CN115100100A - Phase-dependent displacement field acquisition method, electronic device and storage medium - Google Patents

Abstract

The invention discloses a phase correlation displacement field acquisition method, electronic equipment and storage medium. The acquisition method includes reading images before and after deformation, determining the maximum value and the minimum value of an initial sliding window size; taking the maximum value as the initial sliding window size , and use the initial sliding window to divide the image before deformation to obtain the first area window; calculate the variance of the pixels in the first area window, form the variance field matrix from the variance, and normalize the variance field matrix; The optimized sliding window size is calculated from the variance field matrix, the maximum value and the minimum value; the optimized sliding window is used to divide the first image and the second image respectively to obtain the second area window and the third area window; The area window is phase-correlated with the third area window to obtain the displacement of the corresponding pixel position, and the displacement field is formed by the displacement. The displacement field of the invention has high acquisition precision and high efficiency.

Description

Phase-dependent displacement field acquisition method, electronic device and storage medium

technical field

The invention belongs to the structural deformation monitoring technology based on photogrammetry, and in particular relates to a phase correlation fast displacement field acquisition method for image adaptive area division.

Background technique

When using the phase correlation method to obtain the target displacement field, the usual practice is to determine the size of the sliding window according to the possible displacement in the scene, and use a fixed-size sliding window to measure the image (such as the deformation field or displacement of the material structure under compression and tension). Field) is divided, phase correlation is carried out with a fixed-size sliding window, and finally the displacement field is obtained, such as ecompte D, Smits A, Bossuyt S, et al. Quality assessment of speckle patterns for digital image correlation[J].Optics and lasers in Engineering , 2006, 44(11): 1132-1145. and Huang J, Pan X, Peng X, et al. Digital image correlation with self-adaptive Gaussian windows[J]. Experimental Mechanics, 2013, 53(3): 505- 512. When the size of the sliding window is too small, the divided area will be de-correlated due to insufficient texture information, and accurate displacement results cannot be obtained; when the size of the sliding window is too large, the time consumption will be greatly increased. Reduce displacement field acquisition efficiency.

Therefore, it is very necessary to use a sliding window of dynamic size for phase correlation and obtain the displacement field. In fact, some scholars have considered using a sliding window of dynamic size to obtain the displacement field, see Liu, Xiao-Yong, et al. "ASelf-Adaptive Selection of Subset Size Method in Digital ImageCorrelation Based on Shannon Entropy." IEEE Access 8 (2020): 184822-184833., in which the size of the sliding window is first determined, and the size of the sliding window is continuously increased when dividing a certain image until the Shannon entropy in the sliding window is greater than a set threshold T. This method will loop many times at each pixel position, resulting in a lot of time consumption and low efficiency.

In general, the existing displacement field estimation methods based on phase correlation have the problem that the size of the image area is not easily determined. If the area is divided too large, the efficiency is not high, and the area is too small, resulting in inaccurate results.

SUMMARY OF THE INVENTION

The purpose of the present invention is to provide a phase-dependent displacement field acquisition method, electronic device and storage medium, so as to solve the problem that the influence area is too large to cause low efficiency, and that the division is too small to cause inaccurate displacement field acquisition results.

The present invention solves the above-mentioned technical problems through the following technical solutions: a method for obtaining a phase-dependent displacement field, comprising the following steps:

Read the first image before deformation and the second image after deformation, and determine the maximum value w _max and the minimum value w _min of the initial sliding window size;

Taking the maximum value w _max as the size of the initial sliding window, and using the initial sliding window to perform region division at each pixel position of the first image, N first region windows are obtained;

Calculate the variance of all pixels in each of the first area windows, form a variance field matrix from all the variances, and normalize the variance field matrix;

Calculate the optimal sliding window size corresponding to the pixel position according to the normalized variance field matrix, the maximum value w _max and the minimum value w _min ;

The optimized sliding window is used to perform area division at the corresponding pixel position of the first image to obtain N second area windows, and the optimized sliding window is used to perform area division at the corresponding pixel position of the second image to obtain N a third area window;

Phase correlation is performed between the corresponding second area window and the third area window to obtain the displacement of the corresponding pixel position, and a displacement field is formed by all the displacements.

Further, the specific implementation process of using the initial sliding window to perform area division at each pixel position of the first image is as follows: taking the current pixel position as the center, and taking the area intercepted in the first image according to the size of the initial sliding window as The current first area window.

Further, the formula for calculating the variance of all pixels in each of the first area windows is:

为当前第一区域窗口内所有像素点的像素值的均值。Among them, σ is the variance of all pixels in the current first area window, w is the side length of the current first area window, w=w _max , g(i, j) is the pixel position (i, j) in the current first area window The pixel value at j), (i,j) is the position of the pixel in the current first area window,

is the average of the pixel values of all pixels in the current first region window.

