Slope deformation monitoring and dynamic early warning method and system based on multi-sensor data

Abstract

The invention discloses a slope deformation monitoring and dynamic early warning method and system based on multi-sensor data, and relates to the technical field of slope monitoring, and the method comprises the steps: collecting multi-source sensor data by using pre-deployed multi-class sensors, constructing graph structure data according to the sensor distribution and the pre-processed multi-source sensor data, and carrying out the graph structure data; a graph convolutional network is used for modeling, and a slope deformation monitoring model is constructed; introducing a clustering federation learning strategy to carry out joint training on the slope deformation monitoring models of the plurality of sites, and carrying out risk grade division by using the trained slope deformation monitoring models; key influence factors of landslide disasters are extracted, an improved firefly algorithm is introduced to dynamically optimize an early warning threshold value, the optimized early warning threshold value and the current risk level are used for judgment, and early warning information is generated. According to the invention, the reliability of monitoring and the timeliness of early warning are improved through multi-source data fusion and intelligent analysis, and the crossing of slope deformation monitoring from single-point static state to networked intelligence is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of slope monitoring, more particularly, to a slope deformation monitoring and dynamic early warning method and system based on multi-sensor data. BACKGROUND

[0002] Slope stability monitoring is an important issue in the field of geological disaster prevention, especially in mountainous, mining, and transportation areas. Slope instability can cause landslides, collapses, and other disasters, which seriously threaten people's lives and property safety and the stable operation of major infrastructure. Traditional slope monitoring methods mainly rely on single sensors (such as total station, GNSS, inclinometer, etc.) or manual inspection, which have limited data dimensions, insufficient real-time performance, and low early warning accuracy. Traditional early warning is based on static thresholds or empirical models (such as displacement rate thresholds), which are difficult to adapt to the dynamic evolution process of slopes. Currently, dynamic early warning methods are improved through time series prediction models, multi-modal data fusion, and real-time optimization. However, existing systems still need further breakthroughs in model adaptability in small sample scenarios, uncertainty management of multi-source data, and dynamic adjustment of early warning thresholds.

[0003] To overcome the limitations of single sensors, multi-sensor collaborative monitoring has become a development trend. By integrating multi-dimensional data such as surface displacement, deep displacement, and environmental parameters, the slope state can be more comprehensively described. With the rapid development of sensor technology, Internet of Things (IoT), and Artificial Intelligence (AI), multi-sensor data-based slope deformation monitoring and dynamic early warning have gradually become a research hotspot, aiming to improve the reliability of monitoring and the timeliness of early warning through multi-source data fusion and intelligent analysis. SUMMARY

[0004] To solve the above technical problems, the present application proposes a slope deformation monitoring and dynamic early warning method and system based on multi-sensor data, which realizes the leap from single-point static to networked intelligence through the chain of multi-sensor fusion, graph structure modeling, federated learning, and dynamic threshold optimization.

[0005] The present application provides a slope deformation monitoring and dynamic early warning method based on multi-sensor data in the first aspect, comprising the following steps: Using pre-deployed multi-class sensors to collect multi-source sensor data, performing spatio-temporal registration and feature extraction on the multi-source sensor data, and obtaining pre-processed multi-source sensor data; Constructing graph structure data according to sensor distribution and pre-processed multi-source sensor data, modeling the graph structure data using a spatio-temporal graph convolution network, and constructing a slope deformation monitoring model; Introducing a clustering federated learning strategy to jointly train slope deformation monitoring models of multiple sites, and using the trained slope deformation monitoring model for deformation prediction; The key influencing factors of landslide disasters are extracted, a improved firefly algorithm is introduced based on the key influencing factors to dynamically optimize the early warning threshold, and the optimized early warning threshold is used with the deformation prediction result to divide the risk level and generate the early warning information.

[0006] In the application, multi-source sensor data is collected by using pre-deployed multi-class sensors, and the multi-source sensor data is spatio-temporally registered and feature-extracted to obtain pre-processed multi-source sensor data, specifically: Surface displacement monitoring, deep displacement monitoring and environmental parameter monitoring are performed using different sensors to obtain multi-source sensor data, which is transmitted to an edge node cache in real time and subjected to data format standardization processing; The sampling frequencies of different sensor data are obtained, the highest frequency sensor data is selected as the reference data to generate reference data, the dynamic time warping path is calculated for low frequency data and reference data, the optimal time mapping correlation is obtained, the continuous data is generated by using cubic spline interpolation for the missing period, and the sensor data unified to the same time sequence is output to obtain the time-aligned multi-source sensor data; The GNSS parameters and InSAR parameters obtained in the surface displacement monitoring are unified in the coordinate axis, data fusion is performed through Kalman filter iteration, GNSS parameters and InSAR parameters are used as observations, displacement is used as a state variable to construct a state equation, the current displacement is predicted based on the displacement state of the previous period, the predicted value is updated and corrected using the GNSS parameter and InSAR parameter observations, and the fused displacement field is output to obtain the spatially aligned surface displacement monitoring data; The time and space aligned multi-source sensor data is subjected to abnormality detection and missing value processing, and the slope deformation monitoring instance is retrieved and obtained, the SHAP value analysis is performed on the slope deformation monitoring instance, the displacement data features and environmental parameter features that meet the preset requirements for the slope deformation contribution are screened and stored, and the pre-processed multi-source sensor data is obtained.

