A method and system for denoising phase signals of optical cable vibration signals

Abstract

The application discloses a kind of optical cable vibration signal phase signal noise reduction method and system, the phase signal noise reduction method includes: the vibration signal of optical cable is collected, and the phase signal of disturbance point is obtained;Phase signal is decomposed, and high-frequency component and low-frequency component are obtained;High-frequency component is decomposed quantization processing, and the high-frequency component after noise reduction is obtained;Reconstruct low-frequency component and the high-frequency component after noise reduction, and the phase signal after noise reduction is obtained.The vibration intrusion signal on sensing link can be restored more accurately by the application.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of optical fiber sensing, and particularly relates to a phase signal denoising method and system for optical cable vibration signals. BACKGROUND

[0002] Phase-sensitive optical time domain reflectometry (Φ-OTDR) is an advanced new type of distributed optical fiber sensing technology. Φ-OTDR can calculate the waveform of distributed vibration by identifying the differential phase in the sensing system. Due to this outstanding technology, multiple physical parameters can be accurately measured at the same time, thereby realizing the vibration detection of long-distance optical fibers while maintaining a high spatial resolution along the sensing optical fiber. In recent years, it has been widely used in remote monitoring of vibration signals, and is often used in the fields of oil pipeline leakage detection, speed monitoring, perimeter protection, etc.

[0003] However, due to the high sensitivity of backscattered light, it is easily affected by the surrounding environment in practical applications. Furthermore, random noise such as electrical noise and phase noise in the laser can degrade the Φ-OTDR signal, increasing the difficulty of vibration detection. Therefore, to locate interference and improve the signal-to-noise ratio (SNR) of the Φ-OTDR, some denoising methods need to be introduced to improve the overall performance of the system. For example, the paper "AReal-Time Distributed Deep Learning Approach for Intelligent Event Recognition in Long-Distance Pipeline Monitoring with DOFS" by Jiping Chen, Huijuan Wu, Xiangrong Liu, Yao Xiao, and Yunjiang Rao proposes a real-time distributed deep learning model that uses an efficient one-dimensional convolutional neural network (1-D CNN) to learn the identifiable features of different interferences and automatically identifies them by training the original event data (signal). Before constructing and training the 1-D CNN network, they proposed a further wavelet packet decomposition (WPD) denoising method. However, simple wavelet packet decomposition cannot handle nonlinear and non-stationary signals well. The paper "Disturbance pattern recognition based on an ALSTM in a long-distance Φ-OTDR sensing system" by X Chen and C Xu employs an adaptive denoising method based on spectral subtraction to enhance signal features. However, spectral subtraction requires estimation of the noise spectrum, which may lead to over-subtraction and affect the final denoising effect.For example, CN109726642A discloses a denoising method for distributed optical fiber vibration signals based on variational mode decomposition, including: acquiring distributed optical fiber vibration signals; performing variational mode decomposition on the distributed optical fiber vibration signals, which can divide the original signal into components of different modes and filter out high-frequency noise in the vibration signal during the decomposition process, thus achieving good signal-to-noise separation performance; performing wavelet threshold denoising on each mode component, which can further filter out shot noise and low-frequency noise in the environment; and reconstructing the processed signal to obtain denoised optical fiber vibration data. This method can improve the signal-to-noise ratio and optical fiber vibration... While the signal processing accuracy is improved, it targets the vibration signal of the link, whereas this application targets the phase signal of the disturbance point. This is beneficial for reducing the impact of noise on the phase signal of the disturbance point, thereby more accurately reconstructing the disturbance signal on the sensing link. Secondly, in the modal decomposition process of this application, the multiple jumps of some signal local extrema within a very short time interval may lead to mode aliasing, thus affecting the final noise reduction effect. In addition, this application does not provide a detailed description of the wavelet threshold denoising method used. The threshold quantization method used may cause a constant deviation between the estimated wavelet coefficients and the true wavelet coefficients or discontinuities at the threshold, ultimately causing deviations in the reconstructed signal.

[0004] Therefore, how to address the impact of the surrounding environment on Φ-OTDR signals is a problem that urgently needs to be solved in this field. Summary of the Invention

[0005] To address the shortcomings of the existing technology, this invention provides a method and system for denoising the phase signal of optical cable vibration signals. This reduces the impact of noise on the phase signal and more accurately restores the disturbance signal on the sensing link.

[0006] In a first aspect, the present invention provides a method for denoising the phase signal of an optical cable vibration signal, comprising:

[0007] Vibration signals from the optical cable link are collected to obtain the phase signal at the disturbance point;

[0008] The phase signal is decomposed to obtain high-frequency and low-frequency components;

[0009] The high-frequency components are decomposed and quantized to obtain the noise-reduced high-frequency components.

[0010] The low-frequency components and the noise-reduced high-frequency components are reconstructed to obtain the noise-reduced phase signal.

[0011] Furthermore, optical pulse signals are injected into the optical cable link using the Φ-OTDR system, and vibration signals in the optical cable link are collected.

[0012] Furthermore, vibration signals from the optical cable are collected to obtain the phase signals at the disturbance points, including:

[0013] The acquired vibration signal is quadrature demodulated, and the phase signal of the disturbance point is obtained by arctangent operation between the obtained in-phase component and the quadrature component.

