Additive noise suppression method and device for optical fiber Fabry-Perot sensor

Abstract

In view of the noise problem of demodulation of high-precision optical fiber FP sensor, the present invention proposes a method and device for suppressing additive noise of optical fiber Fabry-Perot sensor, including obtaining the reflection spectrum of optical fiber Fabry-Perot sensor under the influence of additive noise; performing fast Fourier transform on the reflection spectrum of optical fiber Fabry-Perot sensor to obtain the imaginary and real parts at each order frequency component; performing arctangent operation on the imaginary and real parts at each order frequency component to establish an additive noise model of demodulation phase change caused by additive noise of characteristic peaks at each order frequency component; utilizing the correlation of demodulation phase change between characteristic peaks at each order frequency component and the non-correlation of additive noise, performing geometric averaging or weighted averaging on the demodulation phases of multiple characteristic peaks to achieve noise suppression. The present invention has strong noise suppression capability and provides basic theoretical guidance for the practical application of high-precision optical fiber FP sensor.

Description

Technical Field

[0001] The invention mainly relates to the technical field of noise suppression of optical fiber Fabry-Perot sensors, in particular to a method and a device for suppressing additive noise of optical fiber Fabry-Perot sensors. Background Art

[0002] The FFT demodulation scheme has the characteristics of simple calculation, large dynamic range and measurement of absolute cavity length, and is widely used in the field of fiber Fabry-Perot sensor (FP sensor) signal processing.

[0003] At present, some people have proposed to use the collected full spectrum information, and the demodulation process only requires non-zero-filled fast Fourier transform (FFT) to complete, successfully realizing real-time demodulation of cavity length. At present, some people have proposed to directly obtain the phase change caused by the change of cavity length by accumulating the FFT of the full spectrum at the signal eigenfrequency. This algorithm takes advantage of the full spectrum information and can improve the demodulation accuracy. However, the above demodulation schemes are limited by the reflectivity of the two ends of the fiber FP cavity, usually referring to low-fineness fiber FP sensors, resulting in low light energy utilization.

[0004] To solve this problem, the reflectivity of the FP cavity end face is improved by coating, which can increase the reflected light intensity and increase the number of multiplexed sensors. Therefore, it is necessary to conduct in-depth research on high-precision fiber FP sensors. At present, there is a proposal to perform FFT on the spectrum of high-precision fiber FP sensors, extract the second-order frequency components of the spectrum and restore the cavity length, thereby achieving optical sensitization, but the impact of noise on demodulation performance is not considered. As fiber FP sensors are widely used in medical detection, underwater acoustic detection, power detection and other fields, the noise suppression capability of the demodulation scheme has attracted much attention.

[0005] At present, there is little research on the noise suppression capability of the demodulation scheme of high-precision fiber FP sensors. Tracing back to the source, fiber FP sensors mainly focus on the light source intensity noise, optoelectronic device noise and noise in the optical transmission process. The FFT demodulation scheme has potential application value in the demodulation of high-precision fiber FP sensors. However, the existing technology is obviously insufficient in the establishment of the noise model and noise suppression measures of the FFT demodulation scheme in this process. Summary of the invention

[0006] In view of the noise problem of demodulation of a high-fineness optical fiber FP sensor, the present invention provides a method and device for suppressing additive noise of an optical fiber Fabry-Perot sensor.

[0007] In order to achieve the above object, the present invention is implemented by the following technical solutions:

[0008] On the one hand, the present invention provides a method for suppressing additive noise of an optical fiber Fabry-Perot sensor, comprising:

[0009] Obtain the reflection spectrum of the fiber optic Fabry-Perot sensor under the influence of additive noise;

[0010] Perform fast Fourier transform on the reflection spectrum of the optical fiber Fabry-Perot sensor to obtain the imaginary and real parts at each frequency component.

[0011] Performing arctangent operation on the imaginary and real parts at each order frequency component, establishing an additive noise model of demodulation phase change caused by characteristic peak additive noise at each order frequency component;

[0012] Based on the additive noise model, the demodulated phases of multiple characteristic peaks are processed by geometric averaging or weighted averaging to achieve noise suppression.