[0007] In the application, a graph structure data is constructed according to the sensor distribution and the pre-processed multi-source sensor data, specifically: Each sensor is regarded as an independent node, the node attributes of each independent node include displacement parameter features, environmental parameter features and sensor deployment static attributes, and the coordinate information and geological identification are regarded as node metadata; The three-dimensional Euclidean distance between all sensors is calculated, and it is judged whether the three-dimensional Euclidean distance is greater than a preset distance threshold, if yes, the spatial distance connection between nodes is established, and the edge weight is set according to the distance; According to the geological exploration data, the rock layer to which the sensor belongs is marked, if two sensors are located in the same rock layer and the rock layer has the same inclination, a geomechanics connection between nodes is established, if the sensors are located in different rock layers but there is a weak interlayer, a cross-layer geomechanics connection between nodes is established, and the edge weight is set according to the shear strength of the rock layer; An adjacency matrix and a feature matrix are constructed according to the sensor nodes and the edge weight, and the adjacency matrix and the feature matrix are used as graph structure data.

[0008] In the present application, a spatio-temporal graph convolution network is used to model the graph structure data, and a slope deformation monitoring model is constructed, specifically: The continuous multi-source sensor data is divided into a plurality of windows using a time sliding window, each window containing a plurality of graph structure data snapshots, the node features are updated according to the latest multi-source sensor data, and the edge weight is updated according to the displacement correlation mutation between sensor nodes, and a dynamic graph sequence is generated by the updated node features and edge weight; According to the dynamic graph sequence, the graph structure data snapshot at the current time and the updated graph structure data snapshot at the last time are obtained, and an update gating mechanism is introduced to obtain the fusion ratio of the graph structure data snapshot at the current time and the updated graph structure data snapshot at the last time; Based on the fusion ratio, the adjacency matrix is dynamically updated, the current adjacency matrix and the updated adjacency matrix are fused to generate the final dynamic graph structure data, and the spatio-temporal graph convolution network is used to model the final dynamic graph structure data to construct the slope deformation monitoring model; The graph convolution network is used to model the spatial dependence of the dynamic graph structure data at each time, the neighborhood node information is aggregated through dynamic graph convolution to obtain the node feature representation at each time, and the spatial features are generated through multi-layer dynamic graph convolution with residual connection; The node feature representation sequence at all times output by the graph convolution network is obtained, the time dependence is modeled through multi-head attention, the different time scale dependence of a single attention head is obtained through the correlation calculation between the current time and the historical time of the node, and the time features are obtained through full connection layer fusion after splicing multi-head attention; The spatial features and the time features are fused and input into a prediction head for prediction, and the slope deformation monitoring model is initially trained through pre-set training data.

[0009] In the present application, a clustering federated learning strategy is introduced to jointly train the slope deformation monitoring models of multiple sites, and the trained slope deformation monitoring model is used for deformation prediction, specifically: Each site trains an independent slope deformation monitoring model based on local data, selects geological environmental features related to slope stability to obtain the geological conditions of each site, obtains the geological similarity between sites according to the geological conditions, and divides the sites into a plurality of clusters using the geological similarity. The stations in the cluster upload the model parameters of the slope deformation monitoring model to the cloud server after encryption, assign weights according to the data volume of each station, perform weighted average aggregation, and distribute the aggregated parameters to all stations in the cluster; According to the preset period, perform federated aggregation, when there is a site offline, skip its participation in the current round of aggregation, each station loads the global model of the cluster, and obtains the displacement increment of the future preset time period as the deformation prediction result.

[0010] In the application, the key influencing factors of landslide disasters are extracted, and the improved firefly algorithm is introduced based on the key influencing factors to dynamically optimize the early warning threshold, specifically: The displacement data features and environmental parameter features corresponding to the multi-source sensor data are combined with the geological environment features as candidate influencing factors, and the historical landslide event labels are used as target variables, a time delay variable is introduced in the candidate influencing factors, and the maximum mutual information time delay of each candidate influencing factor and displacement is calculated; The maximum information coefficient is used to enhance the mRmR algorithm, perform nonlinear redundancy evaluation, calculate the feature importance of the candidate influencing factors through the enhanced mRmR algorithm, sort according to the feature importance, and select a preset number of key influencing factors according to the sorting result; The current parameter value of the key influencing factor is introduced as a weight adjustment factor to minimize the false alarm rate and the false alarm rate for multi-objective optimization, the weight adjustment factor is used to weight the objective function, and the fitness function of the firefly algorithm is constructed through the inverse of the weighted objective function; According to the early warning threshold constraint range, the firefly population is initialized, the brightness of each firefly is calculated using the fitness function, an adaptive step size decreasing with iteration coefficient and a random disturbance vector are introduced to update the firefly position, so that the firefly moves to a brighter neighbor; After iteration to convergence, the brightest firefly is selected as the current dynamic early warning threshold.

