Distributed optical fiber sensing geological disaster monitoring and early warning method and system

By using a distributed fiber optic sensor network and advanced data processing technology, the problems of complex signal processing and insufficient network configuration in geological disaster monitoring have been solved, enabling high-precision, real-time geological disaster monitoring and early warning. It has strong early warning capabilities and adaptability, and improves the sensitivity and accuracy of the monitoring system.

CN118918685APending Publication Date: 2024-11-08广西壮族自治区地质环境监测站 +1
View PDF 0 Cites 14 Cited by

Patent Information

Application Number
CN202411205138.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-30
Publication Date
2024-11-08

AI Technical Summary

Technical Problem

Existing distributed fiber optic sensing technology faces challenges in geological disaster monitoring, including complex signal processing, insufficient data analysis, and a lack of flexibility in monitoring network configuration. This results in limited monitoring range and insufficient sensitivity, making it difficult to provide reliable, high-precision monitoring results and timely early warnings.

Method used

By employing a distributed fiber optic sensor network combined with improved optical frequency domain reflectance technology, spectral regularized autoencoder, spatiotemporal transformer analysis, adaptive multi-level decision tree, and co-evolutionary mechanism, real-time processing and analysis of multimodal physical quantity data are achieved. A high-dimensional dynamic monitoring model is constructed, and the model parameters are adaptively adjusted to identify early anomalies and assess risks of geological disasters. Emergency response is triggered through a multi-level decision tree.

Benefits of technology

It enables comprehensive and real-time monitoring of a wide range of geological environments, improves the accuracy and reliability of monitoring data, has a strong early warning capability, can promptly detect geological disaster anomalies, ensure timely response to disaster risks of different levels, and enhances the comprehensiveness and accuracy of monitoring through multi-source data fusion.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118918685A_ABST
    Figure CN118918685A_ABST
Patent Text Reader

Abstract

The invention discloses a distributed optical fiber sensing geological disaster monitoring and early warning method and system, and the method comprises the following steps: laying an optical fiber sensor network in a monitoring region, and collecting the physical quantity data of strain, temperature, vibration and the like in a geological environment in real time; and the data are processed by utilizing an improved optical technology, so that high precision of long-distance monitoring is ensured. The method comprises the following steps: comprehensively analyzing data through a map regularization technology, constructing a high-dimensional dynamic monitoring model, predicting occurrence time, location and severity of geological disasters, intelligently classifying risk levels in combination with a self-adaptive multi-level decision tree, triggering corresponding emergency response measures, and determining the occurrence time, location and severity of the geological disasters through a co-evolution mechanism and a dynamic genetic algorithm. The sensor network is optimized in real time, and the performance of the monitoring system is continuously improved. The invention aims to provide accurate and real-time geological disaster early warning and guarantee public safety.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of geological disaster monitoring and early warning technology, and in particular to a distributed optical fiber sensing geological disaster monitoring and early warning method and system. Background Art

[0002] Geological disasters, such as landslides, mudslides, and ground fissures, pose a serious threat to human life and property. These disasters are often sudden and unpredictable. Therefore, how to effectively monitor and warn of these disasters has become an important research direction in the field of disaster prevention and mitigation. Traditional geological disaster monitoring methods mainly rely on point sensors, such as strain gauges, inclinometers, accelerometers, etc. These sensors can only monitor local areas of the monitoring points and are difficult to cover large areas of disaster risk. In addition, these methods usually have problems such as limited monitoring range, insufficient sensitivity, and inability to provide real-time warnings, which makes it difficult to obtain comprehensive monitoring information in a timely manner when a disaster occurs, making it impossible to achieve effective disaster warnings.

[0003] The emergence of distributed fiber optic sensing technology provides a new solution for geological disaster monitoring. Unlike traditional point sensors, distributed fiber optic sensors can provide continuous monitoring data along long-distance optical fibers, covering a wide area and providing continuous spatial resolution. This technology can sense the changes in physical quantities such as strain, temperature, and vibration along the fiber optic layout area in real time, thereby achieving real-time monitoring of a large range of geological environments. However, the existing distributed fiber optic sensing technology still faces some challenges when applied to geological disaster monitoring, such as complex signal processing, insufficient data analysis, and lack of flexibility in monitoring network configuration, which leads to many deficiencies in complex and changeable geological environments.

[0004] Existing distributed fiber optic sensing technology often faces problems of signal attenuation and noise interference in long-distance monitoring. Traditional signal processing methods cannot effectively filter out noise while ensuring high spatial resolution, resulting in limited accuracy of monitoring data and difficulty in providing reliable monitoring results. This limits its application in large-scale geological disaster risk areas that require high-precision monitoring. In terms of data analysis, existing technologies mostly use simple statistical analysis methods, which makes it difficult to fully explore and utilize multimodal physical quantity data from different time and space. This leads to insufficient recognition of early abnormal changes in geological disasters, which in turn affects the accuracy and timeliness of early warning. In addition, in the modeling process, existing technologies often cannot adaptively adjust model parameters to adapt to changes in different geological conditions, which further reduces the reliability and sensitivity of the monitoring system. Traditional geological disaster monitoring systems are mostly statically configured and lack the ability to dynamically adjust. When geological conditions change, the configuration of the monitoring network cannot respond in time, resulting in limited monitoring range and sensitivity, and unable to provide continuously optimized monitoring effects. Especially in complex geological environments, fixed monitoring network configurations often find it difficult to cover all risk areas, and there are monitoring blind spots and resource waste.

[0005] Therefore, how to provide a distributed optical fiber sensing geological disaster monitoring and early warning method and system is a problem that technical personnel in this field urgently need to solve. Summary of the invention

[0006] One purpose of the present invention is to propose a distributed optical fiber sensing geological disaster monitoring and early warning method and system. The present invention makes full use of distributed optical fiber sensing technology, improved optical frequency domain reflection technology, spectrum regularization autoencoder, space-time transformer analysis, adaptive multi-level decision tree and collaborative evolution mechanism and other technologies, and describes in detail the whole process from data collection, processing, analysis to early warning and network optimization. It has the advantages of wide monitoring range, high sensitivity, strong real-time performance, high early warning accuracy and strong system adaptability.