[0013] On the other hand, a device for suppressing additive noise of an optical fiber Fabry-Perot sensor is provided, comprising:

[0014] The first module is used to obtain the reflection spectrum of the optical fiber Fabry-Perot sensor under the influence of additive noise;

[0015] The second module is used to perform fast Fourier transform on the reflection spectrum of the optical fiber Fabry-Perot sensor to obtain the imaginary and real parts at each order frequency component;

[0016] The third module is used to perform an inverse tangent operation on the imaginary and real parts at each order frequency component to establish an additive noise model of the demodulation phase change caused by the characteristic peak additive noise at each order frequency component;

[0017] The fourth module is used to perform geometric averaging or weighted averaging on the demodulated phases of multiple characteristic peaks based on an additive noise model to achieve noise suppression.

[0018] In another aspect, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0019] Obtain the reflection spectrum of the fiber optic Fabry-Perot sensor under the influence of additive noise;

[0020] Perform fast Fourier transform on the reflection spectrum of the optical fiber Fabry-Perot sensor to obtain the imaginary and real parts at each frequency component.

[0021] Performing arctangent operation on the imaginary and real parts at each order frequency component, and establishing an additive noise model of demodulation phase change caused by characteristic peak additive noise at each order frequency component;

[0022] Based on the additive noise model, the demodulated phases of multiple characteristic peaks are processed by geometric averaging or weighted averaging to achieve noise suppression.

[0023] In another aspect, a computer readable storage medium stores a computer program, which, when executed by a processor, implements the following steps:

[0024] Obtain the reflection spectrum of the fiber optic Fabry-Perot sensor under the influence of additive noise;

[0025] Perform fast Fourier transform on the reflection spectrum of the optical fiber Fabry-Perot sensor to obtain the imaginary and real parts at each frequency component.

[0026] Performing arctangent operation on the imaginary and real parts at each order frequency component, and establishing an additive noise model of demodulation phase change caused by characteristic peak additive noise at each order frequency component;

[0027] Based on the additive noise model, the demodulated phases of multiple characteristic peaks are processed by geometric averaging or weighted averaging to achieve noise suppression.

[0028] After fast Fourier transforming the reflection spectrum of the optical fiber Fabry-Perot sensor, the present invention establishes a noise model of demodulation phase change at each order characteristic peak frequency component under the influence of additive noise. In order to suppress the influence of the demodulation system additive noise on the demodulation of the high-fineness optical fiber FP sensor signal, the demodulation phases of multiple characteristic peaks are subjected to geometric averaging or weighted averaging based on the additive noise model, using the correlation of the demodulation phase changes between the characteristic peaks at each order frequency component and the non-correlation of the additive noise, to achieve noise suppression.

[0029] The results show that the noise power spectral density is reduced after the phase variation equal ratio averaging processing using the first and second order characteristic peak demodulation, which reflects that the phase variation equal ratio averaging processing can achieve the effect of noise suppression. c 1 is set to a constant value, and the phase change is composed of a fixed geometric average of the first and second order characteristic peak demodulation phase changes, which can only reduce the noise power spectrum density to a certain extent.

[0030] When the light intensity coefficient ratio is 0.21, the noise power spectral density ratio of the weighted average demodulation phase change is 0.9487, which shows that the weighted average processing scheme has a stronger ability to suppress additive noise. The signal demodulation of the high-fineness fiber FP sensor only requires a simple weighted average to enhance the noise suppression capability, providing support for the application and development of noise suppression in high-fineness fiber FP sensors. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the structures shown in these drawings without paying creative work.

[0032] Figure 1 A flow chart of a method for suppressing additive noise in an optical fiber Fabry-Perot sensor provided by an embodiment;

[0033] Figure 2 A curve diagram showing the change in the ratio of geometric average noise power spectrum density as the ratio of light intensity coefficients changes, obtained by simulation of an embodiment;

[0034] Figure 3 A curve diagram showing the change of the weighting coefficient of the first-order characteristic peak when the light intensity coefficient ratio remains unchanged obtained by simulation of an embodiment;

[0035] Figure 4 is a curve diagram showing the change of the optimal weighted noise power spectrum density ratio and the corresponding optimal first-order characteristic peak weighting coefficient as the light intensity coefficient ratio changes in an embodiment;

[0036] Figure 5 is a curve diagram showing a change in the actual optimal weighted noise power spectrum density ratio as the light intensity coefficient ratio changes in an embodiment;

[0037] Figure 6 In one embodiment, c When 1=0.21, k 1. k 2 and relationship diagram. DETAILED DESCRIPTION

