Fiber-based external break event identification method, system, device, and storage medium

Abstract

The application relates to the field of optical fiber sensing, and more particularly to an optical fiber-based external breakage event identification method, system, device and storage medium. The method comprises the following steps: based on a preset target frequency band and frequency domain data of each channel in an optical fiber, obtaining a heat map corresponding to each target frequency band; the target frequency band is preset with a corresponding candidate external breakage event type; the optical fiber is divided into a plurality of channels at a preset distance length; according to the heat map corresponding to the target frequency band, a corresponding target channel is obtained; based on the target channel, corresponding channel information is obtained; based on the channel information, time domain data of the target channel in the optical fiber is obtained; based on the time domain data of the target channel and the preset candidate external breakage event type in the target frequency band, voiceprint information corresponding to the target channel is obtained; and based on the voiceprint information, a target external breakage event type of the target channel is identified. The application is used for realizing accurate and low false alarm optical fiber external breakage event identification.

Description

Technical Field

[0001] This invention relates to the field of fiber optic sensing, and more specifically, to a method, system, device, and storage medium for identifying external damage events based on optical fibers. Background Technology

[0002] Distributed Acoustic Sensing (DAS) technology, based on the principle of phase-sensitive optical time-domain reflectometry (Φ-OTDR), utilizes the interference effect of optical fiber on backscattered Rayleigh light to perform continuous, high-resolution distributed quantitative measurements of external vibrations or acoustic signals. Currently, DAS technology is widely used in engineering scenarios such as external damage monitoring of long-distance optical cable lines, perimeter intrusion alarms, and construction disturbance identification. In actual deployments, the area surrounding optical fiber lines often experiences multiple complex background disturbances simultaneously, including vehicle traffic, personnel activity, wind and rain conditions, mechanical vibrations, and ground construction. This results in a wide variety of signal types received by DAS devices, placing significant pressure on the identification of external damage events and subsequent alarms. Existing common event type identification methods identify signals in a specific spatial channel or area, triggering an alarm when the signal exceeds a preset threshold. While simple to implement, these methods are prone to triggering numerous false alarms because environmental noise, short-term random disturbances, and non-destructive construction activities can all cause occasional increases in signal amplitude. Especially in long-distance, large-scale monitoring scenarios, relying solely on simple threshold strategies is insufficient to balance sensitivity and accuracy: setting the threshold too low will lead to frequent false alarms and reduce operational efficiency; setting the threshold too high may result in the missed detection of truly destructive external damage events such as excavation, drilling, cutting, and mechanical proximity. Therefore, how to fully utilize DAS equipment to achieve accurate identification of external damage events with low false alarm rates has become a pressing technical problem to be solved in current engineering applications. Summary of the Invention

[0003] This invention provides a method, system, device, and storage medium for identifying external damage events based on optical fibers, which can achieve accurate identification of external damage events with low false alarms.

[0004] According to a first aspect of this application, a method for identifying external damage events based on optical fibers is provided, the method comprising: Based on the preset target frequency bands and the frequency domain data of each channel in the optical fiber, a heat map corresponding to each target frequency band is obtained; the target frequency bands are preset with corresponding candidate external damage event types; the optical fiber is divided into several channels with a preset distance length; Based on the heat map corresponding to the target frequency band, obtain the corresponding target channel; Based on the target channel, obtain the corresponding channel information; based on the channel information, obtain the time domain data of the target channel in the optical fiber; based on the time domain data of the target channel and the preset candidate external damage event types in the target frequency band, obtain the acoustic print information corresponding to the target channel. The target channel's external damage event type is identified based on the voiceprint information.

[0005] Understandably, the approach first rapidly identifies the target channel and corresponding candidate external damage event types based on the heatmap of the target frequency band. Then, it adaptively extracts the acoustic signature information of the target channel based on the candidate external damage event types for secondary fine identification. This effectively integrates spatiotemporal and frequency multi-dimensional features, significantly improving the accuracy of external damage event identification and reducing the false alarm rate. Simultaneously, the two-stage progressive identification architecture greatly reduces the amount of data required for subsequent fine analysis, balancing real-time performance and engineering usability in long-distance monitoring scenarios.

[0006] Optionally, obtaining a heatmap corresponding to each target frequency band based on the preset target frequency band and the frequency domain data of each channel in the optical fiber includes: Based on a preset time window, the time-domain data of each channel in the optical fiber is subjected to Fourier transform processing to obtain the frequency-domain data of each channel; wherein, the optical fiber has time-domain data of several channels at several times. The frequency domain data of each channel in the optical fiber is processed by energy integration of the target frequency band to obtain a heat map corresponding to each target frequency band.

[0007] Understandably, by performing short-time Fourier transform on the time-domain data of each channel and integrating the energy over the target frequency band, the time-domain data is transformed into a heatmap reflecting the spatiotemporal distribution of energy in a specific frequency band, thereby effectively suppressing environmental noise and random disturbances. The generated heatmap can intuitively present the energy diffusion and persistence characteristics of the event in the time and distance dimensions, providing a reliable feature basis for quickly locating candidate anomaly regions and initially screening candidate external damage event types.

[0008] Optionally, obtaining the corresponding target channel based on the heatmap corresponding to the target frequency band includes: The heat map corresponding to the target frequency band is input into a preset recognition model for recognition to obtain candidate anomaly regions; wherein the candidate anomaly regions include the candidate channel range where the anomaly occurs; The target channel is determined based on the range of candidate channels.

[0009] Understandably, by inputting the heatmap into the preset recognition model, candidate anomaly regions and their corresponding candidate channel ranges can be automatically and quickly delineated, avoiding computational redundancy caused by full channel traversal. Further determination of the target channel based on the candidate channel range significantly narrows the data range for subsequent voiceprint analysis, effectively improving overall processing efficiency while maintaining detection recall. This lays a crucial spatial positioning foundation for achieving real-time and accurate external damage event recognition in long-distance fiber optic scenarios.

