A DAS system signal processing method based on improved Hilbert-Huang transform

Through the improved Hilbert yellow transformation and EEMD/CEEMD algorithm, combined with sliding average and Hilbert transformation, the adaptability and noise removal problems of DAS system in complex environments are solved, and effective demodulation of audio signals is achieved. and precise restoration.

CN115014497BActive Publication Date: 2024-11-15JINAN UNIVERSITY
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202210538831.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-05-18
Publication Date
2024-11-15
Estimated Expiration
2042-05-18

AI Technical Summary

Technical Problem

In existing DAS systems, there is a problem that audio signals cannot be adaptable and are difficult to demodulate in signal processing, especially in complex environments, which leads to the system not being widely used in industrial applications.

Method used

The improved Hilbert yellow transformation method is used to extract perturbation signals through EEMD and CEEMD algorithms, and combine sliding average and Hilbert transformation to automatically demodulate perturbation signals to reduce dependence on parameters such as basis function and window length.

Benefits of technology

Effectively demodulate the time domain waveform of the disturbed signal, reduce noise interference, is suitable for complex environments in real life, and improves the system's adaptability and signal processing accuracy.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115014497B_ABST
    Figure CN115014497B_ABST
Patent Text Reader

Abstract

The present invention relates to the field of distributed optical fiber sensing technology and distributed optical fiber signal processing, and specifically to a DAS system signal processing method based on improved Hilbert-Huang transform, comprising the following steps: extracting a time domain curve of a disturbance point, processing it using sliding average and improved Hilbert-Huang transform, wherein the improved Hilbert-Huang transform comprises: performing EEMD on its source signal to obtain the required IMF component, then demodulating the IMF component as the source signal through the CEEMD algorithm, and performing Hilbert transform on the demodulation result to perform time-frequency analysis and demonstration. The present invention can meet the need for restoring and demodulating the disturbance audio signal in the actual application process by performing sliding average noise reduction on the original signal and using the improved Hilbert-Huang transform algorithm to extract weak signals, and also makes up for the disadvantage that the existing DAS system cannot be adaptive.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of distributed optical fiber sensing technology and distributed optical fiber signal processing, and in particular to a DAS (full name: Distributed Optic Fiber Acoustic Sensing, Chinese translation: distributed optical fiber acoustic sensing) system signal processing method based on improved Hilbert-Huang transform. Background Art

[0002] The DAS system is a distributed sensing system based on optical fiber. It utilizes the uniformity and continuity of the optical fiber system. When the detected acoustic signal causes the optical signal transmitted in the optical fiber to change, the acoustic signal at any position on the sensing optical fiber line can be inverted by detecting the change in the optical signal.

[0003] At present, based on the phase-sensitive optical time domain reflectometer (Phase-Sensitivity Optical Time Domain Reflectometer, ) technology is one of the research hotspots of fiber optic sensing, which has the advantages of long-distance sensing, high sensitivity and high measurement accuracy. The DAS systems of the prior art still have the disadvantages of being unable to adapt to actual audio signals and being difficult to demodulate in signal processing. The main manifestation is that these systems generally demodulate sound signals through Fourier transform (including fast Fourier transform, short-time Fourier transform and fractional-order domain Fourier transform, etc.), wavelet transform, etc., and the processing process requires artificial selection of basis functions, window lengths, number of decomposition layers and transformation orders, etc. It is difficult to ensure that the selected basis functions or window lengths, number of decomposition layers and transformation orders are the most appropriate parameters, so there is a large uncertainty. In addition, in the prior art, the verification of these systems generally uses piezoelectric ceramics (Piezoelectric Transducer, PZT) simulated vibration as a disturbance signal. PZT uses the piezoelectric effect to generate a vibration signal and transmits it to the sensor optical fiber wrapped around it. This interference signal can be regarded as a single-frequency signal, that is, most of the DAS systems in the prior art use vibration signals simulated by PZT as disturbance signals. However, the interference signals in real life are in complex environments and contain noise that is difficult to remove. This is why most DAS systems are still in the laboratory stage and have not yet been widely used in industry. Summary of the invention

[0004] The object of the present invention is to provide a DAS system signal processing method based on improved Hilbert-Huang transform, by which the DAS system can demodulate actual audio signals.

