A method for structural micro-vibration working modal analysis based on optical flow method

Abstract

The present invention discloses a method for analyzing the working modal of a structural micro-vibration based on an optical flow method, and belongs to the field of structural vibration measurement. The method comprises the following steps: first, collecting a video of the vibration process; second, selecting an area of ​​interest using the video, and extracting the structural contour points of the area; third, extracting the apparent motion response of the structural contour points; fourth, converting the displacement time domain data of the structural contour points into frequency domain data, and analyzing to obtain an approximate value of the modal frequency of the structure; fifth, amplifying the vibration response of the contour points; and sixth, identifying the modal frequency and damping ratio of the structure. Compared with the prior art, the advantages of the present invention are: the analysis method of the present invention realizes non-contact measurement, obtains the modal parameters of the structure with high precision and low cost, and extracts the contour points of the structure as the characteristic points of the structure, and the obtained vibration response data is representative.

Description

Technical Field

[0001] The present invention relates to the field of structural vibration measurement, and more specifically, to a structural micro-vibration working modal analysis method based on an optical flow method. Background Art

[0002] At present, the measurement of structural vibration mostly uses the contact method. For small structures, the mass and additional stiffness of the sensor will have a significant impact on the dynamic characteristics of the measured structure; for large structures, the installation process and subsequent maintenance of the sensor are relatively complicated and costly.

[0003] The non-contact measurement method using the laser Doppler effect can overcome the shortcomings of contact measurement, but its measurement takes a long time, it is impossible to obtain vibration response data of multiple points on the structure at the same time, and the equipment cost is high. By capturing the vibration process of the structure through video, and then using image processing algorithms to process the video images to obtain the response data of the structural vibration, this method can achieve non-contact measurement and has a relatively low cost.

[0004] The prior art discloses the technique of digital image correlation (DIC). The patent with publication number CN105424350 proposes a modal testing method and system for thin-walled parts based on machine vision, which uses a CCD industrial camera to collect the vibration process of thin-walled parts and processes the collected image sequence to obtain the vibration characteristics of the structure; the patent with publication number CN106989812 proposes a modal testing method for large wind turbine blades based on photogrammetry technology, which uses a pair of CMOS / CCD industrial cameras to collect the vibration process of wind turbine blades, and calculates the structural vibration response through dual-camera stereo matching and three-dimensional reconstruction technology; the patent with publication number CN109459127 proposes a non-contact blade wind vibration measurement method based on image processing algorithm, which uses a camera to collect the vibration process of the blade under wind speed, and processes each frame of the image through MATLAB to obtain the vibration response speed of the blade; the patent with publication number CN112146834A proposes a structural vibration displacement measurement method and device, the collected video is the vibration process of the red square target attached to the structure, and the displacement data of the red square target is obtained through a series of image processing. Digital image correlation technology requires spraying marking points on the structure in advance. These four patents use reflective stickers, reflective coded marks, fluorescent targets and red square targets as feature points, respectively, and use the vibration data of the feature points to represent the vibration data of the structure.

[0005] The present invention uses the optical flow method to obtain the vibration response data of the structural contour points, where the contour points are points with large gradient changes in the grayscale value of the structural edge, which can well represent the vibration response of the structure. The modal parameters of the structure can be obtained by processing the time domain data using the random subspace method. Summary of the invention

[0006] In order to solve the shortcomings of the contact measurement method and digital image correlation technology (DIC) in the background technology, the present invention provides a method for analyzing the working modal of structural micro-vibration based on the optical flow method. The method realizes non-contact measurement, obtains the modal parameters of the structure with high precision and low cost, extracts the contour points of the structure as the characteristic points of the structure, and the obtained vibration response data is representative. In addition, the region of interest is selected in the video, and the optical flow method is used to track only the optical flow motion of the edge points of the structure in the region, which has high computational efficiency. The modal frequency and damping ratio of the structure are identified by using the response data including but not limited to the random subspace method, the frequency domain decomposition method, the Hilbert-Huang transform method and the extracted virtual measuring points, with high accuracy.

[0007] To achieve the above objectives, the technical solution of the present invention is as follows.

[0008] A method for analyzing structural micro-vibration working modal based on optical flow method comprises the following steps:

[0009] S1. Use video recording equipment to collect the vibration process of the structure and obtain a video of the vibration process;

[0010] S2. Selecting a region of interest on the first frame of the video using an image processing program and extracting structural contour points of the region of interest;

[0011] S3. Extracting the apparent motion response of structure contour points by visual methods;

[0012] S4. Use fast Fourier transform to convert the displacement time domain data of the structure contour points into frequency domain data, and analyze the peak value of the frequency domain data to obtain the approximate value of the modal frequency of the structure;

[0013] S5. Bandpass filtering the displacement signal of the contour point according to the approximate value of the modal frequency, and mapping the filtered signal to the original video to amplify the vibration response of the contour point;

[0014] S6. Use the structural modal parameter identification method and the filtered displacement time domain data to identify the modal frequency and damping ratio of the structure.