[0010] Optionally, determining the target channel based on the candidate channel range includes: Obtain the candidate channels included in the candidate channel range, and the time-domain data of the candidate channels in the optical fiber; The evaluation metrics for the time-domain data corresponding to the candidate channels are obtained; the evaluation metrics include at least one of the following: energy intensity index, signal-to-noise ratio index, correlation index, and persistence index. Based on the evaluation metrics, obtain the comprehensive score of the candidate channel; The candidate channel with the highest overall score will be selected as the target channel.

[0011] Understandably, by comprehensively evaluating candidate channels and obtaining evaluation indicators, the target channel with the best signal quality is objectively and quantitatively selected from the candidate anomaly regions. Determining the target channel based on the highest comprehensive score ensures that the time-domain data used for subsequent speaker recognition has the strongest effective signal and the highest reliability, thereby improving the accuracy and robustness of secondary recognition.

[0012] Optionally, the candidate external rupture event types include one or more of construction event types, impact event types, and leakage event types; The voiceprint information includes one or more of the following: continuous energy ratio information of a preset first frequency band, instantaneous kurtosis information of a preset second frequency band, and stable bandwidth index information of a preset third frequency band. The process of obtaining the voiceprint information corresponding to the target channel based on the time-domain data of the target channel and the preset candidate external damage event types in the target frequency band includes: If the preset candidate external damage event type in the target frequency band is a construction event type, then the continuous energy ratio information of the first frequency band of the target channel is obtained based on the time domain data of the target channel; If the preset candidate external damage event type in the target frequency band is a knocking event type, then the instantaneous kurtosis information of the second frequency band of the target channel is obtained based on the time domain data of the target channel; If the preset candidate external damage event type in the target frequency band is a leakage event type, then the stable bandwidth index information of the third frequency band of the target channel is obtained based on the time domain data of the target channel.

[0013] Understandably, by adaptively extracting corresponding differentiated voiceprint information from the target channel's time-domain data based on the candidate external damage event types output by the first-level identification, computational redundancy caused by extracting all voiceprint information is avoided, significantly improving the processing efficiency of the second-level identification. Simultaneously, the voiceprint information, which highly matches the physical generation mechanism of each type of external damage event, enhances the neural network's ability to distinguish event types, thereby effectively improving the accuracy of external damage event classification and providing a more targeted basis for subsequent false alarm verification.

[0014] Optionally, identifying the target channel's external damage event type based on the voiceprint information includes: The voiceprint information is input into a preset neural network for processing to determine whether the target channel has experienced an external damage event of the corresponding candidate external damage event type; if so, the corresponding candidate external damage event type is taken as the target external damage event type of the target channel; otherwise, it is determined that the target channel has a misjudgment.

[0015] Understandably, by inputting adaptively extracted voiceprint information into a neural network for binary classification verification, the authenticity of candidate external damage event types identified by the primary heatmap can be determined, effectively filtering false alarms caused by environmental interference or similar textures. If the neural network confirms the event has occurred, it outputs the target external damage event type, achieving high-confidence identification; if it determines it is a misjudgment, it promptly suppresses erroneous alarms. This significantly improves the overall accuracy and reliability of external damage event identification and reduces the cost of manual review.

[0016] Optionally, the method further includes: If the corresponding candidate external damage event type is the target external damage event type of the target channel, the neural network outputs the target external damage event type corresponding to the target channel, as well as one or more of the alarm location, alarm time, and vertical distance judgment results of the external damage event of the target external damage event type. The vertical distance determination result is calculated based on the frequency attenuation characteristics and amplitude attenuation characteristics obtained from the time-domain data of the target channel in the optical fiber.

[0017] Understandably, after confirming that the candidate external damage event type is the target event type, the neural network synchronously outputs the target external damage event type, alarm location, alarm time, and vertical distance judgment result calculated based on frequency attenuation and amplitude attenuation characteristics, providing maintenance personnel with multi-dimensional and accurate information. In particular, the vertical distance estimation effectively reflects the vertical deviation of the event source relative to the optical fiber, helping to determine whether the external damage directly threatens the safety of the optical cable, thus improving the targeting and efficiency of on-site handling. Overall, without increasing additional hardware costs, the alarm information content is significantly enriched, enhancing the practical value of the project.

[0018] According to a second aspect of this application, an optical fiber-based external damage event identification system is provided, the system comprising: A heatmap generation module is used to obtain a heatmap corresponding to each target frequency band based on the frequency domain data of each channel in the optical fiber and the preset target frequency band; the target frequency band has a preset candidate external damage event type; the optical fiber is divided into several channels with a preset distance length; A primary identification module is used to obtain the corresponding target channel based on the heat map corresponding to the target frequency band; The voiceprint information acquisition module is used to acquire corresponding channel information based on the target channel; acquire time-domain data of the target channel in the optical fiber based on the channel information; and acquire voiceprint information corresponding to the target channel based on the time-domain data of the target channel and preset candidate external damage event types in the target frequency band. The secondary identification module is used to identify the target external damage event type of the target channel based on the voiceprint information.

[0019] According to a third aspect of this application, an electronic device is provided, comprising: Memory, used to store one or more computer programs; A processor, when the one or more computer programs are executed by the processor, implements the fiber-optic-based external damage event identification method described in the first aspect above.

[0020] According to a fourth aspect of this application, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions, the computer instructions being configured to cause a processor to execute and implement the fiber optic-based external damage event identification method described in the first aspect above.

