Optical cable vibration event detection method and system based on multi-physical field perception

Abstract

The application discloses a cable vibration event detection method and system based on multi-physical field perception, wherein the method comprises the following steps: collecting and processing the back Rayleigh scattering signal in the optical fiber based on a Φ-OTDR device to obtain the physical position, occurrence timestamp, duration and waveform data of the vibration event on the optical fiber; collecting IoT node original data including the acoustic signal and environmental temperature data of the optical fiber; after identifying the vibration event, performing time-space dimension collaborative screening on the IoT node original data, performing time-domain feature extraction on the acoustic signal, inputting the extracted time-domain features into a lightweight classification model, and outputting the vibration source type and its confidence; and finally determining the vibration event in combination with the energy intensity, temperature change rate and vibration source type. The application solves the problem that the existing Φ-OTDR device cannot effectively distinguish the vibration source type in the cable vibration event identification, resulting in a high false alarm rate.

Description

Technical Field

[0001] This invention belongs to the field of optical cable testing technology, and specifically relates to a method and system for detecting optical cable vibration events based on multi-physics sensing. Background Technology

[0002] A large number of communication optical cables are laid underground, either directly buried or in ducts, and are constantly exposed to multiple risks, including damage from third-party construction, geological subsidence, and environmental erosion. In the event of a cable outage, not only will it cause significant economic losses, but it may also trigger public safety incidents. Therefore, achieving real-time, accurate, and intelligent identification and early warning of the path, depth, and operational status of underground optical cables has become a critical technical challenge that the industry urgently needs to address.

[0003] Current fiber optic cable monitoring technologies mainly rely on traditional manual inspections or automated systems based on single sensing methods. Among them, distributed fiber optic vibration sensing technologies, such as Φ-OTDR (phase-sensitive optical time-domain reflectometer), have been widely used in the field of fiber optic cable security protection due to their fully distributed, high-sensitivity, and long-distance monitoring capabilities. This technology enables continuous sensing and location of vibration events along the cable by emitting laser pulses into the optical fiber and analyzing the phase changes of the backscattered Rayleigh signal.

[0004] Currently, the Φ-OTDR system still has some limitations in practical applications. For example, it can only sense the intensity and location of vibrations and has difficulty effectively distinguishing the type of vibration source (such as mechanical excavation, vehicle crushing, natural disturbance, etc.), resulting in a high false alarm rate. In addition, changes in ambient temperature can cause fiber optic refractive index drift and thermal expansion and contraction effects, causing the positioning reference to shift, which in turn affects the long-term monitoring accuracy.

[0005] For example, Chinese patent CN111912513B discloses a method for identifying excavator construction events along optical cables based on Φ-OTDR. This invention identifies excavator construction events by extracting the basic time-domain and frequency-domain features of the Φ-OTDR vibration signal and combining simple threshold judgment and pattern matching. However, it cannot separate the fiber refractive index drift and thermal expansion and contraction effects caused by changes in ambient temperature from the phase and amplitude changes of the optical signal caused by actual vibration, which easily leads to the continuous shift of the positioning reference with temperature changes. Summary of the Invention

[0006] The purpose of this invention is to provide a method and system for detecting optical cable vibration events based on multi-physics sensing, so as to solve the problem that existing Φ-OTDR equipment is difficult to effectively distinguish the type of vibration source in optical cable vibration event identification, resulting in a high false alarm rate.

[0007] The technical solution of the present invention is as follows: On one hand, the present invention provides a multi-physics sensing method for detecting optical cable vibration events, comprising the following steps: The backscattering Rayleigh signal in the optical fiber is acquired and processed using a Φ-OTDR device to obtain the physical location, timestamp, duration and waveform data of the vibration event on the optical fiber. IoT nodes are deployed at risk points along the optical cable route to collect raw data from the IoT nodes, including acoustic signals from the optical fiber and ambient temperature data. After identifying a vibration event, the original data of the IoT node is screened in a spatiotemporal dimension. The acoustic signals in the screened original data of the IoT node are extracted in the time domain. The extracted time domain features are input into a lightweight classification model, and the acoustic classification results are output. The acoustic classification results include the vibration source type and its confidence level. Energy intensity is obtained by performing energy calculations on waveform data, and the temperature change rate is calculated based on ambient temperature data. Finally, the vibration event is determined by combining the energy intensity, temperature change rate, and vibration source type.

