A Static and Dynamic Parameter Measurement Method for Low-Frequency and Large-Stroke Vibration Tables Based on Binocular Vision

Abstract

The present invention discloses a method for measuring static and dynamic parameters of a low-frequency large-stroke vibration table based on binocular vision. Two different high-contrast measurement markers I and II are respectively installed in the vertical and horizontal directions of the workbench surface of the large-stroke vibration table to represent the movement displacements in the Z direction, X direction, and Y direction. The optical axes of cameras I and II are respectively perpendicular to markers I and II. By dividing the horizontal guide rail of the large-stroke vibration table into several equally spaced positions, the middle position is selected as the reference position where the vertical displacement is zero. The least squares linear fitting principle is used for solution. An external trigger is used to control cameras I and II to simultaneously collect the motion sequence images of markers I and II respectively, and the sub-pixel edge detection method is used to realize the displacement measurement of the workbench surface of the large-stroke vibration table in the X, Y, and Z directions. This method is beneficial to promoting the improvement of the calibration accuracy and efficiency of low-frequency vibration sensors, as well as the improvement of low-frequency vibration calibration and traceability.

Description

Technical Field

[0001] The present invention belongs to the field of spatial motion measurement, and more specifically, particularly relates to a method for measuring static and dynamic parameters of a vibration table. Background Art

[0002] Different types of low-frequency vibration sensors have been increasingly applied in fields such as earthquake early warning, medical assisted diagnosis, bridge and building structure health testing, and rail machine healthy movement. To ensure the reliability of these applications, it is necessary to calibrate the sensitivity of the sensor before or after a period of use. Large stroke vibration tables have been increasingly widely used in low-frequency vibration calibration, and during the calibration process, it can provide specific frequencies and vibration excitations for the sensor. However, the static and dynamic parameters of the large stroke vibration table will inevitably affect the calibration accuracy, and they must be considered in the uncertainty evaluation. Therefore, in order to improve the calibration and traceability of low-frequency vibration sensors, it is very important to develop a measurement method that can accurately and flexibly measure these parameters.

[0003] At present, the common methods for measuring the static parameters of the guide rail of a low-frequency vibration table mainly include the autocollimator measurement method, the accelerometer measurement method, and the model estimation method. The autocollimator measurement method mainly uses an autocollimator to measure the inclination angles of different positions of the vibration table guide rail to obtain the degree of bending, and the accuracy can reach arcseconds. The accelerometer measurement method obtains the degree of bending by using the relationship between the output of the accelerometer and the gravity component. However, both of these methods require additional equipment, which increases the complexity and cost of the low-frequency vibration calibration system. The model estimation method estimates the degree of bending by combining the calibration results with the deviation using the least squares method. Although this method does not require additional equipment, it requires that the calibration sensitivities at different frequencies satisfy a linear relationship. At present, the methods for measuring dynamic parameters mainly include the laser interferometry (LI) and the sensor measurement method (SM). LI based on the Doppler velocity measurement and laser interferometry principle has the advantages of strong anti-interference ability, good dynamic performance, wide measurement frequency range, high linearity, etc., and can achieve high-precision measurement of motion displacement and trajectory. However, its measurement system is complex, costly, less flexible, and it is difficult to be used for low-frequency vibration measurement in a large displacement range. The SM method only requires a biaxial or triaxial vibration sensor to measure the displacement of spatial motion, and has the advantages of simple measurement system, low cost, flexibility, high efficiency, etc. However, due to the limitation of the frequency characteristics of the sensor itself, its measurement accuracy is usually not high. In recent years, the machine vision method has been widely used in many high-precision displacement measurements. Its resolution can reach 1μm, and the uncertainty can reach 0.1%. It has the advantages of high efficiency, flexibility, high precision, and low cost. Therefore, the static and dynamic parameters of a large-stroke vibration table can be measured by the binocular vision method. Therefore, a method for measuring the static and dynamic parameters of a low-frequency large-stroke vibration table based on binocular vision is proposed, which can improve the measurement accuracy while reducing the complexity and cost of the measurement system, and enhancing the flexibility and measurement efficiency.