[0011] In the application, the optimized early warning threshold and the deformation prediction result are used to divide the risk level, and the early warning information is generated, specifically: Get the deformation prediction result of the future preset time period, compare the deformation prediction result with the current early warning threshold, and generate early warning information of different risk levels according to the comparison result; When the early warning information exceeds the preset risk level, trigger multi-modal verification, detect the consistency of the deformation prediction result and the measured displacement trend, and perform similarity calculation according to the current environmental parameter characteristics, and retrieve the historical landslide events with similar degree meeting the requirements; When the consistency verification and the number of historical landslide events meet the preset multi-modal verification standard, the early warning information of the corresponding risk level is generated.

[0012] The second aspect of the present application provides a slope deformation monitoring and dynamic early warning system based on multi-sensor data, which comprises a data acquisition and preprocessing module, a graph structure modeling and intelligent analysis module, a federated learning and model optimization module, a dynamic early warning and decision support module; The data acquisition and preprocessing module uses pre-deployed multi-class sensors to collect multi-source sensor data, performs spatio-temporal registration and feature extraction on the multi-source sensor data, and obtains pre-processed multi-source sensor data; The graph structure modeling and intelligent analysis module constructs graph structure data according to sensor distribution and pre-processed multi-source sensor data, models the graph structure data using a spatio-temporal graph convolution network, constructs a slope deformation monitoring model, and realizes deformation prediction; The federated learning and model optimization module introduces a clustering federated learning strategy to jointly train slope deformation monitoring models of multiple sites, improves the model generalization ability, and protects data privacy; The dynamic early warning and decision support module extracts key influencing factors of landslide disasters, introduces an improved firefly algorithm to dynamically optimize the early warning threshold based on the key influencing factors, uses the optimized early warning threshold and the deformation prediction result to divide the risk level, and generates early warning information.

[0013] Compared with the prior art, the present application has the following advantages: The present application constructs a high-precision, self-adaptive and scalable slope deformation monitoring and early warning system through multi-sensor data fusion, graph convolution network modeling, clustering federated learning and dynamic threshold optimization. Deep fusion of multi-source data is used to improve the monitoring accuracy and robustness. Key features such as displacement rate, strain distribution and environmental lag effect are extracted from the original data. Graph structure data modeling is used to more comprehensively represent the spatial correlation of slope deformation. Graph convolution network modeling is used to capture the spatial correlation of slope deformation and learn the local single sensor data and the global landslide surface sliding trend, avoiding the limitations of traditional point-based methods. The clustering federated learning strategy is introduced to realize multi-site collaborative modeling and privacy protection. Slope monitoring sites with similar geological conditions are divided into the same cluster, and the model is trained jointly within the cluster to avoid overfitting caused by insufficient data of a single site. Dynamic threshold optimization and hierarchical early warning reduce the false alarm rate and the hierarchical early warning mechanism considers real-time and long-term trend prediction. BRIEF DESCRIPTION OF DRAWINGS

[0014] In order to more clearly illustrate the technical solutions in the embodiments or examples of the present application, the drawings needed for the embodiments or example descriptions will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and those skilled in the art can obtain other drawings according to the drawings shown without creating any inventive labor.

[0015] Figure 1 A flowchart of slope deformation monitoring and dynamic early warning based on multi-sensor data is shown. Figure 2 A flowchart of constructing a slope deformation monitoring model in an embodiment is shown. Figure 3 A flowchart of dynamically optimizing early warning thresholds in an embodiment is shown. Figure 4 A block diagram of a slope deformation monitoring and dynamic early warning system based on multi-sensor data is shown. DETAILED DESCRIPTION

[0016] In order to enable a clearer understanding of the above-mentioned purposes, features and advantages of the present application, the present application will be further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0017] In the following description, many specific details are set forth in order to provide a thorough understanding of the present application, however, the present application can also be implemented in other ways different from those described herein, therefore, the scope of protection of the present application is not limited by the specific embodiments disclosed below.

[0018] Figure 1 A flowchart of slope deformation monitoring and dynamic early warning based on multi-sensor data is shown.