[0007] A distributed optical fiber sensing geological disaster monitoring and early warning method and system according to an embodiment of the present invention comprises the following steps:

[0008] S1. In the monitoring area, according to the changing characteristics of geological conditions and the risk distribution of potential geological disasters, a distributed optical fiber sensor network is deployed along key risk paths and prone areas to form a continuous, real-time multi-parameter monitoring chain to achieve all-round perception and coverage of the geological conditions in the entire monitoring area;

[0009] S2. Use the improved distributed optical frequency domain reflectometry technology to process the multi-modal physical quantity data collected in real time by the distributed optical fiber sensor network, and use the dynamic noise filtering algorithm to achieve sub-meter spatial resolution in long-distance monitoring;

[0010] S3. Use graph regularization autoencoder to conduct comprehensive analysis on the processed multimodal physical quantity data, model the spatial and temporal relationship through graph regularization technology, build a high-dimensional dynamic monitoring model, identify early abnormal changes of geological disasters, and adaptively adjust model parameters according to different geological conditions;

[0011] S4. Based on the output of the high-dimensional dynamic monitoring model, the space-time transformer is used to perform trend analysis, and the time, location and severity of possible geological disasters are predicted by combining the space-time dependence of real-time monitoring data. A dynamic risk assessment model is generated to guide the determination of warning levels and the formulation of emergency strategies.

[0012] S5. After detecting potential geological disaster risks, an adaptive multi-level decision tree early warning mechanism is used based on a dynamic risk assessment model to intelligently classify risk levels, trigger emergency response systems at corresponding levels, and generate specific emergency measures to ensure timely response to disaster risks at different levels;

[0013] S6. Combined with the dynamic environment genetic algorithm of the co-evolution mechanism, the monitoring parameters of the distributed fiber optic sensor network are optimized in real time. Through the co-evolution and interaction of different parameter sets, the coverage and resource allocation of the monitoring network are adaptively adjusted to cope with the dynamic changes of geological conditions and continuously optimize the monitoring performance.

[0014] S7. Comprehensively analyze the monitoring data that have been collected, processed, analyzed and optimized with external data from geological radar, satellite remote sensing, and drone monitoring, integrate multi-source data to form comprehensive geological disaster monitoring information, and support dynamic adjustment of monitoring strategies during the early warning process to respond to different disaster types and evolution processes.

[0015] Optionally, the S2 specifically includes:

[0016] S21, using improved distributed optical frequency domain reflection technology, using broadband light source to emit continuous light signal The optical signal is transmitted through the optical fiber sensor network distributed in the monitoring area, and Rayleigh scattering occurs at various points in the optical fiber to form an echo signal containing multi-modal physical quantity data;

[0017] S22, using broadband modulation technology to optimize the spatial resolution of the system, the frequency range covered by the broadband light source, through the formula Achieve sub-meter spatial resolution, where B is the bandwidth of the light source and c is the speed of light. By adjusting the bandwidth of the light source, the system can obtain spatial resolution in long-distance monitoring and adapt to the monitoring needs of complex geological conditions;

[0018] S23, introduce phase modulation technology to optimize the signal's anti-interference ability, and extract useful spectrum information S(f) from the echo signal through phase demodulation technology:

[0019]

[0020] Among them, E mod (t) is the phase modulated optical signal, E0 is the amplitude of the optical signal, f0 is the center frequency of the optical signal, t is the time, is the phase noise term, θ mod (t) is the phase modulation term;

[0021] Through the phase modulation and demodulation process, the nonlinear effect interference is suppressed and the optimized spectrum information S(f) is obtained:

[0022]

[0023] in, is the Fourier transform operator, E demod (t) is the phase demodulated signal;

[0024] S24, collect the echo signal, and use the frequency modulation continuous wave principle to generate a difference frequency signal S by interfering the emitted light signal with the echo signal. df (f) and optimizing the signal-to-noise ratio and detection sensitivity using photodetectors;

[0025] S25, the obtained difference frequency signal S df (f) Performing Fourier transform processing to generate the scattering intensity distribution I(z) at different positions in the optical fiber, applying a dynamic noise filtering algorithm, adjusting the frequency response H(f) in real time through an adaptive filter, optimizing the signal-to-noise ratio of the signal, and generating the final scattering intensity distribution I'(z);

[0026] S26, fusing and analyzing the processed scattering intensity distribution I'(z) with the multimodal physical quantity data collected in real time by the distributed optical fiber sensor network, and adjusting the spatial resolution according to the improved light source bandwidth B to generate the final processed multimodal physical quantity data.

[0027] Optionally, the S3 specifically includes:

[0028] S31, input the final multimodal physical quantity data after generation and processing, including scattering intensity distribution I'(z), strain ε(z), temperature T(z), vibration V(z) and acoustic wave signal A(z), and construct a graph model of space and time relationship. The graph model is constructed through the spatial relationship graph G s and time relationship graph G t Describe the dependencies between data and use the weight matrix W sand W t To characterize;

[0029] S32. Based on the constructed graph model, an automatic encoder is used to extract features from the input multimodal data. The encoder maps the processed final multimodal physical quantity data to the low-dimensional space Z. m , retaining the key feature information and converting the low-dimensional representation Z into m Reconstruct as output data Realize data reconstruction and dimensionality reduction processing;

[0030] S33, the graph regularization term L reg Introduced into the encoder, the spatial and temporal continuity of the data is maintained by minimizing the differences between adjacent nodes in space and the differences between adjacent nodes in time:

[0031]

[0032] Among them, L reg is the graph regularization term, W s (i, j) is the weight between positions i and j in space, ||Z i -Z j || 2 is the square of the Euclidean distance between spatial locations i and j in the high-dimensional representation Z, W t (m,n) is the weight between time points m and n, ||Z m -Z n || 2 is the square of the Euclidean distance between time points m and n in the high-dimensional representation Z;

[0033] S34. By training the graph regularized autoencoder, a high-dimensional dynamic monitoring model is constructed to identify early abnormal changes in geological disasters. During the model training process, the total loss function L is optimized to enable the model to adaptively adjust the parameter θ. e and θ d , adapted to different geological conditions:

[0034] L=L recon +λL reg ;

[0035] Among them, L recon is the reconstruction error, and λ is the hyperparameter that balances the two;

[0036] S35. During model training and application, the model parameters θ are adaptively adjusted according to the changes in real-time monitoring data under different geological conditions. e and θ d , optimize the adaptability and reliability of the monitoring model in various geological environments.