[0005] The object of the present invention is achieved by the following technical solution: A DAS system signal processing method based on improved Hilbert-Huang transform comprises the following steps:

[0006] The time domain curve of the disturbance point (such as the optical fiber position where the speaker is located) is extracted and processed using sliding average and improved Hilbert-Huang transform, where the improved Hilbert-Huang transform includes:

[0007] Perform EEMD on the source signal, that is, add zero-mean normally distributed Gaussian white noise to its source signal, perform EMD (full name: Empiricial Mode Decomposition) to obtain the required IMF component (full name: Intrinsic Mode Function), and then demodulate the IMF component as the source signal through the CEEMD (full name: Complete Ensemble Empiricial Mode Decomposition) algorithm, that is, add white noise to the signal (IMF component) and subtract white noise to get two new signals, and the two signals are processed by EMD at the same time, and the average is calculated to offset the residual noise in the signal.

[0008] The signal processing method also includes a signal verification step, which includes performing HT (full name: Hilbert Transform, Chinese translation: Hilbert Transform) on the IMF component obtained after CEEMD to obtain the Hilbert spectrum, and integrating the Hilbert spectrum over time to further obtain the Hilbert marginal spectrum.

[0009] The signal verification step also includes performing FFT (full name: Fast Fourier Transform, Chinese translation: Fast Fourier Transform) on the IMF component obtained after CEEMD to obtain the Fourier spectrum.

[0010] The sliding average can be used before or after the improved Hilbert-Huang transform. The sliding average can be selected from one or a combination of two or more of the moving average, Gaussian filtering, SG filtering, etc., preferably a combination of the moving average and SG filtering.

[0011] The present invention extracts weak signals through improved Hilbert-Huang transform and reduces noise through sliding average. After three-step noise processing (specifically EEMD, CEEMD and sliding average), even if the disturbance signal source is a speaker, the disturbance signal can be well demodulated.

[0012] The DAS system includes a narrow linewidth laser 1, a first optical coupler 2, a second optical coupler 10, a third optical coupler 7, an optical attenuator 9, a pulse generator 14 with adjustable frequency and pulse width, an acousto-optic modulator 3, a first optical signal amplifier 4, a second optical signal amplifier 6, an optical circulator 5, a sensing optical fiber 8, a photoelectric conversion unit 11, a data acquisition unit 12, a signal processing unit 13 and a disturbance signal source;

[0013] The narrow linewidth laser 1 is connected to the first optical coupler 2 and is divided into two outputs. The optical attenuator 9 is connected to the first optical coupler 2 to control the signal ratio of the two outputs. One output is connected to the acousto-optic modulator 3, the output end of the acousto-optic modulator 3 is connected to the first optical signal amplifier 4, the output end of the first optical signal amplifier 4 is connected to the port 1 of the optical circulator, the other output end of the narrow linewidth laser 1 is connected to an input end of the second optical coupler 10 with a ratio of 50:50, the port 2 of the optical circulator is connected to the third optical coupler 7, and the second optical signal amplifier 6 is connected to the third optical coupler 7. It is also connected to the third optical coupler 7, the third optical coupler 7 is connected to the sensing optical fiber 8, the No. 3 port of the optical circulator 5 is connected to the other input end of the second optical coupler 10 at a ratio of 50:50, the output end of the second optical coupler 10 is connected to the photoelectric conversion unit 11, the output end of the photoelectric conversion unit 11 is connected to the data acquisition unit 12, the output end of the data acquisition unit 12 is connected to the signal processing unit 13, the pulse generator 14 is connected to the acousto-optic modulator 3 and the data acquisition unit 12 respectively, and the disturbance signal source is arranged along the sensing optical fiber 8;