[0008] Preferably, the spatiotemporal collaborative filtering of the raw data from the IoT nodes after identifying the vibration event specifically involves: After a vibration event is detected, the nearest IoT node to the vibration event is queried and a preset number of IoT nodes are retained. The physical location of the vibration event on the optical fiber is converted into geographic location coordinates. The geographic location coordinates are compared and matched with the locations of the retained IoT nodes, and the original data of IoT nodes that meet the preset distance requirement threshold are retained. The system acquires raw IoT node data within a preset time window from the retained raw IoT node data and aligns the occurrence time of the local data carried in the acquired raw IoT node data with the timestamp of the vibration event in the Φ-OTDR device.

[0009] Preferably, the extraction of time-domain features from the acoustic signals in the filtered raw data of IoT nodes specifically involves: The acoustic signal is preprocessed, and the preprocessed acoustic signal is segmented into frames. A short-time Fourier transform (STFT) is performed on each frame of the signal to obtain a complex time-frequency representation. The power spectral density is then calculated to obtain a time-frequency energy spectrum on a linear frequency scale. The time-frequency energy spectrum at the linear frequency scale is mapped to the Mel frequency scale using a Mel filter bank to generate a Mel spectrum. Taking the logarithm of the Mel spectrogram yields the logarithmic Mel spectrogram.

[0010] Preferably, in the lightweight classification model that inputs the extracted time-domain features, the output vibration source type and its confidence level are specifically as follows: The log-Mel spectrogram is first cropped or interpolated in the time and frequency dimensions to obtain a fixed-size two-dimensional array, with rows corresponding to time frames and columns corresponding to Mel frequency channels, forming a standardized grayscale image form of input features that are then processed in the model. The lightweight classification model performs forward inference on the input features and outputs the predicted probability distribution of the vibration source category. It selects the category with the highest probability value in the predicted probability distribution as the final identification result and outputs the highest probability value as the confidence level corresponding to the vibration source type.

[0011] Preferably, the vibration source types include high-risk construction types, vehicles, natural phenomena, and organisms.

[0012] Preferably, the step of calculating the energy intensity from the waveform data, calculating the temperature change rate based on the ambient temperature data, and combining the energy intensity, temperature change rate, and vibration source type to make the final determination of the vibration event specifically involves: First, determine whether the energy intensity of the vibration event is lower than a preset low energy threshold. If it is lower than the low energy threshold, the vibration event is determined to be an invalid event. Secondly, check whether the current temperature change rate exceeds the set temperature threshold, and at the same time determine whether the confidence level in the acoustic classification result is lower than the preset first confidence level threshold. If both conditions are met, the vibration event is judged as a phase drift pseudo-event caused by thermal expansion and contraction. If it is not filtered by the above rules, then the following judgment is made: If the vibration source type is identified as a high-risk construction type and the confidence level in the acoustic classification result is higher than the second confidence level, it is judged as a high-risk construction behavior. If the vibration source type is identified as a vehicle and the confidence level in the acoustic classification result is higher than the second confidence level, then query whether the physical location of the vibration event on the optical fiber is located in a known traffic area; if yes, it is determined to be normal traffic disturbance; if no, it is determined to be abnormal vehicle activity. If the vibration source type is identified as a natural phenomenon and the confidence level in the acoustic classification result is higher than the second confidence level, and the duration of the vibration event in the Φ-OTDR device exceeds the preset time threshold, it is determined to be a natural environmental disturbance. If the vibration source is identified as biological and the confidence level in the acoustic classification result is higher than the second confidence level, and the energy intensity of the vibration event in the Φ-OTDR device is lower than the preset medium threshold, it is considered as harmless biological activity. For vibration events in the acoustic classification results with a confidence level higher than the first confidence threshold but lower than the second confidence threshold, a fallback judgment logic is activated: if the temperature change rate is still higher than the set temperature threshold at this time, it is attributed to an uncertain event caused by thermal interference; otherwise, manual verification is performed.