[0037] Optionally, the S4 specifically includes:

[0038] S41, the output data of the high-dimensional dynamic monitoring model As the input of the spatiotemporal transformer, the output data of the high-dimensional dynamic monitoring model Reflects the reconstruction results of multimodal physical quantity data after being processed by the graph regularized autoencoder, including geological environment characteristics at different time and space positions;

[0039] S42, in the space-time transformer, based on the input reconstruction data Construct a spatiotemporal dependency model, and the spatiotemporal transformer represents the features at time t and spatial position x Embedded into the spatiotemporal dependency matrix T s In (t,x):

[0040]

[0041] in, and Represent the dependency functions in time and space dimensions respectively;

[0042] S43, using the spatiotemporal transformer to generate the spatiotemporal dependency matrix T s (t, x), combined with real-time monitoring data, predict the time, location and severity of possible geological disasters, and generate a geological disaster risk prediction model P through comprehensive analysis of spatiotemporal characteristics. risk (t,x):

[0043] P risk (t,x)=σ(T s (t,x));

[0044] Among them, P risk (t,x) represents the probability of a geological disaster occurring at time t and spatial location x;

[0045] S44, based on risk prediction model P risk (t,x) and the severity of geological disasters S(t,x), and construct a dynamic risk assessment model R(t,x);

[0046] S45. Determine the geological disaster warning level based on the output of the dynamic risk assessment model R(t,x), and formulate corresponding emergency strategies. According to different risk levels, take corresponding emergency response measures to effectively prevent and respond to geological disasters.

[0047] Optionally, the S5 specifically includes:

[0048] S51, using the output of the generated dynamic risk assessment model R(t,x) as the input of an adaptive multi-level decision tree, wherein the dynamic risk assessment model R(t,x) comprehensively considers the probability and severity of occurrence of geological disasters and provides comprehensive risk analysis data as the basis of a multi-level decision tree early warning mechanism;

[0049] S52. In the adaptive multi-level decision tree, define the decision criteria D for each node j , where j represents the level of the decision tree and the decision criterion D j Based on the output value of the dynamic risk assessment model R(t,x) and the adaptively adjusted risk threshold θ j Make a judgment:

[0050]

[0051] Among them, θ j represents the risk threshold of the jth level;

[0052] S53, the multi-level decision tree early warning mechanism determines the risk level layer by layer, and adjusts the threshold θ in real time j , at each level, according to D j The output of the dynamic risk assessment model determines whether to trigger the corresponding emergency response system. When the output of the dynamic risk assessment model meets a certain level of decision criteria D j =1, the emergency response system at this level is triggered;

[0053] S54. Generate specific emergency measures suggestions based on the triggered emergency response system. j (t,x) is generated based on the real-time updated risk assessment model R(t,x) and the output of the decision tree:

[0054] S j (t,x)=g(D j ,R(t,x));

[0055] Among them, g(·) is the emergency measure generation function;

[0056] S55, the adaptive multi-level decision tree early warning mechanism continuously monitors the output of the risk assessment model R(t,x) and updates the decision threshold θ in real time j and emergency response strategies j (t,x).

[0057] Optionally, the S6 specifically includes:

[0058] S61, initializing multiple parameter sets {P1, P2, ..., P n}, each parameter set P iContains key configuration parameters of the monitoring network. The co-evolution mechanism optimizes the coverage and signal quality of the monitoring network by evolving parameter sets in parallel.

[0059] S62. In the dynamic environment genetic algorithm, for each parameter set P i Generate an initial population. The individuals in the population represent different distributed fiber optic sensor network configuration schemes. The fitness function of each individual is F(P i ), the fitness evaluation is based on network coverage, signal quality and adaptive capability, which enables the distributed fiber optic sensor network to achieve stable monitoring under complex geological conditions;

[0060] S63, dynamically adjust the parameter set P through the co-evolution mechanism i In the evolutionary process, the dynamic environment genetic algorithm updates the fitness function F(P) in real time based on environmental changes in each iteration. i ) parameters, enabling the algorithm to adaptively optimize the coverage and resource allocation of the monitoring network to cope with dynamic changes in geological conditions;

[0061] S64, in each iteration process, use selection, crossover and mutation operations to generate a new parameter set, and according to the fitness function F(P i ) value, select the optimal parameter set for the next generation of evolution;

[0062] S65, after multiple generations of iterations, select the fitness function F(P i ) The optimal parameter set P * , and apply it to distributed fiber optic sensor networks to optimize the configuration of the monitoring network in real time.

[0063] Optional modules include:

[0064] Distributed fiber optic sensor network module: Deploy a distributed fiber optic sensor network to cover key risk paths and prone areas, and achieve all-round real-time monitoring of geological conditions;

[0065] Data acquisition and processing module: Process the collected multi-modal data to achieve long-distance monitoring with high spatial resolution;

[0066] Multimodal data analysis and modeling module: Analyze multimodal data, build high-dimensional dynamic monitoring models, identify early anomalies and adaptively adjust model parameters;

[0067] Spatiotemporal trend analysis and risk assessment module: Based on the high-dimensional dynamic monitoring model, it conducts spatiotemporal trend analysis, predicts the time, location and severity of geological disasters, and generates a risk assessment model;

[0068] Adaptive early warning and emergency response module: Based on the risk assessment model, an adaptive multi-level decision tree early warning mechanism is used to classify risk levels, trigger the corresponding emergency response system and generate emergency measures;

[0069] Monitoring network optimization and dynamic adjustment module: Combines the co-evolution mechanism and dynamic environment genetic algorithm to optimize monitoring network parameters in real time and adaptively adjust coverage and resource allocation;

[0070] Multi-source data fusion and feedback module: Comprehensively analyze monitoring data and external data sources to form comprehensive monitoring information and support dynamic adjustment of monitoring strategies.

[0071] The beneficial effects of the present invention are:

[0072] The present invention realizes all-round and real-time monitoring of a large-scale geological environment by deploying a distributed fiber optic sensor network in the monitoring area. Compared with traditional point sensors, the present invention can cover a wider area and provide continuous spatial resolution along the entire optical fiber, effectively solving the problem of limited monitoring range in traditional monitoring methods. By adopting improved distributed optical frequency domain reflection technology and combining it with a dynamic noise filtering algorithm, the spatial resolution in long-distance monitoring is greatly improved, achieving sub-meter accuracy. This combination of technologies enables the system to provide high-quality monitoring data even in complex geological environments, overcomes the problems of signal attenuation and noise interference, and thus ensures the accuracy and reliability of the monitoring data.

[0073] The present invention uses a graph regularized autoencoder to conduct a comprehensive analysis of multimodal physical quantity data, constructs a high-dimensional dynamic monitoring model, and can effectively identify early abnormal changes in geological disasters. By modeling the spatial and temporal relationship, the system has a strong early warning capability, which can detect anomalies in time before a geological disaster occurs, thereby improving the accuracy of the warning. In terms of early warning and emergency response, an adaptive multi-level decision tree early warning mechanism is introduced, which intelligently classifies geological disaster risks based on a dynamic risk assessment model, automatically triggers the corresponding emergency response system, and generates specific emergency measures. Recommendations. This intelligent early warning mechanism can dynamically adjust the warning level according to real-time monitoring data, ensuring a timely response to disaster risks of different levels.

