Structural Modal Recognition Method Based on Digital Image Correlation and Motion Magnification Techniques

Abstract

This application relates to a structural modal recognition method, apparatus, computer device, storage medium, and computer program product based on digital image correlation and motion amplification techniques. The method includes: decomposing the spectrogram corresponding to each frame of an initial structural vibration video into multiple frequency domain sub-bands; performing bandpass filtering on the phase data corresponding to each frequency domain sub-band to obtain target phase data corresponding to each frequency domain sub-band of the structure under test; performing motion amplification on the target phase data based on the optimal amplification factor corresponding to each vibration order of the structure under test to obtain vibration amplification results for each order of the structure under test; obtaining displacement-time history curves corresponding to each measuring point on the structure under test during each vibration order based on the vibration amplification results; and extracting the mode shapes corresponding to each mode of the structure under test based on the displacement-time history curves. This method can improve the accuracy of modal recognition.

Description

Technical Field

[0001] This application relates to the field of structural vibration mode recognition technology, and in particular to a structural mode recognition method, apparatus, computer equipment, storage medium and computer program product based on digital image correlation and motion amplification technology. Background Technology

[0002] Modal characteristics are the inherent vibration features of a structure. The process of calculating or experimentally analyzing modal parameters is called modal analysis, which provides a reliable means for structural design and performance evaluation. Traditional structural modal identification mainly involves analyzing and calculating the raw data of structural vibration (generally acceleration time history information), with data acquisition being the most critical step. However, for some special and complex structures (such as long-span suspension bridges and ultra-long cables), the placement and installation of traditional sensors are severely limited, and traditional data acquisition methods cannot meet the measurement requirements. Therefore, non-contact measurement methods for structural vibration are needed.

[0003] Meanwhile, existing non-contact measurement modal recognition patents mostly use indoor tests for verification and explanation. They require the use of excitation equipment to apply vibration to the structure as an input signal, and finally estimate the modal parameters by the frequency response function, which is difficult to implement and apply in actual engineering. Summary of the Invention

[0004] Therefore, it is necessary to provide a structural modal recognition method, apparatus, computer equipment, computer-readable storage medium, and computer program product based on digital image correlation and motion amplification techniques that can improve the accuracy of modal analysis, addressing the aforementioned technical problems.

[0005] Firstly, this application provides a structural mode recognition method based on digital image correlation and motion magnification techniques. The method includes:

[0006] Obtain initial structural vibration video;

[0007] The spectrogram corresponding to each frame of the initial structural vibration video is decomposed into multiple frequency domain sub-bands.

[0008] Bandpass filtering is performed on the phase data corresponding to each frequency domain sub-band to obtain the target phase data corresponding to each frequency domain sub-band of the structure under test.

[0009] Based on the optimal amplification factor corresponding to each order of vibration of the structure under test, motion amplification processing is performed on the target phase data to obtain the vibration amplification results of each order of the structure under test.

[0010] Based on the vibration amplification results, the displacement-time history curves of each measuring point on the structure under test during each order of vibration are obtained.

[0011] Based on the displacement-time history curves, the mode shapes corresponding to each mode of the structure under test are extracted.

[0012] In one embodiment, after acquiring the initial structural vibration video, and before decomposing the spectrogram corresponding to each frame of the initial structural vibration video, the method further includes:

[0013] The initial structure vibration video is segmented into frames to decompose it into a sequence of initial structure vibration images.

[0014] Two-dimensional discrete Fourier transform processing is performed on each of the initial structural vibration image sequences to obtain the spectrum corresponding to each frame of the initial structural vibration image.

[0015] In one embodiment, the step of decomposing the spectrogram corresponding to each frame of the initial structural vibration video into multiple frequency domain sub-bands includes:

[0016] Based on a preset downsampling ratio, each frame of the initial structural vibration video is downsampled to obtain sub-images at multiple scales.

[0017] Based on the amplitude and phase information contained in each pixel in each of the aforementioned spectrograms, each of the aforementioned spectrograms is decomposed into multiple frequency domain sub-bands.

[0018] In one embodiment, the step of performing bandpass filtering on the phase data corresponding to each of the frequency domain sub-bands to obtain the target phase data corresponding to each of the frequency domain sub-bands of the structure under test includes:

[0019] Based on a time-domain bandpass filter, the phase data corresponding to each frequency domain sub-band is bandpass filtered to obtain the target phase data corresponding to each frequency domain sub-band of the structure under test. The time-domain bandpass filter is set with bandpass filtering parameters of each order.

[0020] In one embodiment, the bandpass filter parameters include a center frequency, an upper limit frequency, and a lower limit frequency, and the method for determining the bandpass filter parameters includes:

[0021] The displacement-time history curve of any measuring point on the structure under test is obtained from the initial structural vibration video; Fourier transform is performed on the displacement-time history curve to obtain the natural frequency information of each order of the structure under test; each natural frequency information is used as the center frequency of the time-domain bandpass filter; based on each center frequency and the bandwidth setting, the upper limit frequency and the lower limit frequency are determined.

[0022] In one embodiment, the method further includes:

[0023] When determining the optimal amplification factor for each order of vibration of the structure under test, the following steps are performed:

[0024] Multiple initial magnification factors are obtained, and each initial magnification factor is determined based on the arithmetic sequence rule;

[0025] Based on the initial amplification coefficients, motion amplification processing is performed on the target phase data to obtain amplified composite structure vibration videos.

[0026] The modal confidence criterion is used as the evaluation standard to evaluate each of the magnified synthetic structure videos and obtain modal confidence curves. The modal confidence curves are the fitting curves of each of the initial magnification coefficients and modal confidence values.