[0038] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0039] High-precision fiber Fabry-Perot sensors (FP sensors) usually use coating to enhance the reflectivity of the two end faces. After the signal light is reflected and transmitted by the first end face, the transmitted light undergoes multiple reflections and transmissions between the two end faces. The N reflected lights interfere with each other, forming the multi-beam interference phenomenon of the high-precision fiber FP sensor. and When two reflected lights interfere, their interference spectrum can be expressed as double-beam interference:

[0040] (1)

[0041] in is the incident light intensity; and For the , The intensity of the reflected light; and are the reflectivities of the first and second end faces of the FP sensor, respectively; is the phase difference generated by the light going back and forth in the FP cavity once, ; is the cavity length of the FP sensor; is the refractive index of the medium, in air Take 1; Refers to the central wavelength. Considering the interference between multiple reflected lights, and The difference between the reflected light Secondary reflection, i.e. , superimposing the reflected light that meets the above conditions can be expressed as:

[0042] (2)

[0043] make , formula (2) can be simplified as:

[0044] (3)

[0045] in is the frequency component corresponding to the characteristic peak of the signal interference model at this time; the phase change , is the change in cavity length.

[0046] The spectral model of the characteristic peak can be simplified to a double-beam interference model for processing. The phase change of the characteristic peak at the order frequency component is exactly the phase change of the low-fineness fiber FP sensor under the same cavity length. m The intensity of the characteristic peak is determined by the phase-dependent AC amplitude To facilitate subsequent analysis, the light intensity coefficient ratio is defined as the ratio that produces the first The AC amplitude ratio of the first-order characteristic peak to the second-order characteristic peak is .

[0047] The phase changes demodulated from the characteristic peaks of each frequency component follow an obvious proportional relationship, that is, , using the characteristic peaks at high-order frequency components to demodulate the signal can achieve the effect of enhancing the sensitivity of the sensor signal. In the demodulation phase change, the phase noise and additive noise are mainly considered. The phase noise increases exponentially with the increase of the phase change, while the additive noise components such as quantization noise and electronic noise are independent of the phase change. Make full use of the phase change after demodulation of the characteristic peaks at different frequency components. The relationship between them helps to reduce the impact of additive noise in the demodulation phase change and improve the signal-to-noise ratio of the demodulated signal.

[0048] For the spectrum of high-fineness fiber FP sensor, considering the error of spectral period caused by wavelength resolution, it is not enough to rely on the traditional full-spectrum FFT noise reduction effect.

[0049] Reference Figure 1 One embodiment provides a method for suppressing additive noise of an optical fiber Fabry-Perot sensor, comprising:

[0050] Obtain the reflection spectrum of the fiber optic Fabry-Perot sensor under the influence of additive noise;

[0051] Perform fast Fourier transform on the reflection spectrum of the optical fiber Fabry-Perot sensor to obtain the imaginary and real parts at each frequency component.

[0052] Performing arctangent operation on the imaginary and real parts at each order frequency component, and establishing an additive noise model of demodulation phase change caused by characteristic peak additive noise at each order frequency component;

[0053] Based on the additive noise model, the demodulated phases of multiple characteristic peaks are processed by geometric averaging or weighted averaging to achieve noise suppression.

[0054] Obtain the reflectance spectrum of a fiber Fabry-Perot sensor under the influence of additive noise:

[0055] (4)

[0056] in express Spectral model of characteristic peaks at the order frequency components; and Respectively represent The DC amplitude and AC amplitude of the order characteristic peak; is the eigenfrequency; It means that under the influence of additive noise, The phase change produced at the characteristic peak of the order; for The frequency component is at the The additive noise of the order characteristic peak, is a positive integer ranging from 1 to N;

[0057] Perform fast Fourier transform on the reflection spectrum of the optical fiber Fabry-Perot sensor, i.e., equation (4), and we get m The actual imaginary and real parts at the order frequency component;

[0058] (5)

[0059] in is the variable associated with the additive noise; is the eigenfrequency; Under ideal conditions (without any noise influence), The phase change produced at the characteristic peak of the order.

