Ultra-precision interlayer displacement measurement method

Abstract

The invention discloses an ultra-precision interlayer displacement measurement method, and relates to the technical field of structural deformation measurement. The method comprises the following steps: installing a high-frame-rate high-definition camera, and arranging a black and white stripe target; recording structure motion video data through a camera; carrying out parameter calibration, and calculating an amplification coefficient of a pixel-physical size; selecting a region of interest, constructing a space-time matrix, extracting a main motion mode by using singular value decomposition, and reconstructing the matrix to eliminate noise interference; and calculating pixel displacement by using an LK optical flow method, converting the pixel displacement into real displacement, and calculating interlayer displacement. Through multi-mode image fusion, singular value decomposition and an optical flow method, the measurement precision and reliability are improved, the method is suitable for low-visibility or complex environments, sub-pixel-level displacement detection is achieved, and an efficient and economical technical means is provided for building structure health monitoring.

Description

Technical Field

[0001] The invention belongs to the technical field of structural deformation measurement, and in particular relates to an ultra-precision interlayer displacement measurement method. Background Art

[0002] As building structures become increasingly complex and safety requirements continue to increase, accurate measurement of inter-story displacement, as an important indicator for evaluating the performance of building structures under earthquakes or other dynamic loads, becomes particularly important. Traditional inter-story displacement measurement methods mainly include contact sensors (such as laser displacement sensors, accelerometers, etc.) and non-contact optical measurement technology. However, these methods have the following shortcomings: Limitations of contact sensors: Contact sensors need to be mounted directly on the structure, which may interfere with the structure itself and are difficult to deploy in complex or dangerous environments. In addition, the wiring is complex and costly when measuring multiple points.

[0003] Challenges of traditional optical measurement: Although non-contact optical measurement avoids physical contact, the measurement accuracy will be significantly reduced in low-visibility environments (such as at night or in smoky conditions) or in scenes with complex lighting. In addition, traditional methods usually rely on a single imaging mode and cannot effectively cope with a variety of working conditions.

[0004] Data processing and noise issues: In actual measurements, the tiny displacement caused by structural vibration is often drowned out by background noise, resulting in inaccurate extracted motion features. Existing data analysis methods still have deficiencies in noise reduction and feature extraction, making it difficult to meet high-precision measurement requirements.

[0005] Real-time and economic constraints: Although high-precision measurement equipment (such as laser interferometers) can provide high measurement accuracy, they are expensive and complex to operate, making them unsuitable for large-scale application. Low-cost equipment is often unable to meet engineering needs due to insufficient accuracy.

[0006] In order to solve the above problems, a method that can achieve high-precision, non-contact interlayer displacement measurement in complex environments is urgently needed. The method should have good anti-interference ability, strong adaptability, and be able to extract micro-motion features from complex backgrounds through advanced image processing algorithms, while ensuring the reliability and economy of the measurement results. Based on this background, the present invention proposes an ultra-precision interlayer displacement measurement method, which aims to overcome the shortcomings of the existing technology and provide a new solution for building structure health monitoring. Summary of the invention

[0007] In order to overcome the shortcomings and deficiencies of the above-mentioned prior art, the present invention adopts the following technical solutions: An ultra-precision interlayer displacement measurement method, the process is as follows: Step 1: Equipment installation and preparation: Select a high-frame-rate HD camera, combine infrared imaging and ordinary optical imaging technology, use multi-mode image fusion measurement technology, install the camera and arrange black and white stripe targets; Step 2: Image acquisition: In actual construction sites or indoor vibration table tests, use a camera to record the movement of the structure under external excitation and save it as video data; Step 3: Parameter calibration: Calculate the pixel-physical size magnification factor k through an object with known geometric length; Step 4: Motion feature extraction and processing: Select the region of interest, construct the space-time matrix and perform singular value decomposition, extract the main motion mode, and reconstruct the matrix to eliminate noise interference; Step 5: Displacement calculation: Use the LK optical flow method to calculate pixel displacement, convert it into real displacement and calculate inter-layer displacement; Step 6: Quantitative model verification: Evaluate the accuracy of the measurement system through theoretical formulas and compare and verify with actual measurement results; Step 7: Example verification: Through laboratory vibration table testing, the measurement results of the method of the present invention are compared with those of the laser displacement sensor to verify the accuracy and effectiveness of the method.

[0008] Preferably, in step one, the camera installation needs to ensure that the shooting range completely covers the target area, and the design and arrangement of the black and white stripe target needs to ensure that the black and white overlap line is parallel to the movement direction.