[0027] When the modal confidence value meets the preset modal confidence value condition, the corresponding initial amplification factor is selected as the amplification factor.

[0028] In one embodiment, extracting the mode shapes corresponding to each order of the structure under test based on the displacement time history curves includes:

[0029] Multiple marker points are selected on the structure to be tested, and one of the marker points is selected as the research point.

[0030] Based on the displacement-time history curves, obtain the time-displacement curve of the research point within the test time.

[0031] Based on the time displacement curve of the research point, each extreme point is selected from the time displacement curve of the research point, and the time of each extreme point is obtained based on the time corresponding to each extreme point.

[0032] Based on the time of each extreme point, the corresponding mode shape diagram at the extreme point time is determined. The mode shape diagram at the extreme point time is determined by the displacement of each of the marked points at the extreme point time.

[0033] The average value of the mode shape diagrams at each extreme point is taken to obtain the mean mode shape diagrams;

[0034] The mean mode shape diagrams are fitted with sinusoidal functions and normalized to extract the mode shapes corresponding to each order of the structure under test.

[0035] Secondly, this application also provides a structural mode recognition device based on digital image correlation and motion magnification techniques, the device comprising:

[0036] The data acquisition module is used to acquire initial structural vibration videos;

[0037] The data processing module is used to decompose the spectrogram corresponding to each frame of the initial structure vibration video into multiple frequency domain sub-bands.

[0038] The filtering module is used to perform bandpass filtering on the phase data corresponding to each frequency domain sub-band to obtain the target phase data corresponding to each frequency domain sub-band of the structure under test.

[0039] The motion amplification module is used to perform motion amplification processing on the target phase data based on the optimal amplification factor corresponding to each order of vibration of the structure under test, so as to obtain the vibration amplification results of each order of the structure under test.

[0040] The curve acquisition module is used to obtain the displacement time history curves of each measuring point on the structure under test under each order of vibration based on the vibration amplification results.

[0041] The mode shape extraction module is used to extract the mode shapes corresponding to each mode of the structure under test based on the displacement time history curves.

[0042] Thirdly, this application also provides a computer device. The computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of the above-described structural modal recognition method based on digital image correlation and motion magnification techniques.

[0043] Fourthly, this application also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program thereon, which, when executed by a processor, implements the steps of the above-described structural modal recognition method based on digital image correlation and motion magnification techniques.

[0044] Fifthly, this application also provides a computer program product. The computer program product includes a computer program that, when executed by a processor, implements the steps of the above-described structural modal recognition method based on digital image correlation and motion magnification techniques.

[0045] The aforementioned structural modal identification method, apparatus, computer equipment, storage medium, and computer program product based on digital image correlation and motion amplification technology acquire an initial structural vibration video; decomposes the spectrum corresponding to each frame of the initial structural vibration video into multiple frequency domain sub-bands; performs bandpass filtering on the phase data corresponding to each frequency domain sub-band to obtain the target phase data corresponding to each frequency domain sub-band of the structure under test, thereby filtering out the phase data of other structures besides the structure under test in the initial structural vibration video, retaining only the phase data of the structure under test; further, performs motion amplification processing on each target phase data to obtain the vibration amplification results of each order of the structure under test; based on each vibration amplification result, obtains the displacement time history curves corresponding to each measuring point on the structure under test during each order of vibration; based on each displacement time history curve, extracts the mode shapes corresponding to each order of the structure under test, making the extracted curves of each order of the structure under test clear and intuitive, thus eliminating the need for frequency response function calculation, and also extracting the identification mode shapes of the structure under test at each order. This method can effectively improve the accuracy of structural modal identification. Attached Figure Description

[0046] Figure 1 This is an application environment diagram of a structural modality recognition method based on digital image correlation and motion amplification technology in one embodiment.

[0047] Figure 2 This is a flowchart illustrating a structural modal recognition method based on digital image correlation and motion amplification techniques in one embodiment.

[0048] Figure 3 This is a time history curve of regional point displacement in a structural modal recognition method based on digital image correlation and motion magnification technology in one embodiment.

[0049] Figure 4 The spectrum diagram of a structural mode recognition method based on digital image correlation and motion amplification technology in another embodiment is shown.

[0050] Figure 5 This is a schematic diagram of the modal confidence criterion values ​​of a structural modal recognition method based on digital image correlation and motion amplification techniques in one embodiment.

[0051] Figure 6 This is a schematic diagram of the modal confidence curve of a structural modal recognition method based on digital image correlation and motion amplification technology in one embodiment;

[0052] Figure 7 This is a flowchart illustrating the process of obtaining the vibration modes in a structural modal recognition method based on digital image correlation and motion amplification techniques in one embodiment.

[0053] Figure 8 This is a flowchart of a structural modal recognition method based on digital image correlation and motion amplification techniques in one embodiment;

[0054] Figure 9 This is a structural block diagram of a structural modality recognition device in one embodiment;

[0055] Figure 10 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0056] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0057] The structural mode recognition method based on digital image correlation and motion magnification technology provided in this application can be applied to, for example... Figure 1 The application environment shown. Among them, Figure 1 The application environment shown includes a data acquisition device 102, a structure under test 104, and an electronic device 106. The data acquisition device 102 can be a DIC (Digital Image Correlation) measurement device (such as a high-speed camera). Specifically, the high-speed camera can acquire free decay vibration video of the structure under test 104 and then transmit it to the electronic device 106. The electronic device 106 processes the decay vibration video using the structural modal recognition method based on digital image correlation and motion amplification technology of this application to obtain the modal parameters of the structure under test.