[0019] As shown in Figure 1 The present embodiment provides a slope deformation monitoring and dynamic early warning based on multi-sensor data, which comprises: S102, using pre-deployed multi-class sensors to collect multi-source sensor data, performing spatio-temporal registration and feature extraction on the multi-source sensor data, and obtaining pre-processed multi-source sensor data; S104, constructing graph structure data according to sensor distribution and pre-processed multi-source sensor data, modeling the graph structure data using a spatio-temporal graph convolution network, and constructing a slope deformation monitoring model; S106, introducing a clustering federated learning strategy to jointly train slope deformation monitoring models of multiple sites, and using the trained slope deformation monitoring model for deformation prediction; S108, extracting key influencing factors of landslide disasters, introducing an improved firefly algorithm to dynamically optimize early warning thresholds based on the key influencing factors, using the optimized early warning thresholds and deformation prediction results to divide risk levels, and generating early warning information.

[0020] It should be noted that different sensors are used for ground surface displacement monitoring, deep displacement monitoring and environmental parameter monitoring. The types of sensors for ground surface displacement monitoring include GNSS, InSAR, crack meter, etc. The types of sensors for deep displacement monitoring include inclinometer, optical fiber sensor, etc. The types of sensors for environmental parameter monitoring include rain gauge, underground water level gauge, vibration sensor, etc. Multi-source sensor data is obtained, and the multi-source sensor data is transmitted in real time to the edge node cache through the 4G / 5G / LoRa network, and data format standardization processing is performed.

[0021] Different sensors have different sampling frequencies, and the time axis needs to be aligned for fusion analysis. The sampling frequencies of different sensor data are obtained, the highest frequency sensor data is selected as the reference time axis to generate reference data, the dynamic time warping path is calculated for low frequency data and reference data, the optimal time mapping correlation is obtained, the continuous data is generated by cubic spline interpolation for the missing period, the sensor data unified to the same time sequence is output, the time-aligned multi-source sensor data is obtained, and the time sequence consistency is ensured. The GNSS parameters and InSAR parameters obtained in the ground surface displacement monitoring are unified in the coordinate axis and converted to the same local coordinate system. Data fusion is performed by Kalman filter iteration. It is assumed that the slope displacement obeys a linear dynamic model, GNSS parameters and InSAR parameters are used as observations, displacement is used as a state variable to construct a state equation, the current displacement is predicted according to the displacement state of the previous period, the predicted value is updated and corrected using the GNSS parameter and InSAR parameter observation value, and the fused displacement field is output. The spatially aligned ground surface displacement monitoring data is obtained. The spatiotemporally aligned multi-source sensor data is subjected to abnormality detection and missing value processing, and slope deformation monitoring instances are retrieved and obtained. SHAP value analysis is performed on the slope deformation monitoring instances, displacement data features and environmental parameter features that meet the preset requirements for slope deformation contribution are selected and stored, and preprocessed multi-source sensor data is obtained. The preferred displacement data features include cumulative displacement, displacement rate, acceleration and spatial gradient, and the environmental data features include rainfall lag effect, underground water level mutation point and vibration energy. Structured features are stored in a time series database for real-time calling by a model, and unstructured data is stored in an object storage system.

[0022] It should be noted that each sensor is taken as an independent node, the node attribute of each independent node includes displacement parameter characteristics, environmental parameter characteristics and sensor deployment static attribute, the sensor deployment static attribute includes sensor type, installation depth, rock mass type, and the longitude and latitude coordinate information and geological identification of sensor deployment are taken as node metadata; the three-dimensional Euclidean distance between all sensors is calculated, whether the three-dimensional Euclidean distance is greater than a preset distance threshold is judged, if greater, the spatial distance connection between nodes is established, and the edge weight is set according to the distance, the closer the distance, the higher the edge weight; according to the geological exploration data (drilling, geological profile), the rock layer (such as the upper loose accumulation body, the lower bedrock) to which the sensor belongs is marked, if two sensors are located in the same rock layer and the rock layer inclination is consistent, the geomechanics connection between nodes is established, reflecting the overall sliding trend of the rock mass; if the sensors are located in different rock layers but there is a weak interlayer, the cross-layer geomechanics connection between nodes is established, reflecting the interlayer sliding risk, and the edge weight is set according to the rock layer shear strength; the adjacency matrix and the feature matrix are constructed according to the sensor nodes and the edge weight, and the adjacency matrix and the feature matrix are used as the graph structure data. In the adjacency matrix, the row and the column represent the sensor nodes, and the matrix element value is the edge weight. In the feature matrix, each row represents a sensor node, and each column represents a feature.

[0023] Figure 2 A flowchart for constructing a slope deformation monitoring model in an embodiment is shown.

