A method of sensor positioning

Abstract

The application provides a sensing positioning method and relates to the technical field of optical fiber sensing. The sensing positioning method comprises the following steps: in response to a vibration event acting on an optical fiber, two signal segments obtained by the phase changes of two detection lights with opposite transmission directions before and after the transmission of the optical fiber are acquired; according to a plurality of first modes of a reference signal segment, the weight coefficients of a plurality of frequencies in the spectrum of each first mode are determined; the weight coefficients represent the correlation between the frequencies and the vibration event, and the reference signal segment is one of the two signal segments with a higher signal-to-noise ratio; the spectrum of a measurement signal segment is reconstructed by using the plurality of weight coefficients corresponding to the first mode, so that a reconstructed spectrum matched with the spectrum of the first mode is obtained; the measurement signal segment is one of the two signal segments with a lower signal-to-noise ratio; and the position of the vibration event acting on the optical fiber is determined by using a plurality of mode pairs. The sensing positioning method of the application realizes high-precision positioning in a super-long distance sensing scene of hundreds of kilometers.

Description

Technical Field

[0001] This invention relates to the field of fiber optic sensing technology, and more specifically, to a sensing and positioning method. Background Technology

[0002] Long-distance vibration monitoring using distributed optical fiber sensing technology has become a key means of ensuring the safety of pipelines, perimeters, and large structures. Traditional distributed acoustic sensing (DAS) systems are mainly based on phase optical time domain reflectometry (Ф-OTDR), relying on backscattered Rayleigh light in the detection fiber. However, due to the inherently weak optical power level of backscattered Rayleigh light and the signal attenuation characteristics during fiber transmission, the effective sensing distance of such systems is usually limited. In addition, the presence of coherent fading noise makes the signal easily submerged in noise, making it difficult to meet the ultra-long-distance monitoring requirements of hundreds of kilometers and above.

[0003] To overcome distance limitations, sensing technologies based on a dual-end optical forward transmission architecture have gradually become a research hotspot. This architecture utilizes forward transmission light with higher optical power as the signal carrier, enabling ultra-long-distance sensing over a single span of hundreds of kilometers without repeaters. It is also easily integrated with existing optical communication links, significantly reducing system deployment and maintenance costs. Despite the advantages of the forward transmission architecture in terms of optical power, extending it to sensing scenarios over hundreds of kilometers still faces severe challenges caused by the asymmetry of the transmission path, leading to the failure of existing positioning algorithms. Summary of the Invention

[0004] In view of this, embodiments of the present invention provide a sensing and positioning method, comprising:

[0005] In response to a vibration event acting on an optical fiber, two signal segments are acquired, obtained from the phase changes of two probe beams with opposite transmission directions before and after transmission through the optical fiber; the signal segments carry vibration event information.

[0006] Based on multiple first modes of the reference signal segment, the weighting coefficients of multiple frequencies in the spectrum of each first mode are determined; the weighting coefficients characterize the correlation between the frequency and the vibration event.

[0007] The spectrum of the measurement signal segment is reconstructed using multiple weighting coefficients corresponding to the first mode to obtain a reconstructed spectrum that matches the spectrum of the first mode; the measurement signal segment is the one with the lower signal-to-noise ratio among the two signal segments.

[0008] The location of a vibration event acting on an optical fiber is determined using multiple mode pairs, wherein the mode pair includes a first mode and a second mode, and the second mode is obtained by reconstructing the spectrum.

[0009] According to an embodiment of the present invention, determining the weighting coefficients of multiple frequencies in the spectrum of each first mode includes:

[0010] For each first mode,

[0011] Determine the maximum value of multiple spectral amplitudes from the spectra of each first mode;

[0012] Based on the maximum value and the spectral amplitude of each frequency in the spectrum of the first mode, the weighting coefficients of each frequency in the spectrum of the first mode are obtained.

[0013] According to an embodiment of the present invention, determining the weighting coefficients of multiple frequencies in the spectrum of each first mode includes:

[0014] For each first mode,

[0015] Determine the spectral amplitude of each frequency in the spectrum of the first mode;

[0016] The weighting coefficients of each frequency in the spectrum of the first mode are obtained by summing the spectral amplitudes of each frequency and the spectral amplitudes of any frequency in the first mode.

[0017] According to an embodiment of the present invention, a method for determining a plurality of first modes includes:

[0018] Variational mode decomposition is performed on the reference signal segment using the target mode number and the target quadratic penalty factor to obtain multiple first modes with different center frequencies. The target mode number represents the number of modes in the reference signal segment, and the target quadratic penalty factor represents the bandwidth of the first mode.

[0019] According to an embodiment of the present invention, a method for determining the target modal number includes:

[0020] Using the initial number of modes N, variational mode decomposition is performed on the reference signal segment to obtain N third modes, where N≥2;

[0021] Remove the third modes from the N third modes whose transient energy ratio is less than the first preset threshold, and obtain the remaining M third modes, 1≤M≤N;

[0022] Sort the energies of the M remaining third modes in descending order, and add them up one by one in the order of sorting until the ratio of the accumulated energy to the total energy of the M remaining third modes meets the preset condition.

[0023] The target number of modes is obtained by counting the number of second modes that participate in the accumulation when the preset conditions are met.

[0024] According to an embodiment of the present invention, the method for determining the target quadratic penalty factor includes:

[0025] Repeat the following steps until the peak-to-sidelobe ratio of the cross-correlation function meets the preset termination condition, and use the candidate quadratic penalty factor that meets the preset termination condition as the target quadratic penalty factor:

[0026] Using the candidate quadratic penalty factor and the target mode number corresponding to the current round, variational mode decomposition is performed on the reference signal segment and the measurement signal segment respectively to obtain multiple fourth modes and multiple fifth modes;

[0027] Cross-correlation is performed on the envelopes of the first target mode among multiple fourth modes and the envelopes of the second target mode among multiple fifth modes to obtain the cross-correlation function corresponding to the current round; the first target mode is the fourth mode with the highest energy or the highest time-frequency energy ratio among multiple fourth modes, and the second target mode is the fifth mode among multiple fifth modes that has the same mode as the first target mode;

[0028] If the peak sidelobe ratio of the cross-correlation function corresponding to the current round does not meet the preset termination condition, a candidate quadratic penalty factor corresponding to the next round is determined.

[0029] According to an embodiment of the present invention, a preset termination condition includes:

[0030] The peak sidelobe ratio obtained in the current round is greater than the first preset threshold, and the growth rate of the peak sidelobe ratio obtained in the current round relative to the peak sidelobe ratio calculated in the previous round is less than the second preset threshold.

[0031] According to an embodiment of the present invention, acquiring two signal segments obtained from the phase changes of two probe beams with opposite transmission directions before and after transmission through an optical fiber includes:

[0032] Two phase signals are obtained by measuring the phase changes of two probe beams with opposite transmission directions before and after transmission through an optical fiber.

[0033] The instantaneous energy sequences of the two phase signals are determined respectively, and the instantaneous energy sequences characterize the change of the energy of the phase signals over time;

[0034] Based on the instantaneous energy sequences of the two phase signals, the one with the higher signal-to-noise ratio is selected as the reference signal. The instantaneous energy sequence of the reference signal is then searched from the maximum energy value in the instantaneous energy sequence of the reference signal in the direction of decreasing time to determine the start time of the vibration event.

