A structural vibration video measurement method and system based on derivative phase optical flow method

Abstract

The application discloses a kind of structural vibration video measurement method and system based on derivative phase optical flow method, utilize camera to the vibration process of measured structure with the frame rate of satisfying Nyquist sampling rate carries out video acquisition, obtains time series variation image sequence rich in vibration information, simultaneously as real part sequence;With every image sequence, two-dimensional first derivative operator is used to do two-dimensional convolution frame by frame, and the imaginary part sequence of analytical signal is obtained;Real part sequence and imaginary part sequence are one-to-one corresponding combination using complex combination operator, and the analytical sequence rich in phase information is obtained;Using four quadrant inverse tangent operation operation above-mentioned analytical sequence, the phase change of adjacent frame rich in relative displacement change field information is obtained;Using amplitude filtering denoising method processing obtains the phase change sequence after denoising;Using the conversion scale factor obtained by calculation and fast fourier transform, the actual vibration displacement and vibration frequency of measured structure are obtained.The application has faster speed, higher precision and anti-noise performance.

Description

Technical Field

[0001] This invention belongs to the field of structural vibration video measurement technology, and relates to a structural vibration video measurement method and system based on derivative phase optical flow method. Background Technology

[0002] Vibration testing and analysis are of great significance in scientific research and engineering applications. Traditional vibration testing and analysis mostly employs contact methods, which suffer from drawbacks such as complex wiring, difficult installation, low spatial resolution, and load effects. With technological advancements, non-contact methods such as laser Doppler vibration measurement and video measurement have been widely applied in vibration measurement, effectively overcoming these shortcomings. However, laser Doppler vibration meters suffer from slow measurement speed, bulky structure, and high cost. Meanwhile, video-based vibration measurement methods, due to their advantages of being non-contact, highly accurate, having high spatial resolution, providing full-field measurement, being easy to operate, and having a simple structure, are gradually becoming an important part of the vibration measurement technology field.

[0003] To address the numerous drawbacks of traditional vibration measurement, digital image correlation (DIC) technology has been widely developed for application in the field of vibration measurement. Motion measurement based on digital video offers advantages such as low cost, high portability, and the provision of global measurements, providing a new perspective for motion measurement technology. Current video measurement technology also includes optical flow methods, which combine image pixel values ​​and image detection methods such as edge detection to achieve vibration response measurement using video methods. Patent CN106989812 proposes a modal measurement method for large wind turbine blades based on photogrammetry. The acquisition process requires dual cameras, and the blade modes are analyzed using DIC's stereo matching technology. Although DIC technology can obtain high-precision data, it requires extremely high computational power, making real-time monitoring and data processing impossible under current technological conditions. Furthermore, it requires pre-coating the measured object with features such as speckle. Patent CN114187330 proposes a structural micro-amplitude vibration working mode analysis method based on optical flow. This invention uses structural contour points obtained by optical flow as the main feature, and the displacement changes of these contour points reflect the actual changes. However, these methods have high requirements for acquisition conditions, needing to strictly ensure that there are no other motion interferences in the vibration area scene. They also have poor noise resistance, high computational load, slow calculation speed, and certain requirements on the camera's frame rate. Patent CN108122248A proposes a phase-based motion estimation method to measure the natural frequency of a dam, using Gabor transform to extract phase information from the dam's edge and calculate the dam's natural displacement. However, common Gabor transform-based phase motion estimation methods require prior knowledge to guide manual wavelength setting, resulting in poor automation. Summary of the Invention

[0004] In view of the above-mentioned prior art, the technical problem to be solved by the present invention is to provide a structural vibration video measurement method and system based on the derivative phase optical flow method, thereby solving the problem of high power consumption in the prior art.

[0005] To address the above problems, the present invention provides a structural vibration video measurement method based on the derivative phase optical flow method, comprising the following steps:

[0006] Step 1: Use a camera to capture video of the vibration process of the structure under test at a frame rate that satisfies the Nyquist sampling rate, and obtain a time-series image sequence rich in vibration information, which is also used as the real part sequence.

[0007] Step 2: Use the two-dimensional first derivative operator to perform two-dimensional convolution with each image sequence frame by frame to obtain the imaginary part sequence of the analytic signal;

[0008] Step 3: Use the complex associative operator to combine the real part sequence and the imaginary part sequence one by one to obtain an analytical sequence rich in phase information;

[0009] Step 4: Use the four-quadrant arctangent operation to operate on the analytical sequence to obtain the phase change of adjacent frames rich in relative displacement change field information;

[0010] Step 5: Use amplitude filtering to denoise the phase change sequence.

[0011] Step 6: Calculate the conversion factor between the actual displacement of the measured structure in the actual space plane and the motion size of the object in the camera plane at the actual measurement distance. Multiply the phase change sequence after amplitude filtering by the conversion factor to obtain the actual amplitude of the structural vibration in time. Obtain the vibration frequency of the measured structure through fast Fourier transform.