[0004] Therefore, aiming at the deficiencies of the current methods for measuring the static and dynamic parameters of the vibration table, such as complex measurement system, poor flexibility, high cost, limited measurement range, and difficulty in being used for low-frequency high-precision measurement, the present invention proposes a low-cost binocular vision measurement method for measuring the static and dynamic parameters of a low-frequency large-stroke vibration table, which is efficient, accurate, and flexible, and is conducive to promoting the improvement of low-frequency vibration calibration and traceability. Summary of the Invention

[0005] Aiming at the deficiencies of the current methods for measuring the displacement of spatial motion, such as complex measurement system, poor flexibility, high cost, limited measurement range, and difficulty in being applicable to low-frequency motion high-precision measurement, the present invention proposes a low-cost binocular vision measurement method (BV) for measuring the static and dynamic parameters of a low-frequency large-stroke vibration table, including:

[0006] Acquisition of two kinds of high-contrast measurement mark motion sequence images: Two kinds of high-contrast measurement marks composed of four circles around a straight line and four circles around a rectangle are respectively installed in the horizontal and vertical directions of the workbench surface of a large-stroke vibration table, representing the motion of the workbench surface of the vibration table in the Z direction, X and Y directions respectively. Cameras I and II acquire a sufficient number of frames of the spatial motion sequence images of measurement marks I and II to ensure that the spatial motion displacement can be accurately measured;

[0007] Sub-pixel extraction of the edge features of linear motion in the X, Y, and Z directions: Use template matching to determine the region of interest on the mark image, and extract the sub-pixel coordinates of the edge points of the rectangle edge of mark II and the straight line edge of mark I within the region of interest through the sub-pixel edge detection method based on Zernike moments. Convert the sub-pixel coordinates into the corresponding world coordinates by using the correspondence between the image pixel coordinates and the world coordinates determined by the camera;

[0008] Displacement solution of the edge features of linear motion in the X, Y, and Z directions: Use the least squares principle to linearly fit the world coordinates of the edge points of the linear motion features in the X, Y, and Z directions to obtain the corresponding fitted straight lines. Select the edge features of the mark image at the zero position of the workbench surface as the zero-displacement reference edge, and obtain the displacements of the edge features in the X, Y, and Z directions by calculating the distances between the edge features of the sequence images and the corresponding edges of the reference image;

[0009] Solution of the static and dynamic parameters of the vibration table: Calculate the curvature of different positions of the vibration table guide rail based on the guide rail bending model and the measured vertical displacement change; linearly fit the curvatures of different positions of the guide rails on the left and right sides centered on the reference position based on the least squares principle, and take the average value of the left and right curvatures obtained by fitting as the curvature of the guide rail of the low-frequency large-stroke vibration table. Realize the displacement measurement of the workbench surface of the large-stroke vibration table in the X, Y, and Z directions through camera calibration and the sub-pixel edge detection method based on Zernike moments. Furthermore, use the measured displacements in the X, Y, and Z directions to calculate the dynamic parameters such as the amplitude-frequency characteristics, repeatability, distortion rate, and lateral ratio of the large-stroke vibration table.

[0010] The technical solution adopted by the present invention is a method for measuring the static and dynamic parameters of a low-frequency large-stroke vibration table based on binocular vision. The measurement method includes the following steps:

[0011] S1: Install the high-contrast measurement mark I, which consists of four circles around a straight line (abbreviated as measurement mark I), and the high-contrast measurement mark II, which consists of four circles around a rectangle (abbreviated as measurement mark II), in the horizontal and vertical directions of the working surface of the large-stroke vibration table respectively. The straight edge of mark I on the working surface of the vibration table represents the movement in the Z direction, and the X and Y direction edges of mark II represent the movement in the X and Y directions respectively. Divide the horizontal guide rail under the working surface of the large-stroke vibration table into several equally spaced positions, select the middle position of the equally spaced positions as the reference position with zero vertical displacement, and use camera I to measure the vertical displacement change of different positions of the horizontal guide rail of the vibration table compared with the reference position;

[0012] S2: Calculate the curvature of different positions of the horizontal guide rail of the vibration table using the bending model of the large-stroke vibration table guide rail and the measured vertical displacement change, and solve the static parameters of the guide rail bending of the large-stroke vibration table based on the least-squares straight line fitting;

[0013] S3: Control cameras I and II to simultaneously collect the motion sequence images of measurement mark I and measurement mark II through an external trigger, and use the sub-pixel edge detection method based on Zernike moments to measure the displacement of the working surface of the large-stroke vibration table in the X, Y, and Z directions;

[0014] S4: Calculate the dynamic parameters of the large-stroke vibration table using the displacements in the X, Y, and Z directions measured in S3;

[0015] S5: Save and display the static and dynamic parameters of the guide rail bending of the measured large-stroke vibration table. The dynamic parameters include amplitude-frequency characteristics, repeatability, distortion rate, and lateral ratio.