[0009] Preferably, in step 2, image acquisition includes video recording of the actual construction site and video recording of the indoor vibration table test to provide image data for subsequent analysis.

[0010] Preferably, in step three, the parameter calibration calculates the pixel-physical size magnification factor k through the formula k=L / N, where L is the known geometric length and N is the number of pixels of the object in the video.

[0011] Preferably, in step 4, motion feature extraction and processing includes: Select the region of interest and construct the space-time matrix , whose dimensions are ;in, Indicates the number of spatial sampling points; Indicates the number of time frames; space-time matrix Each column in represents the spatial intensity distribution at a certain moment; each row represents the time variation of a certain spatial point; The spatial intensity is obtained by reading the RGB of the spatial position of interest in the video and then converting the RGB into grayscale.

[0012] Perform singular value decomposition on the space-time matrix to extract the main motion modes, specifically: ;in, represents the left singular matrix, containing the main spatial features of the space-time matrix, with dimension ; represents the singular value matrix, with dimension , the diagonal elements are singular values, indicating the importance of the features; represents the right singular matrix, containing the main dynamic modes in time, with dimension ; Sort by singular value size and select the top singular values ​​corresponding to the larger singular values mode, Less than and ; After retaining the main motion features, a simplified decomposition is obtained: ;in, Indicates that it contains The matrix of the main spatial modes has the dimension ; Indicates that it contains A diagonal matrix of singular values, with dimension ; Indicates that it contains The matrix of time patterns is ; Through weighted processing and matrix reconstruction, background motion and random noise interference are eliminated. Specifically, the reserved The main characteristic patterns reconstruct the original space-time matrix: ; The reconstructed matrix is the original matrix It is an approximate version of , which retains the most important motion information; at the same time, it removes low-energy noise or irrelevant features.

[0013] Preferably, in step 5, the displacement calculation includes: The LK optical flow method is used to calculate the pixel displacement. Specifically, by Taylor expansion, the light intensity is kept constant and the following is obtained: ;in, , are the spatial gradients of the image grayscale values ​​(derivatives along x and y); is the time gradient (time derivative) of the image gray value; and is the optical flow vector to be solved (the velocity components of the pixel in the x and y directions); in a local window (for or ), all pixels have the same motion ( ); solve the optical flow vector by minimizing the residual sum of squares: ; Taking the derivative and setting it to zero, we get: ; The left matrix is ​​the second-order matrix of the gradient; the right matrix is ​​the product of the gradient and the time derivative; the optical flow vector in the window is calculated ( ); The pixel displacement is averaged and converted into a real displacement; the averaged pixel displacement is converted into a real displacement using the magnification factor obtained by calibration; Based on the calculated displacement of each floor, the inter-story displacement is calculated to obtain the final measurement result.

[0014] Preferably, in step six, the quantitative model verification calculates the theoretical displacement sensitivity limit through a formula and compares it with the actual measurement result. The specific formula is as follows: ; ;in, is the theoretical displacement sensitivity limit, is the bit depth of the camera.

[0015] Preferably, in step seven, the example verification includes: Experimental setup: A three-layer frame structure was used as the experimental object, and the structure movement was recorded using a laser displacement sensor and a camera; Data analysis and processing: Process image data according to the steps of motion feature extraction, matrix reconstruction and optical flow calculation; Result comparison: The calculation results of the method of the present invention are compared with the measurement results of the laser displacement sensor in time and space to verify the accuracy and effectiveness of the method.

[0016] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are: 1. The present invention uses multi-mode image fusion technology (combination of infrared imaging and ordinary optical imaging) and singular value decomposition (SVD) to reduce noise and extract features from data, effectively reducing the impact of background noise and interference information.

[0017] 2. The present invention reduces the data dimension and the calculation complexity through the dimensionality reduction processing of singular value decomposition, thus improving the algorithm operation efficiency. The small vibration and random noise interference in the non-target area are further eliminated through weighted matrix reconstruction, making the signal in the target area more prominent. The LK optical flow method can accurately capture the tiny movements at the pixel level based on the assumption that the light intensity remains constant, thus ensuring the accuracy of the displacement calculation results.