[0012] Furthermore, step two, obtaining the imaginary part sequence of the analytic signal, includes:

[0013] (1) Convert the acquired images into grayscale feature images, and dynamically select the grayscale image features I(x,y,t) and I(x,y,t+Δt) of two adjacent time frames, where t represents the current time, Δt represents the time step, (x,y) are spatial coordinates, and the displacement in the x-direction is δ. x ,but:

[0014]

[0015]

[0016] (2) Select the two-dimensional first-order derivative operator corresponding to the x-direction, and convolve the selected two-dimensional first-order derivative operator with the gray-level feature image in the spatial domain to obtain the first derivative of the image in the spatial domain, which is also used as the imaginary part I. Dx(x,y,t) and I Dx (x,y,t+Δt):

[0017]

[0018]

[0019] Furthermore, the two-dimensional first-order derivative operator is the Sobel operator, the Roberts operator, or the Scharr operator.

[0020] Furthermore, step 3, obtaining the parsed sequence rich in phase information, includes:

[0021] The analytical signal I rich in phase information DA (x,y,t) and I DA (x,y,t+Δt) satisfies:

[0022]

[0023]

[0024] Furthermore, step 4, which involves using the four-quadrant arctangent operation to manipulate the parsed sequence and obtain the phase change of adjacent frames rich in relative displacement change field information, includes:

[0025]

[0026]

[0027] This allows us to obtain the phase change of adjacent frames, which is rich in information about relative displacement changes.

[0028]

[0029] Furthermore, the denoised phase change sequence is as follows:

[0030]

[0031] Where abs[·] represents the magnitude.

[0032] The present invention also includes a structural vibration video measurement system based on the derivative phase optical flow method, comprising an acquisition module, a conversion module, a combination module, an arctangent module, and a result calculation module;

[0033] The acquisition module uses a camera to capture video of the vibration process of the structure under test at a frame rate that meets the Nyquist sampling rate, and obtains a time-series image sequence rich in vibration information, which is then used as the real part sequence.

[0034] The conversion module uses a two-dimensional first-order derivative operator to perform a two-dimensional convolution operation on the acquired time-varying image sequence frame by frame to obtain the imaginary part sequence of the analytical signal;

[0035] The combining module uses a complex associative operator to integrate the real part sequence and the imaginary part sequence to generate an analytical sequence rich in phase changes between adjacent frames;

[0036] The arctangent module uses a four-quadrant arctangent operation to parse the sequence and obtain the phase change of adjacent frames rich in relative displacement change field information.

[0037] The result calculation module calculates the conversion factor between the actual displacement of the measured structure in the actual space plane and the motion size of the object in the camera plane at the actual measurement distance. The phase change sequence after amplitude filtering is multiplied by the conversion factor to obtain the actual amplitude of the structural vibration in time. The vibration frequency of the measured structure is obtained by fast Fourier transform.

[0038] The beneficial effects of this invention are:

[0039] 1. Compared with traditional methods, this invention is always based on spatial-to-spatial phase transformation, which has faster speed and more accurate measurement results, lower hardware requirements and lower cost.

[0040] 2. Compared with the basic optical flow method, it has better noise resistance and more accurate processing results.

[0041] 3. Compared with the currently popular Gabor transformation-based methods, the present invention does not require preset parameters, reduces manual operation, and accelerates the industrial automation process in this field. Attached Figure Description

[0042] Figure 1 This is a schematic diagram of the derivative phase optical flow method.

[0043] Figure 2 This is an image of a cantilever beam taken from the actual camera plane.

[0044] Figure 3 The results are actual time-domain measurements using the derivative phase optical flow method.

[0045] Figure 4 The results are from actual frequency domain measurements using the derivative phase optical flow method. Detailed Implementation

[0046] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0047] Combination Figures 1-4 This invention proposes a video measurement method for structural vibration based on the derivative phase optical flow method, comprising:

[0048] Step 1: Use a camera to capture video of the vibration process of the structure under test at a frame rate that satisfies the Nyquist sampling rate, and obtain a time-series image sequence rich in vibration information, which is also used as the real part sequence.

[0049] Step 2: Use the two-dimensional first derivative operator to perform two-dimensional convolution with each image sequence frame by frame to obtain the imaginary part sequence of the analytic signal;

[0050] Step 3: Use the complex associative operator to combine the real part sequence and the imaginary part sequence one by one to obtain an analytical sequence rich in phase information;

[0051] Step 4: Use the four-quadrant arctangent operation to operate on the above analytical sequence to obtain the phase change of adjacent frames rich in relative displacement change field information;

[0052] Step 5: Use amplitude filtering to denoise the phase change sequence.

