A video vibration calibration method integrating video micro-vibration monitoring and sensor

Abstract

The invention discloses a video vibration calibration method integrating video micro-vibration monitoring and sensors, which comprises dividing a hydraulic monitoring structure into a plurality of partitioned fields of view, setting primary monitoring points and secondary monitoring points, and arranging vibration sensors; applying an excitation signal to the hydraulic monitoring structure, acquiring video information of the hydraulic monitoring structure and a vibration signal time history of the vibration sensor; calculating displacement information of pixel points corresponding to all primary monitoring points and secondary monitoring points according to the video information, and performing Fourier transform to obtain a spectrum; calibrating pixel points corresponding to all primary monitoring points by using ensemble empirical mode decomposition and BP neural network; calibrating pixel points corresponding to secondary monitoring points according to the transfer function and spectrum of the secondary monitoring points; training a neural network by using videos before and after the calibration of the primary monitoring points and the secondary monitoring points, and calibrating pixel points of non-primary monitoring points and secondary monitoring points by using the trained neural network.

Description

Technical Field

[0001] The invention relates to a video calibration technology, and in particular to a video vibration calibration method integrating video micro-vibration monitoring and a sensor. Background Art

[0002] The safety and stability of hydraulic structures are important factors affecting the long-term safe operation of water conservancy projects. In recent years, with the continuous advancement of technology, the health monitoring technology of hydraulic structures based on video micro-vibration has become increasingly necessary as an emerging monitoring method. Video micro-vibration technology can capture tiny vibrations on the surface of the structure through high-resolution cameras, and then analyze the dynamic response and performance status of the structure. However, due to the influence of environmental noise and human operation, the accuracy and reliability of video micro-vibration monitoring data are often affected. Therefore, it is particularly important to calibrate the video micro-vibration monitoring data. Through effective calibration methods, the errors in the monitoring data can be eliminated or reduced, the accuracy and credibility of the data can be ensured, and strong support can be provided for the safety assessment and maintenance decision-making of hydraulic structures.

[0003] The use of vibration sensors for vibration information extraction and calibration of new monitoring methods has always played an important role in the monitoring of hydraulic structures. However, this method has exposed many shortcomings in actual operation. First, the vibration modal analysis of hydraulic structures relies on a large amount of monitoring data. In order to ensure the comprehensiveness and accuracy of the monitoring data, a large number of sensors need to be deployed at multiple key locations, which not only increases the complexity and cost of deployment, but also makes the overall maintenance and management of the monitoring system more difficult; secondly, given the unique wading characteristics of hydraulic structures, some key monitoring points, such as underwater structures and hidden parts, are extremely difficult or even impossible to install sensors due to harsh environments or physical accessibility restrictions, which seriously affects the integrity and representativeness of the monitoring data. In addition, traditional vibration sensors are also susceptible to environmental factors such as electromagnetic interference and temperature changes, which will introduce additional measurement errors, thereby reducing the reliability and accuracy of monitoring data. Summary of the invention

[0004] In view of the above-mentioned deficiencies in the prior art, the present invention provides a method for integrating video micro-vibration monitoring and sensor video vibration calibration, which solves the problems of poor monitoring data integrity and low monitoring accuracy in the prior art using vibration sensors for monitoring.

[0005] In order to achieve the above-mentioned object of the invention, the technical solution adopted by the present invention is:

[0006] A video vibration calibration method integrating video micro-vibration monitoring and sensor is provided, characterized in that it comprises the steps of:

[0007] S1. Divide the hydraulic monitoring structure into several sub-fields of view, set a primary monitoring point and a secondary monitoring point in each sub-field of view, and arrange vibration sensors at the primary monitoring point and the secondary monitoring point of the first sub-field of view;

[0008] S2, applying an excitation signal to the hydraulic monitoring structure, and obtaining video information of the hydraulic monitoring structure and a vibration signal time history of each vibration sensor;

