determination module is connected with the trajectory determination module, and is configured to determine a non-continuous deformation region according to the reference image and the non-continuous boundary trajectory.

[0046] The prediction displacement field determination module is connected with the initial displacement field determination module and the deformation region determination module respectively, and is configured to eliminate data in the non-continuous deformation region in the initial displacement field, and fill the blank region according to data in a continuous region to obtain a prediction displacement field, wherein the continuous region is a region in the initial displacement field except the non-continuous deformation region.

[0047] The subset determination module is connected with the trajectory determination module, the deformation region determination module and the prediction displacement field determination module respectively, and is configured to determine a non-continuous control point subset according to the non-continuous boundary trajectory, the non-continuous deformation region, the prediction displacement field and the reference image.

[0048] The displacement field updating module is connected with the prediction displacement field determination module and the subset determination module respectively, and is configured to update the prediction displacement field by performing digital speckle correlation calculation according to the non-continuous control point subset to obtain a non-continuous deformation displacement field.

[0049] To achieve the above object, the present application further provides the following scheme.

[0050] An electronic device includes a memory and a processor, the memory is used for storing a computer program, and the processor runs the computer program to make the electronic device execute the non-continuous deformation displacement field reconstruction method.

[0051] To achieve the above object, the present application further provides the following scheme.

[0052] A computer readable storage medium stores a computer program, and the computer program is executed by a processor to realize the non-continuous deformation displacement field reconstruction method.

[0053] According to the embodiments of the present application, the following technical effects are provided.

[0054] The application firstly performs digital speckle correlation calculation according to the control points in the reference image and the non-continuous deformation image, establishes an initial displacement field, then draws a ZNCC distribution graph, and determines a non-continuous boundary track and a non-continuous deformation region by using a peak searching algorithm, and then eliminates the data in the non-continuous deformation region in the initial displacement field, so that the region corresponding to the new subset completely can be found in the non-continuous deformation image, the data in the continuous region is used to fill the blank region, and a predicted displacement field is obtained; the non-continuous control point subset is determined according to the non-continuous boundary track, the non-continuous deformation region, the predicted displacement field and the reference image; finally, the digital speckle correlation calculation is performed according to the non-continuous control point subset, the predicted displacement field is updated, and a non-continuous deformation displacement field is obtained. The range and size of the subset participating in the digital speckle correlation calculation are adaptively adjusted, and the non-continuous deformation displacement field reconstruction accuracy is improved. BRIEF DESCRIPTION OF DRAWINGS

[0055] In order to more clearly illustrate the technical solutions of the embodiments of the present application or the prior art, the drawings needed in the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0056] Figure 1 The flow chart of the non-continuous deformation displacement field reconstruction method provided by the present application is shown in the figure.

[0057] Figure 2 The schematic diagram of the non-continuous deformation displacement field reconstruction process is shown in the figure.

[0058] Figure 3 The schematic diagram of the reference image is shown in the figure.

[0059] Figure 4 The schematic diagram of the non-continuous deformation image is shown in the figure.

[0060] Figure 5 The schematic diagram of the digital speckle correlation calculation is shown in the figure.

[0061] Figure 6 The schematic diagram of the first-order shape function deformation of the reference subset is shown in the figure.

[0062] Figure 7 The schematic diagram of the reference subset in the digital speckle correlation calculation is shown in the figure.

[0063] Figure 8 The ZNCC distribution graph is shown in the figure.

[0064] Figure 9 The schematic diagram of the non-continuous control points and the non-continuous deformation region is shown in the figure.

[0065] Figure 10 The schematic diagram of the non-continuous control point subset is shown in the figure.

[0066] Figure 11 A schematic diagram of a non-continuous deformation displacement field reconstruction system provided by the present application.

