Dike underwater piping positioning system based on distributed optical fiber vibration sensing

Abstract

The invention discloses a dike underwater piping positioning system based on distributed optical fiber vibration sensing, and relates to the technical field of positioning systems, and the dike underwater piping positioning system comprises a distributed optical fiber vibration sensing host which is used for transmitting a laser pulse to a sensing optical cable and demodulating a backward Rayleigh scattering signal; the multi-mode sensing optical cable is arranged along the dike underwater area and comprises a vibration sensing optical fiber and a temperature sensing optical fiber; the data processing module is used for executing time-space synchronous analysis of the vibration signal # imgabs0 # and the temperature signal # imgabs1 #; and the data processing module comprises a multi-modal feature fusion unit, and the multi-modal feature fusion unit is used for generating a combined feature factor # imgabs4 # by fusing the intensity # imgabs2 # and the temperature change rate # imgabs3 # of the vibration signal and carrying out weight distribution. According to the invention, through the technical closed loop of multi-mode sensing, intelligent layout and deep learning, the technical blank of the existing distributed optical fiber monitoring system in the aspects of early recognition, accurate positioning and intelligent early warning of underwater piping is filled.

Description

Technical Field

[0001] The invention relates to the technical field of positioning systems, and in particular to a distributed optical fiber vibration sensing embankment underwater piping positioning system. Background Art

[0002] Underwater piping in embankment projects (such as rivers, lakes, and reservoir dams) is a progressive destructive process in which soil particles are lost due to seepage, forming hidden channels. In the early stages, it only manifests as slight seepage and local soil disturbance, which is difficult to detect through traditional manual inspections or surface monitoring. However, once it develops into a large-scale piping, it may cause embankment collapse and cause significant loss of life and property. In order to locate the piping, it is necessary to use an underwater piping location system; However, traditional monitoring methods have problems such as insufficient temporal and spatial coverage, weak concealment identification ability, difficulty in distinguishing multi-source interference and limited positioning accuracy. Therefore, a distributed optical fiber vibration sensing embankment underwater pipe burst positioning system is proposed. Summary of the invention

[0003] The present invention aims to provide a distributed optical fiber vibration sensing system for locating underwater pipe bursts in dikes to solve the problems raised in the above-mentioned background technology.

[0004] In order to solve the above technical problems, the technical solution adopted by the present invention is: A distributed optical fiber vibration sensing system for locating underwater pipe bursts in dikes, comprising: Distributed fiber optic vibration sensing host, used to transmit laser pulses to the sensing optical cable and demodulate the backscattered Rayleigh signal; Multi-mode sensing optical cables, laid along the underwater area of the dike, include vibration sensing optical fibers and temperature sensing optical fibers; Data processing module, executes vibration signal With temperature signal Spatiotemporal synchronization analysis of The warning output module, the data processing module includes a multi-modal feature fusion unit, which fuses the intensity of the vibration signal and temperature change rate Generate joint eigenfactors And distribute the weights.

[0005] Furthermore, the specific calculation process of the multimodal feature fusion unit for calculating the joint feature factor is: ; Among them, α and β are weight coefficients dynamically adjusted according to the environmental noise. is the characteristic frequency band of piping, z is the optical fiber distance coordinate, and t is the time.

[0006] Furthermore, the weight coefficients α and β are adaptively generated through the following steps: Extract the noise floor values of the vibration and temperature signals from historical data and ; Calculate the instantaneous signal-to-noise ratio: ; ; After that, allocate weights according to , Allocate weights.

[0007] Furthermore, the multimodal sensing optical cable adopts a spiral winding three-dimensional layout structure, and the spiral winding three-dimensional layout structure includes: a rigid pile body, which is vertically fixed at the toe of the dike, the sensing optical cable is wound along the surface of the pile body in an equidistant spiral path, and a soil coupling medium is filled between the optical cable and the pile body.

[0008] Furthermore, the data processing module includes a three-dimensional coordinate mapping unit, which converts the fiber optic distance z into geographical coordinates through the following steps ; Establish a path equation according to the spiral layout parameters: ; is the spiral radius, is the single-turn pitch, is the elevation of the pile bottom, is the spiral slope, and it satisfies k = ds / Ls, where ds is the vertical spacing; When a piping event occurs at the position point zp on the fiber optic distance coordinate z, substitute it into the equation to calculate the spatial position.