[0035] Using the start time of the reference signal as the truncation reference, the two phase signals are synchronously truncated for the same time span starting from the reference, resulting in two signal segments.

[0036] According to an embodiment of the present invention, determining the location of a vibration event acting on an optical fiber using multiple mode pairs includes:

[0037] Determine the cross-correlation coefficients and spectral barycenter frequencies of multiple mode pairs;

[0038] Among multiple mode pairs, the mode pair with the largest product of cross-correlation coefficient and barycenter frequency is selected as the target mode pair;

[0039] The location of the vibration event acting on the optical fiber is determined by using the target mode pair.

[0040] According to an embodiment of the present invention, a method for determining the location of a vibration event acting on an optical fiber includes:

[0041] Determine the time delay of the first and second modes in the target mode pair;

[0042] The location of the vibration event acting on the fiber can be obtained based on the time delay, the length of the fiber, the propagation speed of the probe light in the fiber, and the effective refractive index of the fiber core for the probe light.

[0043] The embodiments of the present invention have the following beneficial effects: by using the first mode spectrum of a high signal-to-noise ratio reference channel to determine weighting coefficients (i.e., constructing a spectral mask), and reconstructing the spectrum of a low signal-to-noise ratio measurement signal segment with these weighting coefficients, the measurement channel is forced to share a frequency band strictly consistent with the reference channel, so that the reconstructed second mode and the first mode form a one-to-one correspondence in center frequency and frequency band range. This spectral projection mechanism eliminates frequency misalignment and mode aliasing phenomena in traditional algorithms from the root, ensuring phase consistency of each mode pair in the time-frequency domain.

[0044] This invention relates to a method that projects the weighting coefficients of a reference channel onto the spectrum of a measurement channel, thereby weighting the frequency components of the measurement channel. Since the weighting coefficients characterize the correlation between each frequency and the vibration event, this operation effectively recovers the effective signal components of the measurement signal segment that were previously submerged in noise, enabling time delay estimation for measurement signal segments that were originally unusable for localization due to their low signal-to-noise ratio.

[0045] This invention is implemented entirely at the signal processing level, requiring no additional repeater amplifiers, no changes to the fiber optic link hardware structure, and no increase in the sampling rate of the data acquisition card. Through algorithmic innovation—specifically, by generating a spectral mask using the spectral characteristics of a high signal-to-noise ratio (SNR) reference channel and forcibly projecting it onto a low SNR measurement channel—it effectively overcomes the effects of polarization fading and waveform distortion in long-distance transmission. While retaining the advantages of ultra-long-distance repeaterless sensing, it improves delay estimation and positioning reliability in complex disturbance environments by reconstructing the physical correlation of measurement signal segments, achieving high-precision positioning in ultra-long-distance sensing scenarios of hundreds of kilometers (>100km). Attached Figure Description

[0046] The above and other objects, features and advantages of the present invention will become more apparent from the following description of embodiments of the invention with reference to the accompanying drawings, in which:

[0047] Figure 1 A schematic diagram of a sensing and positioning device according to an embodiment of the present invention is shown.

[0048] Figure 2 A flowchart of a sensing and positioning method according to an embodiment of the present invention is shown.

[0049] Figure 3 A schematic diagram of spectral projection according to an embodiment of the present invention is shown.

[0050] Explanation of reference numerals in the attached figures

[0051] 101: Laser; 102: First coupler; 103: Second coupler; 104: Third coupler; 201: First circulator; 202: Second circulator; 203: First acousto-optic crystal; 204: Second acousto-optic crystal; 205: First acousto-optic driver; 206: Second acousto-optic driver; 207: Optical fiber; 301: Fourth coupler; 302: Fifth coupler; 303: First photodetector; 304: Second photodetector; 400: Data acquisition card; 500: Computer; 600: Arbitrary waveform generator. Detailed Implementation

[0052] In a two-way forward transmission system, the randomness of vibration location leads to extreme asymmetry in the bidirectional probe light transmission path. When vibration approaches one end of the bidirectional probe light link, one probe light path transmits only through the short optical path, resulting in high signal fidelity (high signal-to-noise ratio); while the other reverse probe light path must traverse the long optical path. In ultra-long-distance single-mode fiber transmission, the optical signal not only faces accumulated phase noise and optical power attenuation, but also suffers from signal fading caused by random polarization state evolution (i.e., polarization-induced fading). This polarization-induced fading caused by birefringence leads to severe nonlinear distortion and time-domain warping of the signal envelope in the long optical path channel, resulting in an extremely asymmetric distribution of the signal-to-noise ratio between the two channels.

[0053] In environments with high noise and signal distortion, traditional time-frequency analysis and independent mode decomposition methods have significant limitations. On the one hand, cross-correlation localization algorithms based on waveform similarity fail due to correlation peak ambiguity; on the other hand, although introducing time-frequency analysis tools can improve noise resistance, commonly used wavelet transforms are limited by fixed basis functions and are difficult to adaptively handle non-stationary vibration signals. In contrast, variational mode decomposition (VMD) algorithms are widely used due to their adaptive narrowband decomposition characteristics. However, existing independent variational mode decomposition (IVMD) is essentially an independent optimization for single-channel data. When applied to scenarios with severe signal-to-noise ratio mismatch and waveform distortion in dual channels, if independent decomposition is still performed on the two signals separately, strong noise will interfere with the iterative convergence process of the center frequency of the long optical path channel, leading to severe frequency misalignment or mode aliasing in the local oscillator mode functions (IMFs) decomposed from the dual channels. This loss of physical correspondence at the modal level will further prevent positioning methods based on dual-channel modal delay estimation from obtaining accurate time differences, ultimately severely restricting the positioning accuracy and reliability of ultra-long-distance distributed optical fiber sensing systems.

[0054] This invention aims to solve the problem of extreme signal-to-noise ratio asymmetry in dual-channel systems caused by optical path difference and polarization-induced fading in dual-end forward transmission distributed optical fiber sensing systems in sensing scenarios of hundreds of kilometers (>100km), as well as the problems of mode mismatch and positioning failure caused by traditional independent mode decomposition algorithms. To address this, this invention proposes an ultra-long-distance distributed fiber optic sensing and positioning method and system based on spectral projection variational mode decomposition. The main objectives are: First, to construct a spectral guidance mechanism under asymmetric signal-to-noise ratio (SNR), i.e., to generate a spectral mask using the dominant spectral features of the high SNR reference channel (short optical path) and forcibly project it onto the low SNR measurement channel (long optical path), achieving cross-channel prior feature transfer; Second, to restore the physical consistency of modes, i.e., to forcibly constrain the decomposition boundary of the dual-channel signals within the variational optimization framework, overcoming interference from strong noise and signal distortion, eliminating frequency misalignment and mode aliasing phenomena in traditional algorithms, and ensuring phase consistency of the decomposed modes in the time-frequency domain; Third, to achieve precise positioning, i.e., while retaining the advantages of ultra-long-distance relay-free sensing, to significantly improve the system's time delay estimation and positioning reliability under complex disturbance environments by reconstructing the physical correlation of the dual-channel signals.