[0058] The data acquisition device 102 communicates with the electronic device 106 via a network. The electronic device 102 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. IoT devices can include smart speakers, smart TVs, smart air conditioners, and smart in-vehicle systems. Portable wearable devices can include smartwatches, smart bracelets, and head-mounted devices. The electronic device 106 can also be a server, which can be a standalone server or a server cluster consisting of multiple servers.

[0059] Specifically, the electronic device 106 acquires an initial structural vibration video; decomposes the spectrum corresponding to each frame of the initial structural vibration image in the initial structural vibration video into multiple frequency domain sub-bands; performs bandpass filtering on the phase data corresponding to each frequency domain sub-band to obtain target phase data corresponding to each frequency domain sub-band of the structure under test; based on the optimal amplification factor corresponding to each order of vibration of the structure under test, performs motion amplification processing on the target phase data to obtain vibration amplification results of each order of the structure under test; based on each vibration amplification result, obtains the displacement time history curves corresponding to each measuring point on the structure under test at each order of vibration; and extracts the mode shapes corresponding to each mode of the structure under test based on each displacement time history curve.

[0060] In one embodiment, such as Figure 2 As shown, a structural modality recognition method based on digital image correlation and motion magnification techniques is provided, which is then applied to... Figure 1 Taking electronic device 106 as an example, the explanation includes the following steps:

[0061] Step S202: Obtain the initial structural vibration video.

[0062] The initial structural vibration video can refer to the video obtained by DIC (Digital Image Correlation) measurement equipment (such as a high-speed camera) to capture the free decay vibration of the structure. In addition to the vibration information of the structure under test (such as the cable tower of a bridge), the initial structural vibration video may also contain the vibration information of other structures (such as a moving car, the railing of a bridge, etc.) and the vibration information of passing pedestrians.

[0063] In one embodiment, when acquiring initial structural vibration video using a DIC measurement device, the camera's imaging plane is parallel to the surface of the structure under test, and camera calibration is performed simultaneously. Specifically, the surface of the structure under test is transformed to the camera's imaging plane using a projection matrix. Combined with camera calibration, the correspondence between the actual coordinates and the camera coordinates is determined. By making the surface of the structure under test as parallel as possible to the camera's imaging plane, the matrix can be easily obtained, reducing calculation errors.

[0064] In one embodiment, when processing the initial structural vibration video, speckle processing can be applied to the surface of the structure under test to facilitate DIC target point displacement tracking. If there is obvious texture on the structural surface, speckle processing can be omitted.

[0065] Step S204: Decompose the spectrogram corresponding to each frame of the initial structural vibration video into multiple frequency domain sub-bands.

[0066] The initial structural vibration video is composed of frames of initial structural vibration images. By performing Fast Fourier Transform on each frame of initial structural vibration images, the spectrum of each frame of initial structural vibration images can be obtained. Decomposition processing refers to decomposing the spectrum into various frequency domain sub-bands according to the tangential and radial directions. The image information (such as amplitude, phase, etc.) of the same frequency domain sub-band is similar. By decomposing the spectrum into multiple frequency domain sub-bands, the time domain bandpass filtering effect can be effectively improved because the frequency domain sub-bands with similar image information in the same frequency domain often have similar motion states in the time domain.

[0067] In one embodiment, after acquiring the initial structural vibration video, the process of decomposing the spectrogram corresponding to each frame of the initial structural vibration video before performing the decomposition processing on each frame of the initial structural vibration video includes:

[0068] The initial structure vibration video is segmented into frames to decompose it into a sequence of initial structure vibration images.

[0069] Two-dimensional discrete Fourier transform processing is performed on each of the initial structural vibration image sequences to obtain the spectrum corresponding to each frame of the initial structural vibration image.

[0070] The initial structure vibration image sequence refers to a continuous series of images at different times and directions obtained after the initial structure vibration video is processed into frames. Frame processing can refer to dividing the initial structure vibration video according to time to obtain each initial structure vibration image sequence. Then, performing a two-dimensional discrete Fourier transform on each initial structure vibration image sequence can convert it from the spatial domain to the frequency domain, obtaining the spectrum corresponding to each frame of the initial structure vibration image.

[0071] Step S206: Bandpass filtering is performed on the phase data corresponding to each frequency domain sub-band to obtain the target phase data corresponding to each frequency domain sub-band of the structure under test.

[0072] Here, phase data refers to the time-domain data of phase change along the time distribution corresponding to each frequency sub-band after performing an inverse fast Fourier transform on each frequency sub-band. Bandpass filtering refers to using a time-domain bandpass filter with set filtering parameters to perform bandpass filtering on the phase data. This can retain only the phase data of the structure under test, while filtering out the phase data of other structures besides the structure under test, and finally using the retained phase data as the target phase data corresponding to the frequency sub-band of the structure under test.

[0073] In one embodiment, after obtaining each frequency domain sub-band, the following operations can be performed on any frequency domain sub-band: First, the frequency domain sub-band is transformed into the spatial domain by performing an inverse fast Fourier transform, which constitutes the phase data. For each frame of the initial structural vibration image, there is corresponding phase data for this frequency domain sub-band. By subtracting this phase data from the phase data of the first frame image, the time-domain data of the phase change along time can be obtained.

[0074] Step S208: Based on the optimal amplification factor corresponding to each order of vibration of the structure under test, the target phase data is subjected to motion amplification processing to obtain the vibration amplification results of each order of the structure under test.