[0018] 3. The present invention has verified the accuracy and effectiveness of the method through laboratory vibration table testing and using laser displacement sensors as real value comparison. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0020] Figure 1 A flow chart of an ultra-precision interlayer displacement measurement method of the present invention is shown; Figure 2 A flow chart of motion feature extraction and processing of the present invention is shown. DETAILED DESCRIPTION

[0021] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only 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 ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0022] In addition, the described features, structures or characteristics may be combined in one or more example embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the example embodiments of the present disclosure. However, those skilled in the art will appreciate that the technical solutions of the present disclosure may be practiced while omitting one or more of the specific details, or other methods, components, steps, etc. may be adopted. In other cases, well-known structures, methods, implementations or operations are not shown or described in detail to avoid obscuring various aspects of the present disclosure. Embodiment 1:

[0023] See also Figure 1 As shown, an ultra-precision interlayer displacement measurement method of this embodiment has the following process: Step 1: Equipment installation and preparation.

[0024] Camera installation: Use a high-frame-rate HD camera and install it outside the building structure or model to ensure that the camera's shooting range can fully cover the target area. At the same time, combine infrared imaging and ordinary optical imaging technology, and use multi-mode image fusion measurement technology to improve the measurement applicability in low-visibility or complex light environments.

[0025] Target arrangement: Design black and white striped targets and stick them on the surface of the structure at the location of the floor panel on each floor, ensuring that the black and white coincident lines remain parallel to the direction of movement.

[0026] Step 2: Image acquisition.

[0027] On-site measurement: At the actual construction site, use a mounted camera to record the movement of the structure under external excitation (such as earthquake) and save the data in the form of video.

[0028] Indoor testing: If an indoor vibration table test is conducted, a camera is also used to record video of the vibration process to provide image data for subsequent analysis.

[0029] Step 3: Parameter calibration.

[0030] Determine an object with a known geometric length of L, count the number of pixels of the object in the video, and calculate the pixel-physical size magnification factor k using the formula k=L / N, which is used to subsequently convert pixel displacement into real displacement.

[0031] Step 4: Motion feature extraction and processing.

[0032] See also Figure 2 As shown, the specific process of motion feature extraction and processing is as follows: Select the region of interest: In the initial still frame image, select the floor boundary (usually the pixel where the black and white stripes intersect) as the region of interest. The selection range width should ensure that the black and white regions each contain at least 10 pixels, and the height is the stripe height. Extract the grayscale values ​​of the region of interest in all frames and construct the ultra-sensitive algorithm input space-time matrix , whose dimensions are ;in, Represents the number of spatial sampling points (e.g. pixels or markers); Indicates the number of time frames; space-time matrix Each column represents the spatial intensity distribution at a certain moment; each row represents the time change of a certain spatial point.

[0033] The spatial intensity is obtained by reading the RGB of the spatial position of interest in the video and then converting the RGB into grayscale.

[0034] Singular value decomposition: Perform singular value decomposition (linear principal component analysis PCA) on the input space-time matrix, extract motion feature patterns, decompose the main motion directions and features such as the main deformation mode between floors, and use them to reduce the dimension of the data, reduce interference information, and focus on significant motion patterns. Specifically, For the input matrix Perform singular value decomposition, extract the main motion patterns and features, reduce the dimension of the data, and reduce interference information. Specifically: ; in, represents the left singular matrix, containing the main spatial features of the space-time matrix, with dimension ; represents the singular value matrix, with dimension , the diagonal elements are singular values, indicating the importance of the features; represents the right singular matrix, containing the main dynamic modes in time, with dimension .

[0035] Sort by singular value size and select the top singular values ​​corresponding to the larger singular values. Modes ( Much smaller than and ) After retaining the main motion features, a simplified decomposition is obtained: ; in, Indicates that it contains The matrix of the main spatial modes has the dimension ; Indicates that it contains A diagonal matrix of singular values, with dimension ; Indicates that it contains The matrix of time patterns is .

[0036] Weighted processing and matrix reconstruction: Based on the extracted main motion mode, a weighted matrix matching the motion direction is constructed to correct the original image data. Through weighted processing, background motion (such as small vibrations in non-target areas) and random noise interference are eliminated, the prominence of the target area signal is improved, and the accuracy and robustness of displacement measurement are enhanced.

[0037] Specifically, using the reserved The main characteristic patterns reconstruct the original space-time matrix: ; The reconstructed matrix is the original matrix It is an approximate version of , which retains the most important motion information; at the same time, it removes low-energy noise or irrelevant features.

[0038] In this embodiment, three spatial points, m=3, and the intensity values ​​in four time frames (n=4) are selected to construct the space-time matrix X as follows: ; It should be noted that Row 1: The intensity of point 1 in space fluctuates slightly over time (dominant motion mode).