[0067] Symbol explanation:

[0068] 1 - region of interest, 2 - control point, 3 - subset spacing, 4 - reference subset, 5 - deformation subset, 6 - non-continuous control point, 7 - non-continuous deformation region, 8 - shortest direction vector, 9 - non-continuous control point subset, 10 - initial rectangular shape of subset;

[0069] 21 - image acquisition module, 22 - control point arrangement module, 23 - initial displacement field determination module, 24 - trajectory determination module, 25 - deformation region determination module, 26 - predicted displacement field determination module, 27 - subset determination module, 28 - displacement field update module. DETAILED DESCRIPTION

[0070] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0071] The purpose of the present application is to provide a non-continuous deformation displacement field reconstruction method, system, electronic device and medium, which adaptively adjusts the range and size of the subset participating in digital speckle correlation calculation based on non-continuous boundary trajectory, deformation field gradient and image gray gradient intensity, and improves the non-continuous deformation displacement field reconstruction accuracy.

[0072] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the present application will be further described in detail below with reference to the drawings and specific embodiments.

[0073] Embodiment one

[0074] As shown in Figure 1 and Figure 2 , the present embodiment provides a non-continuous deformation displacement field reconstruction method, which comprises:

[0075] Step 100: acquiring reference image and non-continuous deformation image.

[0076] In the present embodiment, an experimental system is built, and speckle images before and after deformation of an object are acquired by a digital camera. The speckle image before deformation of the object is a reference image, and the speckle image after deformation (crack, dislocation, slip) of the object is a non-continuous deformation image. As shown in Figure 3 and Figure 4 .

[0077] Step 200: arranging control points in the reference image according to a preset region of interest and subset interval.

[0078] The region of interest 1 and the subset interval 3 are parameters that are predefined according to application scenarios and requirements. The region of interest 1 is the region for which the deformation field is to be calculated, and is demarcated by the user. The subsequent steps operate on the pixels in the region of interest 1. The determination of the subset interval 3 takes into account both the calculation accuracy and the calculation cost. A smaller subset interval 3 can improve the algorithm accuracy, but also increases the calculation burden. The subset interval 3 is usually set to be between 1 and 10 pixels, and can be changed according to the calculation results and efficiency in actual operation. The control points 2 are arranged uniformly in the region of interest 1 according to the subset interval 3.

[0079] Step 300: performing digital speckle correlation calculation according to the control points in the reference image and the non-continuous deformation image, to obtain a plurality of reference subsets, a plurality of deformation subsets, a ZNCC value of each reference subset, and an initial displacement field.

[0080] The purpose of the digital speckle correlation calculation is to find the deformation subset that is most similar (best matched) to the reference subset. The process of the digital speckle correlation calculation is an iterative optimization process. In the calculation process, the position and shape of the deformation subset are dynamically adjusted until the deformation subset that makes the ZNCC optimal is found.

[0081] Specifically, the number of the reference subsets and the number of the deformation subsets are the same as the number of the control points. Step 300 includes:

[0082] (31) For any control point in the reference image, determining the reference subset corresponding to the control point and a plurality of preliminary deformation subsets corresponding to the control point in the non-continuous deformation image.

[0083] As shown in FIG. 4, the reference subset 4 is a square centered at the control point P, with a side length of (2M+1), and M is the subset radius. The purpose of this step is to search for the corresponding point P' of the control point P in the deformation image, and then determine the deformation subset 5. Figure 5

[0084] (32) According to the reference subset corresponding to the control point and the plurality of preliminary deformation subsets, using the ZNCC criterion and the first-order shape function, determining the deformation subset corresponding to the control point and the ZNCC value of each reference subset.

[0085] ​In this embodiment, according to the gray scale features of the reference subset, the deformed subset is determined according to certain correlation criterion and shape function. The correlation criterion is used to represent the similarity degree of the deformed subset and the reference subset. The shape function is used to change the shape of the deformed subset. The parameters of the shape function are adjusted to change the shape of the deformed subset until it is most similar to the reference subset. The commonly used correlation criterion is ZNCC, and the commonly used shape function is a first-order shape function.