[0014] Furthermore, the phase signal is decomposed to obtain high-frequency and low-frequency components, including:

[0015] The phase signal is processed based on ensemble empirical mode decomposition to obtain multiple intrinsic mode components;

[0016] The continuous mean square error is obtained by approximate reconstruction and analysis of the intrinsic modal components.

[0017] Based on continuous mean square error, the intrinsic modal components are divided into high-frequency components and low-frequency components.

[0018] Furthermore, the phase signal is processed based on ensemble empirical mode decomposition to obtain multiple intrinsic mode components, including:

[0019] Gaussian white noise with different amplitudes is added to the phase signal a predetermined number of times;

[0020] Empirical mode decomposition is performed on the phase signal after each addition of Gaussian white noise to obtain a predetermined number of intrinsic mode components.

[0021] The intrinsic modal components obtained by decomposing the set by the average predetermined number of times are given the final predetermined number of intrinsic modal components.

[0022] The predetermined number of intrinsic mode components is determined by the phase signal itself; that is, different phase signals can be obtained by empirical mode decomposition with different numbers of intrinsic mode components.

[0023] Furthermore, Gaussian white noise with different amplitudes is added to the phase signal a predetermined number of times, including:

[0024] x i (t)=x(t)+n i (t)

[0025] In the formula, i = 1, 2, 3, ..., M, where M is the number of times Gaussian white noise is added, and n i (t) represents the Gaussian white noise sequence added for the i-th time, x i x(t) represents the phase signal after adding Gaussian white noise for the i-th time, and x(t) represents the original phase signal without adding Gaussian white noise.

[0026] Empirical mode decomposition is performed on the phase signal after each addition of Gaussian white noise to obtain a predetermined number of intrinsic mode components, including:

[0027] Empirical mode decomposition (EMD) is performed on the phase signal after each addition of Gaussian white noise a predetermined number of times to obtain residual components and intrinsic mode components with a quantity corresponding to the predetermined number of times; satisfying the following relationship:

[0028]

[0029] In the formula, c i,j (t) represents the j-th intrinsic mode component obtained after the i-th addition of Gaussian white noise, r i,j (t) represents the j-th residual component obtained after the i-th addition of Gaussian white noise, where J is the number of intrinsic modal components, j = 1, 2, 3, ..., J;

[0030] By averaging the intrinsic modal components obtained from each decomposition, a final predetermined number of intrinsic modal components are given, including:

[0031] Superimpose the intrinsic modal components obtained by the same number of decompositions;

[0032] The superimposed intrinsic mode components are averaged to obtain the average intrinsic mode components after each decomposition.

[0033] Integrate all average intrinsic modal components to obtain the final predetermined number of intrinsic modal components;

[0034] The intrinsic modal components satisfy the following relationship:

[0035]

[0036] In the formula, c j (t) represents the j-th intrinsic mode component obtained by set averaging.

[0037] Furthermore, the continuous mean square error is obtained by approximate reconstruction and analysis of the intrinsic modal components, including:

[0038] All residual components are superimposed and averaged to obtain the final residual components;

[0039] The final predetermined number of intrinsic modal components are sorted according to the decomposition order;

[0040] The approximate reconstructed signal is obtained by superimposing any intrinsic mode component other than the first one with all subsequent intrinsic mode components and the final residual component, specifically satisfying the following relationship:

[0041]

[0042] In the formula, c represents the approximate reconstructed signal of the k-th intrinsic mode component. i(t) represents the i-th intrinsic modal component, r C (t) represents the final residual component, and C represents the number of intrinsic modal components obtained from the final decomposition.

[0043] Each approximate reconstructed signal is analyzed and processed along with its adjacent approximate reconstructed signal to obtain a continuous mean square error, which satisfies the following relationship:

[0044]

[0045] In the formula, Let t represent the continuous mean square error of the k-th approximate reconstructed signal, N represent the sequence length of the approximate reconstructed signal, and t represent the length of the sequence. i c represents the i-th point of the approximate reconstructed signal. k (t i ) represents the value of the k-th intrinsic modal component at the i-th point.

[0046] Furthermore, based on continuous mean square error, the intrinsic modal components are divided into high-frequency components and low-frequency components, including:

[0047] Compare all consecutive mean square errors and give the smallest consecutive mean square error;

[0048] Based on the superposition process of intrinsic modal components and the analysis and processing process of the approximate reconstructed signal, the intrinsic modal component corresponding to the minimum continuous mean square error is obtained;

[0049] The intrinsic modal component corresponding to the minimum continuous mean square error is taken as the dividing point;

[0050] The intrinsic modal components before the boundary point are classified as high-frequency components, and the intrinsic modal components after the boundary point are classified as low-frequency components.

[0051] Furthermore, the high-frequency components are decomposed and quantized to obtain the noise-reduced high-frequency components, including:

[0052] Based on the preset wavelet basis functions, wavelet decomposition is performed on all high-frequency components to obtain their respective wavelet coefficient sets, which satisfy the following relationship:

[0053] I W ={ω1,ω2,…,ω i ,…,ω L}

[0054] In the formula, I w Let ω be the set of wavelet coefficients. i Let L be the i-th wavelet coefficient, and L be the number of wavelet coefficients.