[0055] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the invention. In the following detailed description, numerous specific details are set forth to provide a thorough understanding of the embodiments of the invention for ease of explanation. However, it will be apparent that one or more embodiments may be practiced without these specific details. Furthermore, descriptions of well-known structures and techniques are omitted in the following description to avoid unnecessarily obscuring the concept of the invention.

[0056] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the invention. The terms “comprising,” “including,” etc., as used herein indicate the presence of the stated features, steps, operations, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, or components.

[0057] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein are to be interpreted in a manner consistent with the context of this specification, and not in an idealized or overly rigid way.

[0058] Figure 1 A schematic diagram of a sensing and positioning device according to an embodiment of the present invention is shown.

[0059] like Figure 1 As shown, the sensing and positioning device is used to implement the sensing and positioning method of this embodiment of the invention. Laser 101 emits continuous light with a center wavelength of 1550.12 nm. The light energy output of laser 101 can be controlled by adjusting the current. Laser 101 can be, for example, a semiconductor laser. The laser emitted by laser 101 is split into a first sub-laser and a second sub-laser by a 1:1 optical power ratio via a first coupler 102. Subsequently, the first sub-laser is split again by a 1:1 optical power ratio via a second coupler 103 into a first probe beam and a first local oscillator beam. The second sub-laser is then split again by a 1:1 optical power ratio via a third coupler 104 to form a second local oscillator beam and a second probe beam. The first probe light, traveling clockwise, passes through the first circulator 201, the first acousto-optic crystal 203, the optical fiber 207, and the second acousto-optic crystal 204, reaching the second circulator 202. The second probe light, traveling counterclockwise, passes through the second circulator 202, the second acousto-optic crystal 204, the optical fiber 207, and the first acousto-optic crystal 203, reaching the first circulator 201. The first acousto-optic crystal 203 and the second acousto-optic crystal 204 are modulated and driven by the first acousto-optic driver 205 and the second acousto-optic driver 206, respectively. The frequency shifts of the first acousto-optic crystal 203 and the second acousto-optic crystal 204 are different.

[0060] In this embodiment of the invention, a vibration event refers to an external disturbance acting on the optical fiber, such as mechanical vibration caused by human intrusion, mechanical vibration, seismic waves, or pipe leaks. The vibration event physically couples with the sensing optical fiber, causing an instantaneous change in the fiber's geometric length or refractive index, thereby modulating the phase of the probe light transmitted within it. The sensing and localization method of this embodiment is based on a dual-ended forward transmission type distributed optical fiber sensing system. Probe light is injected from both ends of the optical fiber, and the two probe lights propagate in opposite directions within the fiber. These two probe lights are designated as the first probe light and the second probe light.

[0061] When a vibration occurs at a point on the optical fiber, both probe beams carry information about the phase change caused by the vibration. The optical field of the first probe beam used for heterodyne detection in forward transmission... With the light field of the second book oscillator They can be expressed as equation (1) and equation (2) respectively.

[0062] (1);

[0063] (2).

[0064] in, The intensity amplitude of the first probe light. The intensity amplitude of the second radiant light. The frequency of the laser emitted by the laser; The cumulative frequency shift of the first and second acousto-optic modulators is the sum of the frequency shifts of the first and second acousto-optic modulators. For the first detection light in Phase of time, For the second book of Zhenguang The phase of a moment.

[0065] The first local oscillator light output from the second coupler 103 and the second probe light output from the first circulator 201 are coupled into the fourth coupler 301, and a first beat frequency is performed in the first photodetector 303. The first photodetector 303 is also used to convert the optical signal after the first beat frequency into a first electrical signal. The second local oscillator light output from the third coupler 104 and the first probe light output from the second circulator 202 are coupled into the fifth coupler 302, and a second beat frequency is performed in the second photodetector 304. The second photodetector 304 is also used to convert the optical signal after the second beat frequency into a second electrical signal. When a disturbance is applied at a certain position in the optical fiber 207, the first electrical signal... Represented by equation (3), the second electrical signal It is represented by equation (4).

[0066] (3);

[0067] (4).

[0068] in, These are parameters related to the optical power of the first local oscillator and the second probe beam. These are parameters related to the optical power of the second local oscillator and the first probe beam. and The phase change caused by vibration, This represents the phase change of the second probe light before and after transmission through the optical fiber. This represents the phase change of the first probe light before and after transmission through the optical fiber. The initial phase introduced for the fourth coupler 301 The initial phase introduced for the fifth coupler 302.

[0069] The first and second electrical signals are acquired by the data acquisition card 400. Due to the strong backscattering noise in long-distance fiber optic sensing, the acquired raw signals may contain non-target noise frequency bands. Therefore, the data acquisition card first mixes the two electrical signals with a reference signal at a frequency of (A+B) MHz to obtain two mixed signals. The frequency shift of the first acousto-optic crystal 203 is A MHz, and the frequency shift of the second acousto-optic crystal 204 is B MHz, with (A+B) MHz being the sum of the frequency shifts of the two acousto-optic crystals. Next, specific filtering techniques are used to process the two mixed signals to remove noise caused by far-end backscattering, ensuring high-quality signal transmission. This filtering step effectively suppresses the impact of low signal-to-noise ratio, enhancing the stability and accuracy of the system. Finally, algorithms such as quadrature demodulation, low-pass filtering, arctangent decomposition, and unwinding are used to process the signals, ultimately obtaining two phase signals with highly similar waveforms in the time domain but with a certain time delay difference. and These can be represented as follows.

[0070] (5);

[0071] (6).

[0072] Where c is the speed of light in a vacuum, and n is the refractive index of the fiber core. Where is the total length, and d is the distance the second probe light travels from the vibration point on the optical fiber to the point before being detected. For phase signal functions, For the propagation delay of the second probe light, The propagation delay of the first probe light. This refers to the distance the first probe light travels from the vibration point on the optical fiber to the point before being detected. Afterwards, a computer 500 can be used to... and The location where the vibration event acts on the optical fiber is obtained.

[0073] Figure 2 A flowchart of a sensing and positioning method according to an embodiment of the present invention is shown.

[0074] like Figure 2 As shown, the method includes operations S1 to S4.

[0075] In operation S1, in response to a vibration event acting on the optical fiber, two signal segments are acquired, which are obtained by the phase changes of two probe lights with opposite transmission directions (i.e., the first probe light and the second probe light mentioned above) before and after transmission through the optical fiber; the signal segments carry vibration event information.

[0076] In operation S2, based on the multiple first modes of the reference signal segment, weighting coefficients are determined for multiple frequencies in the spectrum of each first mode. These weighting coefficients characterize the correlation between the frequency and the vibration event; a larger weighting coefficient indicates a stronger correlation. The reference signal segment is the one with the higher signal-to-noise ratio among the two signal segments.

[0077] Because the vibration location is random, the portion of the probe light carrying vibration event information is detected after passing through a short optical path. Therefore, the signal attenuation is small and the noise is low. This signal is defined as a reference signal segment.

[0078] Vibration events acting on optical fibers typically produce phase signals that are not pure signals of a single frequency. In reality, a vibration event itself may contain multiple frequency components. These different frequency components are distributed across different frequency bands in the frequency domain, making it difficult to fully characterize them with a single mode. Therefore, multiple first modes are needed to capture the signal components within different frequency bands.