Further, the normalized variance field matrix is:

Among them, M ^* _σ is the normalized variance field matrix, σ ^* _sr is the element of the sth row and rth column in the variance field matrix M ^* _σ , and σ _sr is the unnormalized variance field matrix M _σ . The element in row s and column r, σ _min is the smallest element in the variance field matrix M _σ , σ _max is the largest element in the variance field matrix M _σ , m is the height of the first image, n is the width of the first image, m×n=N.

Further, the calculation formula of the optimized sliding window size of each pixel position is:

w _sr =w _min +(log ₂ [σ ^* _sr +1])(w _max -w _min )

Among them, w _sr is the optimal sliding window size at the pixel position (s, r), (s, r) is the position of the pixel in the first image or the second image, s=1,2,...,m,r =1,2,...,n, m is the height of the first image or the second image, n is the width of the first image or the second image, m×n=N, σ ^* _sr is the variance after normalization processing The element in the sth row and rth column of the field matrix M ^* _σ .

The present invention also provides an electronic device, comprising a memory and a processor, the memory stores a computer program that can run on the processor, and the processor executes the above-mentioned phase-related displacement when the computer program runs. Field acquisition method steps.

The present invention also provides a computer-readable storage medium, which is a non-volatile storage medium or a non-transitory storage medium, on which a computer program is stored, and the computer program is executed when a processor runs The steps of the phase-dependent displacement field acquisition method are described above.

beneficial effect

Compared with the prior art, the advantages of the present invention are:

In the method, electronic device and storage medium provided by the present invention, the size of the optimized sliding window is determined according to the variance of the pixel points in the initial sliding window, the maximum value and the minimum value of the initial sliding window size, and then using The optimized sliding window divides the images before and after deformation, which can effectively reduce the possibility of phase correlation loss and improve the accuracy of displacement field acquisition. Then, the area is divided to avoid a lot of time consumption caused by multiple cycles at each pixel position to determine the size of the sliding window, and to improve the acquisition efficiency of the displacement field;

Compared with the traditional method, the present invention has good adaptability and can be applied to more scenarios.

Description of drawings

In order to illustrate the technical solutions of the present invention more clearly, the following briefly introduces the accompanying drawings used in the description of the embodiments. Obviously, the accompanying drawings in the following description are only an embodiment of the present invention, which is very important in the art. For those of ordinary skill, other drawings can also be obtained from these drawings without any creative effort.

1 is a flowchart of a method for obtaining a phase-dependent displacement field in an embodiment of the present invention;

2 is a first image and a second image in an embodiment of the present invention, wherein (a) is the first image, and (b) is the first image;

3 is a schematic diagram of an initial sliding window (the size of the sliding window is fixed as w _max ) for region division in an embodiment of the present invention;

Fig. 4 is the variance field matrix M ^* _σ after normalization processing in the embodiment of the present invention;

Fig. 5 is the schematic diagram that adopts dynamic optimized sliding window to carry out area division in the embodiment of the present invention;

6 is the displacement field obtained by adopting different area window division methods in the embodiment of the present invention, wherein (a) is the displacement field obtained by adopting the minimum value w _min , (b) is the displacement field obtained by adopting the method of the present invention, (c) ) is the displacement field obtained using the maximum value w _max .

Detailed ways

The technical solutions in the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the protection scope of the present invention.

The technical solutions of the present application will be described in detail below with specific examples. The following specific embodiments may be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments.

As shown in FIG. 1 , a method for obtaining a phase-dependent displacement field provided by this embodiment includes the following steps:

Step 1: Read the first image before deformation and the second image after deformation, and determine the maximum value w _max and the minimum value w _min of the initial sliding window size.

First take the image of the material structure before deformation to obtain the first image, then use a hydraulic press or other equipment to squeeze or stretch the material, and then use a camera to shoot the extrusion or stretching process to obtain multiple images during the deformation process of the material , and select one image from the multiple images in the deformation process as the second image as needed. In this embodiment, an image of the target before deformation is captured as the first image, and then heated by an electric furnace to deform the target, thereby simulating a displacement field to obtain a second image, as shown in FIG. 2 . In Figure 2, the width of the first image and the second image are both 960 pixels, that is, n=960 pixels, the height is 424 pixels, that is, m= 424 pixels, the shooting distance is about 10m, and the color mode is full color.

The maximum value w _max and the minimum value w _min of the initial sliding window size are determined empirically. In this embodiment, w _max is set to 95 pixels, and w _min is set to 45 pixels. The sliding window is a square, and the maximum value w _max of the sliding window size is 95 pixels, which means that the width and height of the sliding window are both 95 pixels.

Step 2: Taking the maximum value w _max as the size of the initial sliding window, and using the initial sliding window to perform region division at each pixel position of the first image, N first region windows are obtained.