[0086] Specifically, for any preliminary deformed subset corresponding to the control point in the non-continuous deformed image, according to the gray scale values of the pixels in the reference subset and the gray scale values of the pixels in the preliminary deformed subset, the similarity degree of the reference subset and the preliminary deformed subset is determined by using a zero-mean normalized cross-correlation criterion:

[0087]

[0088] wherein C ZNCC is the similarity degree of the reference subset and the preliminary deformed subset, N is the number of the pixels in the reference subset, f(x i ,y i ) is the gray scale value of the pixel (x i ,y i ) in the reference subset, g(x' i ,y' i ) is the gray scale value of the pixel (x' i ,y' i ) in the preliminary deformed subset, f m is the average pixel gray scale value of the reference subset, and g m is the average pixel gray scale value of the preliminary deformed subset.

[0089] According to the similarity degrees of the reference subset and each preliminary deformed subset, the high-quality deformed subset corresponding to the control point is determined. The similarity degree of the reference subset and the high-quality deformed subset is the ZNCC value of the reference subset.

[0090] The shape of the high-quality deformed subset is adjusted by using a first-order shape function to obtain the deformed subset corresponding to the control point. The first-order shape function is:

[0091]

[0092] wherein x' is the adjusted horizontal coordinate, y' is the adjusted vertical coordinate, x is the unadjusted horizontal coordinate, y is the unadjusted vertical coordinate, u x , u y , v x , v y , u, and v are displacement gradients, which can represent translation, stretching, tilting, etc. Figure 6The first-order shape function can well describe the uniform deformation and meet the general deformation measurement requirements.

[0093] In this embodiment, an optimization method such as an inverse compositional Gauss-Newton (IC-GN) algorithm is used to optimize the parameters (u x , u y , v x , v y , u, v) of the first-order shape function, and the deformed subset that best matches the reference subset in the deformed image is found, with the center being the corresponding point P'.

[0094] (33) According to the deformed subset corresponding to each control point, the initial displacement field is determined.

[0095] Step 400: Draw a ZNCC distribution map according to the reference image, the ZNCC values of each reference subset and the deformed subset, and determine a non-continuous boundary trajectory by using a peak searching algorithm according to the ZNCC distribution map.

[0096] The reference subset in the digital speckle correlation calculation is shown in FIG. 4, wherein the reference subset 4 is a square with the control point 2 at the center and a side length of (2M+1) pixels, and M is 3. The reference subset 4 can be expressed as: Figure 7

[0097] S0={(x i ,y j )|i∈[c-M,c+M],j∈[c-M,c+M]};

[0098] wherein S0 is the reference subset, (x c , y c ) represents the pixel coordinates of the control point, and the subscript represents the position relationship of the pixel point in the reference subset with the control point in the pixel coordinate system. When performing the digital speckle correlation calculation, this reference subset 4 is used as a template to perform matching in the deformed image. Through the matching, a deformed subset 5 is obtained. The value of the ZNCC is the matching degree of the deformed subset 5 and the reference subset 4. The higher the value of the ZNCC, the more accurate the matching of the corresponding point of the control point.

[0099] Specifically, in step 400, the size of the ZNCC distribution map is initialized according to the reference image, and an initial ZNCC distribution map is obtained. The size of the initial ZNCC distribution map is the same as the size of the reference image, and the gray value of all pixels in the initial ZNCC distribution map is 0.

[0100] ​Then, according to the ZNCC values of each reference subset, the gray values of each pixel in the initial ZNCC distribution map are determined to obtain a ZNCC distribution map. A peak-seeking algorithm is used to extract a non-continuous boundary trajectory.

[0101] In practical applications, the ZNCC value ranges from 0 to 1, which can be used as the gray value of the image. Since the size of the ZNCC distribution map is the same as that of the reference image, each pixel in each reference subset can find a corresponding position in the ZNCC distribution map. All the pixels in the corresponding area of each reference subset in the ZNCC distribution map are assigned with the corresponding ZNCC value. There is an overlap between the reference subsets, and the average value of the ZNCC values is assigned to the pixel positions in the overlapping area.