[0079] The spectrum of the first mode refers to the spectrum obtained after transforming the first mode from the time domain to the frequency domain. The spectrum can intuitively show which frequencies in the first mode have high energy and which have low energy. The weighting coefficient can also be understood as a confidence or credibility index. Since the vibration signal generated by the vibration event usually has energy concentration characteristics in the local frequency band, the energy distribution of this frequency can be directly mapped to the correlation between the frequency and the vibration event. The higher the weighting coefficient, the greater the correlation between the frequency and the vibration event, or in other words, the greater the probability that the frequency originates from the vibration event.

[0080] In operation S3, the spectrum of the measurement signal segment is reconstructed using multiple weighting coefficients corresponding to the first mode, resulting in a reconstructed spectrum that matches the spectrum of the first mode; the measurement signal segment is the one with the lower signal-to-noise ratio among the two signal segments.

[0081] A measurement signal segment refers to the signal segment with the lower signal-to-noise ratio (SNR) among two signal segments. In a two-way forward transmission system, the longer optical path has a lower SNR than the shorter optical path due to its greater transmission distance, higher accumulated noise, and polarization-induced fading. This signal is defined as the measurement signal segment.

[0082] The set of weighting coefficients determined by the first mode spectrum of the reference signal segment constitutes a spectral mask (filter), the shape of which is entirely determined by the spectral characteristics of the reference signal segment. The process of using this spectral mask to perform weighted filtering on the measurement signal segment in the frequency domain is called spectral projection. Reconstruction in this operation refers to the process of rebuilding or modifying the original spectrum of the measurement signal segment. Specifically, instead of keeping the original spectrum unchanged, the weighting coefficients are used to weight each frequency component of the measurement signal segment, thereby changing the shape and energy distribution of the spectrum.

[0083] The reconstructed spectrum matching the spectrum of the first mode refers to the spectrum of the reconstructed measurement signal segment that is consistent with the spectrum of the first mode of the reference signal segment in terms of frequency domain support (i.e., the frequency band where the main energy is located) and spectral envelope shape. In the reconstructed spectrum, frequency regions with high weighting coefficients (frequency bands in the reference signal segment that belong to vibration events) are preserved or even enhanced; frequency regions with low weighting coefficients (i.e., frequency bands in the reference signal segment that belong to noise) are suppressed or reduced to zero.

[0084] It should be noted that this weighting process is essentially a correction of the spectral amplitude of the measured signal segment. While changing the frequency structure and energy distribution, it completely preserves the original phase information of the frequency band in the measured signal segment, thus providing the physical conditions for subsequent high-precision cross-correlation time delay estimation.

[0085] In operation S4, multiple mode pairs are used to determine the location of the vibration event acting on the optical fiber, wherein the mode pair includes a first mode and a second mode, and the second mode is obtained from the reconstructed spectrum.

[0086] Each mode pair is a signal combination formed by pairing a first mode and a second mode. The first mode comes from a reference signal segment (a high signal-to-noise ratio signal segment), and the second mode comes from a measurement signal segment (a low signal-to-noise ratio signal segment). Due to the preceding spectral projection operation, a one-to-one correspondence between the first and second modes is forced across the frequency band.

[0087] Determining the location of a vibration event on an optical fiber using multiple mode pairs involves using multiple mode pairs, which form a one-to-one correspondence after spectral projection, as the basic unit for localization. Each mode pair contains a first mode from a reference channel and a second mode obtained from the reconstructed spectrum. Since each mode pair corresponds to a dual-channel signal component within the same frequency band, and there is a time difference between the two signals due to the different propagation path lengths, which is related to the vibration location, this time difference can be used to accurately determine the location of the vibration point.

[0088] According to an embodiment of the present invention, weighting coefficients (i.e., constructing a spectral mask) are determined using the spectrum of the first mode of a high signal-to-noise ratio (SNR) reference channel, and the spectrum of a low SNR measurement signal segment is reconstructed using these weighting coefficients. This forces the measurement channel to share a frequency band strictly consistent with the reference channel, ensuring a one-to-one correspondence between the reconstructed second mode and the first mode in terms of center frequency and frequency band. This spectral projection mechanism eliminates frequency misalignment and mode aliasing in traditional algorithms at their source, ensuring phase consistency of each mode pair in the time-frequency domain.

[0089] This invention projects the weighting coefficients (i.e., spectral mask) of the reference channel onto the spectrum of the measurement channel, thereby weighting each frequency component of the measurement channel. Since the weighting coefficients characterize the probability that each frequency belongs to a vibration event, this operation effectively recovers the effective signal components of the measurement signal segment that were submerged in noise, enabling measurement signal segments that were originally unusable for localization due to their low signal-to-noise ratio to have the ability to estimate time delays.

[0090] This invention is implemented entirely at the signal processing level, requiring no additional repeater amplifiers, no changes to the fiber optic link hardware structure, and no increase in the sampling rate of the data acquisition card. Through algorithmic innovation—specifically, by generating a spectral mask using the spectral characteristics of a high signal-to-noise ratio (SNR) reference channel and forcibly projecting it onto a low SNR measurement channel—it effectively overcomes the effects of polarization fading and waveform distortion in long-distance transmission. While retaining the advantages of ultra-long-distance repeaterless sensing, it improves delay estimation and positioning reliability in complex disturbance environments by reconstructing the physical correlation of measurement signal segments, achieving high-precision positioning in ultra-long-distance sensing scenarios of hundreds of kilometers (>100km).

[0091] According to an embodiment of the present invention, in operation S1, two signal segments are obtained by the phase changes of two probe lights with opposite transmission directions before and after transmission through the optical fiber, including operations S11 to S14.

[0092] In operation S11, two phase signals are obtained by the phase changes of two probe lights with opposite transmission directions before and after transmission through the optical fiber.

[0093] In operation S12, the instantaneous energy sequences of the two phase signals are determined respectively. The instantaneous energy sequence characterizes the change of the energy of the phase signal over time.

[0094] In operation S13, based on the instantaneous energy sequences of the two phase signals, the one with the higher signal-to-noise ratio is selected as the reference signal, and the instantaneous energy sequence of the reference signal is searched from the maximum energy value in the instantaneous energy sequence of the reference signal in the direction of decreasing time to determine the start time of the vibration event.

[0095] In operation S14, taking the start time of the reference signal as the truncation reference, synchronous truncation of the two phase signals with the same time span is performed from the reference to obtain the two signal segments.

[0096] For example, in long-distance distributed sensing, accurately identifying the starting point of transient vibration events is a prerequisite for precise localization. This embodiment of the invention preferably introduces a nonlinear Teager energy operator as a pre-processor to extract signal features. It should be noted that the core of this step lies in extracting the effective range of the vibration event; therefore, this invention is not limited to the Teager energy operator algorithm. In practical applications, other transient event detection methods such as short-time energy methods, zero-crossing rate detection methods, or envelope detection methods can also be used, as long as the start and end times of the vibration event can be effectively extracted, they are all within the scope of protection of this invention.