[0013] On the other hand, the present invention provides a multi-physics sensing optical cable vibration event detection system, including a Φ-OTDR acquisition module, an IoT node raw data acquisition module, an acoustic classification module, and a vibration event final determination module; The Φ-OTDR acquisition module is used to acquire and process backscattered Rayleigh signals in optical fibers based on Φ-OTDR devices, and obtain the physical location, timestamp, duration and waveform data of vibration events on the optical fiber. The IoT node raw data acquisition module is used to deploy IoT nodes at risk points in the optical cable route to collect raw data from the IoT nodes. The raw data from the IoT nodes includes the acoustic signal of the optical fiber and the ambient temperature data. The acoustic classification module is used to perform spatiotemporal collaborative filtering of the raw data of IoT nodes after identifying vibration events. It extracts time-domain features from the acoustic signals in the filtered raw data of IoT nodes, inputs the extracted time-domain features into a lightweight classification model, and outputs acoustic classification results, which include the vibration source type and its confidence level. The final vibration event determination module is used to calculate the energy intensity from the waveform data, calculate the temperature change rate based on the ambient temperature data, and combine the energy intensity, temperature change rate and vibration source type to make the final determination of the vibration event.

[0014] In another aspect, the present invention also provides an electronic device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method as described in any embodiment of the present invention.

[0015] In another aspect, the present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in any embodiment of the present invention.

[0016] Compared with the prior art, the present invention has the following technical effects: This invention effectively distinguishes between dangerous construction activities (such as excavators and pile drivers) and harmless disturbances (such as rainfall, vehicles, and animal activities) by integrating Φ-OTDR vibration signals and IoT acoustic sensor data. By using the rate of temperature change as a thermal disturbance discrimination index, it can automatically identify false events caused by thermal expansion and contraction. Through a multi-level, multi-physical field collaborative decision-making mechanism, it can effectively suppress false alarms caused by temperature drift, natural environment, and biological activities while ensuring a high detection rate for high-risk construction events, thus significantly improving the intelligence and reliability of optical cable safety monitoring. Attached Figure Description

[0017] Figure 1This is an overall flowchart of the optical cable vibration event detection method using multi-physics sensing described in this invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below in conjunction with specific embodiments of the present application and with reference to the accompanying drawings.

[0019] Example 1 This embodiment provides a multi-physics sensing method for detecting optical cable vibration events. By efficiently aligning the distributed vibration information of Φ-OTDR with local IoT sensing data in the spatiotemporal dimension, it can integrate distributed fiber optic sensing and IoT multi-physics sensing. This leverages the advantages of Φ-OTDR's full-line, high-precision positioning while utilizing auxiliary information such as acoustics and temperature to identify vibration source types and compensate for environmental interference, thereby improving the intelligence level and operational reliability of underground optical cable status sensing. (See reference...) Figure 1 As shown, it includes the following steps: The Φ-OTDR device collects and processes backscattered Rayleigh signals in the optical fiber to obtain the physical location, timestamp, duration, and waveform data of the vibration event on the optical fiber. Specifically, the Φ-OTDR device is deployed by connecting the Φ-OTDR host to the optical fiber and deploying it in the optical cable central office or aggregation node, and connecting it to the spare fiber core (usually 1-2 cores) of the optical cable to be monitored via a pigtail.

[0020] At the start of identification, the Φ-OTDR device continuously collects backscattered Rayleigh signals in the optical fiber, processes and analyzes these signals to identify vibration events, and assigns a unique vibration event identifier. This identifier, which can be automatically generated by the system, is used to track and associate subsequent processing flows, ensuring that the event is traceable and non-repeating across multiple systems. The physical location of the vibration event on the optical fiber represents the cumulative distance from the Φ-OTDR device's transmitter (usually the central office) along the fiber to the event occurrence point. The timestamp of the vibration event is used to ensure data alignment with IoT nodes and other subsystems under a unified time reference. The duration of the vibration event reflects the length of time the disturbance affects the optical cable, helping to distinguish between instantaneous impacts (such as knocking, typically <100ms) and continuous operations (such as digging, typically >1000ms). The waveform data of a vibration event (raw or pre-processed) is usually a one-dimensional time series array, recording the dynamic process of fiber phase / intensity changes during the event. When the location of the vibration event is determined, the obtained waveform data is a one-dimensional signal value that changes with time. The preprocessing of the waveform data includes filtering, denoising, demodulation, normalization, and continuously recording the intensity of the scattered signal returned by each laser pulse according to the sampling rate to form a two-dimensional data matrix (spatial location + time). The raw waveform data or the pre-processed waveform data can be recorded as needed.