[0102] The ZNCC distribution map drawn by the present application is different from the conventional ZNCC matrix. The size of the conventional ZNCC matrix is the same as that of the control point set. The size of the ZNCC distribution map generated by the present application is the same as that of the reference image, and each pixel point that may participate in the calculation is assigned with a pixel value.

[0103] In order to better understand the technical solutions of the present application, a simple example is provided as follows: Figure 8 As shown in the figure, there are four control points 2 in the reference image, the radius of the reference subset is 3, and the subset spacing is 3. After digital speckle correlation calculation, the ZNCC values of the four reference subsets are 0.99, 0.95, 0.93 and 0.89, respectively. Then, the pixels of the ZNCC distribution map are assigned with values. For the overlapping area, the average value of the ZNCC is assigned.

[0104] The ZNCC value of the non-continuous deformation area will be significantly reduced, and the closer to the non-continuous boundary, the lower the ZNCC value. The peak ridge of the ZNCC distribution map is connected, and is mapped to the reference image to obtain a non-continuous boundary trajectory.

[0105] Step 500: determining a non-continuous deformation area according to the reference image and the non-continuous boundary trajectory.

[0106] Specifically, the non-continuous boundary trajectory should be set in the reference image. As shown in the figure, Figure 9 The non-continuous deformation area 7 is the area in the reference image with a distance from the non-continuous boundary trajectory within a set threshold range. The control points in the non-continuous deformation area 7 are non-continuous control points 6.

[0107] Step 600: removing the data in the non-continuous deformation area in the initial displacement field, and filling the blank area according to the data in the continuous area to obtain a predicted displacement field. The continuous area is the area in the initial displacement field except the non-continuous deformation area.

[0108] Specifically, a continuous surface is fitted with data in the continuous region, and the data is gradually filled in from the edge of the blank region. For a point on the edge of the blank region, a number of points in the neighborhood of the point are selected in the region with data. A surface is fitted based on the data points in the neighborhood, and the data of the blank point is filled according to the surface.

[0109] Step 700: determining a non-continuous control point subset according to the non-continuous boundary trajectory, the non-continuous deformation region, the predicted displacement field, and the reference image.

[0110] The non-continuous control point subset contains non-continuous control points, and the initial shape is a rectangle. The part cut by the non-continuous boundary trajectory is removed, and the pixels on the non-continuous boundary trajectory are retained. The aspect ratio of the non-continuous control point subset is determined according to the displacement field gradient ratio at the non-continuous control point. The size of the non-continuous control point subset is determined according to the gray gradient square sum of the reference image.

[0111] Specifically, step 700 includes:

[0112] (71) As shown in the reference image, the shortest distance of the non-continuous control point to the non-continuous boundary trajectory is determined, and the direction vector of the shortest distance is determined to obtain the shortest direction vector 8. The non-continuous control point corresponding to the shortest direction vector 8 is the target control point. Figure 10 (72) The predicted displacement field is fitted as a continuous surface.

[0113] (73) The subset initial rectangle 10 is determined according to the shortest direction vector and the continuous surface. The edges of the subset initial rectangle 10 are perpendicular or parallel to the shortest direction vector 8. The center of the subset initial rectangle 10 is the target control point.

[0114] Specifically, the displacement surface derivative of the shortest direction vector direction 8 and its normal direction is calculated, and the aspect ratio of the subset initial rectangle 10 is determined according to the displacement surface derivative ratio. The aspect ratio is opposite to the gradient ratio:

[0115]

[0116] wherein, L x is the side length of the subset initial rectangle in the x direction, L y is the side length of the subset initial rectangle in the y direction, and C is the continuous surface.