[0009] Furthermore, the vertical spacing ds satisfies , where is the vibration wave speed in the soil, is the maximum detection frequency of the system.

[0010] Furthermore, the data processing module includes a depth convolutional neural network classifier, whose input is the time-frequency spectrogram of the joint feature factor , and the output is the piping probability . When , trigger an alarm, is the preset threshold.

[0011] Furthermore, the training data of the depth convolutional neural network classifier includes: The vibration-temperature joint data from on-site simulated piping experiments; Negative samples of ship movement, wave, and water flow noise; Synthetic samples generated by time-shifting scaling and noise injection.

[0012] Furthermore, the early warning output module superimposes and displays the three-dimensional piping positions on the GIS visualization product and marks the confidence level.

[0013] Due to the adoption of the above technical solutions, the technical progress achieved by the present invention compared with the prior art is as follows: Through the spatio-temporal synchronous analysis and feature fusion of vibration signals and temperature signals, combined with an adaptive weight allocation mechanism, the accuracy of piping feature recognition and the ability to resist environmental noise interference are effectively improved, solving the problem of insufficient reliability of single-signal detection. Dynamically calculating the weight coefficient based on the noise background value enables the system to automatically optimize the fusion ratio of vibration and temperature signals according to the real-time environmental noise, enhancing the detection robustness in complex underwater environments. Adopting a spiral-wound three-dimensional layout structure and a three-dimensional coordinate mapping algorithm to convert the optical fiber distance coordinates into geospatial coordinates, realizing the precise spatial positioning of underwater piping events on the dike, and solving the problem of single-dimensional positioning in traditional linear layout methods. Introducing a deep convolutional neural network classifier and combining it with data augmentation techniques, using machine learning to automatically extract piping features, improving the ability to distinguish interference signals such as passing ships and water flows, reducing the false alarm rate and increasing the detection efficiency. Superimposing and displaying the three-dimensional positioning results on the GIS visualization product and marking the confidence level, providing intuitive and quantitative early warning information for dike monitoring personnel, assisting in rapid decision-making and emergency handling, making this system more worthy of popularization and use. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required in the embodiments. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings.

[0015] Figure 1 It is the system block diagram of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0016] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0017] As Figure 1 shown, the present invention provides a dike underwater piping positioning system for distributed optical fiber vibration sensing, including: A distributed optical fiber vibration sensing host, which is used to transmit laser pulses to a sensing optical cable and demodulate backward Rayleigh scattering signals; A multimodal sensing optical cable, which is arranged along the underwater area of the dike and includes a vibration sensing optical fiber and a temperature sensing optical fiber; A data processing module, which executes the spatio-temporal synchronous analysis of vibration signals and temperature signals ; An early warning output module, the data processing module includes a multimodal feature fusion unit, which generates a joint feature factor by fusing the intensity of vibration signals and the temperature change rate and performs weight assignment. ;

[0018] The specific calculation process of the multimodal feature fusion unit for calculating the joint feature factor is: ; where α and β are weight coefficients dynamically adjusted according to environmental noise, is the characteristic frequency band of piping, z is the optical fiber distance coordinate, and t is the time; By realizing the multimodal feature fusion of vibration and temperature signals, the signal complementary enhancement advantage is achieved: the change characteristics of both vibration and temperature physical quantities are utilized simultaneously to make up for the limitations of single-signal detection; For example, when piping occurs underwater, the vibration signal reflects the mechanical disturbance caused by water flow scouring, and the temperature signal reflects the local temperature anomaly caused by seepage. The combination of the two can effectively distinguish piping from pure water flow fluctuations (only generating vibration) or natural water temperature changes (only affecting temperature) The setting of adaptive weight assignment can dynamically adjust the α and β coefficients according to environmental noise and optimize the signal fusion strategy in different scenarios Such as in a strong water flow environment: when the vibration noise increases, α decreases and β increases, indicating that the system relies more on temperature changes to identify piping In a temperature-stratified water area: when the temperature fluctuation intensifies, α increases and β decreases, indicating that the system focuses on analyzing vibration characteristics; By the characteristic frequency band focus on the vibration signals related to piping and filter out environmental noise interference For example, the vibration caused by piping usually concentrates in the 50 - 200 Hz low-frequency band The vibration generated by ship navigation is mainly distributed in the 200 - 500 Hz medium-high frequency band The system effectively shields ship noise interference by setting = 50 - 200 Hz Spatio-temporal positioning enhancement Advantages: The combined feature factor J(z,t) contains information in both the spatial (z) and temporal (t) dimensions, improving the positioning accuracy. Example: When an increase in vibration (|V(z,t)| increases) and a sudden change in temperature (|∂T / ∂t| increases) are simultaneously detected at a certain position z. Combined with the dynamic adjustment of α and β weights, the spatial position and starting time of the occurrence of piping can be accurately locked.