[0074] The present invention also realizes real-time optimization of the monitoring parameters of the distributed optical fiber sensor network by combining the co-evolution mechanism with the dynamic environment genetic algorithm, and adaptively adjusts the coverage and resource allocation of the monitoring network to cope with the dynamic changes in geological conditions. This adaptive optimization capability ensures that the monitoring system can still operate efficiently and stably in a complex and changeable environment, overcomes the monitoring blind spots and resource waste problems caused by the static configuration of the traditional system, and realizes the comprehensive analysis of multi-source data. It not only processes and analyzes the data collected by the distributed optical fiber sensor network, but also can integrate it with external data such as geological radar, satellite remote sensing, and drone monitoring to form more comprehensive geological disaster monitoring information. Through the comprehensive analysis of multi-source data, the present invention enhances the system's ability to cope with different disaster types and evolution processes, and further improves the comprehensiveness and accuracy of disaster monitoring and early warning. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0076] Figure 1 A flow chart of a distributed optical fiber sensing geological disaster monitoring and early warning method and system proposed by the present invention;

[0077] Figure 2 It is a schematic diagram of the process of adaptive early warning and emergency response in the present invention. DETAILED DESCRIPTION

[0078] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, which only illustrate the basic structure of the present invention in a schematic manner, and therefore only show the components related to the present invention.

[0079] refer to Figure 1-2 , a distributed optical fiber sensing geological disaster monitoring and early warning method and system, comprising the following steps:

[0080] S1. In the monitoring area, according to the changing characteristics of geological conditions and the risk distribution of potential geological disasters, a distributed optical fiber sensor network is deployed along key risk paths and prone areas to form a continuous, real-time multi-parameter monitoring chain to achieve all-round perception and coverage of the geological conditions in the entire monitoring area;

[0081] S2. Use the improved distributed optical frequency domain reflectometry technology to process the multi-modal physical quantity data collected in real time by the distributed optical fiber sensor network, and use the dynamic noise filtering algorithm to achieve sub-meter spatial resolution in long-distance monitoring;

[0082] S3. Use graph regularization autoencoder to conduct comprehensive analysis on the processed multimodal physical quantity data, model the spatial and temporal relationship through graph regularization technology, build a high-dimensional dynamic monitoring model, identify early abnormal changes of geological disasters, and adaptively adjust model parameters according to different geological conditions;

[0083] S4. Based on the output of the high-dimensional dynamic monitoring model, the space-time transformer is used to perform trend analysis, and the time, location and severity of possible geological disasters are predicted by combining the space-time dependence of real-time monitoring data. A dynamic risk assessment model is generated to guide the determination of warning levels and the formulation of emergency strategies.

[0084] S5. After detecting potential geological disaster risks, an adaptive multi-level decision tree early warning mechanism is used based on a dynamic risk assessment model to intelligently classify risk levels, trigger emergency response systems at corresponding levels, and generate specific emergency measures to ensure timely response to disaster risks at different levels;

[0085] S6. Combined with the dynamic environment genetic algorithm of the co-evolution mechanism, the monitoring parameters of the distributed fiber optic sensor network are optimized in real time. Through the co-evolution and interaction of different parameter sets, the coverage and resource allocation of the monitoring network are adaptively adjusted to cope with the dynamic changes of geological conditions and continuously optimize the monitoring performance.

[0086] S7. Comprehensively analyze the monitoring data that have been collected, processed, analyzed and optimized with external data from geological radar, satellite remote sensing, and drone monitoring, integrate multi-source data to form comprehensive geological disaster monitoring information, and support dynamic adjustment of monitoring strategies during the early warning process to respond to different disaster types and evolution processes.

[0087] In this implementation, S2 specifically includes:

[0088] S21, using improved distributed optical frequency domain reflection technology, using broadband light source to emit continuous light signal The optical signal is transmitted through the optical fiber sensor network distributed in the monitoring area, and Rayleigh scattering occurs at various points in the optical fiber to form an echo signal containing multi-modal physical quantity data;

[0089] S22, using broadband modulation technology to optimize the spatial resolution of the system, the frequency range covered by the broadband light source, through the formula Achieve sub-meter spatial resolution, where B is the bandwidth of the light source and c is the speed of light. By adjusting the bandwidth of the light source, the system can obtain spatial resolution in long-distance monitoring and adapt to the monitoring needs of complex geological conditions;

[0090] S23, introduce phase modulation technology to optimize the signal's anti-interference ability, and extract useful spectrum information S(f) from the echo signal through phase demodulation technology:

[0091]

[0092] Among them, E mod (t) is the phase modulated optical signal, E0 is the amplitude of the optical signal, f0 is the center frequency of the optical signal, t is the time, is the phase noise term, θ mod (t) is the phase modulation term;

[0093] Through the phase modulation and demodulation process, the nonlinear effect interference is suppressed and the optimized spectrum information S(f) is obtained:

[0094]

[0095] in, is the Fourier transform operator, E demod (t) is the phase demodulated signal;

[0096] S24, collect the echo signal, and use the frequency modulation continuous wave principle to generate a difference frequency signal S by interfering the emitted light signal with the echo signal. df (f) and optimizing the signal-to-noise ratio and detection sensitivity using photodetectors;

[0097] S25, the obtained difference frequency signal S df (f) Performing Fourier transform processing to generate the scattering intensity distribution I(z) at different positions in the optical fiber, applying a dynamic noise filtering algorithm, adjusting the frequency response H(f) in real time through an adaptive filter, optimizing the signal-to-noise ratio of the signal, and generating the final scattering intensity distribution I'(z);

[0098] S26, fusing and analyzing the processed scattering intensity distribution I'(z) with the multimodal physical quantity data collected in real time by the distributed optical fiber sensor network, and adjusting the spatial resolution according to the improved light source bandwidth B to generate the final processed multimodal physical quantity data.