[0021] IoT nodes are deployed at risk points along the fiber optic cable route to collect raw data, including acoustic signals from the optical fiber and ambient temperature data. Specifically, IoT nodes are deployed at key risk points along the fiber optic cable route based on the actual scenario. Each IoT node integrates acoustic and temperature sensors. The acoustic sensors are used to collect characteristic acoustic signatures from construction machinery, excavation, and impacts, while the temperature sensors collect ambient temperature data along the fiber optic route. The IoT node data can be aggregated to an edge gateway, which is deployed in the central office equipment room.

[0022] After identifying a vibration event, the original data of the IoT node is screened in a spatiotemporal dimension. The acoustic signals in the screened original data of the IoT node are extracted in the time domain. The extracted time domain features are input into a lightweight classification model, and the acoustic classification results are output. The acoustic classification results include the vibration source type and its confidence level.

[0023] In a preferred embodiment of this practice, the spatiotemporal collaborative filtering of the raw data of the IoT node after identifying the vibration event specifically involves: Upon detecting a vibration event, the system queries the nearest IoT node to the event and retains a preset number (one or more depending on the situation). The physical location of the vibration event on the optical fiber is converted into geographic coordinates. These geographic coordinates are then compared with the locations of the retained IoT nodes. The original data of IoT nodes that meet a preset distance threshold are retained; otherwise, the corresponding original data is considered invalid. The preset distance threshold is determined based on the actual application scenario requirements, such as being set to ≤50m.

[0024] The system acquires raw IoT node data within a preset time window from the retained raw IoT node data, and aligns the occurrence time of the local data carried in the acquired raw IoT node data with the timestamp of the vibration event in the Φ-OTDR device. The preset time window is determined according to the actual application scenario requirements, such as being set to [t]. 2s,t+2s).

[0025] In a preferred embodiment of this invention, the extraction of time-domain features from the acoustic signals in the filtered raw data of IoT nodes specifically involves: The acoustic signal is preprocessed (the purpose of which is to reduce noise and normalize the signal amplitude of the original acoustic signal so that the preprocessed data can be better extracted for features). The preprocessed acoustic signal is then framed (the frame length can be set to N sampling points and the frame shift to M sampling points so that there is partial overlap between adjacent frames). A short-time Fourier transform (STFT) is performed on each frame signal to obtain a complex time-frequency representation, and the power spectral density is calculated to obtain a time-frequency energy spectrum at a linear frequency scale.

[0026] The time-frequency energy spectrum at the linear frequency scale is mapped to the Mel frequency scale using a Mel filter bank to generate a Mel spectrum.

[0027] Taking the logarithm of the Mel spectrogram yields the logarithmic Mel spectrogram, which aims to compress the dynamic range and enhance the distinguishability of low-frequency components.

[0028] In a preferred embodiment of this practice, the extraction of time-domain features is input into a lightweight classification model, and the output vibration source type and its confidence level are specifically as follows: The log-Mel spectrogram is first cropped or interpolated in the time and frequency dimensions to obtain a fixed-size two-dimensional array, with rows corresponding to time frames and columns corresponding to Mel frequency channels, forming a standardized grayscale image form of input features that are then processed in the input model.

[0029] The lightweight classification model performs forward inference on the input features and outputs a predicted probability distribution of vibration source categories. The probability value for each category is between 0 and 1, and the sum of the probabilities of all categories is 1. The category with the highest probability value in the predicted probability distribution is selected as the final identification result, and the highest probability value is output as the confidence level corresponding to the vibration source type. Specifically, the lightweight classification model can be trained using models such as convolutional neural networks for identification.