[0019] The weight coefficients α and β are adaptively generated through the following steps: Extract the noise floor values of the vibration and temperature signals from historical data. And ; Calculate the instantaneous signal-to-noise ratio: ; ; After that, according to , Allocate weights; Dynamically adjust the weights based on the signal-to-noise ratio, automatically suppressing noise interference and enhancing the reliability of the system in complex environments. When the noise floor value of a certain type of signal increases, its weight automatically decreases, ensuring that the overall detection accuracy is not affected by environmental changes, balancing the contribution degrees of the vibration and temperature signals in real time, and giving full play to the complementary advantages of multimodal fusion. The weight allocation can automatically adapt to different monitoring scenarios according to the signal quality, maintaining the optimal detection effect at all stages of the development of piping.

[0020] The weight allocation algorithm is completely based on the real-time signal characteristics and does not require manual intervention, realizing true adaptive detection. The system can automatically identify and cope with short-term interferences (such as a ship passing by) and long-term environmental changes (such as tidal temperature fluctuations), and automatically resume normal weight allocation after the interference. For example, in a calm water environment, the noise levels of both the vibration signal and the temperature signal are low. At this time, the vibration signal-to-noise ratio = 10 dB, and the temperature signal-to-noise ratio = 8 dB. According to the formula It is calculated that α ≈ 0.56 and β ≈ 0.44, and the system balances the use of the two signals for piping detection. When a strong water flow disturbance occurs, the vibration signal is greatly interfered, and the increase in the noise floor value Nv leads to = 3 dB, while the temperature signal still maintains good quality = 7 dB. At this time, α ≈ 0.30 and β ≈ 0.70, and the system automatically adjusts to focus on the temperature signal detection, effectively avoiding the influence of vibration noise.

[0021] If there is a temperature stratification phenomenon in the water area, the noise floor value of the temperature signal increases. = 4 dB, and the vibration signal quality is good = 8 dB. It is calculated that α ≈ 0.67 and β ≈ 0.33, and the system then focuses on vibration signal detection.

[0022] During the typical piping occurrence process, the initial seepage is weak, the vibration signal is not obvious but the temperature change is significant, and at this time the β weight is larger; as the piping develops, the water flow scouring intensifies, the vibration signal increases, and the α weight gradually increases and exceeds β. This dynamic weight adjustment process is automatically matched with the piping development stage to achieve accurate detection throughout the cycle.

[0023] The multi-modal sensing optical cable adopts a spiral winding three-dimensional layout structure, and the spiral winding three-dimensional layout structure includes: a rigid pile body, which is vertically fixed at the toe of the dike slope, the sensing optical cable is wound along the surface of the pile body according to an equal-spacing spiral path, and a soil coupling medium is filled between the optical cable and the pile body; The spiral winding structure enables the sensing optical cable to form a three-dimensional monitoring network in the vertical direction (elevation) and the horizontal direction (circumferential direction), breaking through the limitations of traditional horizontal or vertical single-dimensional layout; A soil coupling medium is filled between the optical cable and the pile body to enhance the conduction efficiency of vibration waves and temperature fields, and avoid signal attenuation caused by rigid contact between the optical cable and the pile body or too large a gap; The fixed and spiral winding design of the rigid pile body enhances the stability of the optical cable in harsh environments such as water flow scouring and sediment deposition, and avoids detection position deviation caused by displacement of traditional flexible layout optical cables; The combination of the spatial sampling density of the spiral structure and the signal processing algorithm realizes a spatial resolution of vibration signals at the millimeter level.

