A method for analyzing the holographic spectrum of a bearing housing based on machine vision

Abstract

The present invention relates to the field of vibration testing technology and vibration signal analysis, and particularly to a method for analyzing the holographic spectrum of a bearing housing based on machine vision. The present invention uses an industrial camera or a machine vision device to acquire the video of the rotor bearing housing during operation, and uses a Gabor filter to perform phase separation on each frame of the acquired video. Based on the separation results, the vibration signal ① in the horizontal direction of the bearing housing and the vibration signal ② in the vertical direction are obtained, thereby realizing non-contact vibration signal acquisition, simplifying the holographic spectrum analysis process, and being easier to implement. The present invention measures vibration with the bearing housing as the object. Compared with the traditional method that uses the rotating shaft as the object, the vibration of the bearing housing is easier to measure.

Description

Technical Field

[0001] The present invention relates to the field of vibration testing technology and vibration signal analysis, and particularly relates to a method for analyzing the holographic spectrum of a bearing housing based on machine vision. Background Art

[0002] When performing vibration analysis, in many cases, vibration data is obtained through a single sensor, and then subsequent processing is carried out using time-domain and frequency-domain analysis methods. However, when the mechanical operating conditions are relatively complex, the data from a single sensor cannot comprehensively reflect the state of the mechanical equipment. Information fusion technology is an effective means to solve this problem, which integrates and converts data from different sensors to provide more comprehensive and accurate information. Holographic spectrum is a commonly used information fusion technology at present. In traditional holographic spectrum analysis methods, it is usually necessary to install multiple displacement sensors on the same cross-section of the equipment and collect vibration data simultaneously. In most cases, fixtures need to be designed by oneself to fix the sensors, and there are relatively strict requirements for the accuracy of the angular arrangement when installing the sensors. At the same time, in order to ensure that the starting sampling times of multiple sensors are consistent, time preprocessing is required, making the entire holographic spectrum analysis process too complex and not conducive to implementation. Summary of the Invention

[0003] The purpose of the present invention is to provide a method for analyzing the holographic spectrum of a bearing housing based on machine vision to simplify the analysis process and facilitate implementation.

[0004] To achieve the above purpose, the present invention adopts the following technical solutions:

[0005] A method for analyzing the holographic spectrum of a bearing housing based on machine vision includes the following steps:

[0006] a. Collect the video of the bearing housing of the rotor system during operation;

[0007] b. Convolve each frame image in the collected video with a Gabor filter to separate the phase Calculate the phase difference Δφ between the current image and the first frame image, and then according to the relational expression between the phase difference and the displacement difference calculate the pixel displacement Δx, where λ is the filter wavelength; adjust the angle of the filter according to requirements to obtain the pixel displacements in the horizontal and vertical directions of the current image;

[0008] c. Calculate the scale factor k according to the actual size of the object and the pixel size of the object in the picture taken by the video acquisition device; multiply the pixel displacement obtained in step b by the scale factor k to convert the pixel displacement into an actual physical displacement, thereby obtaining the horizontal vibration signal ① and the vertical vibration signal ② of the bearing housing;

[0009] d. Perform a fast Fourier transform (FFT) on vibration signals ① and ② to obtain their corresponding spectra; select the main frequency f and each order of multiple frequencies according to the frequency corresponding to the highest amplitude in the spectrum, and correct the amplitudes, frequencies, and phases of the selected multiple frequencies of each order.

[0010] e. Perform holographic spectrum synthesis on the corrected spectrum and calculate the parameters during the operation of the rotor bearing housing. The parameters include the length of the major axis, the length of the minor axis, the inclination angle of the major axis, the eccentricity, etc.

[0011] Furthermore, in step a, an industrial camera or machine vision device is used to collect the video of the rotor system bearing housing during operation.