[0021] Based on any of the above aspects, the fiber optic-based external damage event identification method, system, device, and storage medium provided in this application embodiment can achieve the following technical effects: Based on the spatiotemporal joint constraints of the target frequency band heatmap, high-precision initial screening and identification are achieved: First, short-time Fourier transform is performed on the multi-channel time-domain data collected by the DAS device to obtain the frequency-domain data of each channel, and energy integration is performed on the preset target frequency band to generate the time... Distance heatmap. This heatmap is not merely for visualization; it transforms high-dimensional time-series information from frequency domain data, expanded by distance and time, into a spatiotemporal joint representation that simultaneously expresses the continuous propagation trajectory, energy diffusion range, and duration of the same event across multiple channels. By inputting the heatmap into a pre-defined identification model, candidate anomaly regions can be quickly delineated during the initial screening stage. This fully exploits the joint characteristics of the signal across time, distance, and frequency dimensions, avoiding misjudgments of instantaneous random disturbances by traditional single-channel thresholding methods, and significantly improving the sensitivity and spatial positioning accuracy for weak, continuous external disturbances.

[0022] A two-tiered progressive identification architecture balances accuracy and a low false alarm rate: It employs a progressive identification architecture with a primary heatmap screening and a secondary voiceprint screening. The primary identification uses heatmaps to quickly narrow down abnormal areas, significantly reducing the amount of data required for subsequent detailed analysis, while filtering out a large number of false alarms triggered by non-destructive behaviors such as environmental noise and temporary vehicle passage. The secondary identification focuses only on the target channel output by the primary identification and its corresponding time-domain data. By extracting voiceprint features adaptively matching the candidate external damage event types and using neural networks for authenticity verification, it maintains high real-time processing performance in long-distance, large-scale monitoring scenarios while accurately identifying truly destructive external damage events, effectively reducing false alarm and missed alarm rates.

[0023] Adaptive voiceprint analysis based on candidate external damage event types improves recognition efficiency and classification capabilities: Based on the candidate external damage event types output by the first-level recognition, the corresponding voiceprint feature combinations are adaptively invoked. Compared to indiscriminately extracting the entire set of voiceprint features, this significantly reduces unnecessary computational overhead and improves the processing speed of the second-level recognition. Simultaneously, since the voiceprint information corresponding to each event type is highly correlated with its physical generation mechanism, adaptive selection achieves stronger category discrimination. This allows the neural network to not only verify candidate external damage event types but also further output auxiliary information such as vertical distance and alarm location. This enables the output of reliable target external damage event types, incorporating multiple information such as alarm location, time, and vertical distance, comprehensively enhancing its practicality and intelligence in engineering environments. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 This is a flowchart of a fiber optic-based external damage event identification method provided in this embodiment.

[0026] Figure 2 This is a flowchart for determining the target channel provided in this embodiment.

[0027] Figure 3 This is a schematic diagram of the functional modules of an optical fiber-based external damage event recognition system provided in this embodiment.

[0028] Figure 4 This is a schematic diagram of the structure of the electronic device provided in this embodiment. Detailed Implementation

[0029] The accompanying drawings are for illustrative purposes only and should not be construed as limiting the scope of this application. To better illustrate the following embodiments, some components in the drawings may be omitted, enlarged, or reduced, and do not represent the actual dimensions of the product; it is understandable to those skilled in the art that some well-known structures and their descriptions may be omitted in the drawings.

[0030] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0031] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0032] Existing methods for identifying external damage events in DAS devices based on simple thresholds are prone to triggering numerous false alarms due to environmental noise and non-destructive behaviors under complex background disturbances. Furthermore, they struggle to balance sensitivity and accuracy, leading to the easy underreporting of destructive external damage events. Therefore, there is an urgent need for a scheme that can fully utilize the multidimensional information from DAS devices to achieve accurate identification of external damage events with low false alarm rates.

[0033] This embodiment provides a technical solution that can solve the above problems. The specific implementation of this application will be described in detail below with reference to the accompanying drawings.

[0034] like Figure 1 As shown, this embodiment provides a method for identifying external damage events based on optical fibers. The method can be further divided into the following steps: S100. Based on the preset target frequency band and the frequency domain data of each channel in the optical fiber, obtain the heat map corresponding to each target frequency band; the target frequency band has a preset candidate external damage event type; the optical fiber is divided into several channels according to a preset distance length; In this embodiment, a target frequency band can be pre-set based on historical experience before identifying the type of external damage event. The target frequency band can be used to record the frequency bands corresponding to different candidate external damage event types in the fiber optic frequency domain data. At the same time, each target frequency band will correspond to a pre-set candidate external damage event type, thereby recording the correspondence between the target frequency band and the candidate external damage event type.

[0035] Specifically, the candidate external rupture event types include one or more of the following: construction event types, impact event types, and leakage event types.

[0036] In this embodiment, the construction-related event type is typically a continuous vibration event caused by mechanized or manual engineering construction activities such as excavation, drilling, cutting, crushing, and rolling. It is usually characterized by concentrated energy in the lower frequency band and has temporal continuity or staged evolution. The impact-related event type is an instantaneous vibration event caused by a tool or object impacting the ground, pipeline, or structure near the optical fiber in a single or intermittent manner. Its spectrum is characterized by a higher frequency band, rich components, high kurtosis, and short duration. The leakage-related event type refers to a continuous steady-state sound wave event generated when gas or liquid in a pipeline is ejected or flows out through a broken part. It usually has a stable energy bandwidth in the mid-to-high frequency band, a flat spectral envelope, and a long duration.

[0037] Preferably, a target frequency band of a Hz-b Hz can be preset to correspond to the construction event type based on the typical performance signal of the frequency domain data; a target frequency band of c Hz-d Hz can be preset to correspond to the impact event type based on the typical performance signal of the impact event type in the frequency domain data; and a target frequency band of e Hz-f Hz can be preset to correspond to the leakage event type based on the typical performance signal of the leakage event type in the frequency domain data.