[0055] The threshold is determined based on the intrinsic modal components, satisfying the following relationship:

[0056]

[0057] In the formula, T is the threshold, σ is the variance of the noise, and N is the sequence length of the intrinsic modal components;

[0058] Based on a threshold, a compromise threshold quantization process is performed on all wavelet coefficient sets to obtain the wavelet coefficient set after compromise threshold quantization; the following relationship is satisfied:

[0059] I WT ={ω 1T ,ω 2T ,…,ω iT ,…,ω LT}

[0060]

[0061] In the formula, I WT For the set of wavelet coefficients after compromise threshold quantization, ω iT Let be the wavelet coefficients after the i-th threshold quantization, and α be the added adjustment factor and α∈[0,1].

[0062] The wavelet coefficient set after threshold quantization is reconstructed to obtain the denoised high-frequency components.

[0063] Furthermore, the low-frequency components and the denoised high-frequency components are reconstructed to obtain the denoised phase signal, including:

[0064] Integrating all the noise-reduced high-frequency components yields the overall high-frequency components, which satisfy the following relationship:

[0065]

[0066] In the formula, I WT,i (t) represents the i-th noise-reduced high-frequency component, x H (t) represents the overall high-frequency components.

[0067] The high-frequency and low-frequency components are superimposed to obtain the noise-reduced phase signal, which satisfies the following relationship:

[0068] x D (t)=x H (t)+x L (t)

[0069] In the formula, x D (t) represents the noise-reduced phase signal, x L (t) represents the low-frequency component.

[0070] Secondly, the present invention also provides a phase signal noise reduction system for optical cable vibration signals, employing the aforementioned phase signal noise reduction method for optical cable vibration signals, wherein the identification system includes:

[0071] The acquisition and processing module is used to acquire vibration signals from the optical cable link and obtain the phase signal of the disturbance point.

[0072] The signal decomposition module is used to decompose the phase signal to obtain high-frequency and low-frequency components.

[0073] The high-frequency noise reduction module is used to decompose and quantize high-frequency components to obtain noise-reduced high-frequency components.

[0074] The signal reconstruction module is used to reconstruct the low-frequency components and the noise-reduced high-frequency components to obtain the noise-reduced phase signal.

[0075] Furthermore, the signal decomposition module is also used for:

[0076] The phase signal is processed based on ensemble empirical mode decomposition to obtain multiple intrinsic mode components;

[0077] The continuous mean square error is obtained by approximate reconstruction and analysis of the intrinsic modal components.

[0078] Based on continuous mean square error, the intrinsic modal components are divided into high-frequency components and low-frequency components.

[0079] Furthermore, the signal decomposition module is also used for:

[0080] Add Gaussian white noise of different amplitudes a predetermined number of times to the phase signal;

[0081] Empirical mode decomposition is performed on the phase signal after each addition of Gaussian white noise to obtain a predetermined number of intrinsic mode components.

[0082] The intrinsic modal components obtained by decomposing the set by the average predetermined number of times are given the final predetermined number of intrinsic modal components.

[0083] Furthermore, the high-frequency noise reduction module is also used for:

[0084] Based on the preset wavelet basis functions, all high-frequency components are decomposed into wavelet coefficient sets to obtain their respective sets.

[0085] Thresholds are determined based on intrinsic modal components;

[0086] Based on the threshold, all wavelet coefficient sets are subjected to compromise threshold quantization to obtain the wavelet coefficient set after compromise threshold quantization.

[0087] The wavelet coefficient set after threshold quantization is reconstructed to obtain the denoised high-frequency components.

[0088] Thirdly, the present invention also provides a phase signal noise reduction device for optical cable vibration signals, comprising a transmitting unit, a modulation unit, a detection unit and a processing unit, wherein the transmitting unit and the modulation unit are connected, and the transmitting unit, the detection unit and the processing unit are connected in sequence.

[0089] The transmitting unit injects optical pulse signals into the optical cable under test through the modulation unit;

[0090] The detection unit acquires vibration signals from the optical cable under test;

[0091] The processing unit uses the phase signal noise reduction method of the optical cable vibration signal described above to perform noise reduction processing on the vibration signal obtained by the detection unit.

[0092] Furthermore, the transmitting unit also includes a narrow linewidth laser and a 1*2 fiber coupler. The splitting ratio of the 1*2 fiber coupler is 90:10. The 1*2 fiber coupler divides the signal emitted by the narrow linewidth laser into a first optical path A and a second optical path B. The first optical path A serves as the probe light input modulation unit, and the second optical path B serves as the local oscillator light input detection unit.

[0093] Furthermore, the modulation unit includes an acousto-optic modulator and an erbium-doped fiber amplifier.

[0094] Furthermore, the optical cable under test includes an optical fiber circulator and a sensing optical fiber, and the optical cable under test is connected to an erbium-doped optical fiber amplifier.

[0095] Furthermore, the sensing fiber is a single-mode fiber with a length of 14km. A piezoelectric ceramic tube is used as a disturbance source at a distance of 10km from the fiber under test to apply a standard sinusoidal signal.