[0030] As a preferred embodiment of this invention, the vibration source types include high-risk construction types, vehicles, natural phenomena, and organisms.

[0031] Energy intensity is obtained by performing energy calculations on waveform data, and the temperature change rate is calculated based on ambient temperature data. Finally, the vibration event is determined by combining the energy intensity, temperature change rate, and vibration source type.

[0032] Specifically, the temperature change rate is calculated by extracting ambient temperature data from a temperature sensor within a certain time interval from the current time. The time interval can be set according to actual application requirements, such as setting the time interval to 1 hour. The following calculation of the temperature change rate is based on 1 hour.

[0033] In the formula, The ambient temperature measured at the current time t (in seconds) is expressed in degrees Celsius (°C). 1 hour ago (i.e., t) The temperature measured in 3600 seconds. This represents the difference between the two, i.e., the change in temperature within one hour, expressed in °C. The time interval is defined as follows.

[0034] In a preferred embodiment of this invention, the step of calculating the energy intensity from the waveform data, calculating the temperature change rate based on the ambient temperature data, and finally determining the vibration event by combining the energy intensity, temperature change rate, and vibration source type is as follows: First, determine whether the energy intensity of the vibration event is lower than the preset low energy threshold. If it is lower than the low energy threshold, the vibration event is considered to be caused by weak disturbances such as environmental noise, small animal activity or slight trampling by people, and the vibration event is judged as an invalid event.

[0035] If the energy intensity of the vibration event is higher than the preset low energy threshold, check whether the current temperature change rate exceeds the set temperature threshold, and at the same time determine whether the confidence level in the acoustic classification result is lower than the preset first confidence level threshold. If both conditions are met, it indicates that there is a significant ambient temperature disturbance and no clear acoustic event source is detected. In this case, the vibration event is determined to be a phase drift pseudo-event caused by thermal expansion and contraction.

[0036] If the vibration event is not filtered by the above rules, that is, if the energy intensity of the vibration event is higher than the preset low energy threshold, and the temperature change rate is lower than the set temperature threshold, or the confidence level of the highest category in the acoustic classification result is higher than the preset first confidence threshold, then the following judgment is made: If the vibration source type is identified as a high-risk construction type such as "excavator", "piling machine" or "breaker" and the confidence level in the acoustic classification result is higher than the second confidence level, it is judged as a high-risk construction behavior.

[0037] If the vibration source type is identified as a vehicle and the confidence level in the acoustic classification result is higher than the second confidence level, then query whether the physical location of the vibration event on the optical fiber is located in a known traffic area (such as along a main road or highway); if yes, it is determined to be a normal traffic disturbance and the alarm is ignored; if no, it is determined to be abnormal vehicle activity.

[0038] If the vibration source is identified as a natural phenomenon such as "rain", "strong wind" or "thunder" and the confidence level in the acoustic classification result is higher than the second confidence level, and the duration of the vibration event in the Φ-OTDR device exceeds the preset time threshold, it is determined to be a natural environmental disturbance and is filtered out.

[0039] If the vibration source type is identified as an "animal" or "human" organism and the confidence level in the acoustic classification result is higher than the second confidence level, and the energy intensity of the vibration event in the Φ-OTDR device is lower than the preset medium threshold, it is considered as harmless biological activity.

[0040] For vibration events in the acoustic classification results with a confidence level higher than the first confidence threshold but lower than the second confidence threshold, a fallback judgment logic is activated: if the temperature change rate is still higher than the set temperature threshold at this time, it is attributed to an uncertain event caused by thermal interference; otherwise, it is judged as a lack of sufficient discrimination basis and manual review is performed.

[0041] Example 2 Accordingly, this embodiment provides a multi-physics sensing optical cable vibration event detection system to implement the method described in any embodiment of the present invention, including a Φ-OTDR acquisition module, an IoT node raw data acquisition module, an acoustic classification module, and a vibration event final determination module.

[0042] The Φ-OTDR acquisition module is used to acquire and process backscattered Rayleigh signals in optical fibers based on Φ-OTDR devices, and obtain the physical location, timestamp, duration, and waveform data of vibration events on the optical fiber.