[0024] The data processing module includes a three-dimensional coordinate mapping unit, which converts the optical fiber distance z into geographical coordinates through the following steps ; Establish a path equation according to the spiral layout parameters: ; is the spiral radius, is the single-turn pitch, is the elevation of the pile bottom, is the spiral slope, and satisfies k = ds / Ls, where ds is the vertical spacing; When the piping event occurs at the position point zp on the optical fiber distance coordinate z, substitute it into the equation to calculate the spatial position; Convert the one-dimensional optical fiber distance coordinate (z) into a three-dimensional geographical coordinate (x, y, h), realizing the three-dimensional positioning of the piping event in the horizontal position (x, y) and the vertical depth (h), breaking through the limitation that traditional linear layout can only provide the longitudinal distance; Assume that the spiral radius \(r = 0.5m\), the single - turn pitch \(Ls = 2m\), the pile - bottom elevation \(h0=-5m\) (with the water surface as the reference), and the vertical spacing \(ds = 0.5m\). Then the spiral slope \(k = ds / Ls = 0.25\); When it is detected that a piping event occurs at a fiber - optic distance \(zp = 10m\), substitute it into the path equation: ; It is calculated that the piping position is at the horizontal coordinates \((0.5m,0m)\) and the depth is \(2.5m\) (below the water surface), providing accurate spatial guidance for the emergency rescue personnel; The layout method of the spiral winding on the rigid pile body can prevent the optical cable from being loosened or broken due to water flow scouring or soil displacement, improving the reliability of long - term monitoring.

[0025] For example, at the toe of the dyke where the water flow is rapid, the optical cable is fixed on the surface of the concrete pile by spiral winding. Combined with the filling of the soil - coupling medium, it can effectively resist the water flow impact force and prevent the detection failure caused by the displacement of the optical cable.

[0026] The spiral path increases the contact area between the optical cable and the soil, enhances the transmission efficiency of vibration signals from the soil to the optical fiber, and improves the detection sensitivity of weak piping signals.

[0027] For example, compared with the traditional straight - buried optical cable laid horizontally, the spiral - winding structure increases the contact points between the optical cable per unit length and the soil by 3 - 5 times. When there is a small seepage vibration in the shallow soil layer, the optical cable can capture the signal change more sensitively.

[0028] The vertical spacing \(ds\) satisfies , where is the vibration wave velocity in the soil, is the maximum detection frequency of the system; The above settings can avoid the spatial aliasing effect, ensure that the system satisfies the spatial sampling theorem of vibration signals, prevent the loss of high - frequency vibration information due to too large sampling spacing, and improve the ability to completely capture the characteristics of piping; If the vibration wave velocity in the soil \(= 300m / s\) and the maximum detection frequency of the system \(= 100Hz\), then the theoretical minimum vertical spacing is \(ds=\frac{2\times100}{300}=1.5m\).

[0029] By designing the vertical spacing of the spiral layout as \(ds = 1.0m\) (satisfying \(ds\leq1.5m\)), the highest 100 - Hz vibration signal caused by piping can be completely captured, avoiding the positioning error caused by spatial aliasing; Reducing the vertical spacing enhances the system's ability to resolve spatial details and realizes a more precise piping - position positioning.

[0030] For example, when the vertical distance between two adjacent piping points is 2 m: If ds = 2.0 m (the critical value), the system may only detect one signal peak and cannot distinguish two independent pipings; If ds = 1.0 m, the system can clearly distinguish two peaks and accurately locate each piping point.

[0031] A reasonable vertical spacing ensures that vibration and temperature signals are collected on the same spatial scale, improving the accuracy of spatio-temporal synchronous analysis.

[0032] For example, when seepage causes local temperature changes, if ds is too large (such as 3 m), it may cause the spatial sampling points of vibration signals and temperature signals to be misaligned, affecting the calculation accuracy of the joint characteristic factor J(z,t); However, the optimized ds = 1.0 m can ensure that the two signals are synchronously collected at the same spatial position (error <0.5 m).