[0099] In this implementation, S3 specifically includes:

[0100] S31, input the final multimodal physical quantity data after generation and processing, including scattering intensity distribution I'(z), strain ε(z), temperature T(z), vibration V(z) and acoustic wave signal A(z), and construct a graph model of space and time relationship. The graph model is constructed through the spatial relationship graph G s and time relationship graph G t Describe the dependencies between data and use the weight matrix Ws and W t To characterize;

[0101] S32. Based on the constructed graph model, an automatic encoder is used to extract features from the input multimodal data. The encoder maps the processed final multimodal physical quantity data to the low-dimensional space Z. m , retaining the key feature information and converting the low-dimensional representation Z into m Reconstruct as output data Realize data reconstruction and dimensionality reduction processing;

[0102] S33, the graph regularization term L reg Introduced into the encoder, the spatial and temporal continuity of the data is maintained by minimizing the differences between adjacent nodes in space and the differences between adjacent nodes in time:

[0103]

[0104] Among them, L reg is the graph regularization term, W s (i, j) is the weight between positions i and j in space, ||Z i -Z j || 2 is the square of the Euclidean distance between spatial locations i and j in the high-dimensional representation Z, W t (m,n) is the weight between time points m and n, ||Z m -Z n || 2 is the square of the Euclidean distance between time points m and n in the high-dimensional representation Z;

[0105] S34. By training the graph regularized autoencoder, a high-dimensional dynamic monitoring model is constructed to identify early abnormal changes in geological disasters. During the model training process, the total loss function L is optimized to enable the model to adaptively adjust the parameter θ. e and θ d , adapted to different geological conditions:

[0106] L=L recon +λL reg ;

[0107] Among them, L recon is the reconstruction error, and λ is the hyperparameter that balances the two;

[0108] S35. During model training and application, the model parameters θ are adaptively adjusted according to the changes in real-time monitoring data under different geological conditions. e and θ d , optimize the adaptability and reliability of the monitoring model in various geological environments.

[0109] In this implementation, S4 specifically includes:

[0110] S41, the output data of the high-dimensional dynamic monitoring model As the input of the spatiotemporal transformer, the output data of the high-dimensional dynamic monitoring model Reflects the reconstruction results of multimodal physical quantity data after being processed by the graph regularized autoencoder, including geological environment characteristics at different time and space positions;

[0111] S42, in the space-time transformer, based on the input reconstruction data Construct a spatiotemporal dependency model, and the spatiotemporal transformer represents the features at time t and spatial position x Embedded into the spatiotemporal dependency matrix T s In (t,x):

[0112]

[0113] in, and Represent the dependency functions in time and space dimensions respectively;

[0114] S43, using the spatiotemporal transformer to generate the spatiotemporal dependency matrix T s (t, x), combined with real-time monitoring data, predict the time, location and severity of possible geological disasters, and generate a geological disaster risk prediction model P through comprehensive analysis of spatiotemporal characteristics. risk (t,x):

[0115] P risk (t,x)=σ(T s (t,x));

[0116] Among them, P risk (t,x) represents the probability of a geological disaster occurring at time t and spatial location x;

[0117] S44, based on risk prediction model P risk (t,x) and the severity of geological disasters S(t,x), and construct a dynamic risk assessment model R(t,x);

[0118] S45. Determine the geological disaster warning level based on the output of the dynamic risk assessment model R(t,x), and formulate corresponding emergency strategies. According to different risk levels, take corresponding emergency response measures to effectively prevent and respond to geological disasters.

[0119] In this implementation manner, S5 specifically includes:

[0120] S51, using the output of the generated dynamic risk assessment model R(t,x) as the input of an adaptive multi-level decision tree, wherein the dynamic risk assessment model R(t,x) comprehensively considers the probability and severity of occurrence of geological disasters and provides comprehensive risk analysis data as the basis of a multi-level decision tree early warning mechanism;

[0121] S52. In the adaptive multi-level decision tree, define the decision criteria D for each node j , where j represents the level of the decision tree and the decision criterion D j Based on the output value of the dynamic risk assessment model R(t,x) and the adaptively adjusted risk threshold θ j Make a judgment:

[0122]

[0123] Among them, θ j represents the risk threshold of the jth level;

[0124] S53, the multi-level decision tree early warning mechanism determines the risk level layer by layer, and adjusts the threshold θ in real time j , at each level, according to D j The output of the dynamic risk assessment model determines whether to trigger the corresponding emergency response system. When the output of the dynamic risk assessment model meets a certain level of decision criteria D j =1, the emergency response system at this level is triggered;

[0125] S54. Generate specific emergency measures suggestions based on the triggered emergency response system. j (t,x) is generated based on the real-time updated risk assessment model R(t,x) and the output of the decision tree:

[0126] S j (t,x)=g(D j ,R(t,x));

[0127] Among them, g(·) is the emergency measure generation function;

[0128] S55, the adaptive multi-level decision tree early warning mechanism continuously monitors the output of the risk assessment model R(t,x) and updates the decision threshold θ in real time j and emergency response strategies j (t,x).

[0129] In this implementation manner, S6 specifically includes:

[0130] S61, initializing multiple parameter sets {P1, P2, ..., P n}, each parameter set P iContains key configuration parameters of the monitoring network. The co-evolution mechanism optimizes the coverage and signal quality of the monitoring network by evolving parameter sets in parallel.

[0131] S62. In the dynamic environment genetic algorithm, for each parameter set P i Generate an initial population. The individuals in the population represent different distributed fiber optic sensor network configuration schemes. The fitness function of each individual is F(P i ), the fitness evaluation is based on network coverage, signal quality and adaptive capability, which enables the distributed fiber optic sensor network to achieve stable monitoring under complex geological conditions;

[0132] S63, dynamically adjust the parameter set P through the co-evolution mechanism i In the evolutionary process, the dynamic environment genetic algorithm updates the fitness function F(P) in real time based on environmental changes in each iteration. i ) parameters, enabling the algorithm to adaptively optimize the coverage and resource allocation of the monitoring network to cope with dynamic changes in geological conditions;

[0133] S64, in each iteration process, use selection, crossover and mutation operations to generate a new parameter set, and according to the fitness function F(P i ) value, select the optimal parameter set for the next generation of evolution;

[0134] S65, after multiple generations of iterations, select the fitness function F(P i ) The optimal parameter set P * , and apply it to distributed fiber optic sensor networks to optimize the configuration of the monitoring network in real time.

[0135] In this embodiment, the following modules are included:

[0136] Distributed fiber optic sensor network module: Deploy a distributed fiber optic sensor network to cover key risk paths and prone areas, and achieve all-round real-time monitoring of geological conditions;

[0137] Data acquisition and processing module: Process the collected multi-modal data to achieve long-distance monitoring with high spatial resolution;

[0138] Multimodal data analysis and modeling module: Analyze multimodal data, build high-dimensional dynamic monitoring models, identify early anomalies and adaptively adjust model parameters;

[0139] Spatiotemporal trend analysis and risk assessment module: Based on the high-dimensional dynamic monitoring model, it conducts spatiotemporal trend analysis, predicts the time, location and severity of geological disasters, and generates a risk assessment model;

[0140] Adaptive early warning and emergency response module: Based on the risk assessment model, an adaptive multi-level decision tree early warning mechanism is used to classify risk levels, trigger the corresponding emergency response system and generate emergency measures;

[0141] Monitoring network optimization and dynamic adjustment module: Combines the co-evolution mechanism and dynamic environment genetic algorithm to optimize monitoring network parameters in real time and adaptively adjust coverage and resource allocation;

[0142] Multi-source data fusion and feedback module: Comprehensively analyze monitoring data and external data sources to form comprehensive monitoring information and support dynamic adjustment of monitoring strategies.