[0117]

[0118] ​(74) According to the non-continuous boundary trajectory and the target control point, the subset initial rectangle 10 is divided into a target rectangle and a non-target rectangle. The target rectangle is a region of the subset initial rectangle 10 containing the target control point, and the non-target rectangle is a region of the subset initial rectangle 10 not containing the target control point.

[0119] This step reduces the range of the subset initial rectangle 10, i.e. removes the pixels on the side of the subset initial rectangle 10 where the non-continuous boundary trajectory does not contain the target control point, so that all the pixels in the target rectangle are on the same side of the non-continuous boundary trajectory. That is, the calculation result of the deformation field near the crack is removed because the deformation field near the crack is inaccurate due to the influence of the crack and needs to be recalculated. The way to recalculate is to use a new subset to match the non-continuous control points again and calculate the displacement of the non-continuous control points.

[0120] (75) Calculate the sum of the square of the gray scale gradient of the pixels in the target rectangle.

[0121] (76) Determine whether the sum of the square of the gray scale gradient of the pixels in the target rectangle is greater than a set gray scale threshold. If yes, the target rectangle is the non-continuous control point subset 9, otherwise, the side length of the target rectangle is enlarged by equal ratio, and the enlarged target rectangle is used as a subset initial rectangle for division until the sum of the square of the gray scale gradient of the pixels in the target rectangle is greater than the set gray scale threshold.

[0122] That is, determine whether the following formula is true or not. Where threshold is the set gray scale threshold, f x and f y are the gradients of the image in the x direction and the y direction at the point (x, y), i.e. f x = f(x+1, y) - f(x, y), f y = f(x, y+1) - f(x, y), and f(x, y) is the gray scale value at (x, y).

[0123] Step 800: According to the non-continuous control point subset, perform digital speckle correlation calculation to update the predicted displacement field to obtain a non-continuous deformation displacement field.

[0124] Specifically, the non-continuous control point subset is used to match the non-continuous control points again, the digital speckle correlation calculation is performed, the predicted deformation field is updated, and an accurate non-continuous deformation displacement field is obtained.

[0125] The subset of the standard digital speckle correlation calculation is a square centered on the control point, and the present application uses an adaptive new subset (non-continuous control point subset). The shape of the non-continuous control point subset is usually irregular, and the rest of the calculation steps are the same as the standard digital speckle correlation calculation steps.

[0126] The application uses the ZNCC distribution diagram and the peak searching algorithm to depict the non-continuous boundary trajectory in the reference image, uses the standard digital speckle correlation algorithm to calculate the deformation displacement field, eliminates the error data of the non-continuous deformation region, i.e. eliminates the part in the initial rectangle of the subset which is cut by the non-continuous boundary trajectory, guarantees that the region corresponding to the new subset completely can be found in the non-continuous deformation image; the data of the blank region is predicted and filled based on the data of the continuous region, and the predicted deformation field is obtained; according to the non-continuous boundary trajectory, the predicted deformation field and the reference image gray gradient, the subset range of the non-continuous control point is adaptively set according to a certain rule, i.e. the length-width ratio of the rectangle is set according to the displacement gradient ratio, the pixels in the direction with large deformation gradient are less, and the pixels in the direction with small deformation gradient are more, which simplifies the deformation process of the subset, improves the adaptation degree of the first-order shape function and the subset deformation, the size of the new subset is determined according to the gray gradient square sum, which guarantees that the new subset has enough characteristics, and further improves the reconstruction accuracy of the non-continuous deformation displacement field.

[0127] Embodiment two

[0128] In order to perform the method corresponding to the above-mentioned embodiment one, to realize the corresponding functions and technical effects, the following provides a non-continuous deformation displacement field reconstruction system.

[0129] As shown in Figure 11 The non-continuous deformation displacement field reconstruction system provided by the embodiment includes: an image acquisition module 21, a control point arrangement module 22, an initial displacement field determination module 23, a trajectory determination module 24, a deformation region determination module 25, a predicted displacement field determination module 26, a subset determination module 27 and a displacement field update module 28.