[0015] The above method does not need to use contact sensors such as accelerometers and strain gauges to obtain the vibration response data of the structure, realizing non-contact measurement with low cost; at the same time, it also does not need to spray marks on the surface of the structure, and the contour points of the structure are extracted as the characteristic points of the structure, so the obtained vibration response data is representative.

[0016] Furthermore, the frame rate of the video recording device in step S1 is greater than twice the maximum modal frequency of the structure to be estimated.

[0017] Furthermore, the unit of the structure contour point in step S2 is pixel.

[0018] Furthermore, the processing procedure of the image processing program in step S2 is specifically as follows:

[0019] S21. Manually select the region of interest in the first frame of the video, or automatically identify the structure in the video using image recognition methods;

[0020] S22. performing mean filtering on the region of interest;

[0021] S23. Binarize the filtered region of interest using the Canny algorithm and identify the contour points of the structure in the region;

[0022] S24. Extract the pixel coordinates of the binarized contour points through the cv2.findContours() function and map them to the video image.

[0023] Furthermore, the visual method in step S3 includes but is not limited to sparse optical flow method, dense optical flow method, and phase-based motion amplification technology. An area of ​​interest is selected in the video, and the above visual method is used to track only the optical flow motion of the structural edge points in the area, which has high computational efficiency.

[0024] Furthermore, step S3 is more specifically as follows: the optical flow field of all contour points in the region of interest in the video is calculated by the Gunnar Farneback optical flow algorithm to obtain the velocity vector and displacement vector of the contour points; the velocity vector represents the apparent motion speed of the structure, and the displacement vector can obtain the time domain data of the apparent displacement.

[0025] Furthermore, step S5 is more specifically as follows: bandpass filtering is performed on all contour point displacement signals according to the approximate value of the modal frequency, the filtered signals are converted into changes in the optical flow of the contour points, and the optical flow change values ​​of all contour points in each frame are amplified and mapped to the original video, thereby amplifying the vibration response of the contour points and making the vibration response effect of the structure more obvious.

[0026] The magnification of the optical flow change values ​​of all the above contour points includes but is not limited to 3-10 times.

[0027] Furthermore, the structural modal parameter identification method in step S6 includes but is not limited to the random subspace method, the frequency domain decomposition method, and the Hilbert-Huang transform method. The modal frequency and damping ratio of the structure are identified by using the response data including but not limited to the random subspace method, the frequency domain decomposition method, the Hilbert-Huang transform method, and the extracted virtual measuring points, with high accuracy.

[0028] Furthermore, step S6 is more specifically, according to the approximate value of the modal frequency obtained in step 4 and the time domain data, using the time domain analysis method: the covariance-based random subspace method, to calculate the first two modal frequencies and damping ratios of the structure.

[0029] Compared with the prior art, the advantages and beneficial effects of the structural micro-vibration working modal analysis method of the present invention are as follows: the method of the present invention does not need to use contact sensors such as accelerometers and strain gauges to obtain the vibration response data of the structure, and realizes non-contact measurement with low cost; at the same time, it also does not need to spray marks on the surface of the structure, and the contour points of the structure are extracted as the characteristic points of the structure, and the obtained vibration response data are representative. In addition, the region of interest is selected in the video, and the optical flow method is used to track only the optical flow motion of the edge points of the structure in the region, which has high computational efficiency. The modal frequency and damping ratio of the structure are identified by using, including but not limited to, the random subspace method, the frequency domain decomposition method, the Hilbert-Huang transform method and the response data of the extracted virtual measuring points, with high accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 This is a flow chart of the structural micro-vibration working modal analysis method based on the optical flow method of the present invention;

[0031] Figure 2 To extract the effect of the structure outline in the region of interest;

[0032] Figure 3 It is the time domain diagram of the vibration response displacement of the characteristic contour point;

[0033] Figure 4 It is the frequency domain diagram of vibration response of characteristic contour points;

[0034] Figure 5 This is the algorithm flow chart of the random subspace method;

[0035] Figure 6 Figure 2 shows the experimental setup. DETAILED DESCRIPTION

[0036] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0037] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0038] In this specification, the schematic representation of certain terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, steps, methods or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0039] Combine the following Figures 1 to 6 The technical scheme of the present invention is further illustrated with embodiments.

[0040] Example 1

[0041] A method for structural micro-vibration working modal analysis based on optical flow method, such as Figure 1 As shown, it mainly includes the following steps:

[0042] S1. Use video recording equipment to capture the vibration process of the structure and obtain a video of the vibration process.