[0096] Furthermore, the detection unit includes a 2x2 fiber optic coupler, a balanced photodetector, and a data acquisition card;

[0097] The 2*2 fiber coupler has a splitting ratio of 50:50 and is connected to a narrow linewidth laser and a fiber circulator. Rayleigh scattered light in the optical cable under test is coupled with the second optical path B at the 2*2 fiber coupler.

[0098] The balanced photodetector is connected to a 2*2 fiber optic coupler. The coupled signal is split into two and transmitted into the balanced photodetector, where it is converted into a beat frequency electrical signal.

[0099] The data acquisition card is connected to the balanced photodetector, and the beat frequency electrical signal is acquired by the digital acquisition card to obtain a digital signal.

[0100] Furthermore, after the 1*2 fiber coupler splits the signal emitted by the narrow-linewidth laser into the first optical path A and the second optical path B, it also includes:

[0101] The acousto-optic modulator modulates the first optical path A, which serves as the probe light, into an optical pulse signal and shifts the light frequency to a high frequency of 200MHz. After being amplified by an erbium-doped fiber amplifier, the light is transmitted through an optical fiber circulator into the sensing fiber. The Rayleigh scattering light generated by the optical pulse signal is transmitted back to the optical fiber circulator. After passing through the optical fiber circulator, the light is transmitted to a 2*2 optical fiber coupler and interferes with the local oscillator light of the second optical path B. After being converted by a balanced photodetector, the light is collected by a data acquisition card and finally transmitted to the processing unit for noise reduction.

[0102] The present invention provides a phase signal noise reduction method and system for optical cable vibration signals, which has at least the following beneficial effects:

[0103] (1) The present invention processes the phase signal of the disturbance point. For the problem of mode aliasing that may be caused by ordinary empirical mode decomposition, the ensemble empirical mode decomposition used in the present invention adds Gaussian white noise during the decomposition calculation. It takes advantage of the statistical characteristics of Gaussian white noise having a uniform frequency distribution. By adding Gaussian white noise with different amplitudes each time, the extreme point characteristics of the signal are changed. Then, the corresponding intrinsic mode components obtained by multiple empirical mode decompositions are averaged to cancel the added Gaussian white noise, thereby effectively reducing the influence of mode aliasing and improving the noise reduction effect.

[0104] (2) For phase signals affected by noise, combining the advantages of empirical mode decomposition and wavelet decomposition for noise reduction, wavelet decomposition is performed on high-frequency components and threshold quantization is performed to reconstruct the signal. This further removes the influence of noise on the phase signal and can more accurately restore the vibration intrusion signal on the sensing link.

[0105] (3) When performing wavelet decomposition and threshold quantization on high-frequency components, a compromise threshold quantization process is used, and an adjustment factor is added. The factor is adjusted appropriately during processing, which reduces the impact of constant deviation and discontinuity problems compared with traditional threshold processing, and improves the performance of wavelet reconstructed signals. Attached Figure Description

[0106] Figure 1 A flowchart of a phase signal noise reduction method for optical cable vibration signals provided by the present invention;

[0107] Figure 2 A flowchart for obtaining intrinsic modal components according to one embodiment of the present invention;

[0108] Figure 3 A flowchart illustrating the decomposition and quantization of high-frequency components according to an embodiment of the present invention;

[0109] Figure 4 A comparison diagram of phase signals before and after noise reduction in one embodiment of the present invention;

[0110] Figure 5 A schematic diagram of a phase signal noise reduction system for optical cable vibration signals provided by the present invention;

[0111] Figure 6 This is a schematic diagram of a phase signal noise reduction device for optical cable vibration signals provided by the present invention.

[0112] Explanation of reference numerals in the attached diagram: 100-Transmitting unit, 101-Narrow linewidth laser, 102-1*2 fiber optic coupler, 200-Modulation unit, 201-Acousto-optic modulator, 202-Erbium-doped fiber amplifier, 300-Optical cable under test, 301-Fiber circulator, 302-Sensing fiber, 400-Detection unit, 401-2*2 fiber optic coupler, 402-Balanced photodetector, 403-Data acquisition card, 500-Processing unit. Detailed Implementation

[0113] To better understand the above technical solutions, a detailed description of the solutions will be provided below in conjunction with the accompanying drawings and specific embodiments. Obviously, the described embodiments are merely some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0114] The terminology used in the embodiments of this invention is for the purpose of describing particular embodiments only and is not intended to limit the invention. The singular forms “a,” “the,” and “the” as used in the embodiments of this invention and the appended claims are also intended to include the plural forms, and “multiple” generally includes at least two unless the context clearly indicates otherwise.

[0115] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that an article or device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such an article or device. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the article or device that includes said element.

[0116] like Figure 1 As shown, the present invention provides a phase signal noise reduction method for optical cable vibration signals, comprising:

[0117] Vibration signals from the optical cable link are collected to obtain the phase signal at the disturbance point;

[0118] The phase signal is decomposed to obtain high-frequency and low-frequency components;

[0119] The high-frequency components are decomposed and quantized to obtain the noise-reduced high-frequency components.

[0120] The low-frequency components and the noise-reduced high-frequency components are reconstructed to obtain the noise-reduced phase signal.

[0121] This invention injects optical pulse signals into the optical cable using a Φ-OTDR system and collects vibration signals from the optical cable.