[0016] High-precision extraction of the characteristic edges of the motion sequence images of measurement mark I and measurement mark II specifically includes:

[0017] (1) Determination of the rectangular and straight-line regions of interest in the sequence images of measurement mark I and measurement mark II;

[0018] To improve the stability and accuracy of the extraction of motion characteristic edges, template matching based on the correlation coefficient is used to determine the four circular regions in the sequence images of measurement mark I and measurement mark II. The straight line of mark I and the rectangle of mark II are located within the region of interest formed by the centers of the four circular regions.

[0019] (2) Sub-pixel extraction of the edge points of the long and short sides of the straight line and rectangle;

[0020] To eliminate the interference of background noise and similar edges, extract the straight edge of measurement mark I and the rectangular edge of mark II within the determined region of interest. First, use the Canny operator to extract the motion sequence image {F of measurement mark I L,j(x, y), where j = 1, 2,..., M}, the pixel-level straight edge points, and the moving sequence images {F of the measurement mark II R,j (x, y), where j = 1, 2,..., N}, the long and short side edge points of the rectangle, where M and N are the numbers of the sequence images of the measurement mark I and the measurement mark II collected respectively. Then, by calculating F L,j (x, y) and F R,j (x, y), the sub-pixel coordinates of the straight line and the long and short side edges of the rectangle are solved by the Zernike moments. Finally, using the correspondence between the image pixel coordinates and the world coordinates determined by the camera calibration, the sub-pixel coordinates of the straight edge of the measurement mark I and the long and short side edges of the rectangle of the measurement mark II are converted into the corresponding world coordinates.

[0021] (3) Motion feature edges in the X, Y, and Z directions

[0022] The long and short side edges of the rectangle of the measurement mark II and the straight edge of the measurement mark I represent the motions in the X and Y, Z directions respectively. Based on the least square straight line principle, the world coordinates of the corresponding feature edge points are respectively fitted to obtain the fitted straight lines {l j,X}}、{l j,Y}} and {l j,L} of the motion feature edges in the corresponding X, Y, and Z directions. Among them, l j,X and l j,Y , l j,L are respectively the equivalent long side edge of the rectangle of the measurement mark II and the straight edge of the mark I of F R,j (x, y) and F L,j (x, y).

[0023] For the measurement of the static parameters of the guide rail bending of the low-frequency large-stroke vibration table, the following method is adopted:

[0024] First, divide the horizontal guide rail of the large-stroke vibration table into several equally spaced positions {s X,h , where h = 1, 2,..., Q}, Q is an odd number, and select the middle position (Q + 1) / 2 as the reference position of the horizontal line of the guide rail, and its vertical displacement is denoted as s Z,(Q+1) / 2 . Therefore, the vertical displacement change at any position h is solved by the difference between the straight edge displacement s Z,h and s Z,(Q+1) / 2 in the Z direction. The curvature α h at any position h is expressed as:

[0025]

[0026] Finally, based on the least squares principle, linear fitting is performed on the curvatures at different positions of the left and right guide rails centered on the reference position, and the average values of the left and right curvatures obtained by fitting are used as the guide rail curvature of the low-frequency large-stroke vibration table.

[0027] The X, Y, and Z displacements of the working surface of the low-frequency large-stroke vibration table are obtained by calculating the edge displacements s X (t), s Y (t), and s Z (t) in the X, Y, and Z directions:

[0028]

[0029] where: ω v is the vibration angular frequency, and are the peak values of s X (t), s Y (t), and s Z (t) respectively, and are the initial phases of s X (t), s Y (t), and s Z (t) respectively.

[0030] Using the sine approximation method, the measured displacements s X (t), s Y (t), and s Z (t) in the X, Y, and Z directions are fitted as follows:

[0031]

[0032] When solving for the displacement in the X direction, in the formula: A X and B X are the sine displacement components; C X and D X are the ultra-low-frequency linear interference amount and the offset amount respectively. Solving the overdetermined equations in the form of the first term of formula (3) based on the least squares principle can obtain the parameters A X and B X . Similarly, the parameters A Y and B Y , as well as A Z and B Z can be obtained.