[0024] According to the embodiment of the present application, the spatio-temporal graph convolution network is used to model the graph structure data, and a slope deformation monitoring model is constructed, specifically: S202, using a time sliding window, the continuous multi-source sensor data is divided into a plurality of windows, each window contains a plurality of graph structure data snapshots, the node feature is updated according to the latest multi-source sensor data, and the edge weight is updated according to the displacement correlation mutation between sensor nodes, and a dynamic graph sequence is generated through the updated node feature and edge weight; S204, according to the dynamic graph sequence, the graph structure data snapshot at the current time and the updated graph structure data snapshot at the last time are obtained, and the fusion ratio of the graph structure data snapshot at the current time and the updated graph structure data snapshot at the last time is obtained by introducing an update gating mechanism; S206, based on the fusion ratio, the adjacency matrix is dynamically updated, the current adjacency matrix and the updated adjacency matrix are fused to generate the final dynamic graph structure data, and the spatio-temporal graph convolution network is used to model the final dynamic graph structure data, and a slope deformation monitoring model is constructed; S208, using the graph convolution network, the spatial dependence of the dynamic graph structure data at each time is modeled, the neighborhood node information is aggregated through the dynamic graph convolution, the node feature representation at each time is obtained, and the spatial feature is generated through the multi-layer dynamic graph convolution of the residual connection. S210: Obtain the node feature representation sequence of all moments output by the graph convolutional network, model the time dependency through multi-head attention, calculate the correlation between the current moment and the historical moment of the node to obtain the different time scale dependencies of a single attention head, and after splicing the multi-head attention, fuse it through the fully connected layer to obtain the time feature; S212: The spatial features and the temporal features are integrated and imported into a prediction head for prediction, and the slope deformation monitoring model is initially trained using preset training data.

[0025] It should be noted that the pre-processed multi-source sensor data is input into the time sliding window, the time interval is set according to the monitoring requirements, and the data in the window includes the current time and before Graph data at the moment (node ​​feature matrix , adjacency matrix In dynamic graph updates, an update gating mechanism is introduced Get a snapshot of the graph structure data at the current moment And the graph structure data snapshot updated at the last moment The fusion ratio; , represents the activation function, 、 Represents a learnable parameter, and dynamically updates the adjacency matrix based on the fusion ratio , Represents element-by-element multiplication, fusing the current adjacency matrix and the updated adjacency matrix to generate the final dynamic graph structure data , , 、 Hyperparameters representing the current observation and historical states.

[0026] In the spatiotemporal attention mechanism, spatial dependency modeling is performed through dynamic graph convolution operations, graph convolution is performed on the dynamic graph structure data at each moment, neighborhood node information is aggregated, and residual connections are introduced to prevent the disappearance of deep network gradients. Multi-head attention is used to model temporal dependency, while ordinary attention may implicitly utilize future moment data and is not suitable for real-time warning scenarios. Preferably, the attention parameters Q, K, and V are generated by a one-dimensional causal convolution with a kernel size of k, forcing the model to rely only on historical information while aggregating local context. The attention weight is calculated based on the local sequence fragments of the convolution output to enhance the perception of approaching trends. In the local similarity calculation, for each moment, the local window of Q and K (length is ) Calculate similarity , comparing the overall displacement trends over two time periods (e.g., "continuous acceleration" vs. "steady creep"), rather than instantaneous values, is expressed as:

[0027] where , denotes the current time and historical time, denotes the relative offset in the sliding window, used to traverse each time step in the local window, denotes the query vector at time , containing the displacement trend features of the time and its previous k-1 times, denotes the key vector transpose at time , reflecting the deformation pattern of the historical local window; apply a lower triangular mask matrix before Softmax , forcing , the attention weight is represented as:

[0028] where, denotes the key vector dimension scaling factor, denotes the value matrix.

[0029] Each attention head uses different causal convolution kernels to capture multi-scale time patterns, and multiple attention heads run in parallel, and the local attention outputs of each head are spliced and mapped through a fully connected layer. By comparing the displacement trend segments of length k, single-point noise interference is avoided. The multi-head design with different convolution kernel sizes takes into account both sudden events and long-term evolution, and the causal mask is introduced to strictly follow the principle of predicting the future from the past, meeting the real-time requirements of slope monitoring.

[0030] Each site deploys the same slope deformation monitoring model, which includes dynamic graph updates, spatiotemporal attention, and prediction head modules. The site's graph structure data is input for local model training, and each site trains an independent slope deformation monitoring model based on local data. Geological environmental features related to slope stability are selected to obtain the geological conditions of each site, such as rock and soil type, slope, hydrological conditions, and historical deformation patterns. K-means clustering is used to obtain the geological similarity between sites based on the geological conditions, and the sites are divided into several clusters using the geological similarity. When a new site is added, its Euclidean distance to the center of each cluster is calculated and the site is assigned to the closest cluster. Each site in the cluster encrypts the model parameters of the slope deformation monitoring model and uploads them to the cloud server. Weights are assigned based on the amount of data at each site, and weighted average aggregation is performed. The aggregated parameters are distributed to all sites in the cluster. Federated aggregation is performed according to a preset period. If a site is offline, it is skipped from participating in the current round of aggregation. Each site loads the cluster's global model and obtains the displacement increment for the preset time period in the future as the deformation prediction result. Through cluster federated learning, the generalization ability and warning accuracy of the slope deformation monitoring model have been significantly improved while protecting data privacy. Different geological clusters have customized models, avoiding performance degradation caused by a one-size-fits-all approach.