[0122] The process of acquiring vibration signals from the optical cable to obtain phase signals at the disturbance points may include:

[0123] The acquired vibration signal is quadrature demodulated, and the phase signal of the disturbance point is obtained by arctangent operation between the obtained in-phase component and the quadrature component.

[0124] In this embodiment, the phase signal is decomposed to obtain high-frequency components and low-frequency components, including:

[0125] The phase signal is processed based on ensemble empirical mode decomposition to obtain multiple intrinsic mode components;

[0126] The continuous mean square error is obtained by approximate reconstruction and analysis of the intrinsic modal components.

[0127] Based on continuous mean square error, the intrinsic modal components are divided into high-frequency components and low-frequency components.

[0128] Among them, see Figure 2 As shown, processing the phase signal based on empirical mode decomposition yields multiple intrinsic mode components, which may include:

[0129] Gaussian white noise with different amplitudes is added to the phase signal a predetermined number of times;

[0130] Empirical mode decomposition is performed on the phase signal after each addition of Gaussian white noise to obtain a predetermined number of intrinsic mode components.

[0131] The intrinsic modal components obtained by decomposing the set by the average predetermined number of times are given the final predetermined number of intrinsic modal components.

[0132] By performing ensemble averaging on the corresponding intrinsic mode components obtained from multiple ensemble empirical mode decompositions, the added Gaussian white noise can be canceled out.

[0133] The process of adding Gaussian white noise of different amplitudes a predetermined number of times to the phase signal includes:

[0134] x i (t)=x(t)+n i (t)

[0135] In the formula, i = 1, 2, 3, ..., M, where M is the number of times Gaussian white noise is added, and n i (t) represents the Gaussian white noise sequence added for the i-th time, x i x(t) represents the phase signal after adding Gaussian white noise for the i-th time, and x(t) represents the original phase signal without adding Gaussian white noise.

[0136] Empirical mode decomposition is performed on the phase signal after each addition of Gaussian white noise to obtain a predetermined number of intrinsic mode components, including:

[0137] Empirical mode decomposition (EMD) is performed on the phase signal after each addition of Gaussian white noise a predetermined number of times to obtain residual components and intrinsic mode components with a quantity corresponding to the predetermined number of times; satisfying the following relationship:

[0138]

[0139] In the formula, c i,j (t) represents the j-th intrinsic mode component obtained after the i-th addition of Gaussian white noise, r i,j (t) represents the j-th residual component obtained after the i-th addition of Gaussian white noise, where J is the number of intrinsic modal components, j = 1, 2, 3, ..., J;

[0140] By averaging the intrinsic modal components obtained from each decomposition, a final predetermined number of intrinsic modal components are given, including:

[0141] Superimpose the intrinsic modal components obtained by the same number of decompositions;

[0142] The superimposed intrinsic mode components are averaged to obtain the average intrinsic mode components after each decomposition.

[0143] Integrate all average intrinsic modal components to obtain the final intrinsic modal components;

[0144] The intrinsic modal components satisfy the following relationship:

[0145]

[0146] In the formula, c j (t) represents the j-th intrinsic mode component obtained by set averaging.

[0147] In practical applications, phase signals are processed based on ensemble empirical mode decomposition to obtain multiple intrinsic mode components, including:

[0148] By adding Gaussian white noise to the phase signal and performing multiple empirical mode decompositions, multiple intrinsic mode components are finally obtained through ensemble averaging, which may include:

[0149] Set an overall average number of calculations M, which is the predetermined number of times Gaussian white noise is added;

[0150] Adding Gaussian white noise with a standard normal distribution to the original phase signal generates multiple new phase signals x. i (t)=x(t)+n i (t), where i = 1, 2, 3, ..., M; where n is the value of n. i (t) represents the i-th additive white Gaussian noise sequence, x i (t) represents the signal after the i-th addition of Gaussian white noise;

[0151] The phase signal x of the obtained noisy signal i (t) Perform empirical mode decomposition on each to obtain the form of their respective sums;

[0152] Each decomposition incorporates Gaussian white noise signals of varying amplitudes, and the set of intrinsic mode components obtained after M calculations is {c 1,j (t),c 2,j (t),c 3,j (t),…,c M,j (t)}, where j=1,2,3,…,J;

[0153] The corresponding intrinsic modal components are subjected to ensemble averaging to obtain the final intrinsic modal components after ensemble empirical mode decomposition.

[0154] After obtaining multiple intrinsic mode components, this invention performs approximate reconstruction and analysis on the intrinsic mode components to obtain continuous mean square error, which may include:

[0155] All residual components are superimposed and averaged to obtain the final residual components;

[0156] The final predetermined number of intrinsic modal components are sorted according to the decomposition order;

[0157] The approximate reconstructed signal is obtained by superimposing any intrinsic mode component other than the first one with all subsequent intrinsic mode components and the final residual component, specifically satisfying the following relationship:

[0158]

[0159] In the formula, c represents the approximate reconstructed signal of the k-th intrinsic mode component. i (t) represents the i-th intrinsic modal component, r C (t) represents the final residual component, and C represents the number of intrinsic modal components obtained from the final decomposition.

