Landslide hidden danger point displacement monitoring system using multi-source data fusion

Abstract

The invention discloses a landslide hidden danger point displacement monitoring system applying multi-source data fusion, and the system comprises a multi-source sensing unit which comprises a spaceborne radar sensing module which is used for obtaining large-range ground surface deformation data through a synthetic aperture radar; the earth surface deformation sensing module is used for monitoring earth surface displacement in real time through a double-frequency GNSS receiver and an optical fiber grating sensor; the underground stress sensing module is used for acquiring the stress change of an underground rock-soil body through a micro-seismic monitoring array and a distributed optical fiber sensor; and the edge calculation unit comprises a dynamic weight fusion module which is used for dynamically distributing the weight of the multi-source sensor according to the geological state and generating high-precision fusion displacement field data. According to the invention, through multi-source data fusion, model parameter real-time updating and a closed-loop optimization mechanism, bottlenecks of a traditional monitoring system in aspects of data quality, model precision and response efficiency are broken through, and data support is provided for accurate early warning and intelligent emergency of landslide disasters.

Description

Technical Field

[0001] The present invention relates to the field of monitoring systems, and particularly to a landslide hidden danger point displacement monitoring system using multi-source data fusion. Background Art

[0002] Landslides are common geological disasters that pose a serious threat to people's lives, property and infrastructure. Displacement monitoring of landslide hidden danger points is a key link in geological disaster monitoring and early warning. The core is to use technical means to track the deformation dynamics of potential landslide areas in real time, providing a scientific basis for disaster prediction, risk assessment and emergency response; Existing monitoring systems attempt to improve monitoring capabilities through multi-sensor integration, but there are still problems such as rigid data fusion, fixed model parameters, single warning mechanism and inefficient emergency decision-making. Therefore, a landslide hidden danger point displacement monitoring system using multi-source data fusion is proposed. Summary of the Invention

[0003] The technical problem to be solved by the present invention is: how to solve the problems that existing monitoring systems attempt to improve monitoring capabilities through multi-sensor integration, but still have problems such as rigid data fusion, fixed model parameters, single warning mechanism and inefficient emergency decision-making, and provides a landslide hidden danger point displacement monitoring system using multi-source data fusion.

[0004] The present invention solves the above technical problems through the following technical solutions. The present invention includes: A multi-source sensing unit, the multi-source sensing unit includes: A spaceborne radar sensing module for obtaining large-scale surface deformation data through synthetic aperture radar (InSAR); A surface deformation sensing module for real-time monitoring of surface displacement through a dual-frequency GNSS receiver and a fiber Bragg grating sensor; An underground stress sensing module for obtaining underground rock and soil stress changes through a microseismic monitoring array and a distributed fiber optic sensor; An edge computing unit, the edge computing unit includes: A dynamic weight fusion module for dynamically allocating weights of multi-source sensors according to the geological state to generate high-precision fused displacement field data; A communication protocol adaptation module for realizing protocol conversion and low-latency transmission of multi-source sensing data; A digital twin modeling unit, the digital twin modeling unit includes: A geomechanical parameterized model for constructing a three-dimensional mechanical model of a landslide body based on the finite element method; A real-time data assimilation module for inputting the fused data into the model to invert and update geomechanical parameters; A model self-calibration module for optimizing the model accuracy through parameter perturbation tests; Intelligent early warning unit, which includes: Multi-level early warning trigger module, used to trigger hierarchical early warnings according to abnormal displacement rate, acceleration and underground parameters; Three-dimensional path planning module, used to dynamically plan a safe evacuation path based on the digital twin prediction results.

[0005] Furthermore, the dynamic weight fusion module performs the following operations: Obtain the real-time measurement data of each sensing module and its confidence parameter, and the confidence parameter is jointly determined according to the sensor noise variance and historical error statistics; Calculate the geological state evaluation coefficient , and its expression is: ; In the formula, is the dip angle change influence function, is the real-time dip angle change amount, is the critical dip angle threshold (related to the geological type), is the weight coefficient of the dip angle change, is the displacement rate influence function, is the real-time displacement rate, Critical rate threshold, is the weight coefficient of the displacement rate Dynamically allocate the fusion weights of each sensing module according to the value, and the weight calculation formula is: ; In the formula, is the fusion weight of the i-th sensing module, is the geological state evaluation coefficient, is the confidence parameter of the i-th sensing module, is the total number of sensing modules.