[0009] S3. Calculate the displacement information of the pixels corresponding to all the primary monitoring points and the secondary monitoring points according to the video information, and perform Fourier transform on the displacement information to obtain the frequency spectrum of the primary monitoring points and the secondary monitoring points;

[0010] S4. According to the frequency spectrum and vibration signal time history corresponding to the first-level monitoring points, the pixel points corresponding to all the first-level monitoring points are calibrated using the ensemble empirical mode decomposition and BP neural network;

[0011] S5. Calibrate the spectrum corresponding to the secondary monitoring point according to the transfer function and spectrum of the secondary monitoring point, and then invert the calibrated spectrum to obtain a calibrated video of the secondary monitoring point;

[0012] S6. Use the videos of the primary monitoring points and the secondary monitoring points before and after calibration to train the neural network, and use the trained neural network to calibrate the pixels of the non-primary monitoring points and the non-secondary monitoring points.

[0013] The beneficial effects of the present invention are as follows: this scheme can calibrate some pixels in the video by arranging a small number of vibration sensors and combining them with the transfer function of the constructed secondary monitoring points, and then train the neural network based on the video before and after the pixel calibration, so as to capture the correlation before and after the video calibration, and finally calibrate the remaining pixels in combination with the trained neural network, thereby ensuring the accuracy of the monitoring data of the hydraulic monitoring structure.

[0014] Furthermore, the method for constructing the transfer function of the secondary monitoring point includes:

[0015] S51, applying p excitations to the hydraulic monitoring structure, and obtaining the vibration signal time history of the vibration sensors at the primary monitoring point and the secondary monitoring point under each excitation;

[0016] S52, Fourier transform the vibration signal time history of the primary monitoring point and the secondary monitoring point to obtain the frequency spectrum of the primary monitoring point and the secondary monitoring point, and calculate the cross power spectrum of the primary monitoring point and the secondary monitoring point and the auto power spectrum of the secondary monitoring point:

[0017]

[0018]

[0019] in, is the cross power spectrum between the primary monitoring point and the secondary monitoring point m in the first partitioned field of view; is the auto-power spectrum of the secondary monitoring point m in the first partitioned field of view; F mi (f) is the spectrum of the secondary monitoring point m in the first partitioned field of view under the i-th excitation; F mi (f) is the conjugate; T is 1 / 2 of the sampling time; p is the total number of excitations applied, j = i + 1; X i (f) is the spectrum of the first-level monitoring point in the first partition field of view under the i-th excitation;

[0020] S53, constructing a transfer function of the secondary monitoring point in the first partitioned field of view according to the cross power spectrum and the auto power spectrum of the secondary monitoring point in the first partitioned field of view:

[0021]

[0022] Among them, H 1m (f) is the transfer function of the secondary monitoring point m in the first partition field of view;

[0023] S54, constructing a finite element simulation model of the hydraulic monitoring structure, and using the vibration signal time history of the primary monitoring point and the secondary monitoring point in the first partition field of view as the model boundary condition, and extracting the vibration time history curves of the primary monitoring point and the secondary monitoring point in the remaining partition field of view;

[0024] S55 , according to the vibration time history curves of the primary monitoring points and the secondary monitoring points in the remaining sub-fields of view, the transfer function of each secondary monitoring point in the remaining sub-fields of view is obtained by adopting the method of steps S52 to S53 .

[0025] The beneficial effects of the above technical solution are: the transfer function based on frequency response actually reflects the system's transfer ability to different input signals, and is a non-parametric estimation model that describes the characteristics of a dynamic system. After the primary monitoring point and the secondary monitoring point are determined, the vibration system formed between the two is not affected by changes in the form of vibration source excitation, and therefore has a calibration function. However, due to the inevitable interference of factors such as noise in the actual application process, the accuracy of the transfer function will be affected. This solution proposes to use the transfer function calculated by multiple averaging techniques and cross-spectrum techniques to eliminate the influence of noise and obtain an unbiased, high-precision estimate of the transfer function.