[0097] The two acquired phase signals are abbreviated as follows: and In practical discrete sampling systems, to ensure the strict non-negativity of transient energy for easier threshold determination, ... replace or In this embodiment, a modified absolute value type Teager energy operator is preferably used to calculate the instantaneous energy sequence. The calculation formula is as follows:

[0098] (7);

[0099] Where i represents the sequence number of the sampling point, i = 1, 2, ..., I, and I is the data length or the total number of sampling points. This indicates the phase amplitude at the current sampling point; and These are the phase amplitudes of the preceding and following sampling points of the i-th sampling point, respectively; Represents the Teager energy operator symbol; This represents the calculated instantaneous energy value (here). This represents a nonlinear calculation result that reflects the characteristics of the product of the signal's spectral amplitude and frequency. The larger the value, the stronger the high-frequency transient components of the signal at that moment.

[0100] In ultra-long-distance (e.g., 100-kilometer-scale) sensing, high-frequency vibration components are severely attenuated and easily drowned out by low-frequency noise. The Teager energy operator exhibits joint sensitivity to abrupt changes in the instantaneous amplitude and frequency of the signal.

[0101] To overcome the boundary oscillations and spikes in the instantaneous energy sequence introduced by pre-filtering, this invention proposes an adaptive detection and synchronous truncation strategy based on sliding smoothing and peak backtracking. The specific steps are as follows:

[0102] The first step is to apply a moving average filter to the instantaneous energy sequence, with the window width matching the impact envelope width, to smooth out spikes and oscillations in the sequence.

[0103] The second step is to retrieve the global energy main peak and, based on experience, set a specific proportion of its peak value as an adaptive backtracking threshold. In this embodiment, the preferred proportion is 10%, that is, the adaptive backtracking threshold is taken as 10% of the peak value of the global energy main peak.

[0104] The third step is to start from the position of the global energy main peak and search backward (in reverse time) to find the position where the instantaneous energy value first drops below the adaptive backtracking threshold, and lock this position as the starting index of the vibration event (the starting position in the time domain phase signal).

[0105] To eliminate the time-domain misalignment and frequency resolution mismatch caused by independent detection of two phase signals, this embodiment of the invention forces the origin index of a reference signal segment with a high signal-to-noise ratio to be used as a unified truncation benchmark, while retaining an appropriate amount of pre-trigger buffer to fully extract the precursor wavefront. Finally, two signal segments are obtained by generating region-of-interest signal segments of equal length in the reference signal segment and the measurement signal segment.

[0106] According to an embodiment of the present invention, in operation S2, the method for determining multiple first modes includes: performing variational mode decomposition on a reference signal segment using a target number of modes and a target quadratic penalty factor to obtain multiple first modes with different center frequencies. The target number of modes characterizes the number of first modes in the reference signal segment (the target number of modes is the same as the number of first modes in the reference signal segment), and the target quadratic penalty factor is used to characterize the bandwidth of the frequency band of the first modes.

[0107] Variational Mode Decomposition (VMD) algorithms require two key preset parameters: the target number of modes (or the target number of mode decomposition layers) K and the target second-order penalty factor α. The target number of modes K determines how many local oscillator mode functions (IMFs) the reference signal segment is decomposed into, or how many first modes the reference signal segment is decomposed into. The target second-order penalty factor α controls the bandwidth of each first mode; the larger the α value, the narrower the bandwidth of each first mode, and the higher the bandwidth compactness of the reconstructed waveform. The determination methods for these two parameters are explained in detail below.

[0108] According to an embodiment of the present invention, the method for determining the target number of modes includes: performing a second variational mode decomposition on a reference signal segment using an initial number of modes N to obtain N third modes, where N≥2; removing third modes from the N third modes whose transient energy ratio is less than a first preset threshold to obtain M remaining third modes, where 1≤M≤N; sorting the energy of the M remaining third modes in descending order and accumulating them one by one according to the sorting order until the ratio of the accumulated energy to the total energy of the M remaining third modes satisfies a preset condition; and obtaining the target number of modes based on the number of third modes participating in the accumulation when the preset condition is met.

[0109] To avoid human intervention, this invention proposes the following adaptive strategy for decoupling optimization. A relatively large initial number of modes, N = 6, is set, sufficient to cover low-frequency noise, dominant vibrations, and high-frequency harmonic components in the actual signal. An initial variational mode decomposition is performed on the reference signal segment to obtain N third modes (i.e., the third local oscillator mode functions).

[0110] Considering the high-frequency non-Gaussian impact characteristics of vibration events, the Transient Energy Ratio (TER) is calculated for each third mode obtained from the decomposition. This index combines time-domain energy jumps with high-frequency non-Gaussian impact characteristics, effectively suppressing the probability of misjudgment due to persistent high-frequency Gaussian noise. The calculation formula is as follows:

[0111] (8).

[0112] in, and , representing the data vectors of the m-th third mode in the vibration event interval and the background noise interval, respectively; RMS(·) represents the function for calculating the root mean square value of the vector, and Kurtosis(·) represents the function for calculating the kurtosis ratio of the vector. This comprehensive index combines the time-domain energy jump and high-frequency non-Gaussian impact characteristics, and can effectively suppress the misjudgment probability of continuous high-frequency Gaussian noise.

[0113] Set the filter threshold (For example, it can be 4.0), the third modes (i.e., pure noise modes) with transient energy ratios less than this threshold are removed, resulting in M ​​remaining third modes. The retained third modes are accumulated in descending order of energy, and the number of third modes required when the accumulated energy accounts for a certain proportion (preferably 90%) of the total effective energy is selected. Based on this, the present invention preferably adds 1-dimensional redundant degree of freedom as a residual absorption mode, and the final determined target number of modes K is preferably limited to the range of [2,5]. This redundancy mechanism can effectively absorb the background low-frequency common-mode drift that is not captured by the main mode, and protect the modes with vibration event information from crosstalk.

[0114] According to an embodiment of the present invention, the method for determining the target quadratic penalty factor includes: repeatedly performing the following steps until the peak-to-sidelobe ratio of the cross-correlation function satisfies a preset termination condition, and taking the candidate quadratic penalty factor that satisfies the preset termination condition as the target quadratic penalty factor.

[0115] Step A: Using the candidate quadratic penalty factor and the target mode number corresponding to the current round, perform variational mode decomposition on the reference signal segment and the measurement signal segment to obtain multiple fourth modes and multiple fifth modes.

[0116] Step B involves performing a cross-correlation operation on the envelopes of the first target mode among multiple fourth modes and the second target mode among multiple fifth modes to obtain the cross-correlation function corresponding to the current round. Here, the first target mode is the fourth mode with the highest energy or the highest transient energy ratio among multiple fourth modes, and the second target mode is the fifth mode among multiple fifth modes that shares the same mode as the first target mode.

[0117] Step C: If the peak-to-sidelobe ratio of the cross-correlation function corresponding to the current round does not meet the preset termination condition, determine the candidate quadratic penalty factor corresponding to the next round.

[0118] The aforementioned preset termination conditions include: the peak sidelobe ratio obtained in the current round is greater than the first preset threshold, and the growth rate of the peak sidelobe ratio obtained in the current round relative to the peak sidelobe ratio calculated in the previous round is less than the second preset threshold.