[0006] Furthermore, the dip angle change influence function and the displacement rate influence function are defined as: ; ; Among them, is to prevent zero and tiny amounts, and its value is positively correlated with the sensor measurement accuracy.

[0007] Furthermore, the weight coefficient and are dynamically configured through the geological knowledge base, which specifically includes the following steps: Step 1: Knowledge base construction: Establish the correlation between different geological types and weight coefficients based on geotechnical mechanics theory and historical landslide data; The geological types at least include typical landslide body types, and the typical landslide body types include clayey, sandy, rocky and accumulation layers; Weight coefficient 、 The allocation principle is: Tendency geology (such as clayey): Assign a higher weight to the inclination change ( > ); Rate-sensitive geology (such as rocky): Assign a higher weight to the displacement rate ( > ); Step 2: On-site geological type identification and multi-source data collection: Obtain the physical parameters of the rock and soil mass (such as water content, particle composition, shear strength) through on-site drilling sampling, geological radar detection, and UAV multi-spectral imaging; Input the collected data into a pre-trained geological classification model (such as random forest or convolutional neural network) to output the geological type label of the target area; Step 3: Dynamic loading of weight coefficients. According to the geological type label output in Step 2, retrieve the corresponding 、 values in the knowledge base; Inject the matched weight coefficients into the dynamic weight fusion module in real time to update the calculation logic of the geological state evaluation coefficient ; Step 4: Abnormal handling mechanism. If the on-site geological type is not included in the mapping table, enable the default weights ( = 0.5, = 0.5) and trigger the manual verification process; After that, verify the rationality of the weight configuration through historical data playback. If the deviation between the fusion results and the measured data exceeds the limit for N consecutive times, trigger the update of the knowledge base.

[0008] Furthermore, the real-time data assimilation module executes: Step 1: Receive the fused displacement field data sent by the edge computing unit ; Step 2: Input into the geomechanical parameterization model for inversion calculation to update the slip zone parameter set , where is the shear modulus, is the pore pressure coefficient, is the cohesion; Step 3: Calculate the parameter change rate , when Trigger the model self - calibration module when is a preset sensitivity threshold; is the change rate of the slip zone parameter set, is the updated slip zone parameter set (including shear modulus , pore pressure coefficient , cohesion ), is the slip zone parameter set before update, is a preset sensitivity threshold.

[0009] Furthermore, the inversion calculation adopts the regularized adjoint gradient method, specifically including: Construct the objective function: where, is the finite - element forward operator, is the initial parameter, is the regularization coefficient; By solving the adjoint equation: ; Obtain the parameter gradient ▽J, where is the stiffness matrix, is the adjoint variable; Use the quasi - Newton method to iteratively update the parameters until convergence.

[0010] Furthermore, the model self - calibration module performs the following operations: Save the current parameter snapshot to the historical database; Generate a parameter perturbation set , = 1, 2, 3... m; Parallel - compute the predicted displacement fields corresponding to each perturbed parameter ; Select the parameter group with the highest matching degree with the actual observed data within the subsequent Δt time period as the update benchmark: ; In the formula, is the predicted displacement field of the k - th group of perturbed parameters at time t; is the actual observed displacement data at time t; T is the verification time window length.

[0011] Furthermore, the dynamic weight fusion module and the real - time data assimilation module form a two - way coupling mechanism, specifically manifested as: When the model self - calibration module updates the slip zone parameter set After that, it automatically triggers the recalculation of the geological state evaluation coefficient where the critical dip angle threshold and the critical rate threshold are dynamically adjusted according to the new value: ; ; wherein, and are empirical functions based on geomechanics, is the rock mass density.

[0012] Furthermore, the three-dimensional path planning module performs: Receiving the landslide impact boundary output by the digital twin modeling unit, and then constructing a three-dimensional terrain cost function: ; wherein, is the slope factor, is the surface crack density, is the distance between the current position and the landslide boundary; Using the improved A* algorithm to search for the minimum cost path from the personnel position to the safe area, and the path weight coefficients and and are dynamically adjusted according to the warning level.