[0160] Each approximate reconstructed signal is analyzed and processed along with its adjacent approximate reconstructed signal to obtain a continuous mean square error, which satisfies the following relationship:

[0161]

[0162] In the formula, Let t represent the continuous mean square error of the k-th approximate reconstructed signal, N represent the sequence length of the approximate reconstructed signal, and t represent the length of the sequence. i c represents the i-th point of the approximate reconstructed signal. k (t i ) represents the value of the k-th intrinsic modal component at the i-th point.

[0163] After obtaining the continuous mean square error of all approximate reconstructed signals, this invention divides the intrinsic mode components into high-frequency components and low-frequency components based on the continuous mean square error, which may include:

[0164] Compare all consecutive mean square errors and give the smallest consecutive mean square error;

[0165] Based on the superposition process of intrinsic modal components and the analysis and processing process of the approximate reconstructed signal, the intrinsic modal component corresponding to the minimum continuous mean square error is obtained;

[0166] The k-th intrinsic mode component corresponding to the minimum continuous mean square error is taken as the boundary point; specifically, it is represented as follows:

[0167]

[0168] In the formula, k is the dividing point;

[0169] The intrinsic modal components before the boundary point are classified as high-frequency components, and the intrinsic modal components after the boundary point are classified as low-frequency components.

[0170] In this embodiment, as Figure 3 As shown, the high-frequency components are decomposed and quantized to obtain the noise-reduced high-frequency components, which may include:

[0171] Based on the preset wavelet basis functions, wavelet decomposition is performed on all high-frequency components to obtain their respective wavelet coefficient sets, which satisfy the following relationship:

[0172] I W ={ω1,ω2,…,ω i ,…,ω L}

[0173] In the formula, I W Let ω be the set of wavelet coefficients. i Let L be the i-th wavelet coefficient, and L be the number of wavelet coefficients.

[0174] The threshold is determined based on the intrinsic modal components, satisfying the following relationship:

[0175]

[0176] In the formula, T is the threshold, σ is the variance of the noise, and N is the sequence length of the intrinsic modal components;

[0177] Based on a threshold, a compromise threshold quantization process is performed on all wavelet coefficient sets to obtain the wavelet coefficient set after compromise threshold quantization; the following relationship is satisfied:

[0178] I WT ={ω 1T ,ω 2T ,…,ω iT ,…,ω LT}

[0179]

[0180] In the formula, I WT For the set of wavelet coefficients after compromise threshold quantization, ω iT Let be the wavelet coefficients after the i-th threshold quantization process, and α be the added adjustment factor ∈ [0,1], which is adjusted between [0,1] according to the actual situation;

[0181] The wavelet coefficient set after threshold quantization is reconstructed to obtain the denoised high-frequency components.

[0182] The wavelet coefficient set I after threshold quantization is obtained. WT Then, the components are reconstructed to obtain the denoised high-frequency components I. WT,i (t), integrating all high-frequency components yields the overall high-frequency component x. H (t), that is:

[0183]

[0184] In the formula, I WT,i (t) represents the i-th noise-reduced high-frequency component, x H (t) represents the overall high-frequency components.

[0185] This invention improves noise reduction by selecting a suitable wavelet basis. Specifically, the selected wavelet basis characteristics are consistent with the phase signal characteristics. During wavelet threshold noise reduction, a suitable number of decomposition layers needs to be selected. In practical applications, when using Φ-OTDR for noise reduction of communication optical cables, a four-layer decomposition using the db8 wavelet basis function of the dbN wavelet family can be applied to the acquired vibration signal, resulting in optimal noise reduction performance in the current application scenario.

[0186] In this embodiment, reconstructing the low-frequency component and the denoised high-frequency component to obtain the denoised phase signal includes: superimposing the overall high-frequency component and the low-frequency component to obtain the denoised phase signal; specifically represented as follows:

[0187] x D (t)=x H (t)+x L (t)

[0188] In the formula, x D (t) represents the noise-reduced phase signal, x L (t) represents the low-frequency component.

[0189] See Figure 4 As shown, after denoising using the phase signal denoising method of the present invention, the phase signal waveform after denoising has fewer glitches than before denoising, the waveform is smoother and closer to the sine wave shape of the added disturbance, thus improving the restoration effect of the disturbance signal and resulting in a better phase signal.

[0190] Furthermore, this invention focuses on denoising the phase signal at disturbance points on the link, enabling better reconstruction of the disturbance signal. This invention incorporates Gaussian white noise during decomposition calculations, leveraging the statistical characteristic of Gaussian white noise's uniform frequency distribution. By adding Gaussian white noise with varying amplitudes each time, the extreme point characteristics of the signal are altered. Then, the added Gaussian white noise is offset by an overall average of the corresponding intrinsic mode components obtained from multiple empirical mode decompositions, effectively suppressing mode aliasing. The noise in the phase signal processed by this invention mainly exists in the high-frequency components. Therefore, wavelet thresholding is applied to the high-frequency components, and the final reconstruction achieves denoising, reducing unnecessary data processing and significantly improving the efficiency and real-time performance of the denoising process. In addition, an adjustment factor is introduced during threshold quantization to reduce the impact of constant deviation and discontinuity issues, improving the performance of the wavelet reconstructed signal. Moreover, this invention aims to better recover the disturbance signal for subsequent disturbance type analysis, which is mainly reflected in the changes in the phase signal at the disturbance point; therefore, this invention primarily focuses on phase signal processing.