[0038] In this embodiment, frequency domain data for each channel in the optical fiber is required as the basis for generating the heatmap. It is understood that when collecting time domain data from the optical fiber using a DAS device, the time domain data is collected based on the channels pre-divided by the DAS device. Preferably, the optical fiber can be divided into several channels of equal length, each channel containing a segment of optical fiber. Time domain data is collected for each channel, and then the time domain data of each channel is demodulated and processed to obtain the frequency domain data for each channel.

[0039] In this embodiment, based on the preset target frequency band and the frequency domain data of each channel in the optical fiber, a heatmap corresponding to each target frequency band is obtained, enabling the visualization of the joint high-dimensional information of fiber distance and time in the frequency domain data. It can be understood that one target frequency band corresponds to one heatmap, where the first coordinate of the heatmap is time, and the second coordinate is fiber distance. A certain fiber distance forms a corresponding channel, and the color intensity in the heatmap reflects the signal strength of the target frequency band at that fiber distance or channel at that time point. Generally, in a heatmap, the darker the color or the warmer it is according to the color scale, the stronger the energy; the lighter the color or the cooler it is according to the color scale, the weaker the energy.

[0040] The heat map corresponding to the target frequency band can reflect the signal strength of each fiber distance or channel in the target frequency band over time, which makes it possible to clearly identify the channel range and time range of the candidate external damage event type corresponding to the target frequency band, and provide a data foundation for subsequent accurate identification.

[0041] Specifically, the step of obtaining a heatmap corresponding to each target frequency band based on the preset target frequency band and the frequency domain data of each channel in the optical fiber includes: Based on a preset time window, the time-domain data of each channel in the optical fiber is subjected to Fourier transform processing to obtain the frequency-domain data of each channel; wherein, the optical fiber has time-domain data of several channels at several times. The frequency domain data of each channel in the optical fiber is processed by energy integration of the target frequency band to obtain a heat map corresponding to each target frequency band.

[0042] In this embodiment, the time-domain data of the optical fiber is acquired based on the DAS device, specifically the time-domain data of each channel at each moment. Therefore, the optical fiber has time-domain data for several channels at several moments. After dividing the time-domain data according to a preset time window, Fourier processing is performed on the divided time-domain data to obtain the frequency-domain data of each channel. Preferably, the Fourier processing can be a Short-Time Fourier Transform (STFT) to improve the accuracy of data processing. Subsequently, energy integration processing of the target frequency band is performed on the frequency-domain data of each channel to obtain the heatmap corresponding to each target frequency band.

[0043] S200. Obtain the corresponding target channel based on the heat map corresponding to the target frequency band; In this embodiment, a rough identification of the target channel is performed based on the heatmap corresponding to the target frequency band. This target channel can serve as the data basis for subsequent fine identification. It is understood that the heatmap can simultaneously express the continuous propagation trajectory, energy diffusion range, and duration of the same candidate external damage event type across multiple channels. Therefore, the candidate anomaly region can be narrowed down at the heatmap level first, and then a fine secondary identification can be performed at the original signal level.

[0044] Specifically, obtaining the corresponding target channel based on the heatmap corresponding to the target frequency band includes: The heat map corresponding to the target frequency band is input into a preset recognition model for recognition to obtain candidate anomaly regions; wherein the candidate anomaly regions include the candidate channel range where the anomaly occurs; The target channel is determined based on the range of candidate channels.

[0045] In this embodiment, the recognition model can be a YOLO series model, a Faster Region-based Convolutional Neural Network (Faster R-CNN), a Real-Time Detection Transformer (RT-DETR), or other image region extraction methods based on connected components. Preferably, in this embodiment, the recognition model can be a YOLOv8 model.

[0046] In this embodiment, the recognition model used is pre-trained and tested based on collected heatmap samples. It can be directly used for identifying candidate anomaly regions. The heatmap samples include positive and negative samples. Positive samples do not contain candidate anomaly regions, while negative samples contain them. When the heatmap is identified using the recognition model, if a candidate anomaly region is identified, the target channel needs to be determined within the candidate channel range of that region. It is then determined that the candidate channel range may contain a corresponding candidate external damage event type, requiring subsequent secondary precise identification. If no candidate anomaly region is identified, it indicates that no corresponding candidate external damage event type occurred at any location on the optical fiber within the time period shown in the heatmap; therefore, the secondary precise identification step is not required.

[0047] Specifically, such as Figure 2 As shown, determining the target channel based on the candidate channel range may include the following steps: S210. Obtain the candidate channels included in the candidate channel range, and the time domain data corresponding to the candidate channels in the optical fiber; In this embodiment, the candidate channel range may include several candidate channels. Therefore, it is necessary to obtain the time-domain data corresponding to each candidate channel in the optical fiber as the data basis for subsequently determining the target channel. In another embodiment, if the candidate channel range includes only one candidate channel, then the candidate channel is directly used as the target channel, reducing a large amount of redundant calculation.

[0048] S220. Obtain the evaluation index of the time domain data corresponding to the candidate channel; the evaluation index includes at least one of the following: energy intensity index value, signal-to-noise ratio index value, correlation index value, and persistence index value. In this embodiment, if the candidate channel range may include several candidate channels, it is necessary to obtain evaluation metrics for the time-domain data corresponding to the candidate channels. These evaluation metrics include at least one of energy intensity, signal-to-noise ratio, correlation, and persistence. Preferably, the energy intensity, signal-to-noise ratio, correlation, and persistence metrics of the time-domain data corresponding to the candidate channels can be obtained simultaneously, enabling more accurate determination of the target channel subsequently.