[0026] Furthermore, step S54 further includes:

[0027] According to the geometric characteristics of the hydraulic monitoring structure, a finite element simulation model of the hydraulic monitoring structure is constructed;

[0028] Use concrete with the same mix ratio and raw materials as the hydraulic monitoring structure to carry out physical model tests and obtain the mechanical parameters of concrete;

[0029] The vibration signal time history of the primary and secondary monitoring points in the first partitioned field of view is used as the model boundary condition to perform dynamic calculations and extract the vibration time history curves of the primary and secondary monitoring points in the remaining partitioned fields of view.

[0030] This solution further reduces the need for sensors by calculating the signals of the remaining secondary monitoring points through accurate finite element models.

[0031] Furthermore, the expression for calibrating the spectrum corresponding to the secondary monitoring point is:

[0032] X * nm (f) = X nm (f)*H' n-nm (f)

[0033] Among them, X nm (f) and X * nq (f) are the spectra of the secondary monitoring point m of the nth field of view partition before and after correction; H′ n-nm (f) is the transfer function of the secondary monitoring point m in the nth field of view partition, n>1.

[0034] Furthermore, the expression for calculating the displacement information of the pixel points corresponding to the primary monitoring points and the secondary monitoring points is:

[0035]

[0036]

[0037] Among them, μ and θ are the horizontal and vertical directions of the pixel, respectively. 0 The instantaneous speed of time, is the full video domain phase image at different scales and directions; x and y are the vibration velocities of the pixel in the horizontal and vertical directions, respectively.

[0038] The beneficial effects of the above technical solution are: the non-contact monitoring method of video, due to its characteristics of extracting structural vibration information from the video, has the advantages of not changing the structural dynamic characteristics, being able to perform multi-point measurements, and being adaptable to large-scale long-distance measurements. This solution utilizes the principle that the pixel phase in the video corresponds to the structural vibration, and the theoretical accuracy is the same as that of the vibration sensor.

[0039] Furthermore, step S4 further comprises:

[0040] S41, using ensemble empirical mode decomposition to decompose the frequency spectrum and vibration signal time history corresponding to the primary monitoring point to obtain a number of frequency spectrum waveforms;

[0041] S42, obtaining a waveform diagram of a linear transformation in the frequency spectrum waveform of the vibration signal time course, and performing amplitude fitting on frequency spectrum waveforms corresponding to the waveform diagram of the linear transformation in several frequency spectrum waveforms of the frequency spectrum;

[0042] S43, performing Gabor transformation on the spectrum waveform after amplitude fitting to obtain the vibration time history corresponding to the spectrum waveform;

[0043] S44, inputting the remaining spectrum waveforms among the plurality of spectrum waveforms of the spectrum into the trained BP neural network for calibration, to obtain a calibrated spectrum waveform;

[0044] S45, reconstructing the vibration time history and calibrated spectral waveforms corresponding to all spectral waveforms of the same primary monitoring point to obtain a calibrated video of the pixel points corresponding to the primary monitoring point.

[0045] The beneficial effects of the above technical solution are as follows: under ideal conditions, the vibration signal of the water-excited structure is continuous and smooth, and can be expressed as the superposition of harmonic (sine wave or cosine wave) functions of different frequencies, which can also be regarded as the linear part in the vibration energy transfer and can be calibrated by EEMD; however, although the nonlinear influencing factors mixed in the vibration signal actually obtained have little effect on the structure's own information such as frequency and vibration mode, this part of the energy will affect the structural amplitude, which can be calibrated by the BP neural network.

[0046] Furthermore, before calibrating the pixel points corresponding to the first-level monitoring points, the following steps are also included:

[0047] Calculate the absolute value of the difference between the spectrum of the first-level monitoring point and the vibration signal time course, and determine whether it is less than the deviation threshold. If so, calibrate the pixel points corresponding to the first-level monitoring point. Otherwise, do not calibrate the pixel points corresponding to the first-level monitoring point. This solution can further reduce the workload by determining whether it is necessary to calibrate the spectrum extracted from the corresponding pixel position of the video.