[0043] S2. Selecting a region of interest on the first frame of the video using an image processing program, and extracting structural contour points of the region of interest.

[0044] S3. Extracting the apparent motion response of structure contour points by visual methods.

[0045] S4. Use fast Fourier transform to convert the displacement time domain data of the structure contour points into frequency domain data, and analyze the peak value of the frequency domain data to obtain the approximate value of the modal frequency of the structure.

[0046] S5. Band-pass filtering is performed on the contour point displacement signal according to the approximate value of the modal frequency, and the filtered signal is mapped to the original video to amplify the vibration response of the contour point.

[0047] S6. Use the structural modal parameter identification method and the filtered displacement time domain data to identify the modal frequency and damping ratio of the structure.

[0048] Example 2

[0049] Compared with Example 1, in this embodiment, a method for analyzing the working modal of a structure micro-vibration based on an optical flow method is more specifically provided, and its step flow chart is as follows: Figure 1 As shown, the specific steps include:

[0050] S1. Use a hammer to apply random excitation to the structure under test. The structure under test is a 50 cm long standard steel ruler on which six sensors are evenly installed. Figure 6 Figure a is a schematic diagram of a 50 cm standard steel ruler and six evenly distributed sensors, and Figure b is a schematic diagram of the shooting equipment and shooting distance.

[0051] S2. Use smart phones, sports cameras, high-speed industrial cameras and other video recording devices to collect the vibration process of the structure and obtain the video of the vibration process. At the same time, use traditional detection methods (sensor measurement) to obtain the structural motion response;

[0052] The frame rate of the video equipment used in the acquisition process is greater than 2.56 times the maximum modal frequency of the structure to be estimated. During the acquisition process, the shooting distance is 1.54m.

[0053] S3. Selecting a region of interest on the first frame of the video using an image processing program, and extracting structural contour points of the region of interest in pixels.

[0054] Manually select the region of interest in the first frame of the video, or use image recognition methods to automatically identify the structure in the video; then perform mean filtering on the region of interest, and use the canny algorithm to binarize the filtered region of interest, and identify the contour points of the structure in the region. Use the cv2.findContours() function to extract the pixel coordinates of the binarized contour points and map them to the video image. The effect is as follows: Figure 2 As shown, Figure a is a schematic diagram of selecting an area of ​​interest, and Figure b is a schematic diagram of extracting contour points of the structure in the area.

[0055] S4. Extract the apparent motion response of the structure contour points by visual methods. The visual methods include but are not limited to sparse optical flow method, dense optical flow method, and phase-based motion amplification technology. Selecting a region of interest in the video, using sparse optical flow method and dense optical flow method to track only the optical flow motion of the structure edge points in the region is conducive to high computational efficiency. In this embodiment, the dense optical flow method is used.

[0056] The GF dense optical flow algorithm is used to calculate the displacement vector of the contour pixel point, including the displacement vectors in the horizontal and vertical directions, and the displacement time domain diagram of the vibration response of the contour point is drawn, as shown in Figure 3 As shown;

[0057] S5. Use fast Fourier transform to convert the displacement time domain data into frequency domain data, analyze the peak value to get the approximate value of the modal frequency of the structure, and the amplitude data also reflects the motion characteristics of the structure. The frequency domain diagram is as follows: Figure 4 shown.

[0058] S6. Perform bandpass filtering on the displacement signal according to the approximate value of the modal frequency, retain the signal of ±2 near the peak frequency, and convert the filtered signal into the contour point displacement response data, amplify the vibration response of the contour point by three times, and map it to the original video.

[0059] S7. Use the structural modal parameter identification method and the filtered displacement time domain data to identify the modal frequency and damping ratio of the structure. The structural modal parameter identification method includes but is not limited to the random subspace method, the frequency domain decomposition method, and the Hilbert-Huang transform method. By using the random subspace method, the frequency domain decomposition method, the Hilbert-Huang transform method, and the response data of the extracted virtual measuring points to identify the modal frequency and damping ratio of the structure, the accuracy is high. In this embodiment, the structural modal parameter identification method adopts the random subspace method, and its flow chart is as follows: Figure 5 shown.

[0060] The experimental results of the present invention are compared with the traditional detection method, and the comparison results are shown in Table 1 and Table 2. The frequency resolution of the method proposed in the present invention is 0.05Hz, and the reasonable error range is between ±0.05 of the traditional detection method result. As can be seen from Table 1, the measurement results of the modal frequency are within the reasonable error range. Compared with the traditional detection method, the method proposed in the present invention can not only achieve basically the same accuracy, but also is more convenient in measurement, saving time and cost.