[0049] In this embodiment, the energy strength index is a quantified value of the signal energy calculated based on the time-domain data of the candidate channel, used to reflect the intensity level of the vibration of the candidate channel; a larger value indicates stronger vibration. The signal-to-noise ratio (SNR) index is the ratio of the effective signal energy to the background noise energy calculated based on the time-domain data of the candidate channel, used to measure signal quality; a higher value indicates a more prominent effective signal and stronger noise resistance in the candidate channel. The correlation index is a cross-correlation coefficient or similarity measure calculated based on the time-domain data of the candidate channel and the time-domain data of one or more adjacent channels, used to characterize the consistency between the candidate channel and surrounding channel signals; a higher value indicates continuous signal propagation and conformity to the spatial distribution characteristics of real events. The persistence index is the percentage of time or duration for which the signal energy calculated based on the time-domain data of the candidate channel exceeds a preset threshold, used to measure the continuous stability of the vibration event; a higher value indicates a longer event duration and better conformity to the evolutionary pattern of external damage behavior.

[0050] S230. Based on the evaluation indicators, obtain the comprehensive score of the candidate channel; In this embodiment, the comprehensive score of the candidate channel can be obtained based on energy intensity, signal-to-noise ratio, correlation, and persistence metrics. Each of these metrics has a corresponding weight. Specifically, the comprehensive score of the candidate channel... It can be: in, The time-domain data of the candidate channels. The preset weights for the energy intensity index values, The energy intensity index value of the time-domain data of the candidate channel; The preset weights for the signal-to-noise ratio index value. The signal-to-noise ratio (SNR) value of the time-domain data of the candidate channel; The preset weights for the relevance index values, The correlation index value of the time-domain data of the candidate channel; Preset weights for persistent indicator values, The value represents the persistence index of the time-domain data of the candidate channel.

[0051] S240. Select the candidate channel with the highest comprehensive score as the target channel.

[0052] In this embodiment, the comprehensive scores of all candidate channels are ranked, and the candidate channel with the highest comprehensive score is selected as the target channel. This identifies the channel that performs best in four aspects: energy intensity, signal-to-noise ratio, correlation with adjacent channels, and event persistence, thus best representing the core characteristics of the real vibration event in the candidate anomaly region. Using this candidate channel as the target channel for subsequent acoustic signature analysis maximizes the validity and reliability of the input time-domain data, avoiding secondary identification biases caused by selecting channels with weak signals, high noise, occasional disturbances, or spatial discontinuities, thereby improving the accuracy and robustness of event type determination.

[0053] S300. Obtain the corresponding channel information based on the target channel; obtain the time domain data of the target channel in the optical fiber based on the channel information; and obtain the voiceprint information corresponding to the target channel based on the time domain data of the target channel and the preset candidate external damage event types in the target frequency band. It is understood that the candidate anomaly region not only has a candidate channel range, but also a corresponding time range. By using the time range as the channel information of the target channel, the time domain data required by the target channel can be accurately obtained, the range of secondary identification can be narrowed, and the processing of redundant data can be reduced.

[0054] In this embodiment, the acquired target channel will have corresponding channel information, which can be a channel number and corresponding time information. The channel number and time information can serve as the data basis for subsequently acquiring the time-domain data of the target channel. Since the target frequency band corresponds one-to-one with the candidate external damage event type, after obtaining the corresponding target channel from the heatmap corresponding to the target frequency band, the corresponding candidate external damage event type can be obtained, thus serving as the data basis for acquiring the voiceprint information of the target channel.

[0055] In this embodiment, different candidate external damage event types exhibit different signal characteristics in the time-domain data, resulting in variations in their corresponding voiceprint information. Therefore, it is necessary to calculate the time-domain data of the target channel based on the candidate external damage event types corresponding to the target channel, extract the corresponding voiceprint information, and then use this voiceprint information to perform secondary, refined identification of the external damage event types for the target channel, thereby improving the reliability and accuracy of external damage event type identification.

[0056] Specifically, the voiceprint information includes one or more of the following: continuous energy ratio information of a preset first frequency band, instantaneous kurtosis information of a preset second frequency band, and stable bandwidth index information of a preset third frequency band. The process of obtaining the voiceprint information corresponding to the target channel based on the time-domain data of the target channel and the preset candidate external damage event types in the target frequency band includes: If the preset candidate external damage event type in the target frequency band is a construction event type, then the continuous energy ratio information of the first frequency band of the target channel is obtained based on the time domain data of the target channel; If the preset candidate external damage event type in the target frequency band is a knocking event type, then the instantaneous kurtosis information of the second frequency band of the target channel is obtained based on the time domain data of the target channel; If the preset candidate external damage event type in the target frequency band is a leakage event type, then the stable bandwidth index information of the third frequency band of the target channel is obtained based on the time domain data of the target channel.

[0057] In this embodiment, the acoustic fingerprint features of the time-domain data of the target channel are calculated to obtain the corresponding acoustic fingerprint information. It is understood that the acoustic fingerprint information is a set of feature parameters used to characterize the frequency-domain acoustic fingerprint properties of external damage events, reflecting the unique distribution patterns of different types of external damage events in the spectrum.

[0058] In this embodiment, the duration and pattern of the effect on the optical fiber differ for each candidate external damage event type, resulting in variations in the corresponding acoustic signature information. Therefore, acoustic signature information for key analysis can be obtained based on preset candidate external damage event types in the target frequency band, enabling a more accurate determination of whether the candidate external damage event type has actually occurred in the target channel.

[0059] For construction-related events, these typically manifest as concentrated energy in lower frequency bands and exhibit temporal continuity or phased evolution. Therefore, when the preset candidate external damage event type in the target frequency band corresponding to the target channel is a construction-related event, the sustained energy ratio information of the preset first frequency band in the voiceprint information can be obtained, thereby more accurately identifying whether a construction-related event has actually occurred in the target channel. Preferably, the sustained energy ratio information of the preset first frequency band can be low-frequency sustained energy ratio information.