[0031] Figure 3 A flow chart of dynamically optimizing the warning threshold in an embodiment is shown.

[0032] According to an embodiment of the present invention, key influencing factors of landslide disasters are extracted, and an improved firefly algorithm is introduced based on the key influencing factors to dynamically optimize the warning threshold, specifically: S302: Displacement data features and environmental parameter features corresponding to multi-source sensor data are combined with geological environment features as candidate influencing factors, and historical landslide event labels are used as target variables. A time lag variable is introduced into the candidate influencing factors to calculate the maximum mutual information time lag between each candidate influencing factor and the displacement. S304: Using the maximum information coefficient to enhance the mRmR algorithm, perform nonlinear redundancy evaluation, calculate the feature importance of the candidate influencing factors using the enhanced mRmR algorithm, sort them according to feature importance, and select a preset number of key influencing factors based on the sorting results; S306, introducing the current parameter value of the key influencing factor as a weight adjustment factor to perform multi-objective optimization to minimize the false alarm rate and the missed alarm rate, using the weight adjustment factor to weight the objective function, and constructing the fitness function of the firefly algorithm through the inverse of the weighted objective function; S308, initializing the glowworm population according to the early warning threshold constraint range, calculating the brightness of each glowworm using the fitness function, introducing an adaptive step length decreasing with the iteration coefficient and a random disturbance vector to update the glowworm position, making the glowworm move to a brighter neighbor, iterating to convergence, and selecting the glowworm with the highest brightness as the current dynamic early warning threshold.

[0033] It should be noted that the traditional mRMR selects features by maximizing the correlation between features and target and minimizing the redundancy between features, and introduces lag correlation analysis to enhance its applicability. The time lag mutual information between each candidate influencing factor and the target variable is calculated. For environmental factors such as rainfall and water level, the maximum mutual information time lag between the environmental factors and displacement is calculated, for example, the displacement response is strongest 48 hours after the rainfall. The redundancy of the candidate influencing factors and the selected feature set is calculated, and the features with low correlation with the selected features are preferentially selected. The comprehensive score of the candidate influencing factors is obtained as the feature importance with the goal of maximum correlation and minimum redundancy, and the Top5 key influencing factors are selected according to the feature importance ranking result. The displacement acceleration (lag 0 hours), cumulative rainfall (lag 24 hours), groundwater level mutation (lag 12 hours), and rock-soil type (static) are preferentially selected as key influencing factors.

[0034] The improved glowworm algorithm is used to dynamically optimize the threshold, and the false positive rate and the false negative rate are minimized as optimization objectives. The current parameter value of the key influencing factor is introduced as a weight adjustment factor, and the fitness function is inversely proportional to the weighted objective function , which is expressed as: , represents the false positive rate, represents the false negative rate, represents the weight adjustment factor, represents the glowworm individual. In the iteration process, the glowworm moves to a brighter glowworm, and is dynamically adjusted according to the environmental parameters. If the current rainfall exceeds the threshold, the search range is reduced. After iteration to convergence, the glowworm with the highest brightness is selected as the current dynamic threshold. For example, during continuous rainfall, the algorithm automatically reduces the early warning threshold to give early warning of shallow landslides. The glowworm algorithm adjusts the threshold in combination with real-time environmental data to reduce seasonal or sudden disturbances, and multi-objective optimization avoids the false positive and false negative imbalance problem of traditional threshold methods.

[0035] Obtain the deformation prediction result of the future preset time period, compare the deformation prediction result with the current early warning threshold, and generate early warning information of different risk levels according to the comparison result; when the early warning information exceeds the preset risk level, trigger multi-modal verification to improve the early warning credibility. The consistency of the deformation prediction result and the measured displacement trend is detected, and the similarity is calculated according to the current environmental parameter characteristics, and the historical landslide events with similar similarity are retrieved; when the consistency verification and the number of historical landslide events meet the preset multi-modal verification standard, the early warning information of the corresponding risk level is generated, otherwise, the artificial review is triggered, and the early warning information is visually displayed.

[0036] Figure 4 A slope deformation monitoring and dynamic early warning system based on multi-sensor data is shown.