[0039] Row 2: The intensity of spatial point 2 increases linearly over time (secondary motion mode).

[0040] Row 3: The intensity of spatial point 3 contains random noise (needs to be filtered).

[0041] Perform singular value decomposition on X and get: ; The matrices are: Left singular matrix U (spatial characteristics): ; Column 1: Mainly corresponds to the coordinated motion of spatial points 1 and 2.

[0042] Column 2: Minor modes, describing phase differences between points in space.

[0043] Column 3: Noise correlation.

[0044] Singular value matrix (diagonal matrix): ; The order of the singular value sizes is: 170.2 > 6.1 > 1.3, indicating that the first two modes dominate the motion.

[0045] Right singular matrix (time dynamics): ; Row 1: Time evolution of the main mode.

[0046] Row 2: Temporal dynamics of the sub-mode.

[0047] Reconstruct the original matrix: Keep the first two main modes (r=2), then truncate and reconstruct: The truncated matrix: ; ; ; Reconstructed matrix X_reconstructed: ; The noise in the third row (e.g., 30→29.9) is effectively suppressed, preserving the main motion pattern.

[0048] Step 5: Displacement calculation.

[0049] Pixel displacement calculation: The LK optical flow method is used to calculate the displacement of all pixels in the pixel height direction at the intersection of black and white stripes. The LK optical flow method is based on the assumption that the light intensity remains constant and the displacement is small. Through the optical flow constraint equation and the local window assumption, the normal equation group is constructed and solved to obtain the optical flow vector in the window, that is, the pixel displacement.

[0050] Specifically, by Taylor expansion, keeping the light intensity constant, we get: ; in, , are the spatial gradients of the image grayscale values ​​(derivatives along x and y); is the time gradient (time derivative) of the image gray value; and is the optical flow vector to be solved (the velocity components of the pixel in the x and y directions); In a local window (usually or ), all pixels have the same motion ( ); solve the optical flow vector by minimizing the residual sum of squares: ; Taking the derivative of the above objective function and setting it to zero, we get: ; Among them, the matrix on the left is the second-order matrix of the gradient; the matrix on the right is the product of the gradient and the time derivative.

[0051] Calculate the optical flow vector within the window ( ).

[0052] Averaging and real displacement calculation: The calculated pixel displacement is averaged to further reduce the error. Then, the averaged pixel displacement is converted to the real displacement using the magnification factor obtained by the previous calibration.

[0053] Calculation of inter-story displacement: Based on the displacement calculated for each floor, the inter-story displacement is calculated to obtain the final measurement result.

[0054] Step 6: Quantitative model verification.

[0055] The accuracy of the measurement system is theoretically evaluated and quantitatively analyzed based on the calculation formula for the minimum measurable displacement, and then compared and verified with the actual measurement results.

[0056] The specific formula is as follows: ; ;in, is the theoretical displacement sensitivity limit, is the bit depth of the camera.

[0057] Step 7: Example verification (laboratory vibration table test) Test arrangement: A three-story frame structure was used as the test object, and the scaled Loma Prieta earthquake motion was selected as the earthquake input. Two laser displacement sensors were used to record the motion of the second and third floors of the frame structure, and their measurement results were used as the true value; at the same time, a camera was used to shoot a video of the structure's motion.

[0058] Data analysis and processing: According to the above steps of motion feature extraction, matrix reconstruction and optical flow calculation, the image data collected by the camera is processed to calculate the inter-layer displacement.

[0059] The beneficial effects of this embodiment are as follows: the accuracy and reliability of interlayer displacement measurement are improved through technologies such as multi-mode image fusion, singular value decomposition and optical flow method. It can still maintain high sensitivity in low visibility or complex environments, effectively reduce noise interference, and achieve sub-pixel displacement detection. Laboratory verification shows that compared with laser displacement sensors, the measurement results are highly consistent, proving the accuracy and practicality of the method, and providing an efficient and economical technical means for building structure health monitoring.