[0013] The present invention has the following advantages compared with the prior art: The landslide hidden danger point displacement monitoring system using multi-source data fusion conducts collaborative monitoring of multi-source data and comprehensively improves the monitoring accuracy Through multi-dimensional data collection of satellite-borne radar, surface deformation sensors and underground stress sensors, a comprehensive perception of surface deformation and underground stress changes at landslide risk points is achieved, eliminating the limitations of a single data source. Dynamic weight fusion technology adaptively adjusts sensor weights according to geological conditions, effectively suppresses noise interference, generates high-confidence fusion data, and provides reliable input for subsequent analysis. Based on the geological knowledge base, the weight coefficient is dynamically configured, enabling the system to automatically adapt to different geological types and improve monitoring sensitivity and pertinence. The digital twin model updates geomechanical parameters through real-time data inversion to ensure that the model evolves synchronously with actual geological conditions and avoids prediction deviations caused by parameter solidification. The model self-correction module quickly optimizes the model through parameter perturbation testing and parallel calculation. The model accuracy is improved, and the system's adaptability to complex geological environments is enhanced. The multi-level early warning trigger mechanism combines multi-dimensional indicators such as displacement rate, acceleration and underground parameter anomalies to achieve graded early warning, significantly reducing the risk of false alarms and missed alarms. The three-dimensional path planning module dynamically generates safe evacuation paths based on the landslide impact range predicted by the digital twin, comprehensively considering the slope, crack density and real-time danger distance to improve emergency evacuation efficiency and safety. The improved path search algorithm ensures that the planned path is updated in real time to avoid sudden dangerous areas. The two-way coupling mechanism of dynamic weight fusion and model parameter update forms a closed-loop optimization of data and model, continuously improving monitoring and prediction accuracy. The abnormality handling mechanism ensures the robustness of the system under unknown geological conditions and reduces the need for manual intervention. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 It is a system block diagram of the present invention. DETAILED DESCRIPTION

[0015] The following is a detailed description of an embodiment of the present invention. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process. However, the protection scope of the present invention is not limited to the following embodiment.

[0016] like Figure 1 As shown, this embodiment provides a technical solution: a landslide potential point displacement monitoring system using multi-source data fusion, comprising: Multi-source sensing unit, the multi-source sensing unit includes: Spaceborne radar sensing module, used to obtain large-scale surface deformation data through synthetic aperture radar (InSAR); A surface deformation sensing module, which is used to monitor surface displacement in real time using a dual-frequency GNSS receiver and a fiber Bragg grating sensor; The underground stress sensing module is used to obtain the stress changes of underground rock and soil through microseismic monitoring arrays and distributed optical fiber sensors; Edge computing unit, the edge computing unit includes: The dynamic weight fusion module is used to dynamically allocate the weights of multi-source sensors according to the geological state and generate high-precision fused displacement field data; The communication protocol adaptation module is used to implement protocol conversion and low-latency transmission of multi-source sensing data; The digital twin modeling unit, the digital twin modeling unit includes: The geomechanical parameterization model is used to construct a three-dimensional mechanical model of the landslide body based on the finite element method; The real-time data assimilation module is used to input the fused data into the model to invert and update the geomechanical parameters; The model self-correction module is used to optimize the model accuracy through parameter perturbation testing; The intelligent warning unit, the intelligent warning unit includes: The multi-level warning trigger module is used to trigger hierarchical warnings according to displacement rate, acceleration, and abnormal underground parameters; The three-dimensional path planning module is used to dynamically plan a safe evacuation path based on the digital twin prediction results.