[0037] The second embodiment of the application provides a slope deformation monitoring and dynamic early warning system based on multi-sensor data 4, which comprises a data acquisition and preprocessing module 401, a graph structure modeling and intelligent analysis module 402, a federated learning and model optimization module 403, and a dynamic early warning and decision support module 404. The data acquisition and preprocessing module uses a pre-deployed multi-class sensor to collect multi-source sensor data, performs spatio-temporal registration and feature extraction on the multi-source sensor data, and obtains pre-processed multi-source sensor data. The graph structure modeling and intelligent analysis module constructs graph structure data according to the sensor distribution and the pre-processed multi-source sensor data, models the graph structure data using a spatio-temporal graph convolution network, constructs a slope deformation monitoring model, and realizes deformation prediction. The federated learning and model optimization module introduces a clustering federated learning strategy to jointly train the slope deformation monitoring models of multiple sites, improves the model generalization ability, and protects data privacy. The dynamic early warning and decision support module extracts key influencing factors of landslide disasters, introduces an improved firefly algorithm to dynamically optimize the early warning threshold based on the key influencing factors, uses the optimized early warning threshold to divide the risk level with the deformation prediction result, and generates early warning information.

[0038] The third aspect of the application provides a computer readable storage medium, which comprises a slope deformation monitoring and dynamic early warning method based on multi-sensor data program, and the slope deformation monitoring and dynamic early warning method based on multi-sensor data program is executed by a processor to realize the steps of the slope deformation monitoring and dynamic early warning method based on multi-sensor data.

[0039] In several embodiments provided in the present application, it should be understood that the disclosed method and system can be implemented in other ways. The above-described system embodiments are only illustrative, for example, the division of the modules is only a logical function division, and actual implementation can have another division manner, for example, multiple modules or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the displayed or discussed components can be indirect coupling or communication connection through some interfaces, devices or modules, which can be electrical, mechanical or other forms. In addition, each functional module in each embodiment of the present application can be integrated into a processing module, or each module can be a separate module, or two or more modules can be integrated into one module; the integrated module can be realized in the form of hardware or in the form of hardware plus software functional module.

[0040] Those skilled in the art can understand that all or part of the steps of the above-mentioned method embodiments can be completed by program instruction related hardware, the above-mentioned program can be stored in a computer readable storage medium, and the program executes the steps including the above-mentioned method embodiments when executed; and the above-mentioned storage medium includes a mobile storage device, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk and various storage medium capable of storing program codes.

[0041] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto, any person skilled in the art can easily think of changes or replacements within the technical range disclosed in the present application, which should be covered in the protection scope of the present application.

Claims

1. A slope deformation monitoring and dynamic early warning method based on multi-sensor data, characterized in that: The following steps are involved: Using pre-deployed multiple types of sensors to collect multi-source sensor data, performing spatiotemporal registration and feature extraction on the multi-source sensor data, and obtaining pre-processed multi-source sensor data; Graph structure data is constructed based on sensor distribution and preprocessed multi-source sensor data, and the graph structure data is modeled using a spatiotemporal graph convolutional network to construct a slope deformation monitoring model; A clustering federated learning strategy is introduced to jointly train slope deformation monitoring models at multiple sites, and the trained slope deformation monitoring models are used for deformation prediction. The key influencing factors of landslide disasters are extracted, and the improved firefly algorithm is introduced based on the key influencing factors to dynamically optimize the warning threshold. The optimized warning threshold and deformation prediction results are used to divide the risk level and generate warning information.

2. The slope deformation monitoring and dynamic early warning method based on multi-sensor data according to claim 1 is characterized in that: Use pre-deployed multiple types of sensors to collect multi-source sensor data, perform spatiotemporal registration and feature extraction on the multi-source sensor data, and obtain pre-processed multi-source sensor data, specifically: Use different sensors to monitor surface displacement, deep displacement, and environmental parameters, acquire multi-source sensor data, transmit the multi-source sensor data to edge node cache in real time, and perform data format standardization; The sampling frequencies of different sensor data are obtained, and the highest-frequency sensor data is selected as the base time axis to generate reference data. The dynamic time warping path is calculated for the low-frequency data and the reference data to obtain the optimal time mapping association. For the missing time periods, cubic spline interpolation is used to generate continuous data. The sensor data unified into the same time series is output to obtain time-aligned multi-source sensor data. The GNSS parameters and InSAR parameters obtained from surface displacement monitoring are aligned in coordinate axes, and data fusion is performed through Kalman filtering iteration. GNSS parameters and InSAR parameters are used as observation quantities, and displacement is used as a state quantity to construct a state equation. The current displacement is predicted based on the displacement state of the previous period. The GNSS parameters and InSAR parameter observations are used to update the corrected prediction value. The fused displacement field is output to obtain spatially aligned surface displacement monitoring data. Anomaly detection and missing value processing are performed on the multi-source sensor data after spatiotemporal alignment, and slope deformation monitoring instances are retrieved and obtained. SHAP value analysis is performed on the slope deformation monitoring instances. Displacement data features and environmental parameter features that contribute to slope deformation and meet preset requirements are screened and stored, and the preprocessed multi-source sensor data is obtained.