[0060] The actual demodulated phase change obtained by the inverse tangent operation of formula (5) is compared with the ideal condition. Considering the exponential increase of phase noise, in order to intuitively compare the demodulation effect of each order characteristic peak, The phase change of the demodulated characteristic peak is divided by . Calculate the phase change of the first-order characteristic peak with the ideal one, and get Phase change caused by additive noise of order characteristic peak:

[0061] (6)

[0062] By substituting the actual demodulation phase into equation (5), the inverse tangent operation is used to simplify the calculation to obtain:

[0063] (7)

[0064] In the calculation process ,and It is about A constant for the intensity of light at the characteristic peak of an order.

[0065] Corresponding to the above analysis, the parameters of additive noise are and , if the wavelength If the complete cycle is sampled, the phase change caused by the additive noise is offset. At this time, FFT can suppress the influence of additive noise on the phase change. However, in actual situations, the influence of the spectral wavelength resolution of the high-precision FP sensor causes deviations in the sampling period. If there is an error in the wavelength period , the parameters and Directly Integrating, equation (7) can be rewritten as:

[0066] (8)

[0067] Formula (8) is the additive noise model of demodulation phase change caused by characteristic peak additive noise at each frequency component established by the present invention. Indicates The demodulation phase change caused by the characteristic peak additive noise at the order frequency component, Under ideal conditions (without any noise influence), The phase change produced at the characteristic peak of the order, is the phase change produced at the first-order characteristic peak under ideal conditions; and It is about The constant of the light intensity at the characteristic peak of the order, For the The light intensity at the characteristic peak of the order, is the wavelength period of the spectrum, Indicates wavelength; is the eigenfrequency; for The frequency component is at the The additive noise of the order characteristic peaks.

[0068] To analyze the phase changes caused by additive noise, the power spectral density method is used, which reveals how the energy is distributed as the frequency changes, that is, the distribution of signal power at different frequencies in the frequency domain.

[0069] right Perform FFT to get the corresponding spectrum:

[0070] (9)

[0071] in is the spectrum of phase change caused by additive noise in the corresponding frequency band, specifically: .

[0072] The corresponding power spectral density is calculated by multiplying it with its conjugate:

[0073] (10)

[0074] After taking FFT of the spectrum of the high-fineness optical fiber FP sensor, a noise model of phase change caused by additive noise of characteristic peaks at each order frequency component was established. In order to effectively suppress the additive noise and make full use of the correlation of demodulated phase changes between characteristic peaks of each order and the non-correlation of additive noise, the demodulated phases of multiple characteristic peaks were processed by geometric averaging based on the additive noise model.

[0075] Specifically, in one embodiment, it is proposed to perform geometric averaging processing on the demodulation phase of the first-order characteristic peak and the demodulation phase of the second-order characteristic peak based on the additive noise model to achieve noise suppression, including: performing geometric averaging on the demodulation phase of the first-order characteristic peak and half of the demodulation phase of the second-order characteristic peak, and calculating the phase change caused by the additive noise after the geometric averaging of the demodulation phase. It further includes evaluating the noise suppression effect of the geometric averaging processing by power spectral density, including: calculating the power spectral density corresponding to the phase change caused by the additive noise after the geometric averaging of the demodulation phase; comparing the power spectral density corresponding to the phase change caused by the additive noise after the geometric averaging of the demodulation phase with the noise power spectral density of the first-order characteristic peak, and obtaining the geometric average noise power spectral density ratio; when the geometric average noise power spectral density ratio is less than 1, it means that the noise suppression ability is enhanced after the geometric averaging of the demodulation phase change.

[0076] In order to reduce the influence of additive noise, the demodulated phase of the first-order characteristic peak is geometrically averaged with half of the demodulated phase of the second-order characteristic peak, and the phase change caused by additive noise after geometric averaging of the demodulated phase is calculated:

[0077] (11)

[0078] Substituting the actual demodulation phase of the first and second order characteristic peaks into equation (11), we can obtain:

[0079] (12)

[0080] After the phase change geometric averaging is performed, the phase change caused by the additive noise is exactly the average of the phase changes caused by the first and second order additive noises. The geometric averaging method is conducive to suppressing the additive noise of the demodulated phase change. In order to evaluate the processing effect of this geometric averaging method, The corresponding power spectral density is calculated as:

[0081] (13)

[0082] In the power spectral density calculation of formula (13), since additive noise is usually modeled as Gaussian white noise, its characteristics lead to equal power spectral density at different frequencies. At the same time, this noise is uncorrelated between different frequencies, so the cross-correlation power spectral density at different frequencies is 0. Usually, the first-order characteristic peak demodulation phase contains the least noise information, so in order to measure the noise suppression effect of geometric averaging, the power spectral density of the phase change caused by its additive noise is compared with the noise power spectral density of the first-order characteristic peak. The geometric average noise power spectral density is S 1 is:

[0083] (14)

[0084] in c 1 is the light intensity coefficient ratio, that is c 1= A 2 / A 1= I A2 I A1 By observing from equation (14), we only need to analyze the relationship between the light intensity coefficient ratio and the noise power spectrum density ratio to evaluate the ability of the phase change equal ratio averaging method to suppress additive noise. S When 1 is less than 1, it means that the noise suppression capability is enhanced after the demodulation phase change is processed by proportional averaging.

[0085] The following simulation shows the ratio of the proportional average noise power spectrum density as the light intensity coefficient ratio changes. Figure 2 , a curve diagram of the change of the geometric average noise power spectrum density ratio as the light intensity coefficient ratio changes is obtained by simulation of an embodiment. As the light intensity coefficient ratio increases, the geometric average noise power spectrum density ratio gradually decreases. The increase in the light intensity coefficient ratio indicates that the light intensity of the second-order characteristic peak increases, and thus the amount of available spectral information increases accordingly. At the same time, c When 1=0.21, S 1=1, then c 1 increases the geometric average noise power spectrum density ratio gradually less than 1, reflecting that the phase change geometric average processing can achieve the effect of noise suppression. After the sensor is made, c 1 is set to a constant value, and the phase change is composed of a fixed geometric average of the first and second order characteristic peak demodulation phase changes, which can only reduce the noise power spectrum density to a certain extent.

[0086] In another embodiment of the present invention, the difference of additive noise in demodulation phase in different characteristic peaks and its irrelevance at different frequency components are further considered, and a weighted average processing is performed on the demodulation phase of the first-order characteristic peak and the demodulation phase of the second-order characteristic peak based on the additive noise model to achieve noise suppression, including: setting the weighted average coefficient of the demodulation phase of the first-order characteristic peak k 1, the weighted average coefficient of the demodulation phase of the second-order characteristic peak is 1- k 1. Take the weighted average of the demodulation phase changes of the first and second order characteristic peaks, calculate the phase change caused by the additive noise after the weighted average, and optimize the weighted average coefficient of the demodulation phase of the first order characteristic peak k 1. Get the best noise suppression effect.

[0087] Furthermore, the weighted average coefficient of the demodulation phase of the first-order characteristic peak is optimized k 1. Obtain the weighted average coefficient of the optimal first-order characteristic peak demodulation phase k 1, including:

[0088] Calculate the weighted average noise power spectral density corresponding to the phase change caused by the additive noise after weighted averaging;

[0089] Compare the weighted average noise power spectral density with the power spectral density of the demodulated phase additive noise of the first-order characteristic peak to obtain the weighted average noise power spectral density ratio;

[0090] When the light intensity coefficient ratio of the optical fiber Fabry-Perot sensor remains unchanged, the weighted average coefficient of the demodulation phase of the first-order characteristic peak k 1, the weighted average noise power spectrum density ratio has a minimum value, that is, the optimal noise weighted average power spectrum density ratio, and the weighted average coefficient corresponding to the optimal noise weighted average power spectrum density ratio is k 1 is the weighted average coefficient of the optimal first-order characteristic peak demodulation phase k 1.

[0091] Set the weighted average coefficient of the demodulation phase of the first-order characteristic peak k 1, the weighted average coefficient of the demodulation phase of the second-order characteristic peak is 1- k 1. Take the weighted average of the demodulated phase changes of the first and second order characteristic peaks, and calculate the phase change caused by the additive noise after the weighted average:

[0092] (15)

[0093] Similar to the phase change equal ratio averaging above, the actual demodulated phase change is substituted into equation (15) to simplify it to obtain , the phase change caused by the additive noise after weighted averaging is affected by the weighting weights. k When 1=0.5, the phase change of weighted average caused by the additive noise is consistent with the above-mentioned geometric average processing. The phase change after weighted average shows a wider application range in suppressing noise.