[0033] The data processing module includes a deep convolutional neural network classifier, the input of which is the time-frequency spectrogram of the joint characteristic factor and the output is the piping probability When it triggers an alarm, being a preset threshold; By introducing a deep convolutional neural network (DCNN) classifier, an intelligent piping feature recognition model is constructed. Without the need to manually design complex feature extraction rules, the deep spatio-temporal correlation features in the time-frequency spectrogram of the joint characteristic factor J(z,t) are automatically learned by the neural network, solving the limitation of traditional signal processing methods relying on empirical parameters. Without the need to manually design complex feature extraction rules, the deep spatio-temporal correlation features in the time-frequency spectrogram of the joint characteristic factor J(z,t) are automatically learned by the neural network, solving the limitation of traditional signal processing methods relying on empirical parameters.

[0034] The training data of the deep convolutional neural network classifier includes: Vibration-temperature joint data from on-site simulated piping experiments; Negative samples of ship movement, wave, and water flow noise; Synthetic samples generated by time shift scaling and noise injection; By performing data augmentation operations such as time shift scaling and noise injection on the vibration and temperature joint feature data, the data limitation problem of deep learning models in practical engineering applications is solved. The specific advantages are as follows: Alleviate the small sample learning dilemma. When the measured piping data is scarce, especially when the typical samples under different working conditions are insufficient, diverse training samples are artificially generated to avoid overfitting of the model caused by insufficient data volume and improve the recognition ability for rare piping scenarios, such as extreme water levels and special soil types.

[0035] For example, in a monitoring project of a Yangtze River levee, there were only 200 original effective piping samples. After data augmentation, the equivalent sample size was expanded to 2,000, and the miss detection rate of the model on the test set decreased.

[0036] Improve the anti-interference robustness of the model. By injecting Gaussian noise, salt-and-pepper noise, etc. to simulate noise sources such as electromagnetic interference and water flow impact in the real environment, the model is forced to learn the essential features of the signal, such as the vibration frequency shift caused by seepage, rather than relying on the surface correlation of the data.

[0037] Under the scenario of simulating ship vibration interference in the laboratory, the false alarm rate of the model trained with noise injection enhancement for piping signals is lower than that of the non-enhanced model.

[0038] Reduce the data collection and annotation costs, and reduce the dependence on large-scale on-site measured data. It is especially suitable for scenarios where it is difficult to collect underwater environment data and the annotation cost is high (such as diving operations are required to obtain samples), which can save data collection costs.

[0039] The warning output module superimposes and displays the three-dimensional piping position on the GIS visualization product and marks the confidence level. The above settings achieve the adaptation of differential warning strategies. Set differential safety factors Ks according to the levee grades. For example: for high-risk Grade I levees, lower the warning threshold, such as Pth = 0.6, to achieve earlier warning; For low-risk Grade III levees, increase the threshold, such as Pth = 0.8, to reduce false alarm interference in daily operations.

[0040] For example, in the monitoring of a secondary dam (Grade III levee) of a reservoir, by increasing the threshold, the average daily false alarm times are reduced, significantly reducing the workload of monitoring personnel.

[0041] The ability to adapt to environmental changes. Use the historical false alarm rate α to adjust the threshold in real time, such as the proportion of the number of false alarms in the past week to the total number of warnings. For example: when the water flow disturbance during the flood season causes the false alarm rate to rise, such as α > 0.3, automatically increase the threshold Pth to 0.85 to suppress environmental noise interference; When the water level is stable in winter, lower the threshold to 0.55 to improve the sensitivity to weak piping signals.