[0060] For knocking events, the signal typically exhibits characteristics of high frequency bands, rich components, high kurtosis, and short duration. Therefore, when the preset candidate external damage event type in the target frequency band corresponding to the target channel is a knocking event, the instantaneous kurtosis information of the preset second frequency band in the voiceprint information can be obtained, thereby more accurately identifying whether a knocking event has actually occurred in the target channel. Preferably, the instantaneous kurtosis information of the preset second frequency band can be high-frequency instantaneous kurtosis information.

[0061] For leakage-type events, they typically exhibit stable energy bandwidth in the mid-to-high frequency band, a flat spectral envelope, and a relatively long duration in the signal. Therefore, when the preset candidate external damage event type in the target frequency band corresponding to the target channel is a leakage-type event, the stable bandwidth index information of the preset third frequency band in the voiceprint information can be obtained, thereby more accurately identifying whether a leakage-type event has actually occurred in the target channel. Preferably, the stable bandwidth index information of the preset third frequency band can be mid-to-high frequency stable bandwidth index information.

[0062] S400, Identify the target external damage event type of the target channel based on the voiceprint information.

[0063] In this embodiment, based on the voiceprint information, it is determined whether the candidate external damage event type has actually occurred in the target channel, thereby enabling the identification of the target external damage event type of the target channel.

[0064] Specifically, the step of identifying the target channel's external damage event type based on the voiceprint information includes: The voiceprint information is input into a preset neural network for processing to determine whether the target channel has experienced an external damage event of the corresponding candidate external damage event type; if so, the corresponding candidate external damage event type is taken as the target external damage event type of the target channel; otherwise, it is determined that the target channel has a misjudgment.

[0065] In this embodiment, the voiceprint information is input into a preset neural network. The neural network can identify the voiceprint information and determine whether the corresponding target channel has actually experienced an external damage event of the candidate external damage event type. If so, the corresponding candidate external damage event type is taken as the target external damage event type for the target channel; otherwise, the target channel has not actually experienced an external damage event of the candidate external damage event type. A misjudgment may occur during a single identification, and the identification can be corrected or an alarm message can be issued based on the misjudgment. It is understood that the neural network is pre-trained and tested based on historical voiceprint information samples, enabling direct identification of sound wave information, thereby improving identification efficiency. The historical voiceprint information samples include positive historical voiceprint information samples and negative historical voiceprint information samples. The positive historical voiceprint information samples are historical voiceprint information where the target channel has not experienced an external damage event of the candidate external damage event type, and the negative historical voiceprint information samples are historical voiceprint information where the target channel has actually experienced an external damage event of the candidate external damage event type.

[0066] Specifically, the method further includes: If the corresponding candidate external damage event type is the target external damage event type of the target channel, the neural network outputs the target external damage event type corresponding to the target channel, as well as one or more of the alarm location, alarm time, and vertical distance judgment results of the external damage event of the target external damage event type. The vertical distance determination result is calculated based on the frequency attenuation characteristics and amplitude attenuation characteristics obtained from the time-domain data of the target channel in the optical fiber.

[0067] In this embodiment, when the corresponding candidate external damage event type is used as the target external damage event type for the target channel, the neural network can also output the target external damage event type for the target channel, as well as one or more of the alarm location, alarm time, and vertical distance judgment results of the external damage event of the target external damage event type. The alarm location reflects the location of the external damage event of the target external damage event type in the optical fiber, the alarm event reflects the time of the external damage event of the target external damage event type in the optical fiber, and the vertical distance judgment result reflects the vertical deviation distance of the event source of the target external damage event type relative to the optical fiber laying path, that is, the vertical projection distance from the event source to the optical fiber, which is used to determine whether the external damage behavior directly threatens the safety of the optical cable.

[0068] Understandably, the frequency attenuation characteristic refers to the characteristic parameter of energy attenuation of different frequency components of the signal corresponding to the external damage event during propagation as the propagation distance increases, used to reflect the vertical distance information between the event source and the optical fiber. The amplitude attenuation characteristic refers to the change law of the peak amplitude or root mean square amplitude of the signal corresponding to the external damage event in the time domain gradually decreasing as the propagation distance increases, used to help determine the degree of vertical deviation of the event source relative to the optical fiber. Combining the frequency attenuation characteristic and the amplitude attenuation characteristic, the vertical distance judgment result can be estimated together, used to reflect the vertical projection distance from the event source to the optical fiber.

[0069] like Figure 3 As shown, this application embodiment also provides an external damage event identification system based on optical fiber. Optionally, the system includes: The heatmap generation module 511, the primary recognition module 512, the voiceprint information acquisition module 513, and the secondary recognition module 514, wherein: The heatmap generation module 511 is used to obtain a heatmap corresponding to each target frequency band based on the frequency domain data of each channel in the optical fiber and the preset target frequency band; the target frequency band has a preset candidate external damage event type; the optical fiber is divided into channels according to a preset distance length; In this embodiment, the heatmap generation module 511 can be used to perform... Figure 1For a detailed description of the heatmap generation module 511, please refer to the description of step S100 shown.

[0070] The primary identification module 512 is used to obtain the corresponding target channel based on the heat map corresponding to the target frequency band; In this embodiment, the primary identification module 512 can be used to perform... Figure 1 For a detailed description of step S200 shown, and the specific details of the primary identification module 512, please refer to the description of step S200.

[0071] Specifically, the primary identification module 512 is further configured to acquire candidate channels included in the candidate channel range, and the time-domain data corresponding to the candidate channels in the optical fiber; acquire evaluation indicators for the time-domain data corresponding to the candidate channels; the evaluation indicators include at least one of energy intensity index value, signal-to-noise ratio index value, correlation index value, and persistence index value; based on the evaluation indicators, acquire a comprehensive score for the candidate channels; and select the candidate channel with the highest comprehensive score as the target channel.