[0143] Embodiment 1:

[0144] In order to verify the feasibility of the present invention in implementation, the present invention is applied to the embodiments of the present invention for landslide monitoring and early warning along a mountain highway. The highway passes through a mountainous area with complex geological conditions. It is affected by natural factors such as rainfall and earthquakes all year round, and there are many potential landslide hazards. Since the traditional point monitoring method is difficult to cover a large range of risk areas and cannot reflect the slight geological changes before the landslide in real time, it has resulted in the failure to issue timely early warnings when many landslide disasters occur, causing serious casualties and property losses. In order to improve the accuracy of landslide monitoring and the timeliness of early warning, the local government decided to use the distributed optical fiber sensing geological disaster monitoring and early warning system of the present invention to monitor the highway.

[0145] In the landslide-prone areas along the mountain highway, based on the geological survey results, a distributed fiber optic sensor network was first deployed along the key risk paths and landslide-prone areas of the highway. The fiber optic sensors were buried on the surface and inside the landslide body, and the changes in multimodal physical quantities such as strain, temperature, and vibration in the area were monitored in real time through the optical fibers laid along the route. After the system was started, the distributed fiber optic sensor network began to collect geological environmental data in the monitoring area in real time. Through the improved distributed optical frequency domain reflection technology, the collected optical signals were processed to form echo signals containing multimodal physical quantity data, and a dynamic noise filtering algorithm was used to ensure that high spatial resolution data was obtained in long-distance monitoring. These data were transmitted to the data processing center.

[0146] At the data processing center, the system uses a graph regularization autoencoder to conduct a comprehensive analysis of multimodal physical quantity data and construct a high-dimensional dynamic monitoring model that reflects the geological conditions along the highway. The model can identify early abnormal changes in the landslide body in real time, such as small strain accumulation, temperature fluctuations, or vibration changes. These data will be input into the space-time transformer for trend analysis. By analyzing the space-time dependencies in the monitoring area, the system can predict the specific time, location, and severity of the landslide, and generate a dynamic risk assessment model. When the system detects that a certain area is at high risk, the adaptive multi-level decision tree early warning mechanism will intelligently classify the risk level based on the dynamic risk assessment model. If the risk assessment result of a certain level exceeds the set threshold, the system will automatically trigger the corresponding emergency response system. The system will not only issue an early warning, but also generate specific emergency measures, such as timely evacuating vehicles and pedestrians along the highway, closing relevant sections, and dispatching rescue forces to the scene on standby.

[0147] To further ensure the efficient operation of the monitoring system, the system combines the co-evolution mechanism with the dynamic environment genetic algorithm to optimize the monitoring parameters of the distributed fiber optic sensor network in real time. As time goes by and environmental conditions change, the system will adaptively adjust the coverage and resource allocation of the monitoring network to ensure accurate monitoring and timely warning under dynamic geological conditions.

[0148] The effectiveness of the system has been verified in actual application. After the system was installed and debugged, three potential landslide events along the highway were successfully monitored within the first year of its official operation.

[0149] The first incident occurred in mid-June of a certain year. Continuous heavy rainfall caused obvious geological strain changes along a section of the highway. The system identified the abnormal strain accumulation inside the landslide body through the spectral regularization autoencoder in the early rainy season, and successfully predicted the time and place where the landslide might occur as the rainfall intensity gradually increased through trend analysis of the space-time transformer. Finally, the system issued a red warning signal two hours before the landslide, and the relevant departments quickly took measures to close the road and evacuate people, avoiding potential catastrophic accidents.

[0150] The second incident occurred in August of a certain year. Within a week after the earthquake, the system detected abnormal vibrations in a section of the highway after the earthquake wave passed. Combining historical data with real-time monitoring data, the system predicted that a landslide might occur in the area after the earthquake. Finally, within a day after the warning was issued, a small-scale landslide occurred in the area, but due to preventive measures taken in advance, the highway was not affected.

[0151] The third incident occurred in October of a certain year. An abnormal temperature change caused by nighttime cooling in an area along the highway triggered a system alarm. After further real-time monitoring and data analysis, the system predicted that the risk of landslide in this area was high. Under the guidance of the system's early warning, the local road management department carried out emergency reinforcement in the area and finally successfully avoided the occurrence of landslide.

[0152] After one year of operation, the system has issued 10 warnings of different levels, including 7 yellow warnings and 3 red warnings. All warnings have been responded to and verified in a timely manner, without any missed or false alarms. In addition, the system's real-time optimization mechanism also ensures that the coverage and resource allocation of the monitoring network can be adaptively adjusted under different seasons and environmental conditions, further improving the accuracy of monitoring and the reliability of warnings.

[0153] Table 1 Landslide monitoring event record

[0154]

[0155] Table 2 Early warning system performance and multi-source data fusion effect

[0156]

[0157] It can be seen from the above table that the performance and effect of the distributed optical fiber sensing geological disaster monitoring and early warning system of the present invention in practical application covers the record of landslide monitoring events and the performance of the early warning system and the effect of data fusion. The landslide monitoring event record table shows the system's monitoring and early warning of landslide risks along the highway under different time and environmental conditions. Through triggering factors such as continuous heavy rainfall, earthquakes, and sudden temperature drops, the system successfully predicted the landslide risk in advance and issued an early warning in time before the landslide occurred. The record shows that the system corresponds to different levels of risks through red and yellow warnings, and takes corresponding emergency measures, such as road closures, evacuation of personnel, and reinforcement of road sections, which effectively avoids casualties and property losses. The early warning system performance and multi-source data fusion effect table summarizes the overall early warning performance of the system and the effect of multi-source data fusion. Within one year, the system issued a total of 10 warnings, including 3 red warnings and 7 yellow warnings, without false alarms or missed reports. The early warning response time is 2 hours ahead of schedule on average, and the earliest warning is issued 1 day in advance. The emergency response execution rate reached 100%, successfully avoiding 3 major landslide disasters. By integrating with geological radar, satellite remote sensing and drone monitoring data, the system further enhances the comprehensiveness and accuracy of landslide monitoring, especially in sections B and C. The accuracy of geological disaster warning has been significantly improved through data fusion.