[0017] Furthermore, the dynamic weight fusion module performs the following operations: Obtain the real-time measurement data and its confidence parameter of each sensing module, and the confidence parameter is jointly determined according to the sensor noise variance and historical error statistics; Calculate the geological state evaluation coefficient , and its expression is: ; In the formula, is the inclination change influence function, is the real-time inclination change amount, is the critical inclination threshold (related to the geological type), is the weight coefficient of the inclination change, is the displacement rate influence function, is the real-time displacement rate, Critical rate threshold, is the weight coefficient of the displacement rate; According to value, dynamically allocate the fusion weights of each sensing module, and the weight calculation formula is: ; In the formula, is the fusion weight of the i-th sensing module, is the geological state evaluation coefficient, is the confidence parameter of the i-th sensing module, is the total number of sensing modules; The above process automatically reduces the weight of high-noise sensors by calculating confidence parameters based on sensor noise variance and historical error statistics, suppresses the impact of data anomalies on the fusion result, dynamically adjusts the weight distribution in combination with the geological state evaluation coefficient, makes the fusion process more in line with the actual geological conditions, avoids insufficient sensitivity or misjudgment caused by fixed weights, and captures the deformation trend (dip angle change) and evolution speed (displacement rate) of the landslide body simultaneously through the coupled calculation of the dip angle change influence function and the displacement rate influence function, enhances the recognition ability of landslide stages (such as slow creep, accelerating sliding), and the weight coefficient is dynamically bound to the geological type (e.g., clay focuses on dip angle, rock focuses on displacement rate), enabling the system to automatically adapt to different landslide mechanisms and improving the monitoring pertinence.

[0018] For example, the monitoring of a potential rock landslide point The sensor data is from spaceborne radar ( =0.9): The monthly average displacement rate is monitored =8mm, and the noise is relatively low.

[0019] Surface GNSS ( =0.7): The detected dip angle change =2°, but the noise is relatively high due to weather interference.

[0020] Underground microseismic ( =0.8): The rock mass fracture signal is captured, and the displacement rate =5mm; Dynamic weight distribution, the geological type is rock, and the loaded weights =0.7, =0.3.

[0021] Calculate the geological state evaluation coefficient : =0.35.

[0022] Composite weight calculation: , =28%, =32%; Effect: The weight ratio of the displacement rate data (spaceborne radar, microseismic) is higher (72%), accurately reflecting the rate-sensitive characteristics of rock landslides; the weight of the dip angle data is reduced, avoiding noise interference.

[0023] The dip angle change influence function and the displacement rate influence function are defined as: ; ; Among them, To eliminate zero infinitesimals, its value is positively correlated with the measurement accuracy of the sensor.

[0024] By introducing infinitesimals , to avoid the situation where the denominator of the inclination angle change function becomes zero at ≈ , ensuring stable calculation. The value of

[0025] is positively correlated with the sensor accuracy, automatically adapts to the measurement errors of different sensors, and reduces the influence of noise on the fusion result. reflects linearly the influence of the inclination angle change on stability. When is close to the critical value

[0026] , the function value decreases significantly, warning of potential risks in a timely manner. The exponential decay form sensitively captures high-rate displacements and is more sensitive to the acceleration stage of landslides, avoiding the warning delay that may be caused by linear functions. , is directly related to the geological type, facilitating engineers to adjust the threshold parameters according to the on-site conditions.

[0027] For example, during the monitoring of a potential clay landslide point: The critical inclination angle threshold = 4°, and the sensor accuracy is 0.05°, so ε = 0.05.

[0028] The critical rate threshold = 6 mm / day.

[0029] Example 1: Calculation of the inclination angle change function, real-time inclination angle change = 3.8°: ; Effect: When the inclination angle is close to the critical value, the function value drops sharply, triggering a warning.

[0030] Example 2: Calculation of the displacement rate function Real-time displacement rate = 5.5 mm / day: ; Effect: When the rate is close to the critical value, the function value decreases significantly, reflecting the acceleration trend.

[0031] The weight coefficient and are dynamically configured through the geological knowledge base, specifically including the following steps: Step 1: Knowledge base construction: According to the geomechanics theory and historical landslide data, establish the correlation between different geological types and weight coefficients; The geological types at least include typical landslide types, and the typical landslide types include clay, sand, rock and accumulation layers; Weight coefficient 、 The allocation principle is as follows: Tendency geology (such as clay): Assign a higher weight to the inclination change ( > ); Rate-sensitive geology (such as rock): Assign a higher weight to the displacement rate ( > ); Step 2: On-site geological type identification and multi-source data collection: Through on-site borehole sampling, ground penetrating radar detection, and UAV multispectral imaging, obtain geotechnical physical parameters (such as water content, particle composition, shear strength); Input the collected data into a pre-trained geological classification model (such as random forest or convolutional neural network) to output the geological type label of the target area; Step 3: Dynamic loading of weight coefficients. According to the geological type label output in Step 2, retrieve the corresponding 、 values in the knowledge base; Inject the matched weight coefficients into the dynamic weight fusion module in real time to update the calculation logic of the geological state evaluation coefficient ; Step 4: Abnormal handling mechanism. If the on-site geological type is not included in the mapping table, enable the default weight ( = 0.5, = 0.5) and trigger the manual verification process; After that, verify the rationality of the weight configuration through historical data playback. If the deviation between the fusion results and the measured data exceeds the limit for N consecutive times, trigger the update of the knowledge base; The self-adaptive optimization of geological types dynamically matches weight coefficients through the geological knowledge base, enabling the data fusion process to automatically adapt to different landslide mechanisms (such as clay landslides focusing on inclination changes and rock landslides focusing on displacement rates), improving the monitoring pertinence.