[0043] The IoT node raw data acquisition module is used to deploy IoT nodes at risk points in the optical cable route to collect raw data from the IoT nodes. The raw data from the IoT nodes includes the acoustic signal of the optical fiber and the ambient temperature data.

[0044] The acoustic classification module is used to perform spatiotemporal collaborative filtering of the raw data of IoT nodes after identifying vibration events. It extracts temporal features from the acoustic signals in the filtered raw data of IoT nodes, inputs the extracted temporal features into a lightweight classification model, and outputs acoustic classification results, which include the vibration source type and its confidence level.

[0045] The final vibration event determination module is used to calculate the energy intensity from the waveform data, calculate the temperature change rate based on the ambient temperature data, and combine the energy intensity, temperature change rate and vibration source type to make the final determination of the vibration event.

[0046] Example 3 This embodiment provides an electronic device, which includes: a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the method described in any embodiment of the present invention.

[0047] Example 4 This embodiment provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in any embodiment of the present invention.

[0048] In this application embodiment, "at least one" refers to one or more, and "more than one" refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent the existence of A alone, A and B simultaneously, or B alone. A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one of the following" and similar expressions refer to any combination of these items, including any combination of singular or plural items. For example, at least one of a, b, and c can represent: a, b, c, a and b, a and c, b and c, or a and b and c, where a, b, and c can be single or multiple.

[0049] Those skilled in the art will recognize that the units and algorithm steps described in the embodiments disclosed herein can be implemented using electronic hardware, computer software, or a combination of electronic hardware and software. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0050] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0051] In the several embodiments provided in this application, any function, if implemented as a software functional unit and sold or used as an independent product, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0052] The above description is merely an embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A method for detecting optical cable vibration events using multi-physics sensing, characterized in that, Includes the following steps: The backscattering Rayleigh signal in the optical fiber is acquired and processed using a Φ-OTDR device to obtain the physical location, timestamp, duration and waveform data of the vibration event on the optical fiber. IoT nodes are deployed at risk points along the optical cable route to collect raw data from the IoT nodes, including acoustic signals from the optical fiber and ambient temperature data. After identifying a vibration event, the original data of the IoT node is screened in a spatiotemporal dimension. The acoustic signals in the screened original data of the IoT node are extracted in the time domain. The extracted time domain features are input into a lightweight classification model, and the acoustic classification results are output. The acoustic classification results include the vibration source type and its confidence level. Energy intensity is obtained by performing energy calculations on waveform data, and the temperature change rate is calculated based on ambient temperature data. Finally, the vibration event is determined by combining the energy intensity, temperature change rate, and vibration source type.

2. The method for detecting optical cable vibration events using multi-physics sensing according to claim 1, characterized in that, The specific steps for performing spatiotemporal collaborative filtering of the raw data from IoT nodes after identifying vibration events are as follows: After a vibration event is detected, the nearest IoT node to the vibration event is queried and a preset number of IoT nodes are retained. The physical location of the vibration event on the optical fiber is converted into geographic location coordinates. The geographic location coordinates are compared and matched with the locations of the retained IoT nodes, and the original data of IoT nodes that meet the preset distance requirement threshold are retained. The system acquires raw IoT node data within a preset time window from the retained raw IoT node data and aligns the occurrence time of the local data carried in the acquired raw IoT node data with the timestamp of the vibration event in the Φ-OTDR device.

3. The method for detecting optical cable vibration events using multi-physics sensing according to claim 1, characterized in that, The extraction of time-domain features from the acoustic signals in the filtered raw data of IoT nodes specifically involves: The acoustic signal is preprocessed, and the preprocessed acoustic signal is segmented into frames. A short-time Fourier transform (STFT) is performed on each frame of the signal to obtain a complex time-frequency representation. The power spectral density is then calculated to obtain a time-frequency energy spectrum on a linear frequency scale. The time-frequency energy spectrum at the linear frequency scale is mapped to the Mel frequency scale using a Mel filter bank to generate a Mel spectrum. Taking the logarithm of the Mel spectrogram yields the logarithmic Mel spectrogram.