[0033] By solving the displacement parameters in the X, Y, and Z directions, the peak values and initial phases of the fitted displacements of s X (t), s Y (t), and s Z (t) are obtained:

[0034]

[0035] The dynamic parameters of the low-frequency large-stroke vibration table include amplitude-frequency characteristics, repeatability, distortion rate, lateral ratio, and uniformity. The amplitude-frequency characteristic is the peak value of the displacement in the X direction at different frequencies, that is It reflects the ability of the low-frequency large-stroke vibration table to provide the excitation displacement magnitude for the vibration sensor to be calibrated.

[0036] The repeatability S of the low-frequency large-stroke vibration table Re is the dispersion of the peak values of the displacement in the X direction measured multiple times at the selected test frequency, reflecting the stability of the peak displacement value, and is given by the following formula:

[0037]

[0038] where H is the number of measurements, usually not less than 10; is the average value of the H measurements.

[0039] The distortion rate S of the low-frequency large-stroke vibration table Dr is defined as the ratio of the sum of the amplitudes of three adjacent high-order harmonics to the amplitude of the fundamental harmonic, and is expressed as:

[0040]

[0041] In the formula: is the peak value of the i-th harmonic fitting of the displacement in the X direction.

[0042] The lateral ratio S of the low-frequency large-stroke vibration table Tr is defined as the ratio of the peak value of the axial displacement to the peak value of the lateral displacement, and can be described as:

[0043]

[0044] The uniformity S of the low-frequency large-stroke vibration table NA is the ratio of the maximum absolute displacement peak deviation at the boundary of the worktable surface to the displacement peak value at the center position, and can be expressed as:

[0045]

[0046] where is the difference between the peak displacement at the k-th boundary and the peak displacement at the center position and the value of k is usually 4.

[0047] A static and dynamic parameter measurement device for a low-frequency large-stroke vibration table based on binocular vision. The device mainly includes: a low-frequency large-stroke vibration table (1), a workbench surface (2), a high-contrast measurement mark I (3) containing a straight line, a high-contrast measurement mark II (4) containing a rectangle, a camera I (5), a camera II (6), a trigger (7), and an image processing and display unit (8).

[0048] The workbench surface (2) of the low-frequency large-stroke vibration table (1) provides axial movement in the X direction and lateral movement in the Y and Z directions; the high-contrast measurement marks I (3) and II (4) are respectively installed in the vertical and horizontal directions of the workbench surface (2) of the low-frequency large-stroke vibration table. The displacement of the mark I (3) in the Z direction is consistent with that of the workbench surface (2) of the low-frequency large-stroke vibration table, and the displacements of the mark II (4) in the X and Y directions are consistent with those of the workbench surface (2) of the low-frequency large-stroke vibration table; the optical axes of the camera I (5) and the camera II (6) are respectively perpendicular to the measurement marks I (3) and II (4); the trigger (7) is used to control the cameras (I) and (II) to simultaneously collect the motion sequence images of the measurement marks I (3) and II (4); the image processing and display unit (8) processes the collected motion sequence images, saves and displays the spatial motion displacement and the measurement results.

[0049] The spatial motion displacement measurement method of the present invention has the following advantages:

[0050] ⑴ The method of the present invention is stable, reliable and practical, and can be applied to the measurement of spatial motion displacement in different frequency and displacement ranges at the same time;

[0051] ⑵ The measurement process of the method of the present invention is simple, flexible, efficient and the system cost is low. Only two cameras are required for the motion measurement in different frequency ranges;

[0052] ⑶ The method of the present invention realizes high-precision motion displacement measurement by measuring the displacements in the X, Y and Z directions;

[0053] ⑷ The method of the present invention belongs to the spatial motion measurement method, and can realize high-precision spatial motion displacement measurement in the low-frequency range, and even can reach quasi-static measurement;

[0054] ⑸ The method of the present invention decouples the spatial motion displacement measurement into the motion displacement measurements in the X, Y and Z directions, which greatly promotes the improvement of low-frequency vibration calibration and traceability.