[0191] See Figure 5 As shown, the present invention also provides a phase signal noise reduction system for optical cable vibration signals, employing the aforementioned phase signal noise reduction method for optical cable vibration signals. The identification system includes:

[0192] The acquisition and processing module is used to acquire vibration signals from the optical cable link and obtain the phase signal of the disturbance point.

[0193] The signal decomposition module is used to decompose the phase signal to obtain high-frequency and low-frequency components.

[0194] The high-frequency noise reduction module is used to decompose and quantize high-frequency components to obtain noise-reduced high-frequency components.

[0195] The signal reconstruction module is used to reconstruct the low-frequency components and the noise-reduced high-frequency components to obtain the noise-reduced phase signal.

[0196] The signal decomposition module is also used for:

[0197] The phase signal is processed based on ensemble empirical mode decomposition to obtain multiple intrinsic mode components;

[0198] The continuous mean square error is obtained by approximate reconstruction and analysis of the intrinsic modal components.

[0199] Based on continuous mean square error, the intrinsic modal components are divided into high-frequency components and low-frequency components.

[0200] The signal decomposition module is also used for:

[0201] Add Gaussian white noise of different amplitudes a predetermined number of times to the phase signal;

[0202] Empirical mode decomposition is performed on the phase signal after each addition of Gaussian white noise to obtain a predetermined number of intrinsic mode components.

[0203] The intrinsic modal components obtained by decomposing the set by the average predetermined number of times are given the final predetermined number of intrinsic modal components.

[0204] High-frequency noise reduction modules are also used for:

[0205] Based on the preset wavelet basis functions, all high-frequency components are decomposed into wavelet coefficient sets to obtain their respective sets.

[0206] Thresholds are determined based on intrinsic modal components;

[0207] Based on the threshold, all wavelet coefficient sets are subjected to compromise threshold quantization to obtain the wavelet coefficient set after compromise threshold quantization.

[0208] The wavelet coefficient set after threshold quantization is reconstructed to obtain the denoised high-frequency components.

[0209] See Figure 6 As shown, the present invention also provides a phase signal noise reduction device for optical cable vibration signals, including a transmitting unit 100, a modulation unit 200, a detection unit 400 and a processing unit 500, wherein the transmitting unit 100 and the modulation unit 2000 are connected, and the transmitting unit 100, the detection unit 400 and the processing unit 500 are connected in sequence.

[0210] The transmitting unit 100 injects an optical pulse signal into the optical cable under test 300 through the modulation unit 200;

[0211] The detection unit 400 acquires the vibration signal in the optical cable 300 under test;

[0212] The processing unit 500 uses the phase signal noise reduction method of the optical cable vibration signal described above to perform noise reduction processing on the vibration signal acquired by the detection unit 400.

[0213] Specifically, the transmitting unit 100 also includes a 1*2 fiber coupler 102 with a splitting ratio of 90:10. The 1*2 fiber coupler 102 splits the signal emitted by the narrow-linewidth laser 101 into a first optical path A and a second optical path B. The first optical path A serves as the probe light, and the second optical path B serves as the local oscillator light. The modulation unit 200 includes an acousto-optic modulator (AOM) 201 and an erbium-doped fiber amplifier (EDFA) 202, and is connected to the transmitting unit 100. The optical cable under test 300 includes an optical fiber circulator 301 and a sensing fiber 302, and is connected to the erbium-doped fiber amplifier (EDFA) 202.

[0214] Preferably, the sensing fiber 302 is a single-mode fiber with a length of 14km. A piezoelectric ceramic tube (PZT) is used as a disturbance source at a distance of 10km from the fiber under test to apply a standard sinusoidal signal.

[0215] During operation, the acousto-optic modulator (AOM) 201 modulates the first optical path A, which serves as the probe light, into an optical pulse signal and shifts the light frequency to a high frequency of 200MHz. After amplification by the erbium-doped fiber amplifier (EDFA) 202, the light is transmitted through the fiber optic circulator 301 into the sensing fiber 302. The Rayleigh scattering light generated by the optical pulse signal in the sensing fiber 302 is transmitted back to the fiber optic circulator 301. Specifically, the sensing fiber 302 is the fiber under test. The optical pulse, after passing through the fiber optic circulator 301, is transmitted into the sensing fiber 302. The scattered light returning from the sensing fiber 302 returns to the fiber optic circulator 301, then to the 2*2 fiber coupler 401 for interference. After conversion by the balanced photodetector 402, the light is acquired by the data acquisition card 403 and then transmitted to the processing unit 500 for noise reduction processing.

[0216] The device also includes a detection unit 400, which comprises a 2x2 fiber optic coupler 401, a balanced photodetector (BPD) 402, and a data acquisition card (DAQ) 403. The 2x2 fiber optic coupler 401 has a splitting ratio of 50:50 and is connected to a narrow-linewidth laser 101 and a fiber optic circulator 301. Rayleigh scattered light in the optical cable under test is coupled to the second optical path B at the 2x2 fiber optic coupler 401. The balanced photodetector (BPD) 402 is connected to the 2x2 fiber optic coupler 401, and the coupled signal is split into two and input into the balanced photodetector (BPD) 402, where it is converted into a beat frequency electrical signal. The data acquisition card (DAQ) 403 is connected to the balanced photodetector (BPD) 402, and the beat frequency electrical signal is acquired by the digital acquisition card (DAQ) 403 to obtain a digital signal.