[0075] The optimal amplification factor refers to the most suitable amplification factor for each order of vibration. Motion amplification processing can refer to linearly amplifying the target phase data by a certain factor, i.e., performing phase manipulation processing. Since the target phase data includes the phase data of each order of the structure under test, after amplifying the target phase data, the vibration amplification results of the structure under test at each order can be obtained.

[0076] Step S210: Based on the vibration amplification results, obtain the displacement time history curves of each measuring point on the structure under test during each order of vibration.

[0077] Among them, the measuring point on the structure under test can refer to any pixel point that constitutes the structure under test. The displacement time history curve is a curve fitted by the relevant information of multiple measuring points of the structure under test. The horizontal axis of the displacement time history curve can be the test duration, the vertical axis can be the attribute of the structure under test (such as the length of the tower), and the vertical axis can be the displacement of each marker point at each time point. The displacement time history curve can be obtained based on the vibration amplification results. For each order of vibration amplification results, there is a corresponding displacement time history curve.

[0078] Step S212: Based on the displacement time history curves, extract the mode shapes corresponding to each mode of the structure under test.

[0079] Among them, the mode shape refers to the fitting curve containing the vibration information of the structure under test. The mode shape can be used to extract modal parameters. For the structure under test with multiple modes, there are also multiple mode shapes. After obtaining the mode shape, the structural modal information of the structure under test at each order can be directly extracted from each mode shape.

[0080] In the aforementioned structural modal identification method based on digital image correlation and motion amplification technology, the following steps are taken: First, an initial structural vibration video is acquired. Then, the spectrum corresponding to each frame of the initial structural vibration video is decomposed into multiple frequency domain sub-bands. Next, bandpass filtering is applied to the phase data corresponding to each frequency domain sub-band to obtain the target phase data corresponding to each frequency domain sub-band of the structure under test. This filters out vibration signals from other structures in the initial structural vibration video, retaining only the vibration signal of the structure under test. Furthermore, motion amplification is applied to each target phase data to obtain the vibration amplification results of each order of the structure under test. Based on these vibration amplification results, displacement-time history curves corresponding to each measuring point on the structure under test during each order of vibration are obtained. Based on these displacement-time history curves, the mode shapes corresponding to each order of the structure under test are extracted. This makes the extracted mode shape curves of the structure under test clear and intuitive, eliminating the need for frequency response function calculations. It also allows for the extraction of the identification mode shapes of the structure under test at each order. Finally, the structural modal information of the structure under test at each order is determined through the identification mode shapes. The above methods can effectively improve the accuracy of structural mode recognition.

[0081] In one embodiment, the step of decomposing the spectrogram corresponding to each frame of the initial structural vibration video into multiple frequency domain sub-bands includes:

[0082] Based on a preset downsampling ratio, each frame of the initial structural vibration video is downsampled to obtain sub-images at multiple scales.

[0083] Based on the amplitude and phase information contained in each pixel in each of the aforementioned spectrograms, each of the aforementioned spectrograms is decomposed into multiple frequency domain sub-bands.

[0084] The predicted downsampling ratio refers to the set ratio for downsampling each frame of the initial structural vibration image. The preset downsampling ratio can be adaptively set according to actual needs. When downsampling the initial structural vibration image, a complex steerable pyramid can be used to decompose the frequency domain structural vibration image according to scale, size, and position to obtain the amplitude and phase information after local wavelet transform. Based on the amplitude and phase information, multiple frequency domain sub-bands are determined, thereby achieving a better filtering effect.

[0085] In one embodiment, the step of performing bandpass filtering on the phase data corresponding to each of the frequency domain sub-bands to obtain the target phase data corresponding to each of the frequency domain sub-bands of the structure under test includes:

[0086] Based on a time-domain bandpass filter, the phase data corresponding to each frequency domain sub-band is bandpass filtered to obtain the target phase data corresponding to each frequency domain sub-band of the structure under test. The time-domain bandpass filter is set with bandpass filtering parameters of each order.

[0087] The time-domain bandpass filter is configured with bandpass filtering parameters of various orders, which may include the center frequency, frequency band values, etc. By using the time-domain bandpass filter to filter the phase data, the target phase data can be obtained.

[0088] In one embodiment, the method for determining the bandpass filter parameters includes: acquiring the displacement-time history curve of any measuring point on the structure under test in the initial structural vibration video; performing Fourier transform processing on the displacement-time history curve data to obtain the natural frequency information of each order of the structure under test; using each of the natural frequencies as the center frequency of the time-domain bandpass filter; and determining the upper and lower limits of each order of the time-domain bandpass filter based on each center frequency and the bandwidth setting, wherein the bandpass filter parameters include the upper and lower limits.

[0089] The measurement point is any pixel in the initial structural vibration image of the structure under test. Through any measurement point, the displacement-time history curve of the measurement point as a function of time can be obtained. Then, Fourier transform processing can be performed on the displacement-time history curve of the measurement point to obtain the corresponding spectrum. From the spectrum, the natural frequency information of each order of the structure under test can be determined. Thus, the natural frequency information of each order is used as the center frequency of the time-domain bandpass filter. Based on the center frequency and the bandwidth setting, the upper and lower limits of the time-domain bandpass filter are determined.

[0090] In one embodiment, after the initial structural vibration video is processed into frames, each frame of the initial structural vibration image can be represented as an image sequence (x, t), where x represents the position and t represents the time. Then, any pixel in the region of the structure under test can be obtained from the image sequence as a measurement point. After determining the measurement point, the measurement point is selected as the displacement tracking calculation point at the video information processing end. Combined with the high-speed camera calibration parameters, the displacement time history curve of the measurement point is obtained, such as... Figure 3 As shown, this is the displacement-time history curve of any measuring point, where... Figure 3 The horizontal axis represents Time (seconds), and the vertical axis represents Displacement (millimeters). The vibration amplitude in intervals 2 and 3 is approximately 1 mm.