[0055] The technical scope of the low frequency and large stroke of the present invention:

[0056] (1) The low-frequency range is 0.01 - 20 Hz. In the BV method adopted by the present invention, a comparative experiment is conducted with the LI and SM methods, restricting the lowest frequency to 0.01 Hz. However, in actual situations, this method is still applicable under quasi-static conditions. Additionally, by shortening the relative distance between the camera and the working surface of the vibration table, the upper frequency can be higher than 20 Hz;

[0057] (2) The peak value of the large-stroke vibration table is 400 mm;

[0058] (3) Through comparative verification of the method adopted by the present invention, the measurement accuracy of the bending degree of the vibration table guide rail reaches 0.001%; in the measurement of the dynamic parameters of the vibration table, the maximum relative deviation of the amplitude characteristic is 0.5%, the repeatability deviation is 0.14%, and the maximum distortion rate is 0.245%, which are respectively less than those of the LI and SM methods. Description of the Drawings

[0059] Appendix Figure 1 is a schematic diagram of the device for the specific implementation example of the method of the present invention;

[0060] Appendix Figure 2 is a flowchart of a method for measuring static and dynamic parameters of a low-frequency large-stroke vibration table based on binocular vision;

[0061] Appendix Figure 3 is a flowchart of spatial displacement measurement in the low-frequency vibration calibration of a low-frequency large-stroke vibration table based on binocular vision;

[0062] Appendix Figure 4 is a diagram of the measurement results of the bending degree of the guide rail of a low-frequency large-stroke vibration table for the specific implementation example of the method of the present invention;

[0063] Appendix Figure 5 is a diagram of the measurement results of four dynamic parameters, namely amplitude-frequency characteristics, repeatability, distortion rate, and lateral ratio, of a low-frequency large-stroke vibration table for the specific implementation example of the method of the present invention. Detailed Implementation Manner

[0064] To address the deficiencies of existing measurement methods, such as complex measurement systems, poor flexibility, high costs, limited measurement ranges, and difficulty in low-frequency and high-precision measurements, the present invention proposes a method for measuring static and dynamic parameters of a low-frequency large-stroke vibration table based on binocular vision, achieving high-precision measurement of spatial movement displacement within the low-frequency range. The following provides a detailed description of the present invention in conjunction with the drawings and specific implementation examples.

[0065] Reference Figure 1Schematic diagram of the implementation example device for the method of the present invention. The device mainly includes: a low-frequency large-stroke vibration table (1), a workbench surface (2), a high-contrast measurement mark I (3) containing a straight line, a high-contrast measurement mark II (4) containing a rectangle, a camera I (5), a camera II (6), a trigger (7), and an image processing and display unit (8). The workbench surface (2) of the low-frequency large-stroke vibration table (1) provides axial movement in the X direction and lateral movement in the Y and Z directions; the high-contrast measurement marks I (3) and II (4) are respectively installed in the vertical and horizontal directions of the workbench surface (2) of the low-frequency large-stroke vibration table. The displacement of the mark I (3) in the Z direction is consistent with that of the workbench surface (2) of the low-frequency large-stroke vibration table, and the displacements of the mark II (4) in the X and Y directions are consistent with those of the workbench surface (2) of the low-frequency large-stroke vibration table; the optical axes of the camera I (5) and the camera II (6) are respectively perpendicular to the measurement marks I (3) and II (4); the trigger (7) is used to control the cameras (I) and (II) to simultaneously collect the motion sequence images of the measurement marks I (3) and II (4); the image processing and display unit (8) processes the collected motion sequence images, saves and displays the spatial motion displacement and measurement results.

[0066] Reference Figure 2 Flowchart of a method for measuring static and dynamic parameters of a low-frequency large-stroke vibration table based on binocular vision. The measurement method of the present invention mainly includes the following steps:

[0067] Step S1: A high-contrast measurement mark I formed by surrounding a straight line with four circles, hereinafter referred to as measurement mark I, and a high-contrast measurement mark II formed by surrounding a rectangle with four circles, hereinafter referred to as measurement mark II, are respectively installed in the horizontal and vertical directions of the workbench surface of the large-stroke vibration table. The straight edge of the mark I on the workbench surface of the vibration table represents the movement in the Z direction, and the X and Y direction edges of the mark II respectively represent the movements in the X and Y directions. The horizontal guide rail under the workbench surface of the large-stroke vibration table is divided into several equally spaced positions, and the middle position of the equally spaced positions is selected as the reference position where the vertical displacement is zero. The camera I is used to measure the displacement change in the vertical direction of different positions of the horizontal guide rail of the vibration table compared with the reference position;

[0068] Step S2: The curvature of different positions of the horizontal guide rail of the vibration table is calculated by using the bending model of the horizontal guide rail of the large-stroke vibration table and the measured vertical displacement change, and the static parameters of the guide rail bending of the large-stroke vibration table are solved based on the least squares straight line fitting;

[0069] Step S3: The external trigger is used to control the camera I and the camera II to simultaneously collect the motion sequence images of the measurement mark I and the measurement mark II, and the displacement measurement of the workbench surface of the large-stroke vibration table in the X, Y, and Z directions is realized by using the sub-pixel edge detection method based on Zernike moments;

[0070] Step S4: Calculate the dynamic parameters of the large-stroke vibration table by using the displacements in the X, Y, and Z directions measured in S3.