[0130] The image acquisition module 21 is used to acquire the reference image and the non-continuous deformation image. The reference image is the speckle image before the deformation of the object. The non-continuous deformation image is the speckle image after the deformation of the object.

[0131] The control point arrangement module 22 is connected with the image acquisition module 21, and the control point arrangement module 22 is used to arrange the control points in the reference image according to the pre-set region of interest and the subset interval.

[0132] The initial displacement field determination module 23 is connected with the image acquisition module 21 and the control point arrangement module 22 respectively, and the initial displacement field determination module 23 is used to perform digital speckle correlation calculation according to the control points in the reference image and the non-continuous deformation image, to obtain a plurality of reference subsets, a plurality of deformation subsets, the ZNCC values of each reference subset and the initial displacement field.

[0133] The trajectory determination module 24 is connected with the initial displacement field determination module 23, and the trajectory determination module 24 is configured to draw a ZNCC distribution map according to the reference image, the ZNCC values of each reference subset and the deformation subset, and determine a non-continuous boundary trajectory by using a peak searching algorithm according to the ZNCC distribution map.

[0134] The deformation region determination module 25 is connected with the trajectory determination module 24, and the deformation region determination module 25 is configured to determine a non-continuous deformation region according to the reference image and the non-continuous boundary trajectory.

[0135] The predicted displacement field determination module 26 is connected with the initial displacement field determination module 23 and the deformation region determination module 25 respectively, and the predicted displacement field determination module 26 is configured to remove data in the non-continuous deformation region in the initial displacement field, and fill the blank region according to data in a continuous region to obtain a predicted displacement field, wherein the continuous region is a region in the initial displacement field except the non-continuous deformation region.

[0136] The subset determination module 27 is connected with the trajectory determination module 24, the deformation region determination module 25 and the predicted displacement field determination module 26 respectively, and the subset determination module 27 is configured to determine a non-continuous control point subset according to the non-continuous boundary trajectory, the non-continuous deformation region, the predicted displacement field and the reference image.

[0137] The displacement field updating module 28 is connected with the predicted displacement field determination module 26 and the subset determination module 27 respectively, and the displacement field updating module 28 is configured to perform digital speckle correlation calculation according to the non-continuous control point subset, update the predicted displacement field and obtain a non-continuous deformation displacement field.

[0138] Compared with the prior art, the non-continuous deformation displacement field reconstruction system provided by the embodiment has the same beneficial effects as the non-continuous deformation displacement field reconstruction method provided by the first embodiment, and will not be described here.

[0139] Embodiment three

[0140] The embodiment provides an electronic device, including a memory and a processor, the memory is used for storing a computer program, and the processor runs the computer program to make the electronic device execute the non-continuous deformation displacement field reconstruction method of the first embodiment.

[0141] Optionally, the electronic device can be a server.

[0142] In addition, the embodiment of the application further provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the non-continuous deformation displacement field reconstruction method of the first embodiment.

[0143] The various embodiments described in this specification are presented for the purpose of illustrating the principles of the present application and its best mode of operation. Each of the embodiments described in this specification has been provided for the purpose of illustration and is not intended to limit the application.

[0144] The principles and implementations of the present application have been described in the specification with specific examples. The above description of the embodiments is only for the purpose of helping to understand the method of the present application and its core idea. Meanwhile, for those skilled in the art, the specific implementation and application range of the present application can be changed according to the idea of the present application. In summary, the content of the specification should not be understood as a limitation of the present application.