[0158] These tables reflect the efficiency, reliability and innovation of the present invention in practical applications. Through distributed optical fiber sensing technology and advanced data analysis methods, the system not only realizes wide-area, real-time and high-precision geological disaster monitoring, but also greatly improves the accuracy and timeliness of geological disaster warning through intelligent early warning mechanism and multi-source data fusion, demonstrating significant application value and social benefits.

[0159] In summary, the present invention aims to solve the problems existing in existing geological disaster monitoring methods, such as limited monitoring range, insufficient sensitivity, complex signal processing, insufficient data analysis capabilities, and lack of flexibility in network configuration. By introducing distributed optical fiber sensing technology and advanced data processing, analysis and optimization algorithms, the present invention provides a real-time, comprehensive and efficient geological disaster monitoring and early warning method, which significantly improves the accuracy and timeliness of geological disaster monitoring.

[0160] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A distributed optical fiber sensing geological disaster monitoring and early warning method, characterized in that: The steps include: S1. In the monitoring area, according to the changing characteristics of geological conditions and the risk distribution of potential geological disasters, a distributed optical fiber sensor network is deployed along key risk paths and prone areas to form a continuous, real-time multi-parameter monitoring chain to achieve all-round perception and coverage of the geological conditions in the entire monitoring area; S2. Use the improved distributed optical frequency domain reflectometry technology to process the multimodal physical quantity data collected in real time by the distributed optical fiber sensor network, and use the dynamic noise filtering algorithm to achieve sub-meter spatial resolution in long-distance monitoring; S3. Use graph regularization autoencoder to conduct comprehensive analysis on the processed multimodal physical quantity data, model the spatial and temporal relationship through graph regularization technology, build a high-dimensional dynamic monitoring model, identify early abnormal changes of geological disasters, and adaptively adjust model parameters according to different geological conditions; S4. Based on the output of the high-dimensional dynamic monitoring model, the space-time transformer is used to perform trend analysis, and the time, location and severity of possible geological disasters are predicted by combining the space-time dependence of real-time monitoring data. A dynamic risk assessment model is generated to guide the determination of warning levels and the formulation of emergency strategies. S5. After detecting potential geological disaster risks, an adaptive multi-level decision tree early warning mechanism is used based on a dynamic risk assessment model to intelligently classify risk levels, trigger emergency response systems at corresponding levels, and generate specific emergency measures to ensure timely response to disaster risks at different levels; S6. Combined with the dynamic environment genetic algorithm of the co-evolution mechanism, the monitoring parameters of the distributed fiber optic sensor network are optimized in real time. Through the co-evolution and interaction of different parameter sets, the coverage and resource allocation of the monitoring network are adaptively adjusted to cope with the dynamic changes of geological conditions and continuously optimize the monitoring performance. S7. Comprehensively analyze the monitoring data that have been collected, processed, analyzed and optimized with external data from geological radar, satellite remote sensing, and drone monitoring, integrate multi-source data to form comprehensive geological disaster monitoring information, and support dynamic adjustment of monitoring strategies during the early warning process to respond to different disaster types and evolution processes.

2. A distributed optical fiber sensing geological disaster monitoring and early warning method according to claim 1, characterized in that: The S2 specifically includes: S21, using improved distributed optical frequency domain reflection technology, using broadband light source to emit continuous optical signals The optical signal is transmitted through the optical fiber sensor network distributed in the monitoring area, and Rayleigh scattering occurs at various points in the optical fiber to form an echo signal containing multi-modal physical quantity data; S22, using broadband modulation technology to optimize the spatial resolution of the system, the frequency range covered by the broadband light source, through the formula Achieve sub-meter spatial resolution, where B is the bandwidth of the light source and c is the speed of light. By adjusting the bandwidth of the light source, the system can obtain spatial resolution in long-distance monitoring and adapt to the monitoring needs of complex geological conditions; S23, introduce phase modulation technology to optimize the signal's anti-interference ability, and extract useful spectrum information S(f) from the echo signal through phase demodulation technology: Among them, E mod (t) is the phase modulated optical signal, E0 is the amplitude of the optical signal, f0 is the center frequency of the optical signal, t is the time, is the phase noise term, θ mod (t) is the phase modulation term; Through the phase modulation and demodulation process, the nonlinear effect interference is suppressed and the optimized spectrum information S(f) is obtained: in, is the Fourier transform operator, E demod (t) is the phase demodulated signal; S24, collect the echo signal, and use the frequency modulation continuous wave principle to generate a difference frequency signal S by interfering the emitted light signal with the echo signal. df (f) and optimizing the signal-to-noise ratio and detection sensitivity using photodetectors; S25, the obtained difference frequency signal S df (f) Performing Fourier transform processing to generate the scattering intensity distribution I(z) at different positions in the optical fiber, applying a dynamic noise filtering algorithm, adjusting the frequency response H(f) in real time through an adaptive filter, optimizing the signal-to-noise ratio of the signal, and generating the final scattering intensity distribution I'(z); S26, fusing and analyzing the processed scattering intensity distribution I'(z) with the multimodal physical quantity data collected in real time by the distributed optical fiber sensor network, and adjusting the spatial resolution according to the improved light source bandwidth B to generate the final processed multimodal physical quantity data.

3. A distributed optical fiber sensing geological disaster monitoring and early warning method according to claim 1, characterized in that: The S3 specifically includes: S31, input the final multimodal physical quantity data after generation and processing, including scattering intensity distribution I'(z), strain ε(z), temperature T(z), vibration V(z) and acoustic wave signal A(z), and construct a graph model of space and time relationship. The graph model is constructed through the spatial relationship graph G s and time relationship graph G t Describe the dependencies between data and use the weight matrix W s and W t To characterize; S32. Based on the constructed graph model, an automatic encoder is used to extract features from the input multimodal data. The encoder maps the processed final multimodal physical quantity data to the low-dimensional space Z. m , retaining the key feature information and converting the low-dimensional representation Z into m Reconstruct as output data Realize data reconstruction and dimensionality reduction processing; S33, the graph regularization term L reg Introduced into the encoder, the spatial and temporal continuity of the data is maintained by minimizing the differences between adjacent nodes in space and the differences between adjacent nodes in time: Among them, L reg is the graph regularization term, W s (i, j) is the weight between positions i and j in space, ∥Z i -Z j ∥ 2 is the square of the Euclidean distance between spatial locations i and j in the high-dimensional representation Z, W t (m,n) is the weight between time points m and n, ∥Z m -Z n ∥ 2 is the square of the Euclidean distance between time points m and n in the high-dimensional representation Z; S34. By training the graph regularized autoencoder, a high-dimensional dynamic monitoring model is constructed to identify early abnormal changes in geological disasters. During the model training process, the total loss function L is optimized to enable the model to adaptively adjust the parameter θ. e and θ d , adapted to different geological conditions: L=L recon +λL reg ; Among them, L recon is the reconstruction error, and λ is the hyperparameter that balances the two; S35. During model training and application, the model parameters θ are adaptively adjusted according to the changes in real-time monitoring data under different geological conditions. e and θ d , optimize the adaptability and reliability of the monitoring model in various geological environments.