[0072] In this embodiment, the primary identification module 512 can also be used to perform... Figure 2 For a detailed description of the primary identification module 512, see steps S210-S240 shown below. Further details regarding steps S210-S240 can be found in the description of steps S210-S240.

[0073] The voiceprint information acquisition module 513 is used to acquire corresponding channel information based on the target channel; acquire time domain data of the target channel in the optical fiber based on the channel information; and acquire voiceprint information corresponding to the target channel based on the time domain data of the target channel and preset candidate external damage event types in the target frequency band. In this embodiment, the voiceprint information acquisition module 513 can be used to perform... Figure 1 For a detailed description of the voiceprint information acquisition module 513, please refer to the description of step S300 shown.

[0074] The secondary identification module 514 is used to identify the target external damage event type of the target channel based on the voiceprint information.

[0075] In this embodiment, the secondary recognition module 514 can be used to perform... Figure 1 For a detailed description of the secondary identification module 514, please refer to the description of step S400 shown.

[0076] This application also provides an electronic device, the structure of which is as follows: Figure 4As shown, the electronic device includes a memory 611, a processor 612, a communication module 613, and an input / output interface 614, etc. Optionally, the memory 611, the processor 612, the communication module 613, and the input / output interface 614 can be connected and communicate with each other through a bus 615.

[0077] The memory 611 is used to store one or more computer programs and to transfer the code of the computer programs to the processor 612; when the one or more computer programs are executed by the processor 612, a fiber optic-based external damage event identification method is implemented in this embodiment of the application.

[0078] Optionally, the electronic device can be connected to a network via communication module 613 to communicate with other devices, such as terminals or servers, to achieve data interaction. The electronic device can be various forms of digital computers, exemplarily such as desktop computers, servers, workbenches, mainframes, or other types of computers. The electronic device can also be various forms of mobile terminals, exemplarily such as smartphones, tablets, wearable devices (such as helmets, glasses, watches, etc.), and other similar mobile terminals.

[0079] Optionally, the electronic device can connect to required input / output devices, such as a keyboard or display device, via the input / output interface 614. The electronic device itself may have a display device, and other display devices can also be connected externally via the input / output interface 614. Optionally, a storage device, such as a hard disk, can also be connected via the input / output interface 614 to store data from the electronic device, read data from the storage device, or store data from the storage device in the memory 611. It is understood that the input / output interface 614 can be a wired interface or a wireless interface. Depending on the actual application scenario, the device connected to the input / output interface 614 can be a component of the electronic device or an external device connected to the electronic device when needed.

[0080] Optionally, the memory 611 may be a volatile memory and / or a non-volatile memory. The volatile memory may be a random access memory, etc., and the non-volatile memory may be a read-only memory, a programmable read-only memory, an erasable programmable read-only memory, an electrically erasable programmable read-only memory, or a flash memory, etc.

[0081] Optionally, the computer program stored in the memory 611 can be divided into one or more modules, which are stored in the memory 611 and executed by the processor 612 to perform the method provided in this embodiment. The one or more modules can be a series of computer program instruction segments capable of performing specific functions, which describe the execution process of the computer program in the electronic device.

[0082] Optionally, the processor 612 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the processor 612 include, but are not limited to, a central processing unit, a graphics processing unit, a digital signal processor, various special-purpose artificial intelligence computing chips, various processors running machine learning model algorithms, and can also be any suitable controller, microcontroller, processor, etc. The processor 612 executes the various methods and processes of this embodiment, exemplarily, such as a fiber optic-based external damage event identification method according to an embodiment of this application.

[0083] Optionally, the bus 615 may include a path for transmitting information. Depending on its function, the bus 615 may be divided into an address bus, a data bus, a control bus, etc.

[0084] In an optional implementation, this application embodiment also provides a computer storage medium storing a computer program thereon, which, when executed by a computer, enables the computer to perform the methods described in the above-described method embodiments. Part or all of the computer program can be loaded and / or installed on the memory 611 of an electronic device. When the computer program is executed by the processor 612, one or more steps of a fiber optic-based external damage event identification method according to this application embodiment can be performed.

[0085] Optionally, the computer-readable storage medium may be a random access memory, a read-only memory, a programmable read-only memory, an erasable programmable read-only memory, an electrically erasable programmable read-only memory, etc.

[0086] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the technical solution of the present invention, and are not intended to limit the specific implementation of the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the claims of the present invention should be included within the protection scope of the claims of the present invention.

Claims

1. A method for identifying external damage events based on optical fibers, characterized in that, The method includes: Based on a preset target frequency band and frequency domain data of each channel in the optical fiber, a heat map corresponding to each target frequency band is obtained; the target frequency band has a preset candidate external damage event type; the optical fiber is divided into several channels with a preset distance length; wherein, obtaining a heat map corresponding to each target frequency band based on the preset target frequency band and frequency domain data of each channel in the optical fiber includes: performing Fourier transform processing on the time domain data of each channel in the optical fiber based on a preset time window to obtain the frequency domain data of each channel; wherein, the optical fiber has time domain data of several channels at several times; performing energy integration processing of the target frequency band on the frequency domain data of each channel in the optical fiber to obtain a heat map corresponding to each target frequency band; Based on the heat map corresponding to the target frequency band, obtain the corresponding target channel; Based on the target channel, obtain corresponding channel information; based on the channel information, obtain time-domain data of the target channel in the optical fiber; based on the time-domain data of the target channel and preset candidate external damage event types in the target frequency band, obtain acoustic fingerprint information corresponding to the target channel; wherein the candidate external damage event types include one or more of construction event types, impact event types, and leakage event types; the acoustic fingerprint information includes one or more of the continuous energy ratio information of a preset first frequency band, the instantaneous kurtosis information of a preset second frequency band, and the stable bandwidth index information of a preset third frequency band; the time-domain data of the target channel and the preset candidate external damage event types in the target frequency band are used to obtain the acoustic fingerprint information corresponding to the target channel. Selecting an external damage event type and obtaining the corresponding acoustic signature information for the target channel includes: if the preset candidate external damage event type in the target frequency band is a construction event type, then obtaining the continuous energy ratio information of the first frequency band of the target channel based on the time domain data of the target channel; if the preset candidate external damage event type in the target frequency band is a knocking event type, then obtaining the instantaneous kurtosis information of the second frequency band of the target channel based on the time domain data of the target channel; if the preset candidate external damage event type in the target frequency band is a leakage event type, then obtaining the stable bandwidth index information of the third frequency band of the target channel based on the time domain data of the target channel. The target channel's external damage event type is identified based on the voiceprint information.