4. The method for detecting optical cable vibration events using multi-physics sensing according to claim 3, characterized in that, The extracted time-domain features are input into the lightweight classification model, and the output vibration source type and its confidence level are specifically as follows: The log-Mel spectrogram is first cropped or interpolated in the time and frequency dimensions to obtain a fixed-size two-dimensional array, with rows corresponding to time frames and columns corresponding to Mel frequency channels, forming a standardized grayscale image form of input features that are then processed in the model. The lightweight classification model performs forward inference on the input features and outputs the predicted probability distribution of the vibration source category. It selects the category with the highest probability value in the predicted probability distribution as the final identification result and outputs the highest probability value as the confidence level corresponding to the vibration source type.

5. The method for detecting optical cable vibration events using multi-physics sensing according to claim 1, characterized in that, The vibration sources include high-risk construction activities, vehicles, natural phenomena, and organisms.

6. The multi-physics sensing optical cable vibration event detection method according to claim 5, characterized in that, The process of calculating energy intensity from waveform data, calculating temperature change rate based on ambient temperature data, and combining energy intensity, temperature change rate, and vibration source type to make the final determination of the vibration event is as follows: First, determine whether the energy intensity of the vibration event is lower than a preset low energy threshold. If it is lower than the low energy threshold, the vibration event is determined to be an invalid event. Secondly, check whether the current temperature change rate exceeds the set temperature threshold, and at the same time determine whether the confidence level in the acoustic classification result is lower than the preset first confidence level threshold. If both conditions are met, the vibration event is judged as a phase drift pseudo-event caused by thermal expansion and contraction. If it is not filtered by the above rules, then the following judgment is made: If the vibration source type is identified as a high-risk construction type and the confidence level in the acoustic classification result is higher than the second confidence level, it is judged as a high-risk construction behavior. If the vibration source type is identified as a vehicle and the confidence level in the acoustic classification result is higher than the second confidence level, then query whether the physical location of the vibration event on the optical fiber is located in a known traffic area; if yes, it is determined to be normal traffic disturbance; if no, it is determined to be abnormal vehicle activity. If the vibration source type is identified as a natural phenomenon and the confidence level in the acoustic classification result is higher than the second confidence level, and the duration of the vibration event in the Φ-OTDR device exceeds the preset time threshold, it is determined to be a natural environmental disturbance. If the vibration source is identified as biological and the confidence level in the acoustic classification result is higher than the second confidence level, and the energy intensity of the vibration event in the Φ-OTDR device is lower than the preset medium threshold, it is considered as harmless biological activity. For vibration events in the acoustic classification results with a confidence level higher than the first confidence threshold but lower than the second confidence threshold, a fallback judgment logic is activated: if the temperature change rate is still higher than the set temperature threshold at this time, it is attributed to an uncertain event caused by thermal interference; otherwise, manual verification is performed.

7. A multi-physics sensing optical cable vibration event detection system, characterized in that, The system is used to implement the method as described in any one of claims 1 to 6, and includes a Φ-OTDR acquisition module, an IoT node raw data acquisition module, an acoustic classification module, and a vibration event final determination module; The Φ-OTDR acquisition module is used to acquire and process backscattered Rayleigh signals in optical fibers based on Φ-OTDR devices, and obtain the physical location, timestamp, duration and waveform data of vibration events on the optical fiber. The IoT node raw data acquisition module is used to deploy IoT nodes at risk points in the optical cable route to collect raw data from the IoT nodes. The raw data from the IoT nodes includes the acoustic signal of the optical fiber and the ambient temperature data. The acoustic classification module is used to perform spatiotemporal collaborative filtering of the raw data of IoT nodes after identifying vibration events. It extracts time-domain features from the acoustic signals in the filtered raw data of IoT nodes, inputs the extracted time-domain features into a lightweight classification model, and outputs acoustic classification results, which include the vibration source type and its confidence level. The final vibration event determination module is used to calculate the energy intensity from the waveform data, calculate the temperature change rate based on the ambient temperature data, and combine the energy intensity, temperature change rate and vibration source type to make the final determination of the vibration event.

8. An electronic device, the electronic device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor, when executing the computer program, implements the method as described in any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 6.

Images