[0012] Even further, step a also includes setting the industrial camera before collecting the video of the bearing housing during operation, specifically including the following sub-steps:

[0013] (1) Set the camera parameters according to the actual working conditions of the rotor system. The camera parameters include: frame rate and sampling rate, where the frame rate is set based on the sampling theorem, and the sampling rate is set according to the accuracy requirements.

[0014] (2) Make the camera imaging plane as parallel as possible to the plane where the bearing housing is located.

[0015] Furthermore, the method for correcting the amplitudes, frequencies, and phases of each order of multiple frequencies in step d includes the ratio correction method, the energy centroid correction method, or the phase difference correction method.

[0016] After adopting the above technical solutions, the present invention has the following beneficial effects:

[0017] The present invention uses an industrial camera or machine vision device to obtain the video of the rotor bearing housing during operation, performs phase separation on each frame of the collected video using a Gabor filter, and obtains the horizontal vibration signal ① and the vertical vibration signal ② of the bearing housing based on the separation results. Thus, non-contact vibration signal acquisition is achieved, the holographic spectrum analysis process is simplified, and it is easier to implement.

[0018] The present invention measures vibration with the bearing housing as the object. Compared with the traditional method that takes the rotating shaft as the object, the vibration of the bearing housing is easier to measure. Description of the Drawings

[0019] Figure 1 It is the specific implementation flowchart of the present invention;

[0020] Figure 2 It is the horizontal vibration signal ① and the vertical vibration signal of the bearing housing obtained in the embodiment, where Figure 2 (a) is the horizontal vibration signal ①, Figure 2 (b) is the vertical vibration signal ②;

[0021] Figure 3 For the embodiment, information fusion is performed on the corrected spectrum, and the ellipses of each order of harmonic are drawn;

[0022] Figure 4 The holographic spectrum parameter diagrams of each order of harmonic calculated for the embodiment. Specific Embodiments

[0023] The present invention will be described based on specific embodiments. During the following description process, unless clearly required by the context, the words such as "comprising" and "including" mentioned throughout the specification and claims should be interpreted as having the meaning of inclusion rather than exhaustive or exclusive meaning: that is, it means "including but not limited to".

[0024] Embodiment

[0025] As Figure 1 shown, a method for holographic spectrum analysis of a bearing housing based on machine vision provided in this embodiment includes the following steps:

[0026] a. Collect videos of the bearing housing of the rotor system during operation. The device for collecting videos of the bearing housing of the rotor system during operation can be an industrial camera or other vision devices. In this embodiment, an industrial camera is selected to collect videos of the bearing housing during operation. For the convenience of subsequent unit conversion, the camera should be directly facing the plane where the bearing housing is located. The camera parameters are set according to the actual working conditions of the rotor system. Among them, the resolution will affect the measurement accuracy of vibration signals, and the frame rate is the sampling frequency. When setting the frame rate, the sampling theorem should be satisfied.

[0027] b. Convolve each frame of the image in the video with a Gabor filter to separate the phase Calculate the phase difference Δφ between the current image and the first frame image. According to the relationship between the phase difference and the displacement difference Obtain the relationship of pixel displacement Δx, where λ is the wavelength of the filter. Adjust the angle of the filter according to application requirements to obtain the pixel displacements in the horizontal and vertical directions of the current image.

[0028] c. Place an object with a known size on the bearing housing plane, take pictures and obtain the pixel size of the object, and calculate the scale factor k, where k = actual size / pixel size. Multiply the pixel displacement in step b by the scale factor k to convert the pixel displacement into an actual physical displacement, thereby obtaining the vibration signal ① in the horizontal direction and the vibration signal ② in the vertical direction of the bearing housing.

[0029] d. Perform a fast Fourier transform on vibration signals ① and ② to obtain their corresponding spectra; select the main frequency f and the elliptical shapes of each order of harmonic according to the frequency corresponding to the highest amplitude in the spectrum. Since spectral leakage will cause errors in amplitude, frequency, and phase in the spectrum, it is necessary to correct the amplitude, frequency, and phase of each order of harmonic selected before holographic spectrum synthesis. The methods for correcting the amplitude, frequency, and phase of each order of harmonic include the ratio correction method, the energy centroid correction method, or the phase difference correction method. In this embodiment, the ratio correction method is preferably used.