[0071] Step S5: Save and display the measured static and dynamic parameters of the guide rail bending of the large-stroke vibration table. The dynamic parameters include amplitude-frequency characteristics, repeatability, distortion rate, and lateral ratio.

[0072] Reference Figure 3 It is a flowchart of spatial displacement measurement in the low-frequency vibration calibration of a low-frequency large-stroke vibration table based on binocular vision.

[0073] The spatial motion displacement measurement method of the present invention includes the following steps:

[0074] Step S11: Read in the acquired high-contrast measurement mark motion sequence images.

[0075] Step S12: Use a set of circle template matching methods to match the regions of interest in the images captured at any distance and rotation position.

[0076] Step S13: Obtain the sub-pixel coordinates of the edge points of the linear motion features in the X, Y, and Z directions on the motion sequence images of the acquisition marks I and II by using the sub-pixel edge detection method based on Zernike moments; convert the extracted sub-pixel coordinates into corresponding world coordinates through the correspondence between the image pixel coordinates and the world coordinates determined by camera calibration.

[0077] Step S14: Fit the world coordinates of the feature edges in the X, Y, and Z directions respectively based on the least square linear fitting principle to obtain the corresponding edge lines.

[0078] Step S15: Select the feature edges in the X, Y, and Z directions of the measurement mark image at the zero position of the low-frequency two-component vibration table as the zero displacement reference edges, and calculate the distances between the feature edges in the X, Y, and Z directions of the sequence images and the corresponding reference edges.

[0079] Step S16: The distances of the feature edges in the X, Y, and Z directions of the sequence images are the displacements of the workbench surface in the X, Y, and Z directions.

[0080] The specific parameters of the device in this embodiment are: a low-frequency single-component vibration table with a frequency range of 0.01 - 100 Hz and a maximum peak displacement of 180 mm, two high-contrast measurement marks are a metal plate composed of four circles with a radius of 15 mm surrounding a rectangle with a size of 60 mm x 40 mm and a metal plate composed of four circles with a radius of 15 mm surrounding a straight line with a length of 60 mm, and two IDTOS10-V3-4K industrial cameras with a maximum resolution of nine million pixels and a maximum frame rate of 1000 fps are selected as cameras, and a KOWA lens with a focal length of 16 mm is used.

[0081] To verify the effectiveness of the method for measuring the static and dynamic parameters of a low-frequency large-stroke vibration table based on binocular vision of the present invention, the method of the present invention was used to achieve displacement measurement in the X, Y, and Z directions within the frequency range of 0.01 - 20 Hz. Table 1 shows the measurement results of the bending degree of the guide rails of a low-frequency large-stroke vibration table with loads of 0 kg, 5 kg, and 10 kg by the specific implementation example (BV) of the method of the present invention for static parameter measurement, the accelerometer measurement method (AV), and the model estimation method (ME) respectively within the range of positive and negative displacement peaks of 180 mm. From the results in Table 1, it can be seen that the measurement results of the binocular vision method are highly similar to those of the other two methods. The relative deviations between the measurement results of the BV and ME methods are 1.849×10 -5 rad, 0.987×10 -5 rad, 0.627×10 - 5 rad, slightly less than the relative deviations between the BV and AM methods.

[0082] Table 1 Measurement results of the bending degree of the guide rails of a large-stroke vibration table by the BV, AV, and ME methods

[0083]

[0084] Reference Figure 4 is a graph of the measurement results of the bending degree of the guide rails of a low-frequency large-stroke vibration table based on binocular vision.

[0085] Reference Figure 5 is a graph of the measurement results of four dynamic parameters, namely, amplitude-frequency characteristics (a), repeatability (b), distortion rate (c), and lateral ratio (d), of a low-frequency large-stroke vibration table based on binocular vision.

[0086] The above description is a detailed introduction to the implementation example of the present invention, which is not used to limit the present invention in any form. Those skilled in the relevant art can make a series of optimizations, improvements, and modifications based on the present invention. Therefore, the protection scope of the present invention should be defined by the appended claims.