[0032] Avoid problems of insufficient sensitivity or misjudgment caused by manually preset fixed weights. The geological classification model (such as random forest) quickly identifies the on-site geological type based on multi-source exploration data (drilling, radar, UAV), reducing the manual exploration time. The knowledge base mapping table transforms geomechanics theory into executable parameter rules to ensure the physical rationality of weight allocation.

[0033] The exception handling mechanism (such as default weights and manual verification) ensures the stable operation of the system even when dealing with unknown geological types or missing exploration data. The historical data playback verification automatically triggers the update of the knowledge base, promoting the continuous optimization of the system.

[0034] The real-time data assimilation module performs the following: Step 1: Receive the fused displacement field data sent by the edge computing unit ; Step 2: Input the geomechanical parameterization model for inversion calculation to update the slip zone parameter set , where is the shear modulus, is the pore pressure coefficient, is the cohesion; Step 3: Calculate the parameter change rate , when it triggers the model self-correction module, where is the preset sensitivity threshold; is the change rate of the slip zone parameter set, is the updated slip zone parameter set (including the shear modulus , the pore pressure coefficient , the cohesion ), is the slip zone parameter set before update, is the preset sensitivity threshold; Real-time parameter update to improve model accuracy By inversely updating the geomechanical parameters (such as shear modulus μ, pore pressure coefficient λ, cohesion c) through the fused displacement field data, it ensures the dynamic synchronization of the digital twin model with the real geological conditions, avoiding prediction deviations caused by the fixed parameters of traditional models. The calculation of the parameter change rate can quantify the degree of model deviation, trigger the self-correction mechanism in a timely manner, prevent error accumulation, and the update of the slip zone parameter set directly reflects the changes in the internal mechanical properties of the landslide body (such as slip zone softening and pore water pressure increase), providing a physical basis for early warning decision-making. The setting of the sensitivity threshold η enables the system to distinguish normal fluctuations from significant anomalies, balancing the response speed and stability.

[0035] The regularized adjoint gradient method is used to accelerate parameter inversion, reducing the computational cost while ensuring accuracy, and is suitable for edge-cloud collaborative deployment.

[0036] For example, the monitoring of a potential soil landslide point Initial state:

[0037] Preset sensitivity threshold = 5%; First, receive the fusion data, and the edge computing unit sends the fusion displacement field data , indicating that the displacement rate in the toe region increases abnormally.

[0038] Inverse calculation updates the parameters, and the real-time data assimilation module inversely obtains new parameters: ; Calculate the real-time change rate, ; Because X > Y , trigger the model self-correction module; Model self-correction, save the current parameter snapshot = Z ; Generate a set of perturbation parameters ± A , and calculate the predicted displacement field in parallel; Select the parameter group that best matches the measured data in the next 2 hours as the new benchmark model.

[0039] The inverse calculation adopts the regularized adjoint gradient method, which specifically includes: Construct the objective function:

[0040] Among them, F is the finite element forward operator, m0 is the initial parameter, α is the regularization coefficient; By solving the adjoint equation: ; Obtain the parameter gradient ▽J, where K is the stiffness matrix, p is the adjoint variable; Adopt the quasi-Newton method to iteratively update the parameters until convergence; Efficient gradient calculation, significantly reducing the calculation cost, The adjoint method can obtain the complete gradient of the objective function through one forward calculation (solving the finite element equation) and one adjoint calculation (solving the adjoint equation), avoiding the huge calculation amount of multiple forward calculations required by the traditional finite difference method. Regularization ensures the stability of the inversion. A regularization term is introduced into the objective function to constrain the parameter update direction, preventing parameter drift (such as physically unreasonable values like negative shear modulus) caused by observation data noise or model errors. By balancing the data fitting term and the parameter prior information, overfitting is suppressed, and the generalization ability of the model is improved.