4. A distributed optical fiber sensing geological disaster monitoring and early warning method according to claim 1, characterized in that: The S4 specifically includes: S41, the output data of the high-dimensional dynamic monitoring model As the input of the spatiotemporal transformer, the output data of the high-dimensional dynamic monitoring model Reflects the reconstruction results of multimodal physical quantity data after being processed by the graph regularized autoencoder, including geological environment characteristics at different time and space positions; S42, in the space-time transformer, based on the input reconstruction data Construct a spatiotemporal dependency model, and the spatiotemporal transformer represents the features at time t and spatial position x Embedded into the spatiotemporal dependency matrix T s In (t,x): in, and Represent the dependency functions in time and space dimensions respectively; S43, using the spatiotemporal transformer to generate the spatiotemporal dependency matrix T s (t, x), combined with real-time monitoring data, predict the time, location and severity of possible geological disasters, and generate a geological disaster risk prediction model P through comprehensive analysis of spatiotemporal characteristics. risk (t,x): P risk (t,x)=σ(Ts(t,x)); Among them, P risk (t,x) represents the probability of a geological disaster occurring at time t and spatial location x; S44, based on risk prediction model P risk (t,x) and the severity of geological disasters S(t,x), and construct a dynamic risk assessment model R(t,x); S45. According to the output of the dynamic risk assessment model R(t,x), determine the geological disaster warning level and formulate corresponding emergency strategies. According to different risk levels, take corresponding emergency response measures to effectively prevent and respond to geological disasters.

5. A distributed optical fiber sensing geological disaster monitoring and early warning method according to claim 1, characterized in that: The S5 specifically includes: S51, using the output of the generated dynamic risk assessment model R(t,x) as the input of an adaptive multi-level decision tree, wherein the dynamic risk assessment model R(t,x) comprehensively considers the probability and severity of occurrence of geological disasters and provides comprehensive risk analysis data as the basis of a multi-level decision tree early warning mechanism; S52. In the adaptive multi-level decision tree, define the decision criteria D for each node j , where j represents the level of the decision tree and the decision criterion D j Based on the output value of the dynamic risk assessment model R(t,x) and the adaptively adjusted risk threshold θ j Make a judgment: Among them, θ j represents the risk threshold of the jth level; S53, the multi-level decision tree early warning mechanism determines the risk level layer by layer, and adjusts the threshold θ in real time j , at each level, according to D j The output of the dynamic risk assessment model determines whether to trigger the corresponding emergency response system. When the output of the dynamic risk assessment model meets a certain level of decision criteria D j =1, the emergency response system at this level is triggered; S54. Generate specific emergency measures suggestions based on the triggered emergency response system. j (t,x) is generated based on the real-time updated risk assessment model R(t,x) and the output of the decision tree: S j (t,x)=g(D j ,R(t,x)); Among them, g(·) is the emergency measure generation function; S55, the adaptive multi-level decision tree early warning mechanism continuously monitors the output of the risk assessment model R(t,x) and updates the decision threshold θ in real time j and emergency response strategies j (t,x).

6. A distributed optical fiber sensing geological disaster monitoring and early warning method according to claim 1, characterized in that: The S6 specifically includes: S61, initializing multiple parameter sets {P1, P2, ..., P n }, each parameter set P i Contains key configuration parameters of the monitoring network. The co-evolution mechanism optimizes the coverage and signal quality of the monitoring network by evolving parameter sets in parallel. S62. In the dynamic environment genetic algorithm, for each parameter set P i Generate an initial population. The individuals in the population represent different distributed fiber optic sensor network configuration schemes. The fitness function of each individual is F(P i ), the fitness evaluation is based on network coverage, signal quality and adaptive capability, which enables the distributed fiber optic sensor network to achieve stable monitoring under complex geological conditions; S63, dynamically adjust the parameter set P through the co-evolution mechanism i In the evolutionary process, the dynamic environment genetic algorithm updates the fitness function F(P) in real time based on environmental changes in each iteration. i ) parameters, enabling the algorithm to adaptively optimize the coverage and resource allocation of the monitoring network to cope with dynamic changes in geological conditions; S64, in each iteration process, use selection, crossover and mutation operations to generate a new parameter set, and according to the fitness function F(P i ) value, select the optimal parameter set for the next generation of evolution; S65, after multiple generations of iterations, select the fitness function F(P i ) The optimal parameter set P * , and apply it to distributed fiber optic sensor networks to optimize the configuration of the monitoring network in real time.

7. A distributed optical fiber sensing geological disaster monitoring and early warning system, characterized in that: Includes the following modules: Distributed fiber optic sensor network module: Deploy a distributed fiber optic sensor network to cover key risk paths and prone areas, and achieve all-round real-time monitoring of geological conditions; Data acquisition and processing module: Process the collected multi-modal data to achieve long-distance monitoring with high spatial resolution; Multimodal data analysis and modeling module: Analyze multimodal data, build high-dimensional dynamic monitoring models, identify early anomalies and adaptively adjust model parameters; Spatiotemporal trend analysis and risk assessment module: Based on the high-dimensional dynamic monitoring model, it conducts spatiotemporal trend analysis, predicts the time, location and severity of geological disasters, and generates a risk assessment model; Adaptive early warning and emergency response module: Based on the risk assessment model, an adaptive multi-level decision tree early warning mechanism is used to classify risk levels, trigger the corresponding emergency response system and generate emergency measures; Monitoring network optimization and dynamic adjustment module: Combines the co-evolution mechanism and dynamic environment genetic algorithm to optimize monitoring network parameters in real time and adaptively adjust coverage and resource allocation; Multi-source data fusion and feedback module: Comprehensively analyze monitoring data and external data sources to form comprehensive monitoring information and support dynamic adjustment of monitoring strategies.

Citation Information

Cited By

  • Lower reservoir high slope landslide vibration monitoring and early warning system

    CN119992796A

  • Geological disaster monitoring method and system based on distributed optical fiber vibration sensing system

    CN120195722A

  • Solid waste storage yard safety risk early warning method

    CN120542256A

  • System and method for detecting alien species intrusion

    CN120564387A

  • Multi-source sensing fusion highway slope geological disaster real-time early warning method

    CN120636106A