[0042] Through the spatio-temporal synchronization analysis and feature fusion of vibration signals and temperature signals, combined with an adaptive weight allocation mechanism, the present invention effectively improves the accuracy of pipe burst feature recognition and the ability to resist environmental noise interference, solves the problem of insufficient reliability of single-signal detection, dynamically calculates the weight coefficient based on the noise background value, enables the system to automatically optimize the fusion ratio of vibration and temperature signals according to the real-time environmental noise, enhances the detection robustness in complex underwater environments, adopts a spiral-wound three-dimensional layout structure and a three-dimensional coordinate mapping algorithm to convert the optical fiber distance coordinates into geographical space coordinates, realizes the precise spatial positioning of underwater pipe burst events on the dike, solves the problem of single positioning dimension of traditional linear layout methods, introduces a deep convolutional neural network classifier and combines data augmentation techniques, automatically extracts pipe burst features using machine learning, improves the ability to distinguish interference signals such as passing ships and water flows, reduces the false alarm rate and improves the detection efficiency, superimposes and displays the three-dimensional positioning results on the GIS visualization product and marks the confidence level, provides intuitive and quantitative early warning information for dike monitoring personnel, and assists in rapid decision-making and emergency handling.

[0043] The above is only the specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present application can easily think of changes or substitutions, which should all be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A positioning system for underwater piping in dikes using distributed fiber optic vibration sensing, characterized in that, Comprising: A distributed optical fiber vibration sensing host for transmitting laser pulses to a sensing optical cable and demodulating backscattered Rayleigh signals; A multimodal sensing optical cable laid along the underwater area of the dike, including a vibration sensing optical fiber and a temperature sensing optical fiber; A data processing module that performs spatio-temporal synchronization analysis of vibration signals and temperature signals ; Early warning output module, the data processing module includes a multi-modal feature fusion unit, which generates a joint feature factor by fusing the intensity of the vibration signal and the temperature change rate and performs weight allocation; ​ The specific calculation process of the multimodal feature fusion unit for calculating the joint feature factor is as follows: ； where α and β are weight coefficients dynamically adjusted according to environmental noise, is the characteristic frequency band of piping, z is the fiber optic distance coordinate, and t is the time.

2. The underwater pipe gushing positioning system for dike based on distributed optical fiber vibration sensing according to claim 1, characterized in that: The weight coefficients α and β are adaptively generated through the following steps: Extract the noise floor values of vibration and temperature signals from historical data and ; Calculate the instantaneous signal-to-noise ratio: ； ； After that, press , to assign weights.

3. The underwater pipe burst location system for dike based on distributed optical fiber vibration sensing according to claim 1, characterized in that: The multimodal sensing optical cable adopts a spiral winding three-dimensional layout structure, and the spiral winding three-dimensional layout structure includes: a rigid pile body vertically fixed at the toe of the dike slope, the sensing optical cable winding along the surface of the pile body in an equidistant spiral path, and a soil coupling medium filled between the optical cable and the pile body.

4. A positioning system for underwater piping of dikes in distributed optical fiber vibration sensing according to claim 1, characterized in that: The data processing module includes a three-dimensional coordinate mapping unit that converts the optical fiber distance z into geographical coordinates through the following steps ; Establish a path equation according to the spiral layout parameters: ； is the spiral radius, is the pitch of a single turn, is the elevation of the pile bottom, is the spiral slope and satisfies k = ds / Ls, where ds is the vertical spacing; When a piping event occurs at the position point zp on the fiber distance coordinate z, substitute it into the equation to calculate the spatial position.

5. The underwater piping location system for dike based on distributed fiber optic vibration sensing according to claim 4, wherein: The vertical spacing ds satisfies , where is the vibration wave velocity in the soil, is the maximum detection frequency of the system.

6. The underwater pipe gushing positioning system for dike based on distributed optical fiber vibration sensing according to claim 1, characterized in that: The data processing module includes a deep convolutional neural network classifier, whose input is the joint feature factor of the time-frequency spectrogram, and the output is the probability of piping . When , an alarm is triggered, where is a preset threshold.

7. A positioning system for underwater pipe gushing of dike based on distributed optical fiber vibration sensing according to claim 1, characterized in that: The training data of the deep convolutional neural network classifier includes: Vibration-temperature joint data from on-site simulated piping experiments; Negative samples of ship traffic, waves, and water flow noise; Synthetic samples generated by time shift scaling and noise injection.

8. A system for locating underwater pipe gushing in a dike by distributed optical fiber vibration sensing according to any one of claims 1-7, characterized in that: The early warning output module superimposes and displays the three-dimensional piping position on the GIS visualization product and marks the confidence level.

Images