[0053] Step 6: Using the calculated conversion scaling factor and fast Fourier transform, obtain the actual vibration displacement and vibration frequency of the measured structure.

[0054] The two-dimensional first-order derivative operators mentioned in step 2 include, but are not limited to, the Sobel operator, the Roberts operator, and the Scharr operator.

[0055] The process of obtaining the imaginary part sequence specifically includes:

[0056] (1) Convert the acquired image into a grayscale feature image, and dynamically select the grayscale image features I(x,y,t) and I(x,y,t+Δt) of two adjacent time frames, where t represents the current time, Δt represents the time step, (x,y) are spatial coordinates, and the displacement δ in the x-direction is used. x For example,

[0057]

[0058]

[0059] A two-dimensional first-order derivative operator corresponding to the x-direction is selected, and convolved with the aforementioned grayscale feature image in the spatial domain using the selected two-dimensional first-order derivative operator, which is also used as the imaginary part I. Dx (x,y,t) and I Dx (x,y,t+Δt),

[0060]

[0061]

[0062] Step 3 specifically involves:

[0063] The real and imaginary parts are combined using a complex associative operator to generate an analytical signal I rich in phase information. DA (x,y,t) and I DA (x,y,t+Δt),

[0064]

[0065]

[0066] The calculation formula for the above analytic signal using the four-quadrant arctangent arg[·] operation in step 4 is as follows:

[0067]

[0068]

[0069] This allows us to obtain the phase change of adjacent frames, which is rich in information about relative displacement changes.

[0070]

[0071] The specific formula for calculating the denoised phase change Ф(x,y) obtained by amplitude filtering and denoising in step 5 is as follows:

[0072]

[0073] Where abs[·] represents the magnitude.

[0074] In step 6, during the conversion between actual distance and pixel distance, the conversion factor between the camera coordinate system and the world coordinate system changes with the camera's acquisition distance.

[0075] In step 6, during the calculation of frequency domain information using the obtained actual displacement field, the use of the time-series actual displacement field needs to ensure at least one period of the time-series actual displacement field.

[0076] This invention proposes a structural vibration video measurement system based on the derivative phase optical flow method. The system specifically includes:

[0077] Acquisition module: Uses a camera to acquire video of the vibration process of the structure under test at a frame rate that meets the Nyquist sampling rate, obtains a time-series image sequence rich in vibration information, and uses it as the real part sequence;

[0078] Conversion module: It uses a two-dimensional first-order derivative operator to perform two-dimensional convolution operation on the acquired time-varying image sequence frame by frame to obtain the imaginary part sequence of the analytical signal;

[0079] Combination module: The real part sequence and the imaginary part sequence are integrated using the complex number associativity operator to generate an analytical sequence rich in phase changes between adjacent frames;

[0080] Arctangent module: Uses four-quadrant arctangent operation to parse sequences and obtain phase changes of adjacent frames rich in relative displacement change field information;

[0081] Results calculation module: Using the calculated conversion scaling factor and fast Fourier transform, the actual vibration displacement and vibration frequency of the measured structure are obtained.

[0082] To better describe the method described in this application, the following embodiments are used to illustrate the complete process of the fast full-field structural vibration measurement method based on derivative phase optical flow in practical applications.

[0083] In this embodiment, a fast full-field structural vibration measurement method based on the derivative phase optical flow method is specifically provided, and its flowchart is as follows. Figure 1 As shown, the specific steps include:

[0084] (1) Prepare a cantilever beam with a length of 20cm, such as Figure 2 As shown, a hammer is used to excite the bottom of the cantilever beam to induce vibration modes.

[0085] (2) The vibration process of the cantilever beam under test is collected and recorded using a smartphone, ordinary camera, or high-speed industrial camera to obtain a video containing the vibration information of the cantilever beam. At the same time, the vibration data is measured using a traditional contact method (accelerometer measurement method) as a comparison reference value. During the acquisition process, the acquisition frame rate is set to 180Hz, and LED light source is used for supplementary lighting.

[0086] (3) Select the Sobel image operator in the x-direction, which is generally applicable. Convolve the acquired image, i.e., the real part sequence, frame by frame using the Sobel operator to obtain the imaginary part result, thus obtaining the aforementioned I. Dx (x,y,t) and I Dx(x,y,t+Δt), where t represents the current time, Δt represents the time step, and (x,y) are spatial coordinates.

[0087] (4) The real part sequence and the imaginary part sequence are combined using the complex associative operator to generate an analytical signal I rich in phase information. DA (x,y,t) and I DA (x,y,t+Δt),

[0088]

[0089]

[0090] (5) The four-quadrant arctangent operator arg[·] is used to process the analytical sequence features containing phase differences to obtain the phase information of adjacent frames.