2. The method according to claim 1, characterized in that, The step of obtaining the corresponding target channel based on the heatmap corresponding to the target frequency band includes: The heat map corresponding to the target frequency band is input into a preset recognition model for recognition to obtain candidate anomaly regions; wherein the candidate anomaly regions include the candidate channel range where the anomaly occurs; The target channel is determined based on the range of candidate channels.

3. The method according to claim 2, characterized in that, Determining the target channel based on the candidate channel range includes: Obtain the candidate channels included in the candidate channel range, and the time-domain data of the candidate channels in the optical fiber; The evaluation metrics for the time-domain data corresponding to the candidate channels are obtained; the evaluation metrics include at least one of the following: energy intensity index, signal-to-noise ratio index, correlation index, and persistence index. Based on the evaluation metrics, obtain the comprehensive score of the candidate channel; The candidate channel with the highest overall score will be selected as the target channel.

4. The method according to any one of claims 1 to 3, characterized in that, The identification of the target channel's external damage event type based on the voiceprint information includes: The voiceprint information is input into a preset neural network for processing to determine whether the target channel has experienced an external damage event of the corresponding candidate external damage event type; if so, the corresponding candidate external damage event type is taken as the target external damage event type of the target channel; otherwise, it is determined that the target channel has a misjudgment.

5. The method according to claim 4, characterized in that, The method further includes: If the corresponding candidate external damage event type is the target external damage event type of the target channel, the neural network outputs the target external damage event type corresponding to the target channel, as well as one or more of the alarm location, alarm time, and vertical distance judgment results of the external damage event of the target external damage event type. The vertical distance determination result is calculated based on the frequency attenuation characteristics and amplitude attenuation characteristics obtained from the time-domain data of the target channel in the optical fiber.

6. A fiber optic-based external damage event identification system, characterized in that, The system includes: A heatmap generation module is used to obtain a heatmap corresponding to each target frequency band based on a preset target frequency band and frequency domain data of each channel in the optical fiber; the target frequency band has a preset candidate external damage event type; the optical fiber is divided into several channels by a preset distance length; wherein, obtaining a heatmap corresponding to each target frequency band based on the preset target frequency band and frequency domain data of each channel in the optical fiber includes: performing Fourier transform processing on the time domain data of each channel in the optical fiber based on a preset time window to obtain the frequency domain data of each channel; wherein, the optical fiber has time domain data of several channels at several times; performing energy integration processing of the target frequency band on the frequency domain data of each channel in the optical fiber to obtain a heatmap corresponding to each target frequency band; A primary identification module is used to obtain the corresponding target channel based on the heat map corresponding to the target frequency band; The acoustic signature information acquisition module is used to acquire corresponding channel information based on the target channel; acquire time-domain data of the target channel in the optical fiber based on the channel information; and acquire acoustic signature information corresponding to the target channel based on the time-domain data of the target channel and preset candidate external damage event types in the target frequency band; wherein the candidate external damage event types include one or more of construction event types, impact event types, and leakage event types; the acoustic signature information includes one or more of the continuous energy ratio information of a preset first frequency band, the instantaneous kurtosis information of a preset second frequency band, and the stable bandwidth index information of a preset third frequency band; the acquisition module is used to acquire the corresponding channel information based on the time-domain data of the target channel and the target frequency band. The system obtains the voiceprint information corresponding to the target channel based on preset candidate external damage event types, including: if the preset candidate external damage event type in the target frequency band is a construction event type, then the system obtains the continuous energy ratio information of the first frequency band of the target channel based on the time domain data of the target channel; if the preset candidate external damage event type in the target frequency band is a knocking event type, then the system obtains the instantaneous kurtosis information of the second frequency band of the target channel based on the time domain data of the target channel; if the preset candidate external damage event type in the target frequency band is a leakage event type, then the system obtains the stable bandwidth index information of the third frequency band of the target channel based on the time domain data of the target channel. The secondary identification module is used to identify the target external damage event type of the target channel based on the voiceprint information.

7. The system according to claim 6, characterized in that, The primary identification module also includes: The heat map corresponding to the target frequency band is input into a preset recognition model for recognition to obtain candidate anomaly regions; wherein the candidate anomaly regions include the candidate channel range where the anomaly occurs; The target channel is determined based on the range of candidate channels.

8. The system according to claim 7, characterized in that, The primary identification module also includes: Obtain the candidate channels included in the candidate channel range, and the time-domain data of the candidate channels in the optical fiber; The evaluation metrics for the time-domain data corresponding to the candidate channels are obtained; the evaluation metrics include at least one of the following: energy intensity index, signal-to-noise ratio index, correlation index, and persistence index. Based on the evaluation metrics, obtain the comprehensive score of the candidate channel; The candidate channel with the highest overall score will be selected as the target channel.

9. An electronic device, characterized in that, include: Memory, used to store one or more computer programs; A processor, when the one or more computer programs are executed by the processor, implements a fiber-optic-based external damage event identification method as described in any one of claims 1-5.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that cause a processor to execute and implement the fiber optic-based external damage event identification method as described in any one of claims 1-5.

Images