[0060] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

[0061] The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to only specific implementation methods. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. An ultra-precision interlayer displacement measurement method, characterized in that: The method flow is as follows: Step 1: Equipment installation and preparation: Select a high-frame-rate HD camera, install the camera and arrange black and white stripe targets; Step 2: Image acquisition: In actual construction sites or indoor vibration table tests, use a camera to record the movement of the structure under external excitation, and save the video data in combination with multi-mode image fusion measurement technology; Step 3: Parameter calibration: Calculate the pixel-physical size magnification factor k through an object with known geometric length; Step 4: Motion feature extraction and processing: Select the region of interest, construct the space-time matrix and perform singular value decomposition, extract the main motion mode, and reconstruct the matrix to eliminate noise interference; Step 5: Displacement calculation: Use the LK optical flow method to calculate pixel displacement, convert it into real displacement and calculate inter-layer displacement; Step 6: Quantitative model verification: Evaluate the accuracy of the measurement system through theoretical formulas and compare and verify with actual measurement results.

2. The ultra-precision interlayer displacement measurement method according to claim 1 is characterized in that: In step 1, the camera installation must ensure that the shooting range completely covers the target area, and the design and arrangement of the black and white striped target must ensure that the black and white overlap line is parallel to the movement direction.

3. The ultra-precision interlayer displacement measurement method according to claim 1 is characterized in that: In the step 2, the image acquisition includes video recording of the actual construction site or video recording of the indoor vibration table test to provide image data for subsequent analysis.

4. The ultra-precision interlayer displacement measurement method according to claim 1 is characterized in that: In the step three, the parameter calibration calculates the pixel-physical size magnification factor k through the formula k=L / N, where L is the known geometric length and N is the number of pixels of the object in the video.

5. The ultra-precision interlayer displacement measurement method according to claim 1 is characterized in that: In the step 4, the motion feature extraction and processing includes: Select the region of interest and construct the space-time matrix , whose dimensions are ;in, Indicates the number of spatial sampling points; Indicates the number of time frames; space-time matrix Each column in represents the spatial intensity distribution at a certain moment; each row represents the time change of a certain spatial point; wherein, the spatial intensity is obtained by reading the RGB of the spatial position of interest in the video, and then converting the RGB into grayscale; Perform singular value decomposition on the space-time matrix to extract the main motion modes, specifically: ;in, represents the left singular matrix, containing the main spatial features of the space-time matrix, with dimension ; represents the singular value matrix, with dimension , the diagonal elements are singular values, indicating the importance of the features; represents the right singular matrix, containing the main dynamic modes in time, with dimension ; Sort by singular value size and select the top singular values ​​corresponding to the larger singular values mode, Less than and ; After retaining the main motion features, a simplified decomposition is obtained: ;in, Indicates that it contains The matrix of the main spatial modes has the dimension ; Indicates that it contains A diagonal matrix of singular values, with dimension ; Indicates that it contains The matrix of time patterns is ; Through weighted processing and matrix reconstruction, background motion and random noise interference are eliminated. Specifically, the reserved The main characteristic patterns reconstruct the original space-time matrix: ; The reconstructed matrix is the original matrix It is an approximate version of , which retains the most important motion information; at the same time, it removes low-energy noise or irrelevant features.

6. The ultra-precision interlayer displacement measurement method according to claim 1 is characterized in that: In the step 5, the displacement calculation includes: The LK optical flow method is used to calculate the pixel displacement. Specifically, based on the assumption that the light intensity is constant, the light intensity is kept constant through Taylor expansion, and the result is: ;in, , are the spatial gradients of the image grayscale values, which are the derivatives along x and y; It is the time gradient of the gray value of the image, that is, the time derivative; and is the optical flow vector to be solved, that is, the velocity component of the pixel in the x and y directions; in a local window All pixels have the same motion ( ); solve the optical flow vector by minimizing the residual sum of squares: ; Taking the derivative and setting it to zero, we get: ; The left matrix is ​​the second-order matrix of the gradient; the right matrix is ​​the product of the gradient and the time derivative; the optical flow vector in the window is calculated ( ); The pixel displacement is averaged and converted into a real displacement; the averaged pixel displacement is converted into a real displacement using the magnification factor obtained by calibration; Based on the calculated displacement of each floor, the inter-story displacement is calculated to obtain the final measurement result.

7. The ultra-precision interlayer displacement measurement method according to claim 1 is characterized in that: In step 6, the quantitative model verification calculates the theoretical displacement sensitivity limit through a formula and compares it with the actual measurement result. The specific formula is as follows: ; ;in, is the theoretical displacement sensitivity limit, is the bit depth of the camera.

8. The ultra-precision interlayer displacement measurement method according to claim 1 is characterized in that: Based on the quantitative model and on-site testing requirements, the camera bit depth and number of effective pixels can be adjusted to ensure accurate and effective measurements.

Images