[0091] In one embodiment, the natural frequencies of each order are the frequencies corresponding to the extreme points of amplitude in the vibration information of each order of the structure under test, such as... Figure 4 As shown, this is the vibration information of each order obtained after performing a Fourier transform on the displacement-time history curve of the measuring point. Figure 4 The horizontal axis represents frequency in Hz, and the vertical axis represents amplitude. Figure 4 The test structure contains four extreme points, meaning it exhibits four orders of vibration. The four extreme points are 7.559, 11.60, 15.53, and 19.34. The frequencies corresponding to these four extreme points can be used as the first, second, third, and fourth natural frequencies of the test structure. Finally, the natural frequencies of each order can be used as the center frequency of the time-domain bandpass filter.

[0092] In one embodiment, the method further includes:

[0093] When determining the optimal amplification factor for each order of vibration of the structure under test, the following steps are performed:

[0094] Multiple initial magnification factors are obtained, and each initial magnification factor is determined based on the arithmetic sequence rule;

[0095] Based on the initial amplification coefficients, motion amplification processing is performed on the target phase data to obtain amplified composite structure vibration videos.

[0096] The modal confidence criterion is used as the evaluation standard to evaluate each of the magnified synthetic structure videos and obtain modal confidence curves. The modal confidence curves are the fitting curves of each of the initial magnification coefficients and modal confidence values.

[0097] When the modal confidence value meets the preset modal confidence value condition, the corresponding initial amplification factor is selected as the amplification factor.

[0098] The initial amplification factor refers to the coefficient that linearly amplifies the target phase data, increasing its amplitude. The arithmetic progression rule ensures that the differences between initial amplification factors are the same. For example, when selecting the initial amplification factor, it can be determined according to a preset arithmetic progression rule, with increments of 10 and a maximum of 100. Thus, the initial amplification factor can be 10, 20, 30, 40...100, etc. However, since there are target phase data for each order of the structure under test, and the dominant mode in the structural vibration may differ after each excitation, using a fixed amplification factor is not feasible. Therefore, for each order of vibration of the structure under test, different initial amplification factors are selected for comparative experiments. The amplification effect is evaluated to determine the most suitable amplification factor for each order.

[0099] The preset modal confidence value condition can refer to whether the modal confidence value is in a stable state. When it is in a stable state, the preset modal confidence value condition is met. After obtaining the vibration video of each amplified synthetic structure, the modal confidence criterion can be used as the evaluation standard to evaluate each amplified synthetic structure video and obtain the modal confidence curve. The modal confidence curve is the fitting curve between each initial amplification factor and the modal confidence value. The maximum initial amplification factor corresponding to the modal confidence value when in a stable state is taken as the optimal amplification factor.

[0100] As can be seen from formula (1), different motions in an image can be represented by a series of signals with different amplitudes and phases:

[0101]

[0102] Among them, A ω Indicates amplitude, φ ω Let f(x) represent the phase, ω represent the angular frequency, and x represent a pixel in any frame of an image. f(x) can represent the signal representation of a frame of an image.

[0103] As shown in formula (1), the image I(x) at position x and time t can be obtained by substituting variables in the above formula, thus transforming formula (1) from a sine and cosine function into an exponential function, as shown in formula (2), for the convenience of subsequent calculations:

[0104]

[0105] The image difference between two moments can be represented by the phase difference. Therefore, phase difference amplification is used to amplify the motion of the structure, resulting in the amplified image, as shown in the formula below, where α is the magnification factor:

[0106]

[0107] The modal confidence curve is a curve fitted by each initial magnification factor and the MAC (modal confidence criterion) value. It can be used to express the change of the MAC value with each initial magnification factor. The range of the MAC value is

[01] . The larger the value, the closer the test mode shape is to the theoretical mode shape. Figure 5 As shown, this is a detailed diagram illustrating the MAC addresses at various points on the two curves. Figure 6As shown, the confidence curves of one of the modes of the structure under test after amplification are obtained using phase-based (magnification factor) and Riess transform-based methods, respectively. The MAC value shows little change between 0 and 60 times the amplification factor, but drops sharply beyond 60 times. This indicates that for this mode of the structure under test, a magnification factor of 60 times achieves a relatively perfect MAC value and video amplification effect. Analogous to other vibration orders, the MAC value is tested by increasing the amplification factor by 10 times to obtain the optimal amplification factor for each vibration order.

[0108] In one embodiment, after amplifying the phase data of each target, the high-pass residual, low-pass residual, and vibration amplification results can be combined to perform video synthesis processing to obtain multiple target structure vibration videos. At this time, the vibration part of the structure under test in the target structure vibration video has been amplified. Finally, based on the vibration video of each target structure, the displacement time history curves of the structure under test at each order of vibration are obtained.

[0109] In one embodiment, when obtaining the displacement time history curves of the structure under test at each vibration order based on the vibration video of the target structure, displacement tracking processing can be performed on the vibration video of the target structure to obtain the displacement time history curves of the structure under test at each vibration order. Specifically, during processing, a sub-pixel matching algorithm can be used. The sub-pixel matching algorithm is the essence of DIC deformation testing. Based on the iterative initial value provided by the integer pixel initial value search algorithm, the sub-pixel matching algorithm can converge the integer pixel initial value to a local optimum solution at the sub-pixel level, that is, to achieve sub-pixel level deformation testing, thereby improving the accuracy of the structural displacement time history curves.