[0091]

[0092]

[0093] The phase change field in the video is obtained by subtracting the two phases.

[0094]

[0095] Then, amplitude filtering is used to denoise the phase change field.

[0096]

[0097] (7) The conversion factor of the actual displacement of the measured object in the actual space plane and the motion size (pixel) of the object in the camera plane at the actual measurement distance is 5.2mm / pixel.

[0098] (8) Multiply the amplitude-filtered temporal relative motion displacement field by the conversion scaling factor to obtain the actual motion field. The time-domain measurement results are as follows: Figure 3 As shown, the frequency domain field of the vibration displacement signal is obtained after Fourier transform, and the result is as follows. Figure 4 As shown.

[0099] (9) The overlap between the actual measurement data and the data of the present invention was calculated to be 99.1%.

[0100] The present invention provides a detailed description of a structural vibration video measurement method and system based on derivative phase optical flow. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A video measurement method for structural vibration based on derivative phase optical flow, characterized in that, Includes the following steps: Step 1: Use a camera to capture video of the vibration process of the structure under test at a frame rate that satisfies the Nyquist sampling rate, and obtain a time-series image sequence rich in vibration information, which is also used as the real part sequence. Step 2: Use the two-dimensional first derivative operator to perform two-dimensional convolution with each image sequence frame by frame to obtain the imaginary part sequence of the analytic signal; Step 3: Use the complex associative operator to combine the real part sequence and the imaginary part sequence one by one to obtain an analytical sequence rich in phase information; Step 4: Use the four-quadrant arctangent operation to operate on the analytical sequence to obtain the phase change of adjacent frames rich in relative displacement change field information; Step 5: Use amplitude filtering to denoise the phase change sequence. Step 6: Calculate the conversion factor between the actual displacement of the measured structure in the actual space plane and the motion size of the object in the camera plane at the actual measurement distance. Multiply the phase change sequence after amplitude filtering by the conversion factor to obtain the actual amplitude of the structural vibration in time. Obtain the vibration frequency of the measured structure through fast Fourier transform.

2. The structural vibration video measurement method based on derivative phase optical flow method according to claim 1, characterized in that: Step two, obtaining the imaginary part sequence of the analytic signal, includes: (1) Convert the acquired images into grayscale feature images and dynamically select grayscale image features from two adjacent time frames. as well as ,in Representing the current moment, Indicates the time step. These are spatial coordinates. The displacement in the direction is ,but: ； (2) Select the corresponding A two-dimensional first-order derivative operator is used to convolve the grayscale feature image in the spatial domain, yielding the first derivative of the image in the spatial domain, which also serves as the imaginary part. as well as : 。 3. A structural vibration video measurement method based on derivative phase optical flow method according to claim 1 or 2, characterized in that: The two-dimensional first-order derivative operator is Operator, Operator or Operator.

4. The structural vibration video measurement method based on derivative phase optical flow method according to claim 1, characterized in that: Step 3, obtaining the parsed sequence rich in phase information, includes: The analytical signal rich in phase information as well as satisfy: ； 。 5. The structural vibration video measurement method based on derivative phase optical flow method according to claim 1, characterized in that: Step 4, which involves using the four-quadrant arctangent operation to manipulate the analytical sequence and obtain the phase changes of adjacent frames rich in relative displacement change field information, includes: ； ； This allows us to obtain the phase change of adjacent frames, which is rich in information about relative displacement changes. : 。 6. The structural vibration video measurement method based on derivative phase optical flow method according to claim 1, characterized in that: The denoised phase change sequence is as follows: ； in This indicates the calculation of the amplitude.

7. A structural vibration video measurement system based on the derivative phase optical flow method, characterized in that: It includes a data acquisition module, a conversion module, a combination module, an arctangent module, and a result calculation module; The acquisition module uses a camera to capture video of the vibration process of the structure under test at a frame rate that meets the Nyquist sampling rate, and obtains a time-series image sequence rich in vibration information, which is then used as the real part sequence. The conversion module uses a two-dimensional first-order derivative operator to perform a two-dimensional convolution operation on the acquired time-varying image sequence frame by frame to obtain the imaginary part sequence of the analytical signal; The combining module uses a complex associative operator to integrate the real part sequence and the imaginary part sequence to generate an analytical sequence rich in phase changes between adjacent frames; The arctangent module uses a four-quadrant arctangent operation to parse the sequence and obtain the phase change of adjacent frames rich in relative displacement change field information. The result calculation module calculates the conversion factor between the actual displacement of the measured structure in the actual space plane and the motion size of the object in the camera plane at the actual measurement distance. The phase change sequence after amplitude filtering is multiplied by the conversion factor to obtain the actual amplitude of the structural vibration in time. The vibration frequency of the measured structure is obtained by fast Fourier transform.

Images