[0119] The target quadratic penalty factor α determines the bandwidth compactness of the reconstructed waveform. To directly serve the final positioning target, this embodiment of the invention breaks with the conventional single-channel frequency domain extremum and proposes optimization based on the peak side lobe ratio (PSR) of the dual-channel spatial envelope cross-correlation.

[0120] Before performing PSR calculation, it is necessary to first determine the two signals used for cross-correlation. From the K first modes obtained by decomposing the reference signal segment, the mode with the highest energy or the highest transient energy ratio (TER) is selected and defined as the first target mode. This first target mode contains the dominant vibration characteristics and best represents the true distribution of vibration events in the frequency domain. At the same time, variational mode decomposition is performed on the measurement signal segment to obtain K fifth modes. To prevent mode misalignment or aliasing caused by low signal-to-noise ratio and to avoid the measurement signal segment from mistakenly selecting noise modes due to fading, this embodiment adopts a mode number locking mechanism: directly based on the mode of the first target mode in the reference channel, the fifth mode that is the same as the first target mode is extracted from these K fifth modes and strictly defined as the second target mode. The envelope of the first target mode of the reference channel and the second target mode corresponding to the measurement channel are extracted, and then the cross-correlation function of the two is calculated. The PSR calculation formula is expressed as Equation (9).

[0121] (9);

[0122] in, Let be the cross-correlation function of the envelopes of the first target mode of the reference signal segment and the second target mode of the measured signal segment. The PSR is the time delay index corresponding to the main peak, and the denominator represents the absolute spectral amplitude of the largest sidelobe. A larger PSR indicates a sharper cross-correlation main peak. The amplitude of the main cross-correlation peak reflects the degree of matching between the two envelopes under optimal time delay.

[0123] In practice, α is scanned within a predefined candidate set (e.g., α∈[100, 300, 500, ..., 2000]) that includes multiple candidate secondary penalty factors. When an increase in α causes the PSR improvement rate to fall below a predefined threshold (e.g., 5%), the iteration is terminated early, thereby determining the target secondary penalty factor. This method avoids wavefront distortion caused by excessive bandpass contraction.

[0124] Using the optimized target mode number and target quadratic penalty factor, the reference signal segment is re-decomposed into variational mode, and the first mode with the highest energy or the highest transient energy ratio is extracted as the reference target spectrum, so as to specify an absolute standard reference template for the subsequent generation of masks (whether soft masks or hard masks).

[0125] According to embodiments of the present invention, in order to transfer the high signal-to-noise ratio characteristics of a reference signal segment to a measurement signal segment, multiple filters, i.e., multiple spectral masks, can be constructed using the spectral characteristics of each of the multiple first modes of the reference signal segment. The spectral mask consists of weighting coefficients for multiple frequencies, the magnitude of which characterizes the probability that the frequency belongs to a vibration event. This embodiment provides two methods for determining the weighting coefficients, which can be adaptively selected according to the actual noise environment. The first is a binarized hard spectral mask method, suitable for ideal Gaussian white noise scenarios, which divides frequency points in the spectrum into two categories by setting an amplitude threshold; the second is a normalized soft spectral mask method, suitable for real scenarios with severe dispersion in long-distance optical fibers, which calculates the energy proportion of each frequency in all first modes as the weight. These two methods are described in detail below.

[0126] In one embodiment of the present invention, in operation S2, the weighting coefficients of each of the multiple frequencies in the spectrum of the first mode are determined. For example, operations S21 to S22 can be performed for each of the first modes.

[0127] In operation S21, the maximum value of multiple spectral amplitudes is determined from the spectra of each first mode.

[0128] In operation S22, the weighting coefficients of each frequency in the spectrum of the first mode are obtained based on the maximum value and the spectral amplitude of each frequency in the spectrum of the first mode.

[0129] According to an embodiment of the present invention, this determination method is applicable to an ideal Gaussian white noise scenario. It can also be called a binary hard mask. The value of is called the weighting coefficient of frequency f. An amplitude threshold coefficient γ is set (preferably γ = 0.03 in this embodiment). The maximum spectral amplitude is expressed as... , This represents the spectral function of frequency f in the first mode. This represents the spectral amplitude of frequency f in the first mode.

[0130] For the frequency f in the spectrum of the first mode, its spectral amplitude is... With γ and maximum spectral amplitude The products are compared, and the weighting coefficients can be determined based on the comparison results. Specifically, it can be expressed as equation (10).

[0131] (10).

[0132] In formula (10), express hour, The value of . When the weighting coefficient obtained according to equation (10) is equal to 1, it indicates that the frequency is very likely to belong to a vibration event. When the weighting coefficient is equal to 0, it indicates that the frequency is very likely to belong to noise or interference.

[0133] According to another embodiment of the present invention, in operation S2, determining the weighting coefficients of each of the multiple frequencies in the spectrum of the first mode may include, for example, operations S23 to S24 for each of the first modes.

[0134] In operation S23, the spectral amplitude of each of the multiple frequencies in the spectrum of the first mode is determined.

[0135] In operation S24, the weighting coefficients of each frequency in the spectrum of the first mode are obtained based on the sum of the spectral amplitudes of each of the multiple frequencies and the spectral amplitudes of any frequency in the multiple first modes.

[0136] According to an embodiment of the present invention, this method corresponds to a normalized soft spectral mask and is applicable to real-world scenarios where long-distance optical fibers exhibit severe dispersion. This strategy utilizes the spectral information of all K modes of the reference channel to construct a normalized weighted mask, calculated using the following formula.

[0137] (11).

[0138] The meanings of each symbol are as follows: f represents frequency; This represents a normalized soft spectral mask. The value represents the weighting coefficient for that frequency f. For the reference signal segment, the spectrum of the k-th first mode at frequency f is... for The spectral amplitude; represents the spectral amplitude of the reference signal segment at frequency f in the j-th mode; K is the total number of the first modes obtained by variational mode decomposition of the reference signal segment. This is a small constant used to prevent the denominator from being zero. Unlike hard threshold truncation, this soft masking strategy suppresses dominant noise while preserving the edge transition bands of high-frequency transient components, avoiding the Gibbs effect caused by hard truncation, and ensuring the physical continuity of the reconstructed phase in the dual channels. It is particularly suitable for compensating for group velocity dispersion and polarization state evolution in long-distance optical fibers. When the weighting coefficient is close to 1 (or equal to 1), it indicates that the frequency is highly likely to be a vibration event. When the weighting coefficient is close to 0 (or equal to 0), it indicates that the frequency is highly likely to be noise or interference.

[0139] It should be noted that the spectral mask construction methods proposed in this embodiment are not limited to the two methods described above. Both the binarized mask and the normalized soft spectral mask can be uniformly represented by the spectral mask. express.

[0140] Any method that utilizes the spectral distribution characteristics (such as center frequency and bandwidth range) of a reference signal segment to construct a filter or weighting matrix and applies it to a measurement channel to constrain its decomposition or reconstruction process falls within the scope of the spectral projection concept of this invention.

[0141] According to an embodiment of the present invention, in operation S3, after the spectral mask is constructed, the next step is to perform a spectral projection operation, that is, to apply the constructed spectral mask to the spectrum of the measurement signal segment in order to achieve cross-channel feature transfer and signal reconstruction.