Claims

1. A method for reconstructing a discontinuous deformation displacement field, characterized in that, The non-continuous deformation displacement field reconstruction method comprises: collecting a reference image and a non-continuous deformation image; the reference image is a speckle image before deformation of an object; the non-continuous deformation image is a speckle image after deformation of the object; arranging control points in the reference image according to a preset region of interest and a preset subset interval; performing digital speckle correlation calculation according to the control points in the reference image and the non-continuous deformation image to obtain a plurality of reference subsets, a plurality of deformation subsets, a ZNCC value of each reference subset and an initial displacement field; drawing a ZNCC distribution map according to the reference image, the ZNCC value of each reference subset and the deformation subsets, and determining a non-continuous boundary track by using a peak searching algorithm according to the ZNCC distribution map; determining a non-continuous deformation region according to the reference image and the non-continuous boundary track; removing data in the non-continuous deformation region in the initial displacement field, and filling a blank region according to data in a continuous region to obtain a predicted displacement field; the continuous region is a region other than the non-continuous deformation region in the initial displacement field; determining a non-continuous control point subset according to the non-continuous boundary track, the non-continuous deformation region, the predicted displacement field and the reference image; performing digital speckle correlation calculation according to the non-continuous control point subset to update the predicted displacement field and obtain a non-continuous deformation displacement field.

2. The method of claim 1, wherein, The number of the reference subsets and the number of the deformation subsets are the same as the number of the control points; performing digital speckle correlation calculation according to the control points in the reference image and the non-continuous deformation image to obtain a plurality of reference subsets, a plurality of deformation subsets, a ZNCC value of each reference subset and an initial displacement field, specifically comprising: for any control point in the reference image, determining a reference subset corresponding to the control point and a plurality of preliminary deformation subsets corresponding to the control point in the non-continuous deformation image; determining a deformation subset corresponding to the control point and a ZNCC value of each reference subset by using a zero-mean normalized cross-correlation criterion and a first-order shape function according to the reference subset corresponding to the control point and the plurality of preliminary deformation subsets; determining an initial displacement field according to the deformation subset corresponding to each control point.

3. The method of claim 2, wherein, determining a deformation subset corresponding to the control point and a ZNCC value of each reference subset by using a zero-mean normalized cross-correlation criterion and a first-order shape function according to the reference subset corresponding to the control point and the plurality of preliminary deformation subsets, specifically comprising: for any preliminary deformation subset corresponding to the control point in the non-continuous deformation image, determining a similarity between the reference subset and the preliminary deformation subset by using a zero-mean normalized cross-correlation criterion according to a gray value of each pixel in the reference subset and a gray value of each pixel in the preliminary deformation subset; determining a high-quality deformation subset corresponding to the control point according to the similarity between the reference subset and each preliminary deformation subset; the similarity between the reference subset and the high-quality deformation subset is the ZNCC value of the reference subset; adjusting a shape of the high-quality deformation subset by using a first-order shape function to obtain the deformation subset corresponding to the control point.

4. The method of claim 3, wherein, The similarity between the reference subsets and the preliminary deformation subsets is determined by using the following formula: Wherein, C ZNCC is the similarity degree between the reference subset and the preliminary deformation subset, N is the number of pixel points in the reference subset, f(x i ,y i ) is the gray value of the pixel (x i ,y i ) in the reference subset, g(x' i ,y' i ) is the gray value of the pixel (x' i ,y' i ) in the preliminary deformation subset, f m is the average pixel gray value of the reference subset, and g m is the average pixel gray value of the preliminary deformation subset.

5. The method of claim 1, wherein, A ZNCC distribution map is drawn according to the reference image, the reference subsets, the ZNCC values of the reference subsets and the deformation subset, and specifically includes: The size of the ZNCC distribution map is initialized according to the reference image, and an initial ZNCC distribution map is obtained; the size of the initial ZNCC distribution map is the same as the size of the reference image; The gray values of the pixels in the initial ZNCC distribution map are determined according to the ZNCC values of the reference subsets, and a ZNCC distribution map is obtained.

6. The method of claim 1, wherein, A non-continuous deformation region is determined according to the reference image and the non-continuous boundary trajectory, and specifically includes: The non-continuous boundary trajectory is set in the reference image; the non-continuous deformation region is a region in the reference image with a distance from the non-continuous boundary trajectory within a set threshold range.