[0110] In one embodiment, extracting the mode shapes corresponding to each order of the structure under test based on the displacement time history curves includes:

[0111] Multiple marker points are selected on the structure to be tested, and one of the marker points is selected as the research point.

[0112] Based on the displacement-time history curves, obtain the time-displacement curve of the research point within the test time.

[0113] Based on the time displacement curve of the research point, each extreme point is selected from the time displacement curve of the research point, and the time of each extreme point is obtained based on the time corresponding to each extreme point.

[0114] Based on the time of each extreme point, the corresponding mode shape diagram at the extreme point time is determined. The mode shape diagram at the extreme point time is determined by the displacement of each test point at the extreme point time.

[0115] The average value of the mode shape diagrams at each extreme point is taken to obtain the mean mode shape diagrams;

[0116] The mean mode shape diagrams are fitted with sinusoidal functions and normalized to extract the mode shapes corresponding to each order of the structure under test.

[0117] Among these methods, any one of the test points on the structure under test can be selected as the research point. Based on the obtained displacement time history curve, the research point time displacement curve can be extracted from the displacement time history curve. For the research point time displacement curve, there are extreme points (maximum and minimum points). Then, the time corresponding to each extreme point is determined to obtain the time of each extreme point.

[0118] After determining the extreme point time, the displacement of all test points at the extreme point time can be extracted from the displacement-time history curve, thus forming the mode shape diagram at the extreme point time.

[0119] In theory, the vibration of all points on a structure should be the same. However, during measurement, the vibration of each point may differ. Therefore, multiple marker points are selected to reduce variance. The average value of the mode shape diagrams at each extreme point is taken to obtain the mean mode shape diagrams. Finally, the mean mode shapes are normalized and fitted to determine the identification mode shapes of the structure under test at each order, thereby improving the accuracy of mode shape identification.

[0120] In one embodiment, such as Figure 7 The diagram shown is a flowchart of obtaining the identified vibration mode in a specific embodiment:

[0121] In this method, after performing motion magnification processing on the initial structural vibration video, the magnified vibration of each order of the structure was directly obtained, such as... Figure 7 The specific identification process is as follows: First, 7(a) is the structural displacement time history curve. A point can be selected from all the marker points in the structural displacement time history curve as the research point. Then, based on the structural displacement time history curve, a time-displacement curve of the research point within the test time is generated, and each extreme point in the time-displacement curve of the research point is selected, along with the time of occurrence of each extreme point. Then, the displacement of all marker points at each extreme point is obtained, resulting in a preliminary mode shape diagram (e.g., ...). Figure 7 (b)) The mode shape diagram can be processed to take the average value (e.g. Figure 7 (c) shows the line where the solid line is derived from the positive extreme point and the dashed line is derived from the negative extreme point; Figure 7 (c) Perform sine function fitting and normalization, the result is as follows: Figure 7 (d), where Figure 7 (d) is the final identified vibration mode.

[0122] In one embodiment, such as Figure 8The diagram shown is a flowchart illustrating a structural modal recognition method based on digital image correlation and motion magnification techniques in a specific embodiment.

[0123] First, the original video collected by DIC measurement and identification is obtained. The initial structural vibration video can refer to the video obtained by DIC (Digital Image Correlation) measurement equipment (such as a high-speed camera) for structural free decay vibration video acquisition. In addition to the vibration information of the structure under test (such as the cable tower of a bridge), the original video may also contain vibration information of other structures (such as moving cars, bridge railings, etc.) and vibration information of passing pedestrians.

[0124] When acquiring raw video using DIC measurement equipment, the camera's imaging plane is parallel to the surface of the structure under test. Simultaneously, camera calibration is performed. Specifically, the surface of the structure under test is transformed to the camera's imaging plane using a projection matrix. Combined with camera calibration, the correspondence between the actual coordinates and the camera coordinates is determined. By making the surface of the structure under test as parallel as possible to the camera's imaging plane, the matrix can be easily obtained, reducing calculation errors.

[0125] The original video is then processed. Specifically, the original video is divided into frames to decompose it into individual original structural vibration images. Fourier transform is performed on each original vibration image to obtain the corresponding frequency domain structural vibration images. Then, a complex steerable pyramid is used to decompose the video according to spatial scale (pixels), orientation, and position to obtain the amplitude and phase information after local wavelet transform. Finally, the phase change time domain data corresponding to each frequency domain sub-band is determined.

[0126] Furthermore, based on the values ​​of each filtering parameter, time-domain bandpass filtering is performed on the phase change time-domain data corresponding to each frequency sub-band. Since the values ​​of the filtering parameters are determined based on the vibration information of the structure under test, time-domain bandpass filtering can retain only the phase change data of the frequency band of the structure under test, while filtering out the phase change data of other structures besides the structure under test, and the finally retained phase change data is used as the target phase data corresponding to the structure under test.

[0127] After time-domain bandpass filtering, noise reduction and amplification can be performed. Specifically, when performing motion amplification calculations with each natural frequency as the center frequency, different amplification factors are selected for comparative experiments. The amplification effect is evaluated to determine the most suitable amplification factor for each frequency to achieve a better amplification effect.

[0128] After amplifying the phase data of each target, the high-pass residual, low-pass residual, and amplification results of each structural mode can be combined to perform video synthesis processing to obtain multiple target structure vibration videos. At this time, the vibration part of the structure under test in the target vibration video has been amplified. Finally, based on the vibration videos of each target structure, the structural displacement time history curves of the structure under test at each order can be obtained.