[0041] The quasi - Newton method uses gradient information for efficient iterative optimization, accelerates convergence to the global optimal solution, ensures the physical rationality of inversion parameters, accurately calculates the gradient with the adjoint equation, avoids the truncation error of the numerical difference method, improves the accuracy of the parameter update direction, adapts to complex non - linear problems, and can handle strong non - linear relationships in geomechanical models (such as the coupling effect between pore pressure and displacement rate), and is applicable to the parameter inversion requirements of different types of landslide bodies such as soil and rock masses; For example, the shear modulus of a certain landslide body inversion initial parameter = 50 MPa, and the regularization coefficient γ = 0.1.

[0042] Observation data: The integrated displacement field shows local displacement anomalies in the slope body.

[0043] After that, construct the objective function, ; Forward - calculate the displacement field = , solve the adjoint equation, ; Calculate the gradient ; The quasi - Newton method iterates along the gradient direction and finally converges to = 42 MPa, which is in good agreement with the borehole test result (41 MPa).

[0044] The model self - calibration module performs the following operations: Save the current parameter snapshot to the historical database; Generate a set of parameter perturbations , = 1, 2, 3... m; Parallel - calculate the predicted displacement fields corresponding to each perturbed parameter ; Select the parameter group with the highest matching degree with the actual observation data in the subsequent Δt time period as the update benchmark: ; In the formula, is the predicted displacement field of the k - th group of perturbed parameters at time t; is the actual observed displacement data at time t; [[ID=6�]]T is the length of the verification time window; By generating multiple sets of parameter perturbation combinations, different regions of the parameter space are explored to avoid the traditional optimization algorithm from falling into local optimal solutions, ensuring the global optimality of model parameter updates. Meanwhile, the predicted displacement fields of multiple sets of parameters are calculated, significantly shortening the parameter screening time to meet the real-time monitoring requirements. The predicted results of the perturbation parameters are verified through subsequent observation data, and the parameter group that best matches the actual landslide evolution is selected to ensure the dynamic synchronization of model parameters with the real geological conditions. Parameter snapshots are saved to form a historical database, providing data support for long-term landslide mechanism research. Parameter diversity coverage: Through the parallel testing of multiple sets of perturbation parameters, the system can adapt to sudden changes in the internal mechanical properties of the landslide body (such as slip zone fracture and sudden rise in pore water pressure), enhancing the model's adaptability to complex geological conditions. The parameter group with the highest matching degree is directly fed back to the digital twin model to optimize the subsequent prediction accuracy, forming a closed-loop of prediction, verification, and optimization.

[0045] The dynamic weight fusion module and the real-time data assimilation module form a two-way coupling mechanism, which is specifically manifested as: When the model self-correction module updates the slip zone parameter set it automatically triggers the recalculation of the geological state evaluation coefficient where the critical dip threshold and the critical rate threshold are dynamically adjusted according to the new value: ; ; where , are empirical functions based on geotechnical mechanics, and is the rock mass density; Dynamic closed-loop optimization to enhance the system's self-adaptability. After the model self-correction module updates the slip zone parameter set, it automatically triggers the recalculation of the geological state evaluation coefficient, realizing a closed-loop process from data fusion to model update to parameter feedback to re-optimization. The critical dip threshold and the critical rate threshold are dynamically adjusted according to the updated parameters, enabling the warning conditions to evolve synchronously with the actual geological state, avoiding misjudgments caused by static thresholds. By recalculating using the updated Ψ value, the weight allocation of sensors is dynamically adjusted, giving priority to key parameters (such as displacement rate or inclination change). The empirical functions , are designed based on geotechnical mechanics theory to ensure the physical rationality of threshold adjustment, enhancing the response speed to the landslide acceleration stage. High-precision model parameters provide physical constraints for data fusion, and high-quality fused data feeds back to model optimization, forming a positive cycle. The automated closed-loop mechanism reduces the dependence on manual parameter tuning and adapts to the long-term monitoring requirements in complex geological environments.