[0048] Furthermore, the camera's field of view only accommodates one partitioned field of view. When collecting video information of the hydraulic monitoring structure, the camera is used to record the video of each partitioned field of view in turn. For large-volume structures, if the camera is used for full-area monitoring, the resolution will be poor, and the phase change of the pixel will not be enough to reflect the small vibration of the structure. Therefore, partitioned video monitoring and improving the relative resolution of the video can increase the image's detailed information, improve the frequency response function signal-to-noise ratio and vibration information accuracy.

[0049] Furthermore, each partitioned field of view is provided with 4 secondary monitoring points.

[0050] The beneficial effect of the above technical solution is that the secondary monitoring points are arranged around each partition. Therefore, when the correction of the secondary monitoring points is completed, it means that the vibration conditions of the boundary of the area have been mastered. When the vibration conditions of other video pixel points in the partition are obtained, the reliability can be preliminarily judged based on the results of the secondary monitoring points. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 Flowchart of the method for fusion video micro-vibration monitoring and video vibration calibration of sensors.

[0052] Figure 2 This is a schematic diagram of the arrangement of primary and secondary monitoring points in the hydraulic monitoring structure. DETAILED DESCRIPTION

[0053] The specific implementation modes of the present invention are described below so that those skilled in the art can understand the present invention. However, it should be clear that the present invention is not limited to the scope of the specific implementation modes. For those of ordinary skill in the art, as long as various changes are within the spirit and scope of the present invention as defined and determined by the attached claims, these changes are obvious, and all inventions and creations utilizing the concept of the present invention are protected.

[0054] refer to Figure 1 , Figure 1 A flow chart of a method for integrating video micro-vibration monitoring and sensor video vibration calibration is shown; Figure 1 As shown, the method S includes steps S1 to S6.

[0055] In step S1, the hydraulic monitoring structure is divided into several sub-fields of view, each sub-field of view is provided with a primary monitoring point and a secondary monitoring point, and vibration sensors are arranged at the primary monitoring point and the secondary monitoring point of the first sub-field of view.

[0056] This scheme preferably sets up 4 secondary monitoring points in the partitioned field of view, and arranges the views of the primary and secondary monitoring points. You can refer to Figure 2 , among which Mo1~Mo6 are primary monitoring points, and S-Mo1~S-Mo4 are secondary monitoring points.

[0057] In step S2, an excitation signal is applied to the hydraulic monitoring structure, and video information of the hydraulic monitoring structure and the vibration signal time course of each vibration sensor are obtained; when collecting video information of the hydraulic monitoring structure, the field of view of the camera only accommodates one partitioned field of view, and the camera is used to record video for each partitioned field of view in turn.

[0058] In step S3, displacement information of pixel points corresponding to all primary monitoring points and secondary monitoring points is calculated based on the video information, and Fourier transform is performed on the displacement information to obtain the frequency spectrum of the primary monitoring points and secondary monitoring points.

[0059] During implementation, the solution preferably calculates the displacement information of the pixel points corresponding to the primary monitoring points and the secondary monitoring points as follows:

[0060]

[0061]

[0062] Among them, μ and θ are the horizontal and vertical directions of the pixel, respectively. 0 The instantaneous speed of time, is the full video domain phase image at different scales and directions; x and y are the vibration velocities of the pixel in the horizontal and vertical directions, respectively.

[0063] During implementation, the solution preferably further includes the following steps before calibrating the pixel points corresponding to the primary monitoring points:

[0064] Calculate the absolute value of the difference between the spectrum of the first-level monitoring point and the vibration signal time history, and determine whether it is less than the deviation threshold. If so, calibrate the pixel points corresponding to the first-level monitoring point. Otherwise, do not calibrate the pixel points corresponding to the first-level monitoring point.