[0030] e. Perform holographic spectrum synthesis on the corrected spectrum and calculate parameters such as the major axis length, minor axis length, major axis inclination angle, eccentricity, etc.

[0031] The horizontal vibration signal ① and the vertical vibration signal ② of the bearing housing obtained in this embodiment according to the above steps are as Figure 2 shown, where Figure 2 (a) is the horizontal vibration signal ①, Figure 2 (b) is the vertical vibration signal ②. After performing a fast Fourier transform (FFT) on vibration signal ① and vibration signal ②, the corresponding spectra are obtained. Select the frequency corresponding to the peak point in the spectrum as the main frequency, and use the ratio correction method to correct the frequency, amplitude, and phase of the main frequency and its harmonics. Perform information fusion on the corrected spectrum and draw the elliptical shapes of each order of harmonic as Figure 3 shown. At the same time, calculate the holographic spectrum parameters of each order of harmonic as Figure 4 shown, where S x and C x respectively represent the sine coefficient and cosine coefficient in the horizontal direction, and S y and C y respectively represent the sine coefficient and cosine coefficient in the vertical direction.

[0032] In summary, a method for holographic spectrum analysis of a bearing housing based on machine vision provided in this embodiment realizes non-contact vibration signal acquisition, simplifies the holographic spectrum analysis process, and is easier to implement.

Claims

1. A method for analyzing the holographic spectrum of a bearing housing based on machine vision, characterized in that, It includes the following steps: a. Collect the video of the bearing housing of the rotor system during operation; b. Convolve each frame image in the collected video with a Gabor filter to separate the phase Calculate the phase difference Δφ between the current image and the first frame image, and then according to the relationship between the phase difference and the displacement difference Calculate the pixel displacement Δx, where λ is the filter wavelength; Adjust the angle of the filter according to the requirement to obtain the pixel displacements in the horizontal and vertical directions of the current image; c. Calculate the scale factor k according to the actual size of the object and the pixel size of the object in the picture taken by the acquisition video device; Multiply the pixel displacement obtained in step b by the scale factor k to convert the pixel displacement into an actual physical displacement, thereby obtaining the horizontal vibration signal ① and the vertical vibration signal ② of the bearing housing; d. Perform a fast Fourier transform on the vibration signals ① and ② to obtain their corresponding spectra; Select the main frequency f and each order of multiple frequencies according to the frequency corresponding to the highest amplitude in the spectrum, and correct the amplitudes, frequencies and phases of the selected multiple frequencies of each order; e. Perform holographic spectrum synthesis on the corrected spectrum and calculate the parameters during the operation of the rotor bearing housing. The parameters include the long axis length, short axis length, long axis inclination angle, and eccentricity parameter.

2. The method for analyzing the holographic spectrum of a bearing block based on machine vision according to claim 1, wherein: In step a, an industrial camera or a machine vision device is used to collect the video of the bearing housing of the rotor system during operation.

3. The method for analyzing the holographic spectrum of a bearing housing based on machine vision according to claim 2, characterized in that: Step a also includes setting the industrial camera before collecting the video of the bearing housing during operation, specifically including the following sub-steps: (1) Set the camera parameters according to the actual working conditions of the rotor system. The camera parameters include: frame rate and sampling rate, where the frame rate is set based on the sampling theorem, and the sampling rate is set according to the accuracy requirement.

4. The method for analyzing the holographic spectrum of a bearing housing based on machine vision according to claim 1, characterized in that: The method for correcting the amplitudes, frequencies and phases of each order of multiple frequencies in step d includes the ratio correction method, the energy centroid correction method or the phase difference correction method.

Images