[0046] The 3D path planning module performs: Receive the landslide impact boundary output by the digital twin modeling unit and then construct a three-dimensional terrain cost function: ; in, is the slope factor, is the surface crack density, is the distance between the current position and the landslide boundary; The improved A* algorithm is used to search for the minimum cost path from the personnel location to the safe zone, and the path weight coefficient 、 and Dynamic adjustment according to the warning level; Real-time landslide boundary prediction: Dynamically update evacuation routes based on the landslide impact boundary output by the digital twin modeling unit to avoid real-time evolving danger zones and ensure the timeliness and safety of escape routes.

[0047] Multi-factor comprehensive assessment: The cost function integrates slope, surface crack density, and distance to the landslide boundary to avoid path risks caused by a single factor.

[0048] Adaptive warning level response, dynamic weight adjustment: path weight coefficient 、 and Automatically adjust according to the warning level. For example: During a red alert, avoid the landslide front first (increase weight); During a yellow warning, choose a low-slope route (increase weight).

[0049] Hierarchical risk avoidance strategy: Different warning levels match different path optimization goals to balance safety and traffic efficiency.

[0050] Based on the traditional A* algorithm, a dynamic update mechanism of the cost function is introduced to quickly respond to terrain changes, reduce calculation delays, and meet the real-time requirements of emergency scenarios. Based on the high-precision three-dimensional terrain data of digital twins, it supports path search in complex terrain (such as cliffs and valleys) and avoids the limitations of two-dimensional plane planning.

[0051] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of the technical features being referred to. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. In the description of the present invention, "plurality" means at least two, such as two, three, etc., unless otherwise specifically defined.

[0052] In the description of this specification, the descriptions referring to terms such as "one embodiment", "some embodiments", "examples", "specific examples", or "some examples", etc., mean that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, without contradiction, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0053] Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention.

Claims

1. A landslide hidden danger point displacement monitoring system using multi-source data fusion, characterized in that, including a multi-source sensing unit, which includes: a spaceborne radar sensing module for obtaining large-scale surface deformation data through synthetic aperture radar; a surface deformation sensing module for real-time monitoring of surface displacement through dual-frequency GNSS receivers and fiber Bragg grating sensors; an underground stress sensing module for obtaining stress changes in underground rock and soil masses through microseismic monitoring arrays and distributed fiber optic sensors; an edge computing unit, which includes: a dynamic weight fusion module for dynamically allocating weights of multi-source sensors according to the geological state to generate high-precision fused displacement field data; a communication protocol adaptation module for realizing protocol conversion and low-latency transmission of multi-source sensing data; a digital twin modeling unit, which includes: a geomechanical parameterized model for constructing a three-dimensional mechanical model of a landslide body based on the finite element method; a real-time data assimilation module for inputting the fused data into the model to invert and update geomechanical parameters; a model self-correction module for optimizing the model accuracy through parameter perturbation tests; an intelligent early warning unit, which includes: a multi-level early warning trigger module for triggering hierarchical early warnings according to displacement rate, acceleration, and abnormal underground parameters; a three-dimensional path planning module for dynamically planning a safe evacuation path based on the prediction results of the digital twin.

2. The landslide hidden danger point displacement monitoring system using multi-source data fusion according to claim 1, characterized in that: The dynamic weight fusion module performs the following operations: Obtain the real-time measurement data of each sensing module and its confidence parameter, where the confidence parameter is jointly determined according to the sensor noise variance and historical error statistics; Calculate the geological state evaluation coefficient , and its expression is: ； In the formula, is the influence function of inclination angle change, is the real-time inclination angle change amount, is the critical inclination angle threshold, is the weight coefficient of inclination angle change, is the influence function of displacement rate, is the real-time displacement rate, is the critical rate threshold, is the weight coefficient of displacement rate According to dynamically allocate the fusion weights of each sensing module, and the weight calculation formula is: ； wherein, is the fusion weight of the i-th sensing module, is the geological state evaluation coefficient, is the confidence parameter of the i-th sensing module, is the total number of sensing modules.

3. The landslide hidden danger point displacement monitoring system using multi-source data fusion according to claim 2, characterized in that: The inclination change influence function and the displacement rate influence function are defined as: ； ； Among them, To eliminate zero and tiny amounts, its value is positively correlated with the measurement accuracy of the sensor.