[0065] In step S4, based on the frequency spectrum and vibration signal time history corresponding to the first-level monitoring points, the pixel points corresponding to all the first-level monitoring points are calibrated using the ensemble empirical mode decomposition and BP neural network;

[0066] In one embodiment of the present invention, step S4 further comprises:

[0067] S41, using ensemble empirical mode decomposition to decompose the frequency spectrum and vibration signal time history corresponding to the primary monitoring point to obtain a number of frequency spectrum waveforms;

[0068] S42, obtaining a waveform diagram of a linear transformation in the frequency spectrum waveform of the vibration signal time course, and performing amplitude fitting on frequency spectrum waveforms corresponding to the waveform diagram of the linear transformation in several frequency spectrum waveforms of the frequency spectrum;

[0069] S43, performing Gabor transformation on the spectrum waveform after amplitude fitting to obtain the vibration time history corresponding to the spectrum waveform;

[0070] S44, inputting the remaining spectrum waveforms among the plurality of spectrum waveforms of the spectrum into the trained BP neural network for calibration, to obtain a calibrated spectrum waveform;

[0071] S45, reconstructing the vibration time history and calibrated spectral waveforms corresponding to all spectral waveforms of the same primary monitoring point to obtain a calibrated video of the pixel points corresponding to the primary monitoring point.

[0072] In step S5, the spectrum corresponding to the secondary monitoring point is calibrated according to the transfer function and spectrum of the secondary monitoring point, and then the calibrated spectrum is inverted to obtain the calibrated video of the secondary monitoring point;

[0073] The expression for calibrating the spectrum corresponding to the secondary monitoring point is:

[0074] X * nm (f) = X nm (f)*H′ n-nm (f)

[0075] Among them, X nm (f) and X * nq (f) are the spectra of the secondary monitoring point m of the nth field of view partition before and after correction; H′ n-nm (f) is the transfer function of the secondary monitoring point m in the nth field of view partition, n>1.

[0076] The inversion of the calibrated spectrum is specifically to perform an inverse Fourier transform on the calibrated spectrum to obtain displacement information, and then use an expression of the displacement information to obtain a video after pixel calibration.

[0077] In one embodiment of the present invention, a method for constructing a transfer function of a secondary monitoring point includes:

[0078] S51, applying p excitations to the hydraulic monitoring structure, and obtaining the vibration signal time history of the vibration sensors at the primary monitoring point and the secondary monitoring point under each excitation;

[0079] S52, Fourier transform the vibration signal time history of the primary monitoring point and the secondary monitoring point to obtain the frequency spectrum of the primary monitoring point and the secondary monitoring point, and calculate the cross power spectrum of the primary monitoring point and the secondary monitoring point and the auto power spectrum of the secondary monitoring point:

[0080]

[0081]

[0082] in, is the cross power spectrum between the primary monitoring point and the secondary monitoring point m in the first partitioned field of view; is the auto-power spectrum of the secondary monitoring point m in the first partitioned field of view; F mi (f) is the spectrum of the secondary monitoring point m in the first partitioned field of view under the i-th excitation; F mi (f) is the conjugate; T is 1 / 2 of the sampling time; p is the total number of excitations applied, j = i + 1; X i (f) is the spectrum of the first-level monitoring point in the first partition field of view under the i-th excitation;

[0083] S53, constructing a transfer function of the secondary monitoring point in the first partitioned field of view according to the cross power spectrum and the auto power spectrum of the secondary monitoring point in the first partitioned field of view:

[0084]

[0085] Among them, H 1m (f) is the transfer function of the secondary monitoring point m in the first partition field of view;

[0086] S54, constructing a finite element simulation model of the hydraulic monitoring structure, and using the vibration signal time history of the primary monitoring point and the secondary monitoring point in the first partition field of view as the model boundary condition, and extracting the vibration time history curves of the primary monitoring point and the secondary monitoring point in the remaining partition field of view;

[0087] S55 , according to the vibration time history curves of the primary monitoring points and the secondary monitoring points in the remaining sub-fields of view, the transfer function of each secondary monitoring point in the remaining sub-fields of view is obtained by adopting the method of steps S52 to S53 .