[0061] Table 1 Modal frequencies

[0062]

[0063] Table 2 Modal damping ratio

[0064]

[0065] Compared with the prior art, the beneficial effects of the above embodiment are as follows: the method of the above embodiment does not need to use contact sensors such as accelerometers and strain gauges to obtain the vibration response data of the structure, realizing non-contact measurement with low cost; at the same time, it is also not necessary to spray marks on the surface of the structure, and the contour points of the structure are extracted as the characteristic points of the structure, and the obtained vibration response data are representative. In addition, the region of interest is selected in the video, and the optical flow method is used to track only the optical flow motion of the edge points of the structure in the region, which has high computational efficiency. The modal frequency and damping ratio of the structure are identified by using, including but not limited to, the random subspace method, the frequency domain decomposition method, the Hilbert-Huang transform method and the response data of the extracted virtual measuring points, with high accuracy.

[0066] Example 3

[0067] Compared with Example 2, the difference of this embodiment is that: in step S4, the visual method adopts the sparse optical flow method, and the structural modal parameter identification method in step S7 adopts the frequency domain decomposition method.

[0068] In this embodiment, the sparse optical flow method can achieve the beneficial effect of the dense optical flow method in embodiment 2, that is, only the optical flow motion of the edge points of the structure in the region is tracked, and the calculation efficiency is high. The frequency domain decomposition method can achieve the effect of the random subspace method in embodiment 2, that is, the response data of the extracted virtual measurement points can identify the modal frequency of the structure with high accuracy.

[0069] Furthermore, those skilled in the art may combine and associate different embodiments or examples and features of different embodiments or examples described in this specification without mutual contradiction.

[0070] It is worth noting that the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the embodiments here. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the claims of the present invention.

Claims

1. A method for structural micro-vibration working modal analysis based on optical flow method, characterized in that: The following steps are involved: S1. Use video recording equipment to collect the vibration process of the structure and obtain a video of the vibration process; S2. Selecting a region of interest on the first frame of the video using an image processing program and extracting structural contour points of the region of interest; S3. Extracting the apparent motion response of structure contour points by visual methods; S4. Use fast Fourier transform to convert the displacement time domain data of the structure contour points into frequency domain data, and analyze the peak value of the frequency domain data to obtain the approximate value of the modal frequency of the structure; S5. Bandpass filtering the displacement signal of the structure contour point according to the approximate value of the modal frequency, and mapping the filtered signal to the original video to amplify the vibration response of the structure contour point; S6. Use the structural modal parameter identification method and the filtered displacement time domain data to identify the modal frequency and damping ratio of the structure.

2. The method according to claim 1, characterized in that: The frame rate of the video recording device in step S1 is greater than twice the maximum modal frequency of the structure to be estimated.

3. The method according to claim 1, characterized in that: The unit of the structure contour points in step S2 is pixel.

4. The method according to claim 1, characterized in that: The processing process of the image processing program in step S2 is specifically as follows: S21. Manually select the region of interest in the first frame of the video, or automatically identify the structure in the video using image recognition methods; S22. performing mean filtering on the region of interest; S23. Binarize the filtered region of interest using the Canny algorithm and identify the structural contour points of the structure in the region; S24. Extract the pixel coordinates of the binary structure contour points through the cv2.findContours() function and map them to the video image.

5. The method according to claim 1, characterized in that: The visual method in step S3 includes but is not limited to a sparse optical flow method, a dense optical flow method, and a phase-based motion magnification technique.

6. The method according to claim 5, characterized in that: Step S3 is more specifically: the optical flow field of all structural contour points in the region of interest in the video is calculated by the Gunnar Farneback optical flow algorithm to obtain the velocity vector and displacement vector of the structural contour points; the velocity vector represents the apparent motion speed of the structure, and the displacement vector can obtain the time domain data of the apparent displacement.

7. The method according to claim 1, characterized in that: Step S5 is more specific as follows: bandpass filtering is performed on the displacement signals of all structural contour points according to the approximate value of the modal frequency, the filtered signals are converted into changes in the optical flow of the structural contour points, and the optical flow change values ​​of all structural contour points in each frame are amplified and mapped to the original video to amplify the vibration response of the structural contour points.

8. The method according to claim 7, characterized in that: The magnification of the optical flow change values ​​of all structure contour points includes but is not limited to 3-10 times.

9. The method according to claim 1, characterized in that: The structural modal parameter identification method in step S6 includes but is not limited to a random subspace method, a frequency domain decomposition method, and a Hilbert-Huang transform method.

10. The method according to claim 9, characterized in that: Step S6 is more specifically: according to the approximate value of the modal frequency obtained in step 4 and the time domain data, the first two modal frequencies and damping ratios of the structure are calculated using the time domain analysis method: the covariance-based random subspace method.

Images