[0129] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0130] Based on the same inventive concept, this application also provides a structural modal recognition device for implementing the structural modal recognition method based on digital image correlation and motion magnification technology described above. The solution provided by this device is similar to the implementation described in the above method; therefore, the specific limitations of one or more structural modal recognition device embodiments provided below can be found in the limitations of the structural modal recognition method based on digital image correlation and motion magnification technology described above, and will not be repeated here.

[0131] In one embodiment, such as Figure 9 As shown, a structural modal identification device is provided, comprising: a data acquisition module, a data processing module, a filtering module, a motion amplification module, a curve acquisition module, and a mode shape extraction module, wherein:

[0132] The data acquisition module 902 is used to acquire the initial structural vibration video.

[0133] The data processing module 904 is used to decompose the spectrogram corresponding to each frame of the initial structure vibration video into multiple frequency domain sub-bands.

[0134] The filtering module 906 is used to perform bandpass filtering on the phase data corresponding to each frequency domain sub-band to obtain the target phase data corresponding to each frequency domain sub-band of the structure under test.

[0135] The motion amplification module 908 is used to perform motion amplification processing on the target phase data based on the optimal amplification factor corresponding to each order of vibration of the structure under test, so as to obtain the vibration amplification results of each order of the structure under test.

[0136] The curve acquisition module 910 is used to obtain the displacement time history curves of each measuring point on the structure under test during each order of vibration based on the vibration amplification results.

[0137] The mode shape extraction module 912 is used to extract the mode shapes corresponding to each order of the structure under test based on the displacement time history curves.

[0138] In one embodiment, the apparatus further includes a frame processing module;

[0139] The frame-segmentation processing module is used to perform frame-segmentation processing on the initial structural vibration video, decomposing the initial structural vibration video into a sequence of initial structural vibration images; and performing two-dimensional discrete Fourier transform processing on each initial structural vibration image sequence to obtain the spectrum corresponding to each frame of the initial structural vibration image.

[0140] In one embodiment, the data processing module is further configured to perform downsampling processing on each frame of the initial structural vibration image of the initial structural vibration video based on a preset downsampling ratio to obtain sub-images at multiple scales; and to decompose each of the spectrum maps into multiple frequency domain sub-bands according to the amplitude information and phase information contained in each pixel point in each spectrum map.

[0141] In one embodiment, the filtering module is used to perform bandpass filtering on the phase data corresponding to each frequency sub-band based on a time-domain bandpass filter to obtain the target phase data corresponding to each frequency sub-band of the structure under test. The time-domain bandpass filter is configured with bandpass filtering parameters of each order.

[0142] In one embodiment, the filtering module is used to determine the natural frequencies of the structure under test based on the vibration information of the structure under test; and to use each natural frequency as the center frequency of the time-domain bandpass filter, wherein the filtering parameter values ​​include the center frequency values.

[0143] In one embodiment, the bandpass filter parameters include a center frequency, an upper limit frequency, and a lower limit frequency. The filtering module is used to acquire the displacement-time history curve of any measuring point on the structure under test in the initial structural vibration video; perform Fourier transform processing on the displacement-time history curve to obtain the natural frequency information of each order of the structure under test; use each of the natural frequency information as the center frequency of the time-domain bandpass filter; and determine the upper limit frequency and the lower limit frequency based on each center frequency and the bandwidth setting.

[0144] In one embodiment, the device further includes: an optimal amplification factor determination module;

[0145] The optimal amplification factor determination module acquires multiple initial amplification factors, each determined based on an arithmetic sequence rule. Based on each initial amplification factor, motion amplification processing is performed on the target phase data to obtain amplified composite structure vibration videos. The modal confidence criterion is used as an evaluation standard to evaluate each amplified composite structure video, obtaining a modal confidence curve, which is a fitting curve between each initial amplification factor and the modal confidence value. When the modal confidence value meets a preset modal confidence value condition, the corresponding maximum initial amplification factor is selected as the optimal amplification factor.

[0146] In one embodiment, the mode shape extraction module is used to select multiple marker points on the structure under test, and select one point from each marker point as a study point; obtain the study point time-displacement curve within the test time according to each displacement-time history curve; filter out each extreme point from the study point time-displacement curve based on the study point time-displacement curve, and obtain the time of each extreme point based on the time corresponding to each extreme point; determine the corresponding mode shape diagram of each extreme point time according to each extreme point time, the mode shape diagram of each extreme point time is determined by the displacement of each marker point at the extreme point time; take the average value of each mode shape diagram of each extreme point time to obtain each mean mode shape diagram; perform sine function fitting and normalization on each mean mode shape diagram to extract the mode shape corresponding to each mode of the structure under test.

[0147] Each module in the aforementioned structural modality recognition device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0148] In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 10As shown, the computer device includes a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, NFC (Near Field Communication), or other technologies. When executed by the processor, the computer program implements a structural modal recognition method based on digital image correlation and motion magnification techniques. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0149] Those skilled in the art will understand that Figure 10 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0150] In one embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps of the above-described structural modal recognition method based on digital image correlation and motion magnification techniques.

[0151] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, which, when executed by a processor, implements the steps of the above-described structural modal recognition method based on digital image correlation and motion magnification techniques.

[0152] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps of the above-described structural modal recognition method based on digital image correlation and motion magnification techniques.

[0153] It should be noted that the data involved in this application (including but not limited to data used for analysis, stored data, and displayed data) are all information and data authorized by the user or fully authorized by all parties.