[0088] In step S6, the videos of the primary monitoring points and the secondary monitoring points before and after calibration are used to train the neural network, and the trained neural network is used to calibrate the pixels of the non-primary monitoring points and the non-secondary monitoring points.

[0089] During implementation, the preferred step S54 of this solution further includes:

[0090] According to the geometric characteristics of the hydraulic monitoring structure, a finite element simulation model of the hydraulic monitoring structure is constructed;

[0091] Use concrete with the same mix ratio and raw materials as the hydraulic monitoring structure to carry out physical model tests and obtain the mechanical parameters of concrete;

[0092] The vibration signal time history of the primary and secondary monitoring points in the first partitioned field of view is used as the model boundary condition to perform dynamic calculations and extract the vibration time history curves of the primary and secondary monitoring points in the remaining partitioned fields of view.

[0093] In summary, when the calibration method of this scheme is used to calibrate the video, only a small number of vibration sensors need to be deployed to calibrate the monitoring data of the hydraulic monitoring structure with high accuracy.

Claims

1. A video vibration calibration method integrating video micro-vibration monitoring and sensor, characterized in that: Includes steps: S1. Divide the hydraulic monitoring structure into several sub-fields of view, set a primary monitoring point and a secondary monitoring point in each sub-field of view, and arrange vibration sensors at the primary monitoring point and the secondary monitoring point of the first sub-field of view; S2, applying an excitation signal to the hydraulic monitoring structure, and obtaining video information of the hydraulic monitoring structure and a vibration signal time history of each vibration sensor; S3. Calculate the displacement information of the pixels corresponding to all the primary monitoring points and the secondary monitoring points according to the video information, and perform Fourier transform on the displacement information to obtain the frequency spectrum of the primary monitoring points and the secondary monitoring points; S4. According to the frequency spectrum and vibration signal time history corresponding to the first-level monitoring points, the pixel points corresponding to all the first-level monitoring points are calibrated using the ensemble empirical mode decomposition and BP neural network; S5. Calibrate the spectrum corresponding to the secondary monitoring point according to the transfer function and spectrum of the secondary monitoring point, and then invert the calibrated spectrum to obtain a calibrated video of the secondary monitoring point; S6, using the videos before and after calibration of the primary monitoring points and the secondary monitoring points to train the neural network, and using the trained neural network to calibrate the pixels of the non-primary monitoring points and the non-secondary monitoring points; Step S4 further comprises: S41, using ensemble empirical mode decomposition to decompose the frequency spectrum and vibration signal time history corresponding to the primary monitoring point to obtain a number of frequency spectrum waveforms; S42, obtaining a waveform diagram of a linear transformation in the frequency spectrum waveform of the vibration signal time course, and performing amplitude fitting on frequency spectrum waveforms corresponding to the waveform diagram of the linear transformation in several frequency spectrum waveforms of the frequency spectrum; S43, performing Gabor transformation on the spectrum waveform after amplitude fitting to obtain the vibration time history corresponding to the spectrum waveform; S44, inputting the remaining spectrum waveforms among the plurality of spectrum waveforms of the spectrum into the trained BP neural network for calibration, to obtain a calibrated spectrum waveform; S45, reconstructing the vibration time history and calibrated spectral waveforms corresponding to all spectral waveforms of the same primary monitoring point to obtain a calibrated video of the pixel points corresponding to the primary monitoring point.