4. A landslide hidden danger point displacement monitoring system using multi-source data fusion according to claim 2 or 3, characterized in that: The weight coefficient and is dynamically configured through the geological knowledge base, specifically including the following steps: Step 1: Knowledge base construction: Establish the correlation between different geological types and weight coefficients according to geotechnical mechanics theory and historical landslide data; The geological types at least include typical landslide body types, and the typical landslide body types include clayey, sandy, rocky, and accumulation layers; Weight coefficient and are allocated according to the following principles: Tendency geology: Assign a higher weight to the inclination change; Rate-sensitive geology: Assign a higher weight to the displacement rate; Step 2: On-site geological type identification and multi-source data collection: Obtain the physical parameters of the rock and soil mass through on-site borehole sampling, geological radar detection, and UAV multi-spectral imaging; Input the collected data into a pre-trained geological classification model and output the geological type label of the target area; Step 3: Dynamically load the weight coefficient. According to the geological type label output in Step 2, retrieve the corresponding , value in the knowledge base; Inject the weight coefficient of the match into the dynamic weight fusion module in real time to update the geological state evaluation coefficient Calculation logic; Step 4: Abnormal handling mechanism. If the on-site geological type is not included in the mapping table, enable the default weight and trigger the manual verification process; After that, verify the rationality of the weight configuration through historical data playback. If the deviation between the fusion results and the measured data exceeds the limit for N consecutive times, trigger the knowledge base update.

5. The landslide hidden danger point displacement monitoring system using multi-source data fusion according to claim 1, wherein: The real-time data assimilation module performs: Step 1: Receive the fused displacement field data sent by the edge computing unit ; Step 2: Input the into the geomechanical parameterized model for inversion calculation to update the set of slip zone parameters , where is the shear modulus, is the pore pressure coefficient, is the cohesion; Step 3: Calculate the parameter change rate When is triggered, the model self-correction module is triggered, where is the preset sensitivity threshold; is the change rate of the slip zone parameter set, is the updated slip zone parameter set, is the slip zone parameter set before update, is the preset sensitivity threshold.

6. The landslide hidden danger point displacement monitoring system using multi-source data fusion according to claim 5, characterized in that: The inversion calculation uses the regularization adjoint gradient method, which specifically includes: Construct an objective function: Among them, is the finite element forward operator, is the initial parameter, is the regularization coefficient; Solve the adjoint equation: ； Obtain the parameter gradient ▽J, where is the stiffness matrix, is the adjoint variable; Use the quasi-Newton method to iteratively update the parameters until convergence.

7. A landslide hidden danger point displacement monitoring system using multi-source data fusion according to claim 1, characterized in that: The model self-correction module performs the following operations: Save the snapshot of the current parameters to the historical database; Generate a set of parameter perturbations , = 1, 2, 3... m; Parallelly calculate the predicted displacement fields corresponding to each perturbation parameter ; Select the parameter group with the highest matching degree with the actual observed data within the subsequent Δt time period as the update benchmark: ；； wherein, is the predicted displacement field of the k-th group of perturbation parameters at time t; is the actual observed displacement data at time t; T is the length of the verification time window.

8. A landslide hidden danger point displacement monitoring system using multi-source data fusion according to claim 2 or 5, characterized in that: The dynamic weight fusion module and the real-time data assimilation module form a two-way coupling mechanism, which is specifically manifested as: When the model self-calibration module updates the slip zone parameter set it automatically triggers the recalculation of the geological state evaluation coefficient where the critical dip angle threshold and the critical rate threshold are dynamically adjusted according to the new values: ； ； Among them, and are empirical functions based on geomechanics, is the density of rock mass.

9. A landslide hidden danger point displacement monitoring system using multi-source data fusion according to claim 1, characterized in that: The three-dimensional path planning module performs: Receive the landslide influence boundary output by the digital twin modeling unit, and then construct a three-dimensional terrain cost function: ； Among them, is the slope factor, is the surface crack density, is the distance between the current position and the landslide boundary; Use the improved A* algorithm to search for the minimum-cost path from the personnel location to the safe area, and the path weight coefficients , and are dynamically adjusted according to the warning level.

Images