[0014] The acousto-optic modulator 3 modulates a part of the continuous light signal emitted by the narrow linewidth laser 1 into a pulse light signal, which is input from port 1 of the optical circulator 5 and output from port 2 after being amplified by the first light signal amplifier 4, amplified by the second light signal amplifier 6 to increase the sensing distance, and coupled to the sensing optical fiber 8 by the third optical coupler 7. Rayleigh scattering occurs in the sensing optical fiber 8, and the backscattered light generated by the Rayleigh scattering is collected by port 2 of the optical circulator 5 and output from port 3 thereof, coupled with another part of the local oscillator light emitted by the narrow linewidth laser 1, and input to the second optical coupler 10. The photoelectric conversion unit 11 is used to convert the optical signal output by the second optical coupler 10 into an electrical signal, the data acquisition unit 12 is used to collect the electrical signal and output it to the signal processing unit 13, and the pulse generator 14 is used to provide a driving signal for the acousto-optic modulator 3 and synchronously provide an external trigger signal for the signal acquisition unit 12, so as to ensure that the signal collected by the data acquisition unit 12 and the signal generated by the optical path system have the same start and end time.

[0015] The disturbance signal source is a loudspeaker.

[0016] The recommended setting ratio of the two outputs of the narrow linewidth laser 1 is 10:90, 10% is output to the acousto-optic modulator 3 , and 90% is output to the second optical coupler 10 .

[0017] The narrow linewidth laser 1 has a linewidth of 1 kHz. The narrow linewidth laser has good frequency stability, ultra-long coherence length and ultra-low noise.

[0018] The pulse generator 14 is a pulse signal generator based on FPGA, and its frequency and pulse width are adjustable within 1-10 kHz and 10-100 ns respectively.

[0019] The sensing optical fiber 8 is a common single-mode optical fiber, that is, a standard single-mode optical fiber.

[0020] The first optical signal amplifier 4 is an erbium-doped fiber amplifier, and the second optical signal amplifier 6 is a Raman fiber amplifier, which are used to amplify optical signals.

[0021] The photoelectric conversion unit 11 is a balanced detector with high sensitivity.

[0022] Compared with the prior art, the present invention has the following beneficial effects:

[0023] The signal processing method of the present invention extracts weak signals through improved Hilbert-Huang transform and reduces noise through sliding average. After three-step noise processing, even when a disturbance signal source closer to real life, such as a loudspeaker, is used, the time domain waveform of the disturbance signal can be well demodulated. Moreover, in the process of using the algorithm to demodulate the disturbance signal, there is no need to manually set parameters such as basis functions and window lengths, which can meet the needs of restoring and demodulating the disturbance audio signal in the actual application process, and also make up for the disadvantage that the existing DAS system cannot be adaptive when processing audio disturbance sensor signals. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 A framework diagram of a DAS system used in a preferred embodiment of the present invention;

[0025] In the figure, 1 is a narrow linewidth laser, 2 is a first optical coupler (10:90), 3 is an acousto-optic modulator, 4 is a first optical signal amplifier, which uses an erbium-doped fiber amplifier, 5 is an optical circulator, 6 is a second optical signal amplifier, which uses a Raman fiber amplifier, 7 is a third optical coupler, 8 is a sensing fiber, which uses an ordinary single-mode fiber, 9 is an optical attenuator, 10 is a second optical coupler (50:50), 11 is a photoelectric conversion unit, 12 is a data acquisition unit, 13 is a signal processing unit, which uses a DSP, and 14 is a pulse generator with adjustable frequency and pulse width;

[0026] Figure 2 The audio time domain signal is restored by using EEMD and CEEMD-HHT;

[0027] Figure 3 Frequency domain verification of audio signals recovered by EEMD and CEEMD-HHT;

[0028] Figure 4 Hilbert spectrum of the audio signal restored by EEMD and CEEMD-HHT;

[0029] Figure 5 This is the Hilbert marginal spectrum of the audio signal restored using EEMD and CEEMD-HHT. DETAILED DESCRIPTION

[0030] The present invention is further described in detail below in conjunction with specific embodiments and accompanying drawings:

[0031] A DAS system signal processing method based on improved Hilbert-Huang transform, such as Figure 1 As shown, Figure 1 This is a framework diagram of the DAS system on which the signal processing method is based.