[0094] The corresponding weighted average noise power spectral density is further calculated as:

[0095] (16)

[0096] The weighted average noise power spectral density is compared with the power spectral density of the demodulated phase additive noise of the first-order characteristic peak, and the weighted average noise power spectral density ratio is obtained:

[0097] (17)

[0098] In order to analyze the noise suppression ability after weighted averaging, the light intensity coefficient ratio with the same phase change ratio is c When 1=0.21, the weighted average noise power spectrum density ratio is simulated. Figure 3 , which is the curve obtained by simulation when the light intensity coefficient ratio remains unchanged, with the change of the weighted coefficient of the first-order characteristic peak, the weighted average noise power spectrum density ratio has a minimum value, that is, the optimal noise weighted average power spectrum density ratio, such as Figure 3 As shown in Figure 1, it shows that there is an optimal first-order characteristic peak weighting coefficient to achieve the best noise suppression effect. That is, compared with the phase change proportional average, the first-order peak weighted average coefficient is k When 1=0.75, the minimum value of the weighted average noise power spectrum density ratio is 0.9487, which is significantly better than the noise suppression capability of the direct geometric average mentioned above. k When 1 is in the range of [0.5-1], the weighted average noise power spectral density ratio is always below 1, and k 1 is greater than 0.5, indicating that the first-order characteristic peak demodulation phase is dominant in the phase change after weighted average processing. c Under the condition of 1=0.21, the noise content in the first-order characteristic peak is relatively low, which helps to achieve better noise suppression effect.

[0099] It is clearly observed in the above phase change analysis that the light intensity coefficient ratio has a significant impact on the noise power spectrum density ratio. When the reflectivity of the two end faces of the high-fineness fiber FP sensor is determined, the corresponding light intensity coefficient ratio will also be determined accordingly. The following simulation analysis of the noise suppression effect after weighted averaging is carried out for different light intensity coefficient ratios to further explore its influence mechanism. Figure 4 As shown in the figure, the optimal weighted noise power spectrum density ratio and the corresponding optimal first-order characteristic peak weighting coefficient change curve obtained by simulation are as the light intensity coefficient ratio changes. k When 1 is 1 or 0, the phase change weighted average processing method has no effect on reducing noise, and the light intensity coefficient is c 1 shows a certain scope of application. c 1 changes in the interval [0.16, 0.38], the minimum value of the optimal weighted noise power spectrum density ratio gradually decreases, and the corresponding k 1 also decreases from 1 to 0. In this change process, it is noted that when c 1=0.25, k 1 = 0.5, the weighted average phase change is equivalent to the direct geometric average phase change, and the minimum value of the optimal weighted noise power spectrum density ratio is 0.8333. c1 further increases, the weighted proportional coefficient of the second-order characteristic peak begins to exceed the first order. This phenomenon shows that the second-order characteristic peak helps to suppress the noise after weighted averaging. That is, compared with the first-order characteristic peak, the signal strength of the second-order characteristic peak has exceeded its internal noise level, and the signal-to-noise ratio of the second-order characteristic peak is higher. At this time, in order to accurately measure the noise suppression ability of the weighted average processing method, the weighted average noise power spectral density is compared with the second-order characteristic peak noise power spectral density:

[0100] (18)

[0101] As the light intensity coefficient ratio changes, in order to accurately obtain the actual optimal weighted power spectrum density ratio noise suppression effect, the weighted average noise power spectrum density is compared and analyzed with the power spectrum density with the best noise suppression effect among the characteristic peaks participating in the weighted average. As the light intensity coefficient ratio increases, the actual geometric average noise power spectrum density does not always decrease. The actual optimal weighted noise power spectrum density ratio simulation result should be as follows: Figure 5 As shown, in c 1=0.25, k When 1=0.5, the optimal weighted noise power spectrum density ratio has a minimum value of 0.8333, at which time the additive noise suppression capability is most significant. Therefore, by reasonably designing the light intensity coefficient ratio, the optimal noise suppression effect can be further achieved.

[0102] In order to utilize more characteristic peak demodulation phase information to perform weighted averaging, thereby further enhancing the ability to suppress noise, the third-order characteristic peak demodulation phase is introduced to participate in the weighted average calculation. After this adjustment, the weighted average is compared with the noise power spectrum density of the first-order characteristic peak demodulation phase:

[0103] (19)

[0104] in k 2 is the weighted average coefficient of the demodulation phase of the second-order characteristic peak, c 2 is the intensity ratio of the third-order characteristic peak to the first-order characteristic peak, and .