[0217] The device also includes a processing unit 500 connected to a data acquisition card (DAQ) 403. Preferably, the processing unit can be a computer.

[0218] During use, the data acquisition card 403 collects various types of disturbance and intrusion signals and transmits them to the processing unit 500. The processing unit 500 processes and identifies the signals collected by the digital acquisition card 403.

[0219] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention. Clearly, those skilled in the art can make various alterations and modifications to the invention without departing from its spirit and scope. Thus, if these modifications and modifications of the invention fall within the scope of the claims and their equivalents, the invention is also intended to include these modifications and modifications.

Claims

1. A method for identifying a phase signal denoising of an optical cable vibration signal, characterized in that, The method comprises the following steps: adding a predetermined number of Gaussian white noises with different amplitudes to the phase signal, comprising: performing a predetermined number of empirical mode decompositions on the phase signal after adding the Gaussian white noise each time to obtain residual components and intrinsic mode components corresponding to the predetermined number of times; x i (t)=x(t)+n i (t) In the formula, i = 1, 2, 3,..., M, M is the number of times of adding Gaussian white noise, n i (t) represents the phase signal after adding the Gaussian white noise for the i-th time, and x(t) represents the original phase signal without adding the Gaussian white noise. i (t) represents the phase signal after adding the Gaussian white noise for the i-th time, and x(t) represents the original phase signal without adding the Gaussian white noise. superimposing the intrinsic mode components obtained by the same decomposition number; performing average processing on the superimposed intrinsic mode components to obtain average intrinsic mode components after each decomposition; integrating all the average intrinsic mode components to obtain a final predetermined number of intrinsic mode components; performing approximate reconstruction and analysis processing on the intrinsic mode components to obtain a continuous mean square error; based on the continuous mean square error, dividing the intrinsic mode components into high-frequency components and low-frequency components; based on a preset wavelet basis function, performing wavelet decomposition on all the high-frequency components to obtain respective sets of wavelet coefficients; determining a threshold value based on the intrinsic mode components; based on the threshold value, performing compromise threshold quantization processing on all the sets of wavelet coefficients to obtain sets of wavelet coefficients after compromise threshold quantization processing; and reconstructing the sets of wavelet coefficients after threshold quantization processing to obtain respective high-frequency components after noise reduction; reconstructing the low-frequency components and the high-frequency components after noise reduction to obtain a phase signal after noise reduction. performing approximate reconstruction and analysis processing on the intrinsic mode components to obtain a continuous mean square error, comprising:

2. The phase signal noise reduction method of claim 1, wherein, superimposing and performing average processing on all the residual components to obtain a final residual component; sorting the final predetermined number of intrinsic mode components according to the decomposition order; selecting any one intrinsic mode component except the first one and superimposing the intrinsic mode component with all the intrinsic mode components and the final residual component after the intrinsic mode component to obtain an approximate reconstruction signal; performing analysis processing on each approximate reconstruction signal and a neighboring approximate reconstruction signal after the approximate reconstruction signal to obtain a continuous mean square error. based on the continuous mean square error, dividing the intrinsic mode components into high-frequency components and low-frequency components, comprising:

3. The phase signal noise reduction method of claim 2, wherein, comparing all the continuous mean square errors to give a minimum continuous mean square error; according to the superimposition process of the intrinsic mode components and the analysis processing process of the approximate reconstruction signal, obtaining an intrinsic mode component corresponding to the minimum continuous mean square error; taking the intrinsic mode component corresponding to the minimum continuous mean square error as a demarcation point; dividing the intrinsic mode components before the demarcation point into high-frequency components and the intrinsic mode components after the demarcation point into low-frequency components. the sets of wavelet coefficients after compromise threshold quantization processing satisfy the following relationship:

4. The phase signal noise reduction method of claim 3, wherein, reconstructing the low-frequency components and the high-frequency components after noise reduction to obtain a phase signal after noise reduction, comprising: ； wherein is a set of wavelet coefficients after the threshold quantization process, ω i is the i-th wavelet coefficient, ω iT is the i-th wavelet coefficient after the threshold quantization process, α is an adjustment factor and α ∈ [0, 1], and T is a threshold.

5. The phase signal noise reduction method of claim 4, wherein, integrating all the high-frequency components after noise reduction to obtain a total high-frequency component; superimposing the total high-frequency component and the low-frequency component to obtain a phase signal after noise reduction. The system comprises:

6. A system for reducing noise in a phase signal of an optical cable vibration signal, employing the method for reducing noise in a phase signal of an optical cable vibration signal according to any one of claims 1 to 5, characterized by, a collection and processing module configured to collect vibration signals of an optical cable link to obtain phase signals of disturbance points; a signal decomposition module configured to decompose the phase signals to obtain high-frequency components and low-frequency components; ​ A high-frequency noise reduction module is configured to perform decomposition and quantization on the high-frequency component to obtain a noise-reduced high-frequency component; A signal reconstruction module is configured to reconstruct the low-frequency component and the noise-reduced high-frequency component to obtain a noise-reduced phase signal.

Images