[0032] The DAS system includes a narrow linewidth laser 1, a first optical coupler 2, a second optical coupler 10, a third optical coupler 7, an optical attenuator 9, a pulse generator 14 with adjustable frequency and pulse width, an acousto-optic modulator 3, a first optical signal amplifier 4, a second optical signal amplifier 6, an optical circulator 5, a sensing optical fiber 8, a photoelectric conversion unit 11, a data acquisition unit 12, a signal processing unit 13 and a speaker (not shown).

[0033] The narrow linewidth laser 1 is connected to the first optical coupler 2 and is divided into two outputs. The optical attenuator 9 is connected to the first optical coupler 2 to control the signal ratio of the two outputs. One output is connected to the acousto-optic modulator 3, the output end of the acousto-optic modulator 3 is connected to the first optical signal amplifier 4, the output end of the first optical signal amplifier 4 is connected to port 1 of the optical circulator, the other output end of the narrow linewidth laser 1 is connected to an input end of the second optical coupler 10 with a ratio of 50:50, the port 2 of the optical circulator is connected to the third optical coupler 7, and the second optical signal amplifier 6 is also connected to the first optical coupler 8. The optical coupler 7 is connected to three optical couplers, the third optical coupler 7 is connected to the sensing optical fiber 8, the port No. 3 of the optical circulator 5 is connected to the other input end of the second optical coupler 10 at a ratio of 50:50, the output end of the second optical coupler 10 is connected to the photoelectric conversion unit 11, the output end of the photoelectric conversion unit 11 is connected to the data acquisition unit 12, the output end of the data acquisition unit 12 is connected to the signal processing unit 13, the pulse generator 14 is connected to the acousto-optic modulator 3 and the data acquisition unit 12 respectively, and the speaker is arranged along the sensing optical fiber 8, and the optical fiber position where it is located is the disturbance point.

[0034] The signal transmission process in the DAS system is as follows:

[0035] The acousto-optic modulator 3 modulates a part of the continuous light signal emitted by the narrow linewidth laser 1 into a pulse light signal, which is input from port 1 of the optical circulator 5 and output from port 2 after being amplified by the first light signal amplifier 4, amplified by the second light signal amplifier 6 to increase the sensing distance, and coupled to the sensing fiber 8 by the third optical coupler 7. Rayleigh scattering occurs in the sensing fiber 8, and the backscattered light generated by the Rayleigh scattering is collected by port 2 of the optical circulator 5 and output from port 3 thereof, coupled with another part of the local oscillator light emitted by the narrow linewidth laser 1, and input to the second optical coupler 10. The photoelectric conversion unit 11 is used to convert the optical signal output by the second optical coupler 10 into an electrical signal, the data acquisition unit 12 is used to collect the electrical signal and output it to the signal processing unit 13, and the pulse generator 14 is used to provide a driving signal for the acousto-optic modulator 3 and synchronously provide an external trigger signal for the signal acquisition unit 12.

[0036] The data acquisition unit 12 is of external trigger fixed-point acquisition type. It uses an external trigger source and detects the rising edge of the modulated pulse generated by the FPGA to start collecting data. The signal sampling rate can reach 125MSa / s. The DAS system can only obtain a complete Rayleigh scattering curve in one pulse cycle at a time. Here, a pulse generator 14 is used to synchronously control the acousto-optic modulator 3 and the data acquisition unit 12 to ensure that the signal collected by the data acquisition unit 12 and the signal generated by the optical path system have the same start and end time, thereby ensuring that the data collected each time is not lost or overlapped, ensuring the correctness of the data, and facilitating subsequent data processing.

[0037] The signal collected by the data collection unit 12 is input to the signal processing unit 13. The processing method of the signal processing unit 13 is as follows:

[0038] The time domain curve of the disturbance point is extracted and processed using sliding average and improved Hilbert-Huang transform, where the improved Hilbert-Huang transform includes:

[0039] EEMD is performed on the source signal, and modal aliasing is eliminated by adding noise, that is, zero-mean normally distributed Gaussian white noise is added to the source signal, and EMD is performed to obtain the required IMF component. The IMF component is then demodulated as the source signal through the CEEMD algorithm to obtain the time domain waveform of the disturbance signal.