[0105] exist c 1=0.21, the noise power spectrum density ratio of the third-order weighted average is simulated. Since the sum of the weighted coefficients of the demodulation phase of the first, second, and third-order characteristic peaks is 1, Figure 6 There are missing parts in the 3D image. Figure 6 As shown in the figure, the minimum value of the weighted average noise power spectrum density ratio of the three characteristic peaks calculated by simulation is 0.9787, and at this time k3 is 0. It is noted that when the third-order weighted average noise power spectrum density ratio takes the minimum value, the weighting coefficient is consistent with the result of the weighted average of the two characteristic peaks. This phenomenon shows that under the current conditions, the third-order characteristic peak is seriously interfered by noise, so that its signal is submerged by noise, which leads to the weighting coefficient of the third-order peak tending to 0 in the actual calculation, and fails to contribute to the overall noise suppression effect. In the future, the high-order characteristic peaks can be further utilized by studying the light intensity coefficient ratio to improve the noise suppression effect.

[0106] The present invention is aimed at the research on noise suppression of the demodulation scheme of high-precision optical fiber FP sensor, and analyzes the phase change of characteristic peak demodulation under the influence of additive noise in the frequency domain. Aiming at the problem that the spectral resolution of high-precision FP sensor causes sampling period deviation and increases demodulation noise, an additive noise model of characteristic peak demodulation phase change at each order frequency component under the influence of additive noise is established, and an innovative method of equal-ratio averaging of the demodulation phase of the first and second order characteristic peaks is proposed to suppress the influence of additive noise to a certain extent, and the noise suppression effect of this method is evaluated by power spectral density. In order to make full use of the information of characteristic peaks of each order, the idea of ​​weighted averaging is introduced, and the weighted average noise power spectral density is compared with the noise power spectral density with the strongest noise suppression effect among the characteristic peaks participating in the averaging. The results show that after weighted averaging, the phase change after weighted averaging shows a wider application range in noise suppression, and has stronger noise suppression ability, which provides certain basic theoretical guidance for the practical application of high-precision optical fiber FP sensor.

[0107] One embodiment provides an additive noise suppression device for an optical fiber Fabry-Perot sensor, comprising:

[0108] The first module is used to obtain the reflection spectrum of the optical fiber Fabry-Perot sensor under the influence of additive noise;

[0109] The second module is used to perform fast Fourier transform on the reflection spectrum of the optical fiber Fabry-Perot sensor to obtain the imaginary and real parts at each order frequency component;

[0110] The third module is used to perform an inverse tangent operation on the imaginary and real parts at each order frequency component to establish an additive noise model of the demodulation phase change caused by the characteristic peak additive noise at each order frequency component;

[0111] The fourth module is used to perform geometric averaging or weighted averaging on the demodulated phases of multiple characteristic peaks based on an additive noise model to achieve noise suppression.

[0112] The implementation methods of the above modules and the construction of the model can adopt the methods described in any of the above embodiments, which will not be repeated here.

[0113] On the other hand, the present invention provides a computer device, including a memory and a processor, the memory storing a computer program, and the processor implementing the steps of the additive noise suppression method of the optical fiber Fabry-Perot sensor provided in any of the above embodiments when executing the computer program. The computer device may be a server. The computer device includes a processor, a memory, a network interface and a database connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store sample data. The network interface of the computer device is used to communicate with an external terminal via a network connection.

[0114] On the other hand, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the additive noise suppression method for an optical fiber Fabry-Perot sensor provided in any of the above embodiments.

[0115] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0116] Matters not covered by the present invention are known technologies.

[0117] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0118] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention. It should be pointed out that, for a person of ordinary skill in the art, several modifications and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

[0119] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for suppressing additive noise in an optical fiber Fabry-Perot sensor, characterized in that: include: Obtain the reflection spectrum of the fiber optic Fabry-Perot sensor under the influence of additive noise; Perform fast Fourier transform on the reflection spectrum of the optical fiber Fabry-Perot sensor to obtain the imaginary and real parts at each frequency component. Perform arc tangent operation on the imaginary and real parts at each order frequency component, and establish an additive noise model of demodulation phase change caused by characteristic peak additive noise at each order frequency component, where the additive noise model is as follows: ； in Indicates The demodulation phase change caused by the characteristic peak additive noise at the order frequency component, Under ideal conditions, The phase change produced at the characteristic peak of the order, is the phase change produced at the first-order characteristic peak under ideal conditions; and is a constant related to the intensity of light. For the The light intensity at the characteristic peak of the order, is the wavelength period of the spectrum, Indicates wavelength; is the eigenfrequency of the signal; for The frequency component is at the Additive noise of the order characteristic peak; Based on the additive noise model, the demodulated phases of multiple characteristic peaks are processed by geometric averaging or weighted averaging to achieve noise suppression.