[0142] Spectrum of mask projected onto low signal-to-noise ratio measurement channel The above can be represented as equation (12).

[0143] (12).

[0144] in, For measuring signal segments The spectrum, The spectrum of the reconstructed measurement signal segment. express and Pointwise multiplication in the frequency domain (i.e., element-wise multiplication). Perform an inverse fast Fourier transform (IFFT) to obtain the denoised measurement signal segment. .

[0145] From a physical perspective, due to the spectral mask It is a pure real non-negative coefficient matrix, which mathematically acts as an absolute zero-phase filter. When reconstructing the measurement signal segment, it not only forces it to inherit the bandwidth limitation of the reference signal segment, but also eliminates the relative group delay bias caused by independent filtering (or independent VMD decomposition) of the two signal segments from the physical source, laying the foundation for subsequent phase alignment.

[0146] Figure 3 A schematic diagram of spectral projection according to an embodiment of the present invention is shown.

[0147] like Figure 3 As shown, it intuitively demonstrates the principle of spectral projection, which is to construct a bandpass mask using the pure spectrum of the reference signal segment (high signal-to-noise ratio), and then forcibly multiply it by the dot product (i.e. project it onto) the spectrum of the measurement signal segment (containing noise), thereby filtering out out-of-band noise and obtaining the spectrum of the reconstructed measurement signal segment. This ensures that the spectrum of the reference signal segment and the spectrum of the reconstructed measurement signal segment have completely consistent frequency band support, thus eliminating mode decomposition misalignment from a physical source.

[0148] According to an embodiment of the present invention, in operation S4, the location of the vibration event acting on the optical fiber is determined using multiple mode pairs, including operations S41 to S42.

[0149] In operation S41, the cross-correlation coefficients and spectral barycenter frequencies of multiple mode pairs are determined.

[0150] In operation S42, the mode pair with the largest square of the product of the cross-correlation coefficient and the spectral barycenter frequency is selected as the target mode pair from multiple mode pairs.

[0151] In order to select the mode pair with the highest positioning accuracy from the multiple decomposed modes, this embodiment introduces a comprehensive evaluation system based on the Cramér-Rao Lower Bound (CRLB) theory.

[0152] Specifically, according to the CRLB theory, the lower bound of the variance of the time delay estimate satisfies equation (13).

[0153] (13).

[0154] in, This represents the variance of the time delay estimate. Here, SNR represents the signal-to-noise ratio, and SNR represents the effective bandwidth of the signal. This formula shows that the theoretically highest accuracy of time delay estimation depends on the combined support of a high signal-to-noise ratio and a high effective bandwidth.

[0155] Based on the above theory, in order to approximate the lower bound of the theory, this invention defines a comprehensive scoring index. , which is expressed as equation (14).

[0156] (14).

[0157] in, This represents the overall score of the k-th modality pair. It represents the cross-correlation coefficient between mode pairs, which is used to equivalently characterize the consistency and joint signal-to-noise ratio of two signals; The spectral barycenter frequency of a mode, under the narrowband assumption, is related to the effective bandwidth of the signal. It shows a strong positive correlation.

[0158] In practice, all candidate modal pairs are traversed, and their respective comprehensive scores are calculated. Select The mode pair with the largest value is used as the optimal positioning signal (i.e., the target mode pair) for high-precision time delay calculation.

[0159] Furthermore, it should be noted that this embodiment preferably uses a comprehensive index based on CRLB for mode selection. However, within the scope of protection of this invention, the mode selection criteria are not limited to this. Those skilled in the art can also select the optimal mode based solely on cross-correlation coefficients, modal energy proportions, or other signal quality evaluation indicators such as spectral kurtosis. In other words, as long as the core idea is to utilize prior information from the reference channel to select high-quality components from the decomposition results, it should be covered within the scope of protection of this invention.

[0160] In operation S42, the location of the vibration event acting on the optical fiber is determined using the target mode pair.

[0161] Specifically, the target mode pairs selected through operations S41-S42 can be used for two-stage time delay estimation to calculate the location of the vibration event acting on the optical fiber. The first mode in the target mode pair is represented as... The second mode is represented as .

[0162] First, a rough time delay estimate is made using the first and second modes.

[0163] To address the potential whole-cycle ambiguity issue in phase measurements, a coarse time delay estimate of the phase time-domain signal is first performed based on the envelope. This is then applied to the following: The first mode and its representation are as follows The second mode is subjected to Hilbert transform to extract its signal envelope; then the cross-correlation function of the two envelope signals is calculated, and the time lag corresponding to the cross-correlation peak is searched as a coarse time delay window. This step determines the approximate range of the delay.

[0164] Secondly, a fine time delay estimation based on subsampling point-level high-resolution phase cross-correlation technology is adopted.

[0165] Obtaining a rough delay window Then, near this time delay window, from the first mode Second mode The corresponding original bandpass phase signal segment is extracted from the data. The original bandpass phase signal refers to the original phase signal that has not undergone envelope extraction and retains the complete high-frequency carrier characteristics.

[0166] The truncated signal is processed using a frequency-domain zero-padding interpolation resampling technique. Specifically, the spectra of the two signals are zero-paddinged in the frequency domain to extend the spectral length, and then the signals are mapped to a higher virtual operating frequency (e.g., boosted to 10 MHz) through an inverse Fourier transform. This operation significantly improves the time resolution of the time-domain signal without altering the original sampled data.

[0167] Cross-correlation is directly performed on the two resampled phase signals. Compared to the envelope-based cross-correlation in the coarse time delay estimation step, directly using the resampled phase signals that retain the complete high-frequency carrier characteristics for cross-correlation yields a sharper cross-correlation peak.

[0168] Parabolic fitting is performed at the maximum peak of the cross-correlation function and discrete points on both sides. Specifically, a parabola is fitted using the cross-correlation values ​​of the peak and its adjacent points, and the vertex position of this parabola is taken as the peak position of the subsampling precision. This method breaks through the time resolution limit of the discrete sampling grid and achieves fine time delay calculation with subsampling precision, denoted as Δτ.

[0169] Finally, based on the total length L of the optical fiber, the speed of light propagation in the optical fiber c, and the effective refractive index n of the fiber core, the precise vibration position d is calculated using the following formula.

[0170] (15).

[0171] To verify the feasibility and effectiveness of the technical solution of this invention in ultra-long-distance scenarios, the applicant used the aforementioned system to build a distributed optical fiber sensing experimental platform with a single span length exceeding 200 km. The experiment focused on testing the extreme asymmetric signal-to-noise ratio environment caused by optical path differences at the end of the optical fiber link.

[0172] Verification results show that when traditional independent variational mode decomposition algorithms fail to locate due to mode mismatch, the spectral projection variational mode decomposition localization algorithm described in this embodiment successfully recovers the signal characteristics submerged in noise. Multiple tests have verified that this system maintains good linearity and spatial consistency across the entire link range and achieves meter-level high-precision positioning at ultra-long sensing distances of up to 200 km. These results fully demonstrate that this invention effectively overcomes polarization fading and waveform distortion in long-distance transmission without adding relay equipment, significantly enhancing the system's engineering application value.