The specific implementation process of using the initial sliding window to divide the area at each pixel position of the first image is as follows: taking the current pixel position as the center, and taking the area intercepted according to the size of the initial sliding window as the current first area window. Take Fig. 3 as an example, Fig. 3 shows a part of the first image, each grid □ represents a pixel position, that is, a pixel point. ) as the center, and the area intercepted according to the initial sliding window size (for the convenience of drawing, the initial sliding window size is 11 pixels) is used as the first area window of the pixel position (that is, the area framed by the thick black line in Figure 3) . Therefore, each pixel position corresponds to a first region window, thereby obtaining N first region windows, N=m×n=424×960.

Step 3: Calculate the variance of all pixel points in each first region window, form a variance field matrix by all variances, and normalize the variance field matrix.

The formula for calculating the variance of all pixels in each first region window is:

为当前第一区域窗口内所有像素点的像素值的均值。由方差构成的方差场矩阵为：Among them, σ is the variance of all pixels in the current first area window, w is the side length of the current first area window, w=w _max =95, g(i,j) is the pixel position in the current first area window ( The pixel value at i,j), the pixel position (i,j) refers to the position of the pixel in the current first area window, i=1,2,...,95, j=1,2,...,95,

is the average of the pixel values of all pixels in the current first region window. The variance field matrix composed of variance is:

Each element σ _sr in the variance field matrix M _σ has a one-to-one correspondence with the pixel position (s, r) in the first image. In order to facilitate the subsequent calculation of the optimal sliding window size to map to [w _min , w _max ], the variance field matrix M _σ is normalized, and the specific formula is:

Among them, M ^* _σ is the normalized variance field matrix, σ ^* _sr is the element of the sth row and rth column in the variance field matrix M ^* _σ , and σ _sr is the sth row and rth column of the variance field matrix M _σ . element, σ _min is the smallest element in the variance field matrix M _σ , and σ _max is the largest element in the variance field matrix M _σ . Each element σ ^* _sr in the variance field matrix M ^* _σ has a one-to-one correspondence with the pixel position (s, r) in the first image. The normalized variance field matrix M ^* _σ is shown in Figure 4.

Step 4: Calculate the optimized sliding window size corresponding to the pixel position according to the normalized variance field matrix, the maximum value w _max and the minimum value w _min . The specific calculation formula is:

w _sr =w _min +(log ₂ [σ ^* _sr +1])(w _max -w _min )

Among them, w _sr is the optimal sliding window size at the pixel position (s, r), (s, r) is the position of the pixel in the first image or the second image, s=1,2,...,m,r =1,2,...,n. The optimized sliding window matrix is thus obtained as:

Each element w _sr in the optimized sliding window matrix M _w corresponds to the pixel position (s, r) in the first image or the second image, that is, each pixel position (s, r) has its corresponding optimization Sliding window size.

Step 5: Use the optimized sliding window calculated in step 4 to perform area division at the corresponding pixel position of the first image, to obtain N second area windows, and use the optimized sliding window calculated in step 4 to perform the corresponding pixel position of the second image. The area is divided to obtain N third area windows.

Exemplarily, with the pixel position (1, 1) as the center, the areas captured in the first image and the second image according to the optimized sliding window size w ₁₁ are used as the first second area window and the first third area, respectively. Window: Taking the pixel position (s, r) as the center, the regions intercepted in the first image and the second image according to the optimized sliding window size w _sr are respectively used as the second region window and the third region window. As shown in Fig. 5, the optimized sliding window sizes corresponding to different pixel positions are different, that is, the obtained regional window sizes are different.

For each pixel position of the first image or the second image, the size of the optimized sliding window used is not fixed, but each pixel position corresponds to an optimized sliding window, and the size of the optimized sliding window is suitable for the pixel position. It not only ensures that the texture information in each region window is sufficient, avoids the loss of correlation, improves the accuracy of displacement field acquisition, but also avoids the increase in time consumption caused by the excessive size of the sliding window corresponding to different pixel positions. At the same time, the present invention first determines the optimal sliding window size of each pixel position before dividing the first image and the second image, and then divides the area at the pixel position according to the determined optimal sliding window size to avoid In order to determine the sliding window size of a certain pixel position, the time consumption is huge and the efficiency is low due to multiple loops at the pixel position.