2. The additive noise suppression method of an optical fiber Fabry-Perot sensor according to claim 1, characterized in that: Based on the additive noise model, geometric averaging is performed on the demodulation phases of the first-order and second-order characteristic peaks to achieve noise suppression, including: geometric averaging the demodulation phase of the first-order characteristic peak and half of the demodulation phase of the second-order characteristic peak, and calculating the phase change caused by the additive noise after the geometric averaging of the demodulation phase.

3. The additive noise suppression method of the optical fiber Fabry-Perot sensor according to claim 2, characterized in that: It also includes the evaluation of the noise suppression effect of the equal-ratio averaging process through power spectral density, including: Calculate the power spectral density corresponding to the phase change caused by the additive noise after the demodulated phase is uniformly averaged; The power spectrum density corresponding to the phase change caused by the additive noise after the demodulation phase geometric average is compared with the noise power spectrum density of the first-order characteristic peak to obtain the geometric average noise power spectrum density ratio; When the geometric average noise power spectrum density ratio is less than 1, it means that the noise suppression capability is enhanced after the geometric average processing of the demodulation phase change.

4. The additive noise suppression method of the optical fiber Fabry-Perot sensor according to claim 1, characterized in that: Based on the additive noise model, the first-order and second-order characteristic peak demodulation phases are weighted averaged to achieve noise suppression, including: Set the weighted average coefficient of the demodulation phase of the first-order characteristic peak k 1, the weighted average coefficient of the demodulation phase of the second-order characteristic peak is 1- k 1. Take the weighted average of the demodulation phase changes of the first and second order characteristic peaks, calculate the phase change caused by the additive noise after the weighted average, and optimize the weighted average coefficient of the demodulation phase of the first order characteristic peak by calculating the phase noise level. k 1. Get the best noise suppression effect.

5. The additive noise suppression method for an optical fiber Fabry-Perot sensor according to claim 4, characterized in that: Optimizing the weighted average coefficient of the demodulation phase of the first-order characteristic peak k 1. Obtain the weighted average coefficient of the optimal first-order characteristic peak demodulation phase k 1, including: Calculate the weighted average noise power spectral density corresponding to the phase change caused by the additive noise after weighted averaging; Compare the weighted average noise power spectral density with the power spectral density of the demodulated phase additive noise of the first-order characteristic peak to obtain the weighted average noise power spectral density ratio; When the light intensity coefficient ratio of the optical fiber Fabry-Perot sensor remains unchanged, the weighted average coefficient of the demodulation phase of the first-order characteristic peak k 1, the weighted average noise power spectrum density ratio has a minimum value, that is, the optimal noise weighted average power spectrum density ratio, and the weighted average coefficient corresponding to the optimal noise weighted average power spectrum density ratio is k 1 is the weighted average coefficient of the optimal first-order characteristic peak demodulation phase k 1.

6. An additive noise suppression device for an optical fiber Fabry-Perot sensor, characterized in that: include: The first module is used to obtain the reflection spectrum of the optical fiber Fabry-Perot sensor under the influence of additive noise; The second module is used to perform fast Fourier transform on the reflection spectrum of the optical fiber Fabry-Perot sensor to obtain the imaginary and real parts at each order frequency component; The third module is used to perform inverse tangent operation on the imaginary and real parts at each order frequency component, and establish an additive noise model of demodulation phase change caused by characteristic peak additive noise at each order frequency component, wherein the additive noise model is as follows: ； in Indicates The demodulation phase change caused by the characteristic peak additive noise at the order frequency component, Under ideal conditions, The phase change produced at the characteristic peak of the order, is the phase change produced at the first-order characteristic peak under ideal conditions; and is a constant related to the intensity of light. For the The light intensity at the characteristic peak of the order, is the wavelength period of the spectrum, Indicates wavelength; is the eigenfrequency of the signal; for The frequency component is at the Additive noise of the order characteristic peak; The fourth module is used to perform geometric averaging or weighted averaging on the demodulated phases of multiple characteristic peaks based on an additive noise model to achieve noise suppression.

Images