[0154] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0155] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0156] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A structural mode recognition method based on digital image correlation and motion magnification techniques, characterized in that, The method includes: Obtain initial structural vibration video; The spectrogram corresponding to each frame of the initial structural vibration video is decomposed into multiple frequency domain sub-bands. Bandpass filtering is performed on the phase data corresponding to each frequency domain sub-band to obtain the target phase data corresponding to each frequency domain sub-band of the structure under test. Based on the optimal amplification factor corresponding to each order of vibration of the structure under test, motion amplification processing is performed on the target phase data to obtain the vibration amplification results of each order of the structure under test. Based on the vibration amplification results, the displacement-time history curves of each measuring point on the structure under test during each order of vibration are obtained. Based on the displacement-time history curves, the vibration modes corresponding to each mode of the structure under test are extracted. When determining the optimal amplification factor for each order of vibration of the structure under test, the following steps are performed: Multiple initial magnification factors are obtained, and each initial magnification factor is determined based on the arithmetic sequence rule; Based on the initial amplification coefficients, motion amplification processing is performed on the target phase data to obtain amplified composite structure vibration videos. The modal confidence criterion is used as the evaluation standard to evaluate each of the magnified synthetic structure videos and obtain modal confidence curves. The modal confidence curves are the fitting curves of each of the initial magnification coefficients and modal confidence values. When the modal confidence value meets the preset modal confidence value condition, the corresponding maximum initial amplification factor is selected as the optimal amplification factor.

2. The method according to claim 1, characterized in that, After acquiring the initial structural vibration video, and before decomposing the spectrogram corresponding to each frame of the initial structural vibration video, the process further includes: The initial structure vibration video is segmented into frames to decompose it into a sequence of initial structure vibration images. Two-dimensional discrete Fourier transform processing is performed on each of the initial structural vibration image sequences to obtain the spectrum corresponding to each frame of the initial structural vibration image.

3. The method according to claim 1, characterized in that, The process of decomposing the spectrogram corresponding to each frame of the initial structural vibration video into multiple frequency domain sub-bands includes: Based on a preset downsampling ratio, each frame of the initial structural vibration video is downsampled to obtain sub-images at multiple scales. Based on the amplitude and phase information contained in each pixel in each of the aforementioned spectrograms, each of the aforementioned spectrograms is decomposed into multiple frequency domain sub-bands.

4. The method according to claim 1, characterized in that, The step of performing bandpass filtering on the phase data corresponding to each frequency domain sub-band to obtain the target phase data corresponding to each frequency domain sub-band of the structure under test includes: Based on a time-domain bandpass filter, the phase data corresponding to each frequency domain sub-band is bandpass filtered to obtain the target phase data corresponding to each frequency domain sub-band of the structure under test. The time-domain bandpass filter is set with bandpass filtering parameters of each order.

5. The method according to claim 4, characterized in that, The bandpass filter parameters include a center frequency, an upper limit frequency, and a lower limit frequency. The determination of the bandpass filter parameters includes: Obtain the displacement-time history curve of any measuring point on the structure under test in the initial structural vibration video; The displacement-time history curves of the measuring points are processed by Fourier transform to obtain the natural frequency information of each order of the structure under test. The natural frequency information of each of the above is used as the center frequency of the time-domain bandpass filter; Based on the respective center frequencies and the bandwidth settings, the upper limit frequency and the lower limit frequency are determined.

6. The method according to claim 1, characterized in that, The step of extracting the mode shapes corresponding to each order of the structure under test based on the displacement time history curves includes: Multiple marker points are selected on the structure to be tested, and one of the marker points is selected as the research point. Based on the displacement-time history curves, obtain the time-displacement curve of the research point within the test time. Based on the time displacement curve of the research point, each extreme point is selected from the time displacement curve of the research point, and the time of each extreme point is obtained based on the time corresponding to each extreme point. Based on the time of each extreme point, the corresponding mode shape diagram at the extreme point time is determined. The mode shape diagram at the extreme point time is determined by the displacement of each of the marked points at the extreme point time. The average value of the mode shape diagrams at each extreme point is taken to obtain the mean mode shape diagrams; The mean mode shape diagrams are fitted with sinusoidal functions and normalized to extract the mode shapes corresponding to each order of the structure under test.

7. A structural modal recognition device based on digital image correlation and motion magnification technology, characterized in that, The device includes: The data acquisition module is used to acquire initial structural vibration videos; The data processing module is used to decompose the spectrogram corresponding to each frame of the initial structure vibration video into multiple frequency domain sub-bands. The filtering module is used to perform bandpass filtering on the phase data corresponding to each frequency domain sub-band to obtain the target phase data corresponding to each frequency domain sub-band of the structure under test. The motion amplification module is used to perform motion amplification processing on the target phase data based on the optimal amplification factor corresponding to each order of vibration of the structure under test, so as to obtain the vibration amplification results of each order of the structure under test. The curve acquisition module is used to obtain the displacement time history curves of each measuring point on the structure under test under each order of vibration based on the vibration amplification results. The mode shape extraction module is used to extract the mode shapes corresponding to each mode of the structure under test based on the displacement time history curves. The optimal amplification factor determination module is used to obtain multiple initial amplification factors, each of which is determined based on an arithmetic sequence rule; based on each initial amplification factor, motion amplification processing is performed on the target phase data to obtain amplified composite structure vibration videos; the modal confidence criterion is used as an evaluation standard to evaluate each amplified composite structure video to obtain a modal confidence curve, which is a fitting curve between each initial amplification factor and the modal confidence value; when the modal confidence value meets the preset modal confidence value condition, the corresponding maximum initial amplification factor is selected as the optimal amplification factor.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images