Claims

1. A method for measuring static and dynamic parameters of a low-frequency large-stroke vibration table based on binocular vision, characterized in that: The measurement method includes the following steps: S1: Install a high-contrast measurement mark I formed by four circles around a straight line, abbreviated as measurement mark I, and a high-contrast measurement mark II formed by four circles around a rectangle, abbreviated as measurement mark II, in the horizontal and vertical directions of the workbench surface of the large-stroke vibration table respectively; the straight edge of mark I on the workbench surface of the vibration table represents the movement in the Z direction, and the X and Y direction edges of mark II represent the movement in the X and Y directions respectively; divide the horizontal guide rail under the workbench surface of the large-stroke vibration table into several equally spaced positions, select the middle position of the equally spaced positions as the reference position where the vertical displacement is zero, and use camera I to measure the vertical displacement change of different positions of the horizontal guide rail of the vibration table compared with the reference position. S2: Calculate the curvature of different positions of the horizontal guide rail of the vibration table using the bending model of the large-stroke vibration table guide rail and the measured vertical displacement change, and solve the static parameters of the guide rail bending of the large-stroke vibration table based on the least squares straight line fitting. S3: Control camera I and camera II to simultaneously collect the motion sequence images of measurement mark I and measurement mark II through an external trigger, and use the sub-pixel edge detection method based on Zernike moments to realize the displacement measurement of the workbench surface of the large-stroke vibration table in the X, Y, and Z directions. S4: Use the X, Y, and Z direction displacements measured in S3 to calculate the dynamic parameters of the large-stroke vibration table. S5: Save and display the measured static and dynamic parameters of the guide rail bending of the large-stroke vibration table, and the dynamic parameters include amplitude-frequency characteristics, repeatability, distortion rate, and lateral ratio.

2. A method for measuring the static and dynamic parameters of a low-frequency large-stroke vibration table based on binocular vision according to claim 1, characterized in that: High-precision extraction of the characteristic edges of the motion sequence images of measurement mark I and measurement mark II specifically includes: (1) Determination of the rectangular and straight-line regions of interest of the sequence images of measurement mark I and measurement mark II; To improve the stability and accuracy of the motion characteristic edge extraction, template matching based on the correlation coefficient is used to determine the four circular regions of the sequence images of measurement mark I and measurement mark II. The straight line of mark I and the rectangle of mark II are located within the region of interest formed by the centers of the four circular regions. (2) Sub-pixel extraction of the edge points of the long and short sides of the straight line and rectangle; To eliminate the interference of background noise and similar edges, the straight edge of measurement mark I and the rectangular edges of mark II are extracted within the determined region of interest. First, the Canny operator is used to extract the pixel-level straight edge points of the moving sequence images {F L,j (x, y), j = 1, 2,..., M} of measurement mark I, and the long and short edge points of the rectangular shape of the moving sequence images {F R,j (x, y), j = 1, 2,..., N} of measurement mark II, where M and N are the numbers of the sequence images of measurement mark I and measurement mark II collected respectively. Then, the sub-pixel coordinates of the straight line and the long and short edges of the rectangle are solved by calculating the Zernike moments of F L,j (x, y) and F R,j (x, y). Finally, the sub-pixel coordinates of the straight edge of measurement mark I and the long and short edges of the rectangle of measurement mark II are converted into the corresponding world coordinates by using the correspondence between the image pixel coordinates and the world coordinates determined by camera calibration; (3) Motion characteristic edges in the X, Y, and Z directions; The long and short edges of the rectangle of the measurement mark II and the straight edge of the measurement mark I respectively characterize the movements in the X and Y, Z directions. Based on the least squares straight line principle, the world coordinates of the corresponding characteristic edge points are respectively fitted to obtain the fitted straight lines {l j,X}, {l j,Y}, and {l j,L} of the movement characteristic edges in the X, Y, and Z directions; among them, l j,X and l j,Y , l j,L are respectively the equivalent long edges of the rectangle of the measurement mark II and the straight edge of the mark I of F R,j (x, y) and F L,j (x, y).

3. A method for measuring the static and dynamic parameters of a low-frequency large-stroke vibration table based on binocular vision according to claim 1, characterized in that: The static parameter measurement of the guide rail bending of the low-frequency large-stroke vibration table is carried out in the following manner: First, divide the horizontal guide rail of the large-stroke vibration table into a number of equally spaced positions {s X,h , h = 1, 2, …, Q}, where Q is an odd number, and select the middle position (Q + 1) / 2 as the reference position of the guide rail horizontal line, and its vertical displacement is denoted as s Z,(Q+1) / 2 ; Therefore, the vertical displacement change at any position h is solved by the difference between the linear edge displacement s Z,h in the Z direction and s Z,(Q+1) / 2 ; The curvature α h at any position h is expressed as: Finally, based on the least squares principle, a straight line fitting is performed on the curvatures of different positions of the left and right guide rails centered on the reference position, and the average value of the left and right curvatures obtained by fitting is used as the guide rail curvature of the low-frequency large-stroke vibration table.