7. The method of claim 1, wherein, A non-continuous control point subset is determined according to the non-continuous boundary trajectory, the non-continuous deformation region, the predicted displacement field and the reference image, and specifically includes: In the reference image, the shortest distance of the non-continuous control point to the non-continuous boundary trajectory is determined, and the direction vector of the shortest distance is determined, to obtain a shortest direction vector; the non-continuous control point corresponding to the shortest direction vector is a target control point; the non-continuous control point is a control point in the non-continuous deformation region; The predicted displacement field is fitted as a continuous surface; A subset initial rectangle is determined according to the shortest direction vector and the continuous surface; the edges of the subset initial rectangle are perpendicular or parallel to the shortest direction vector; The subset initial rectangle is divided into a target rectangle and a non-target rectangle according to the non-continuous boundary trajectory and the target control point; the target rectangle is a region in the subset initial rectangle containing the target control point, and the non-target rectangle is a region in the subset initial rectangle not containing the target control point; The sum of the square of the gray gradient of the pixels in the target rectangle is calculated; It is judged whether the sum of the square of the gray gradient of the pixels in the target rectangle is greater than a set gray threshold, if yes, the target rectangle is a non-continuous control point subset, otherwise, the length of the edge of the target rectangle is enlarged by equal ratio, and the enlarged target rectangle is taken as a subset initial rectangle for division until the sum of the square of the gray gradient of the pixels in the target rectangle is greater than the set gray threshold.

8. A non-continuously deformed displacement field reconstruction system, characterized by, The non-continuous deformation displacement field reconstruction system includes: An image acquisition module is configured to acquire a reference image and a non-continuous deformation image; the reference image is a speckle image before deformation of an object; the non-continuous deformation image is a speckle image after deformation of the object; A control point arrangement module is connected with the image acquisition module and configured to arrange control points in the reference image according to a pre-set region of interest and a subset spacing; An initial displacement field determination module is connected with the image acquisition module and the control point arrangement module and configured to perform digital speckle correlation calculation according to the control points in the reference image and the non-continuous deformation image to obtain a plurality of reference subsets, a plurality of deformation subsets, ZNCC values of the reference subsets and an initial displacement field; The trajectory determination module is connected with the initial displacement field determination module, and is configured to draw a ZNCC distribution map according to the reference image, the ZNCC values of each reference subset and the deformation subset, and determine a non-continuous boundary trajectory by using a peak searching algorithm according to the ZNCC distribution map; The deformation region determination module is connected with the trajectory determination module, and is configured to determine a non-continuous deformation region according to the reference image and the non-continuous boundary trajectory; The prediction displacement field determination module is connected with the initial displacement field determination module and the deformation region determination module respectively, and is configured to eliminate data in the non-continuous deformation region in the initial displacement field, and fill the blank region according to data in a continuous region to obtain a prediction displacement field; the continuous region is a region in the initial displacement field except the non-continuous deformation region; The subset determination module is connected with the trajectory determination module, the deformation region determination module and the prediction displacement field determination module respectively, and is configured to determine a non-continuous control point subset according to the non-continuous boundary trajectory, the non-continuous deformation region, the prediction displacement field and the reference image; The displacement field updating module is connected with the prediction displacement field determination module and the subset determination module respectively, and is configured to update the prediction displacement field by performing digital speckle correlation calculation according to the non-continuous control point subset to obtain a non-continuous deformation displacement field.

9. An electronic device, comprising: The electronic device comprises a memory and a processor, the memory is configured to store a computer program, and the processor is configured to run the computer program to enable the electronic device to perform the non-continuous deformation displacement field reconstruction method in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer program is stored in the memory and is executed by the processor to implement the non-continuous deformation displacement field reconstruction method in any one of claims 1 to 7.

Images