2. The video vibration calibration method of the fusion video micro-vibration monitoring and sensor according to claim 1 is characterized in that: The construction method of the transfer function of the secondary monitoring point includes: S51. Apply to hydraulic monitoring structures p excitation, obtaining the vibration signal time history of the vibration sensors at the primary monitoring point and the secondary monitoring point under each excitation; S52, Fourier transform the vibration signal time history of the primary monitoring point and the secondary monitoring point to obtain the frequency spectrum of the primary monitoring point and the secondary monitoring point, and calculate the cross power spectrum of the primary monitoring point and the secondary monitoring point and the auto power spectrum of the secondary monitoring point: in, The first and second monitoring points in the first partition field of view m The cross power spectrum of Secondary monitoring point in the first partition field of view m The autopower spectrum of Secondary monitoring point in the first partition field of view m In the i Spectrum under sub-excitation; for conjugation of; T 1 / 2 of the sampling time; p is the total number of excitations applied, j = i +1; The first level monitoring point in the first partition field of view is i Spectrum under sub-excitation; S53, constructing a transfer function of the secondary monitoring point in the first partitioned field of view according to the cross power spectrum and the auto power spectrum of the secondary monitoring point in the first partitioned field of view: in, Secondary monitoring point in the first partition field of view m The transfer function of S54, constructing a finite element simulation model of the hydraulic monitoring structure, and using the vibration signal time history of the primary monitoring point and the secondary monitoring point in the first partition field of view as the model boundary condition, and extracting the vibration time history curves of the primary monitoring point and the secondary monitoring point in the remaining partition field of view; S55 , according to the vibration time history curves of the primary monitoring points and the secondary monitoring points in the remaining sub-fields of view, the transfer function of each secondary monitoring point in the remaining sub-fields of view is obtained by adopting the method of steps S52 to S53 .

3. The video vibration calibration method of fusion video micro-vibration monitoring and sensor according to claim 2 is characterized in that: Step S54 further comprises: According to the geometric characteristics of the hydraulic monitoring structure, a finite element simulation model of the hydraulic monitoring structure is constructed; Use concrete with the same mix ratio and raw materials as the hydraulic monitoring structure to carry out physical model tests and obtain the mechanical parameters of concrete; The vibration signal time history of the primary and secondary monitoring points in the first partitioned field of view is used as the model boundary condition to perform dynamic calculations and extract the vibration time history curves of the primary and secondary monitoring points in the remaining partitioned fields of view.

4. The video vibration calibration method of fusion video micro-vibration monitoring and sensor according to claim 1 is characterized in that: The expression for calibrating the spectrum corresponding to the secondary monitoring point is: in, and Respectively n Secondary monitoring points in each field of view m Spectra before and after correction; For the n Secondary monitoring points in each field of view m The transfer function of n >1.

5. The video vibration calibration method of fusion video micro-vibration monitoring and sensor according to claim 2 is characterized in that: The expression for calculating the displacement information of the pixel points corresponding to the primary monitoring points and the secondary monitoring points is: in, and The pixels in the horizontal and vertical directions are t The instantaneous speed at time 0, It is the full video domain phase image at different scales and directions; x and y are the vibration speeds of the pixel in the horizontal and vertical directions respectively.

6. The method for fusion video micro-vibration monitoring and sensor video vibration calibration according to any one of claims 1 to 5, characterized in that: Before calibrating the pixel points corresponding to the first-level monitoring points, the following steps are also included: Calculate the absolute value of the difference between the spectrum of the first-level monitoring point and the vibration signal time history, and determine whether it is less than the deviation threshold. If so, calibrate the pixel points corresponding to the first-level monitoring point. Otherwise, do not calibrate the pixel points corresponding to the first-level monitoring point.

7. The method for fusion video micro-vibration monitoring and sensor video vibration calibration according to any one of claims 1 to 5, characterized in that: The field of view of the camera only accommodates one partitioned field of view. When collecting video information of the hydraulic monitoring structure, the camera is used to record video for each partitioned field of view in turn.

8. The method for fusion video micro-vibration monitoring and sensor video vibration calibration according to any one of claims 1 to 5, characterized in that: Each partitioned field of view is set up with 4 secondary monitoring points.

Images