4. A method for measuring the static and dynamic parameters of a low-frequency large-stroke vibration table based on binocular vision according to claim 1, characterized in that: The X, Y, and Z displacements of the workbench surface of the low-frequency large-stroke vibration table are obtained by calculating the edge displacements s X (t), s Y (t), and s Z (t): where: ω v is the angular frequency of vibration, and are respectively the peaks of s X (t), s Y (t) and s Z (t); and are respectively the initial phases of s X (t), s Y (t) and s Z (t). Using the sine approximation method to fit the measured displacements s X (t), s Y (t) and s Z (t) as follows: When solving for the displacement in the X direction, in the formula: A X and B X are the sine displacement components; C X and D X are respectively the ultra-low frequency linear interference amount and the offset amount; Solving the overdetermined equations in the form of the first term of formula (3) based on the least squares principle can obtain the parameters A X and B X , obtaining the parameters A Y and B Y as well as A Z and B Z ; By solving the displacement parameters in the X, Y, and Z directions, s is obtained. X s Y s Z (t), the fitting displacement peak values and initial phases of s 5. A method for measuring the static and dynamic parameters of a low-frequency large-stroke vibration table based on binocular vision according to claim 1, characterized in that: The dynamic parameters of the low-frequency large-stroke vibration table include amplitude-frequency characteristics, repeatability, distortion rate, lateral ratio, and uniformity; the amplitude-frequency characteristics are the peak displacement values in the X direction at different frequencies, that is reflect the ability of the low-frequency large-stroke vibration table to provide the excitation displacement magnitude for the vibration sensor to be calibrated; Repeatability S of the low-frequency large-stroke vibration table Re It is the dispersion of the peak displacement in the X direction for multiple measurements at the selected test frequency, reflecting the stability of the peak displacement, and is given by the following formula: Among them, H is the number of measurements, not less than 10; is the average value of H measurements; Distortion rate S of the low-frequency large-stroke vibration table Dr It is defined as the ratio of the sum of the amplitudes of three adjacent higher harmonics to the amplitude of the fundamental harmonic, and is expressed as: Where: is the peak value of the i-th harmonic fitting of the displacement in the X direction; Transverse ratio S of low-frequency large-stroke vibration table Tr Defined as the ratio of the peak axial displacement to the peak transverse displacement and described as: Uniformity S of the low-frequency large-stroke vibration table NA It is the ratio of the peak deviation of the maximum absolute displacement at the boundary of the worktable surface to the peak displacement at the center position, and is expressed as: Among them, is the difference between the peak boundary displacement at the k-th position and the peak displacement at the center position and the value of k is 4.

6. A static and dynamic parameter measurement device for a low-frequency large-stroke vibration table based on binocular vision according to any one of claims 1-5, characterized in that: The device includes: a low-frequency large-stroke vibration table (1), a workbench surface (2), a high-contrast measurement mark I (3) including a straight line, a high-contrast measurement mark II (4) including a rectangle, a camera I (5), a camera II (6), a trigger (7), and an image processing and display unit (8); The workbench surface (2) of the low-frequency large-stroke vibration table (1) provides axial movement in the X direction and lateral movement in the Y and Z directions; the high-contrast measurement mark I (3) and the high-contrast measurement mark II (4) are respectively installed in the vertical and horizontal directions of the workbench surface (2) of the low-frequency large-stroke vibration table. The displacement of the mark I (3) in the Z direction is consistent with that of the workbench surface (2) of the low-frequency large-stroke vibration table, and the displacements of the mark II (4) in the X and Y directions are consistent with those of the workbench surface (2) of the low-frequency large-stroke vibration table; the optical axes of the camera I (5) and the camera II (6) are respectively perpendicular to the measurement mark I (3) and the measurement mark II (4); the trigger (7) is used to control the camera (I) and the camera II to simultaneously collect the motion sequence images of the measurement mark I (3) and the measurement mark II (4); the image processing and display unit (8) processes the collected motion sequence images, saves and displays the spatial motion displacement and the measurement results.

Images