[0040] The CEEMD algorithm aims to add complementary noise (white noise plus and white noise minus) to eliminate modal aliasing while eliminating the residual noise of EEMD to obtain a purer signal and extract new IMF components. Then, it can be further Hilbert transformed to obtain the Hilbert spectrum. On the Hilbert spectrum, the change of signal frequency with time transformation can be observed from the time-frequency aspect, that is, time-frequency analysis can be performed, and then the Hilbert spectrum is integrated in time to obtain the Hilbert marginal spectrum. The frequency of the Hilbert marginal spectrum indicates the instantaneous frequency, indicating the frequency that actually exists in the signal, which can be compared and analyzed with the Fourier spectrum obtained after the fast Fourier transform, so as to better verify the demodulation effect of the method of the present invention on the signal.

[0041] The sliding average can be used before or after the improved Hilbert-Huang transform. The sliding average can be selected from one or a combination of two or more of the moving average, Gaussian filtering, SG filtering, etc., preferably a combination of the moving average and SG filtering.

[0042] The specific method of extracting the time domain curve of the disturbance point is as follows:

[0043] The data collected by the data acquisition unit 12 is saved in the form of a binary file, and its data format is a binary complement code or a binary offset code. The signal processing unit 13 reads the data and then performs a slice process, and slices the data with a single sampling length as the length, and lists them into a matrix. The principle is to accumulate the data collected M times in a short time to obtain the data of N measurement points of the optical fiber at a time point. The matrix A is expressed as follows:

[0044]

[0045] in, i=1,2,…,T,j=1,2,…,N, each element a in the matrix A ij represents the average of M measurements at each measurement point in a short period of time. If there are 1000 measurement points, we take the average of the data of 5 measurement points and record it as a ij , at this time M=5, which can effectively compress the data volume to 1 / 5 of the original amount. Because we use a high-speed data acquisition card (i.e., data acquisition unit 12) for data acquisition, there will be two problems for the high-speed data acquisition card. One is that there are several meaningless 0 or sign bits in every 2 bytes of the data stream transmitted inside the acquisition card, which wastes the limited data bandwidth. The other is that when the collected data is uploaded to the PC, i.e., the memory, through the PCIE interface, a part of the limited transmission rate is wasted. In order to make full use of the data transmission bandwidth, an internal data compression module is required, and the above method can compress the data well. As shown in the above formula, vector A i =[a i1 ,a i2 ,…,a iN ] T is the disturbance vector of each measuring point of the DAS system sensor fiber. According to the disturbance vector, the deformation of each position of the sensor fiber can be obtained; vector B i =[a 1j ,a 2j ,…,a Tj ] T is the disturbance vector of the jth measurement point on the sensing optical fiber of the DAS system. According to the disturbance vector, different time curves of the fixed position points of the optical fiber can be obtained.

[0046] The pulse generator 14 is a pulse signal generator based on FPGA, and its frequency can be adjusted arbitrarily within the range of 1-10 kHz and pulse width of 10-100 ns.

[0047] In this embodiment, the narrow linewidth laser 1 has a center frequency of 1550 nm and a linewidth of 1 kHz.

[0048] The ratio of the two outputs of the narrow linewidth laser 1 is recommended to be set to 10:90, 10% is output to the acousto-optic modulator 3, and 90% is output to the second optical coupler 10. This does not mean that the two outputs of the narrow linewidth laser 1 can only be set according to the above ratio, and can actually be set to 20:80, 15:85, etc. The recommended setting is 10:90 to obtain the best coherence effect.

[0049] The sensing optical fiber 8 is a common single-mode optical fiber.

[0050] The photoelectric conversion unit 11 is a balanced detector with high sensitivity.