Step 6: Perform phase correlation between the corresponding second area window and the third area window to obtain the displacement of the corresponding pixel position, and form a displacement field by all the displacements. The specific expression of the displacement field is:

为x方向位移场，

为y方向位移场，x_sr为像素位置(s,r)处在x方向的位移，y_sr为像素位置(s,r)处在y方向的位移。第二区域窗口与第一影像的像素位置一一对应，第三区域窗口与第二影像的像素位置一一对应，因此，第二区域窗口与第三区域窗口一一对应，例如第一个第二区域窗口与第一个第三区域窗口一一对应，将对应的第二区域窗口与第三区域窗口进行相位相关，即可得到当前像素位置的位移。in,

is the displacement field in the x direction,

is the displacement field in the y direction, x _sr is the displacement of the pixel position (s, r) in the x direction, and y _sr is the displacement of the pixel position (s, r) in the y direction. The second area window has a one-to-one correspondence with the pixel positions of the first image, and the third area window has a one-to-one correspondence with the pixel positions of the second image. Therefore, the second area window has a one-to-one correspondence with the third area window. The second area window is in one-to-one correspondence with the first third area window, and the phase correlation between the corresponding second area window and the third area window is performed to obtain the displacement of the current pixel position.

Figure 6 shows the displacement field obtained by the method of the present invention and the displacement field obtained by using the maximum value w _max and the minimum value w _min . Table 1 shows the displacement field obtained by the method of the present invention and the maximum value w _max , The time used for the displacement field obtained by the minimum value w _min . It can be seen from Fig. 6 and Table 1 that the time consumption of the present invention is greatly reduced when the displacement field with the same accuracy obtained by using the maximum value w _max is obtained; meanwhile, the accuracy of the displacement field obtained by the minimum value w _min is poor.

Table 1. Different sliding window sizes to obtain the time used by the displacement field

The above disclosure is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited to this. should be included within the protection scope of the present invention.

Claims (7)

1. a phase correlation displacement field acquisition method, is characterized in that, comprises the following steps: Read the first image before deformation and the second image after deformation, and determine the maximum value w _max and the minimum value w _min of the initial sliding window size; Taking the maximum value w _max as the size of the initial sliding window, and using the initial sliding window to perform region division at each pixel position of the first image, N first region windows are obtained; Calculate the variance of all pixels in each of the first area windows, form a variance field matrix from all the variances, and normalize the variance field matrix; Calculate the optimal sliding window size corresponding to the pixel position according to the normalized variance field matrix, the maximum value w _max and the minimum value w _min ; The optimized sliding window is used to perform area division at the corresponding pixel position of the first image to obtain N second area windows, and the optimized sliding window is used to perform area division at the corresponding pixel position of the second image to obtain N a third area window; Phase correlation is performed between the corresponding second area window and the third area window to obtain the displacement of the corresponding pixel position, and a displacement field is formed by all the displacements. 2. The method for obtaining a phase-dependent displacement field according to claim 1, wherein the specific implementation process of using the initial sliding window to perform region division at each pixel position of the first image is: using the current pixel position as the center, and the area captured in the first image according to the size of the initial sliding window is used as the current first area window. 3. The phase correlation displacement field acquisition method as claimed in claim 1, wherein the formula for calculating the variance of all pixel points in each of the first area windows is:

Among them, σ is the variance of all pixels in the current first area window, w is the side length of the current first area window, w=w _max , g(i, j) is the pixel position (i, j) in the current first area window The pixel value at j), (i,j) is the position of the pixel in the current first area window,

is the average of the pixel values of all pixels in the current first region window. 4. The method for obtaining a phase-dependent displacement field as claimed in claim 1, wherein the normalized variance field matrix is:

Among them, M ^* _σ is the normalized variance field matrix, σ ^* _sr is the element of the sth row and rth column in the variance field matrix M ^* _σ , and σ _sr is the unnormalized variance field matrix M _σ . The element in row s and column r, σ _min is the smallest element in the variance field matrix M _σ , σ _max is the largest element in the variance field matrix M _σ , m is the height of the first image, n is the width of the first image, m×n=N. 5. The method for obtaining a phase-dependent displacement field according to any one of claims 1 to 4, wherein the calculation formula of the optimized sliding window size for each pixel position is: w _sr =w _min +(log ₂ [σ ^* _sr +1])(w _max -w _min ) Among them, w _sr is the optimal sliding window size at the pixel position (s, r), (s, r) is the position of the pixel in the first image or the second image, s=1,2,...,m,r =1,2,...,n, m is the height of the first image or the second image, n is the width of the first image or the second image, m×n=N, σ ^* _sr is the variance after normalization processing The element in the sth row and rth column of the field matrix M ^* _σ . 6. An electronic device, comprising a memory and a processor, wherein the memory stores a computer program that can be run on the processor, wherein the processor executes claims 1 to 1 when the processor runs the computer program. Steps of the method for obtaining a phase-dependent displacement field according to any one of 5. 7. A computer-readable storage medium, the computer-readable storage medium being a non-volatile storage medium or a non-transitory storage medium, on which a computer program is stored, characterized in that: the computer program is run by a processor When executing the steps of the method for obtaining the phase-dependent displacement field according to any one of claims 1 to 5.

Images