[0173] The spectral projection variational mode decomposition optimization algorithm in this invention not only solves the problems of mode mismatch and frequency misalignment caused by the asymmetry of signal-to-noise ratio in two phase signals, and eliminates the interference of strong noise and signal distortion on cross-correlation calculation; but also, through Cramer-Rhodes lower bound optimization and subsampling technology, it ensures a significant improvement in the peak resolution of cross-correlation without increasing the cost of the acquisition card, making up for the limited accuracy under the sampling rate limitation, and thus improving the positioning performance of ultra-long-distance systems.

[0174] Specifically, in this embodiment, the computer 500 is used for editing and storing data; the arbitrary waveform generator 600 has two functions: first, to provide a standard sine wave signal of a fixed frequency to the two acousto-optic driving modules; and second, to control the timing of the entire distributed acoustic sensing system, ensuring the continuity and accuracy of signal acquisition.

[0175] Furthermore, those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware (such as a processor, controller, etc.). The program can be stored in a computer-readable storage medium, and when executed, it can include the processes described in the above method embodiments. The computer-readable storage medium can be a memory, magnetic disk, or optical disk.

[0176] The embodiments of the present invention have been described above. However, these embodiments are merely illustrative and not intended to limit the scope of the invention. Although various embodiments have been described above, this does not mean that the measures in the various embodiments cannot be used advantageously in combination. Various substitutions and modifications can be made by those skilled in the art without departing from the scope of the invention, and all such substitutions and modifications should fall within the scope of the invention.

Claims

1. A sensing and positioning method, characterized in that, include: In response to a vibration event acting on the optical fiber, two signal segments are obtained by measuring the phase changes of two probe lights with opposite transmission directions before and after transmission through the optical fiber. The signal segment carries vibration event information; Based on multiple first modes of a reference signal segment, weighting coefficients are determined for multiple frequencies in the spectrum of each first mode; the weighting coefficients characterize the correlation between the frequency and the vibration event, and the reference signal segment is the one with the higher signal-to-noise ratio among two signal segments. The spectrum of the measurement signal segment is reconstructed using multiple weighting coefficients corresponding to the first mode to obtain a reconstructed spectrum that matches the spectrum of the first mode; the measurement signal segment is the one with the lower signal-to-noise ratio among the two signal segments. The location of the vibration event acting on the optical fiber is determined using multiple mode pairs, wherein the mode pairs include a first mode and a second mode, and the second mode is obtained from the reconstructed spectrum.

2. The sensing and positioning method according to claim 1, characterized in that, Determine the weighting coefficients for multiple frequencies in the spectrum of each first mode, including: For each first mode, Determine the maximum value of multiple spectral amplitudes from the spectra of each of the first modes; Based on the maximum value and the spectral amplitude of each frequency in the spectrum of the first mode, the weighting coefficients of each frequency in the spectrum of the first mode are obtained.

3. The sensing and positioning method according to claim 1, characterized in that, Determine the weighting coefficients for multiple frequencies in the spectrum of each first mode, including: For each first mode, Determine the spectral amplitude of each of the multiple frequencies in the spectrum of the first mode; The weighting coefficients of each frequency in the spectrum of the first mode are obtained by summing the spectral amplitudes of each of the multiple frequencies and the spectral amplitudes of any one of the multiple frequencies in the multiple first modes.

4. The sensing and positioning method according to claim 1, characterized in that, The method for determining the plurality of first modes includes: Variational mode decomposition is performed on the reference signal segment using a target mode number and a target quadratic penalty factor to obtain multiple first modes with different center frequencies. The target mode number represents the number of modes in the reference signal segment, and the target quadratic penalty factor represents the bandwidth of the first mode.

5. The sensing and positioning method according to claim 4, characterized in that, The method for determining the target modal number includes: The reference signal segment is subjected to variational mode decomposition using the initial number of modes N to obtain N third modes, where N≥2; Remove the third modes from the N third modes whose transient energy ratio is less than the first preset threshold, and obtain the remaining M third modes, 1≤M≤N; The energies of the M remaining third modes are sorted from largest to smallest, and then accumulated one by one in the order of sorting until the ratio of the accumulated energy to the total energy of the M remaining third modes meets a preset condition. The target number of modes is obtained by counting the number of second modes that participate in the accumulation when the preset conditions are met.

6. The sensing and positioning method according to claim 5, characterized in that, The method for determining the target secondary penalty factor includes: Repeat the following steps until the peak-to-sidelobe ratio of the cross-correlation function meets the preset termination condition, and use the candidate quadratic penalty factor that meets the preset termination condition as the target quadratic penalty factor: Using the candidate quadratic penalty factor corresponding to the current round and the target mode number, variational mode decomposition is performed on the reference signal segment and the measurement signal segment respectively to obtain multiple fourth modes and multiple fifth modes; Cross-correlation is performed on the envelopes of the first target mode among multiple fourth modes and the envelopes of the second target mode among multiple fifth modes to obtain the cross-correlation function corresponding to the current round; the first target mode is the fourth mode with the highest energy or the highest time-frequency energy ratio among the multiple fourth modes, and the second target mode is the fifth mode among the multiple fifth modes that has the same mode as the first target mode; If the peak sidelobe ratio of the cross-correlation function corresponding to the current round does not meet the preset termination condition, the candidate quadratic penalty factor corresponding to the next round is determined.

7. The sensing and positioning method according to claim 6, characterized in that, The preset termination conditions include: The peak sidelobe ratio obtained in the current round is greater than the first preset threshold, and the growth rate of the peak sidelobe ratio obtained in the current round relative to the peak sidelobe ratio calculated in the previous round is less than the second preset threshold.

8. The sensing and positioning method according to claim 1, characterized in that, Two signal segments are obtained by acquiring the phase changes of two probe beams with opposite transmission directions before and after transmission through an optical fiber, including: Two phase signals are obtained by measuring the phase changes of two probe beams with opposite transmission directions before and after transmission through an optical fiber. The instantaneous energy sequences of the two phase signals are determined respectively, and the instantaneous energy sequences characterize the change of the energy of the phase signals over time; Based on the instantaneous energy sequences of the two phase signals, the one with the higher signal-to-noise ratio is selected as the reference signal, and the instantaneous energy sequence of the reference signal is searched from the maximum energy value in the instantaneous energy sequence of the reference signal in the direction of decreasing time to determine the start time of the vibration event. Using the start time of the reference signal as the truncation reference, the two phase signals are synchronously truncated for the same time span starting from the reference to obtain the two signal segments.

9. The sensing and positioning method according to claim 1, characterized in that, Determining the location where the vibration event acts on the optical fiber using multiple mode pairs includes: Determine the cross-correlation coefficients and spectral barycenter frequencies of each of the multiple mode pairs; Among multiple mode pairs, the mode pair with the largest product of cross-correlation coefficient and barycenter frequency is selected as the target mode pair; The location where the vibration event acts on the optical fiber is determined using the target mode pair.

10. The sensing and positioning method according to claim 9, characterized in that, The method for determining the location of the vibration event acting on the optical fiber includes: Determine the time delay of the first and second modes in the target mode pair; The location where the vibration event acts on the optical fiber is obtained based on the time delay, the length of the optical fiber, the propagation speed of the probe light in the optical fiber, and the effective refractive index of the fiber core for the probe light.

Images