[0051] Figure 1 The DAS system uses heterodyne coherent detection. There are two detection modes in the DAS system, direct detection and coherent detection. The signal obtained by direct detection only has intensity information, and the phase information cannot be obtained. This is because the frequency of the photodetector is much lower than the frequency of light. The photodetector can only respond to the envelope of the light signal, and its phase information cannot be extracted. Therefore, the direct detection sensing system can only detect intensity changes and cannot measure the phase information of the disturbance signal. There are many types of coherent detection, such as zero difference method, heterodyne method, use of interferometer, etc. Here, heterodyne coherent detection is used, which not only makes the system structure simple, but also can realize phase demodulation. However, this does not mean that the signal processing method of the present invention must be bound to this DAS system. The signal processing method of the present invention also has certain reference value for the noise reduction processing of signals of other systems in industrial control, such as DTS (distributed fiber optic temperature sensing system) and the like.

[0052] Effect experiment

[0053] In the actual experiment, the frequency of the disturbance signal is an 800 Hz audio disturbance signal played by a speaker on the sensing fiber. The time domain of the disturbance signal obtained by EEMD and CEEMD is as follows: Figure 2 As shown in (before using EEMD, moving average and SG filtering were used to smooth the extracted time domain signal to reduce noise interference), the frequency domain analysis of the disturbance signal obtained is as follows Figure 3 As shown, the result is Hilbert transformed to obtain the Hilbert spectrum as Figure 4 As shown, integrating it over time, we get the Hilbert marginal spectrum as Figure 5 shown.

[0054] After CEEMD, the restored sound signal is obtained. Since the accuracy of the restored sound signal cannot be seen from the results of CEEMD, further demonstration is needed. Figure 3This is an FFT method demonstration, but it only shows that the correct frequency may exist in the results obtained by CEEMD, and the Hilbert transform Figure 4 It is argued from the time-frequency aspect that when time changes, there is energy concentrated at that frequency. Figure 5 It is further proved that the signal frequency obtained by CEEMD is the required frequency using Hilbert marginal spectrum.

[0055] Depend on Figure 3-5 , the EEMD algorithm combined with the CEEMD-HHT algorithm is used to demodulate the audio signal (sent by the speaker) acting on the sensing optical fiber. Although the complex environmental factors lead to great noise interference, the disturbed 800Hz audio signal frequency can still be demodulated well.

[0056] The disturbance signal source for the above effect experiment is the speaker. In fact, when PZT is used as the interference signal source, compared with Figure 2-Figure 3 , the signal recovery is better.

[0057] The signal processing method of the present invention extracts weak signals through improved Hilbert-Huang transform and reduces noise through sliding average. After three-step noise processing, even when a disturbance signal source closer to real life, such as a loudspeaker, is used, the time domain waveform of the disturbance signal can be well demodulated. Moreover, in the process of using the algorithm to demodulate the disturbance signal, there is no need to manually set parameters such as basis functions and window lengths, which can meet the needs of restoring and demodulating the disturbance audio signal in the actual application process, and also make up for the disadvantage that the existing DAS system cannot be adaptive when processing audio disturbance sensor signals.

[0058] The above specific implementation modes are only used to further explain the present invention in detail, and the implementation modes of the present invention cannot be considered to be limited to the above. Simple improvements or substitutions made on the basis of the concept of the present invention should be regarded as falling within the protection scope of the present invention.

Claims

1. A DAS system signal processing method based on improved Hilbert-Huang transform, It is characterized in that The steps include: The time domain curve of the disturbance point is extracted and processed using sliding average and improved Hilbert-Huang transform, where the improved Hilbert-Huang transform includes: Perform EEMD on the source signal to obtain the required IMF component, and then demodulate the IMF component as the source signal through the CEEMD algorithm; The DAS system comprises a narrow linewidth laser (1), a first optical coupler (2), a second optical coupler (10), a third optical coupler (7), an optical attenuator (9), a frequency and pulse width adjustable pulse generator (14), an acousto-optic modulator (3), a first optical signal amplifier (4), a second optical signal amplifier (6), an optical circulator (5), a sensing optical fiber (8), a photoelectric conversion unit (11), a data acquisition unit (12), a signal processing unit (13) and a disturbance signal source; The narrow linewidth laser (1) is connected to the first optical coupler (2) and is divided into two output paths. The optical attenuator (9) is connected to the first optical coupler (2) to control the signal ratio of the two output paths. One output path is connected to the acousto-optic modulator (3). The output end of the acousto-optic modulator (3) is connected to the first optical signal amplifier (4). The output end of the first optical signal amplifier (4) is connected to port 1 of the optical circulator. The other output end of the narrow linewidth laser (1) is connected to an input end of the second optical coupler (10) with a ratio of 50:

50. The port 2 of the optical circulator is connected to the third optical coupler (7). The second optical signal amplifier (6) is also connected to the output end of the optical circulator. The third optical coupler (7) is connected to the sensing optical fiber (8), the third optical coupler (7) is connected to the sensing optical fiber (8), the No. 3 port of the optical circulator (5) is connected to the other input end of the second optical coupler (10) at a ratio of 50:50, the output end of the second optical coupler (10) is connected to the photoelectric conversion unit (11), the output end of the photoelectric conversion unit (11) is connected to the data acquisition unit (12), the output end of the data acquisition unit (12) is connected to the signal processing unit (13), the pulse generator (14) is connected to the acousto-optic modulator (3) and the data acquisition unit (12) respectively, and the disturbance signal source is arranged along the sensing optical fiber (8); The acousto-optic modulator (3) modulates a portion of the continuous optical signal emitted by the narrow linewidth laser (1) into a pulsed optical signal, which is amplified by the first optical signal amplifier (4) and inputted from port 1 of the optical circulator (5) and outputted from port 2. After being amplified by the second optical signal amplifier (6) to increase the sensing distance, the signal is coupled and outputted to the sensing optical fiber (8) by the third optical coupler (7). Rayleigh scattering occurs in the sensing optical fiber (8). Backscattered light generated by the Rayleigh scattering is collected by port 2 of the optical circulator (5) and outputted from port 3 thereof to be coupled to the backscattered light emitted by the narrow linewidth laser (1). The other part of the local oscillation light emitted by the optical modulator (1) is coupled to the second optical coupler (10), the photoelectric conversion unit (11) is used to convert the optical signal output by the second optical coupler (10) into an electrical signal, the data acquisition unit (12) is used to acquire the electrical signal and output it to the signal processing unit (13), the pulse generator (14) is used to provide a driving signal for the acousto-optic modulator (3) and synchronously provide an external trigger signal for the data acquisition unit (12), so as to ensure that the signal acquired by the data acquisition unit (12) and the signal generated by the optical path system have the same start and end time.

2. The signal processing method according to claim 1, It is characterized in that The method also includes a signal verification step, which includes performing HT on the IMF component obtained after CEEMD to obtain a Hilbert spectrum, and integrating the Hilbert spectrum over time to further obtain a Hilbert marginal spectrum.

3. The signal processing method according to claim 2, It is characterized in that The signal verification step also includes performing fast Fourier transform on the IMF component obtained after CEEMD to obtain a Fourier spectrum.

4. The signal processing method according to claim 1, It is characterized in that The sliding average is used before or after the improved Hilbert-Huang transform, and the sliding average selects one of the moving average, Gaussian filtering, SG filtering, or a combination of two or more thereof.

5. The signal processing method according to claim 1, It is characterized in that The disturbance signal source is a loudspeaker.

6. The signal processing method according to claim 1, It is characterized in that The ratio of the two outputs of the narrow linewidth laser (1) is recommended to be set to 10:90, 10% is output to the acousto-optic modulator (3), and 90% is output to the second optical coupler (10).

7. The signal processing method according to claim 1, It is characterized in that The linewidth of the narrow linewidth laser (1) is 1 kHz.

8. The signal processing method according to claim 1, It is characterized in that The pulse generator (14) is a pulse signal generator based on FPGA, and can be adjusted arbitrarily within the range of frequency 1-10 kHz and pulse width 10-100 ns.

9. The signal processing method according to claim 1, It is characterized in that The first optical signal amplifier (4) is an erbium-doped optical fiber amplifier, and the second optical signal amplifier (6) is a Raman optical fiber amplifier.