3. The slope deformation monitoring and dynamic early warning method based on multi-sensor data according to claim 1 is characterized in that: Graph structure data is constructed based on sensor distribution and preprocessed multi-source sensor data, specifically: Each sensor is treated as an independent node. The node attributes of each independent node include displacement parameter characteristics, environmental parameter characteristics and sensor deployment static attributes, and the coordinate information and geological identification are used as node metadata. Calculate the three-dimensional Euclidean distance between all sensors and determine whether the three-dimensional Euclidean distance is greater than a preset distance threshold. If so, establish a spatial distance connection between the nodes and set edge weights based on the distance. The rock layers to which the sensors belong are marked according to geological exploration data. If the two sensors are located in the same rock layer and the rock layer dips are consistent, a geomechanical connection is established between the nodes. If the sensors are located in different rock layers but there are weak interlayers, a cross-layer geomechanical connection is established between the nodes, and the edge weights are set according to the shear strength of the rock layer. An adjacency matrix and a feature matrix are constructed according to the sensor nodes and edge weights, and the adjacency matrix and the feature matrix are used as graph structure data.

4. The slope deformation monitoring and dynamic early warning method based on multi-sensor data according to claim 1 is characterized in that: The graph structure data is modeled using a spatiotemporal graph convolutional network to construct a slope deformation monitoring model, specifically: A time sliding window is used to divide continuous multi-source sensor data into several windows, each of which contains several snapshots of graph structure data. Node features are updated based on the latest multi-source sensor data, and edge weights are updated based on the sudden change in displacement correlation between sensor nodes. A dynamic graph sequence is generated based on the updated node features and edge weights. Obtaining a graph structure data snapshot at the current moment and a graph structure data snapshot updated at the previous moment according to the dynamic graph sequence, and introducing an update gating mechanism to obtain a fusion ratio of the graph structure data snapshot at the current moment and the graph structure data snapshot updated at the previous moment; Dynamically updating the adjacency matrix based on the fusion ratio, fusing the current adjacency matrix and the updated adjacency matrix to generate final dynamic graph structure data, using the final dynamic graph structure data of the spatiotemporal graph convolutional network for modeling, and constructing a slope deformation monitoring model; Use graph convolutional networks to model the spatial dependencies of dynamic graph structure data at each moment, aggregate neighborhood node information through dynamic graph convolution, obtain node feature representation at each moment, and generate spatial features through multi-layer dynamic graph convolution with residual connections; Obtain the node feature representation sequence of all moments output by the graph convolutional network, model the temporal dependency through multi-head attention, calculate the correlation between the current moment and the historical moment of the node to obtain the different time scale dependencies of a single attention head, and after splicing the multi-head attention, fuse it through the fully connected layer to obtain the temporal features; The spatial features and temporal features are fused and imported into a prediction head for prediction, and the slope deformation monitoring model is initially trained using preset training data.

5. The slope deformation monitoring and dynamic early warning method based on multi-sensor data according to claim 1 is characterized in that: A cluster federated learning strategy is introduced to jointly train the slope deformation monitoring models of multiple sites, and the trained slope deformation monitoring models are used for deformation prediction. Specifically: Each site trains an independent slope deformation monitoring model based on local data, selects geological environmental features related to slope stability to obtain the geological conditions of each site, obtains the geological similarity between sites based on the geological conditions, and divides the sites into several clusters using the geological similarity; Each site in the cluster encrypts the model parameters of the slope deformation monitoring model and uploads them to the cloud server. A weight is assigned based on the data volume of each site, and weighted average aggregation is performed. The aggregated parameters are then distributed to all sites in the cluster. Federated aggregation is performed according to the preset period. If a site is offline, it is skipped from participating in the current round of aggregation. Each site loads the cluster global model and obtains the displacement increment in the future preset time period as the deformation prediction result.

6. The slope deformation monitoring and dynamic early warning method based on multi-sensor data according to claim 1 is characterized in that: The key influencing factors of landslide disasters are extracted, and the improved firefly algorithm is introduced based on the key influencing factors to dynamically optimize the warning threshold. Specifically: The displacement data characteristics and environmental parameter characteristics corresponding to the multi-source sensor data are combined with the geological environment characteristics as candidate influencing factors, and the historical landslide event labels are used as target variables. Time lag variables are introduced into the candidate influencing factors, and the maximum mutual information lag between each candidate influencing factor and the displacement is calculated. The mRmR algorithm is enhanced using the maximum information coefficient to perform nonlinear redundancy evaluation. The feature importance of candidate influencing factors is calculated using the enhanced mRmR algorithm, and they are sorted according to feature importance. A preset number of key influencing factors are selected based on the sorting results. Introducing the current parameter values ​​of the key influencing factors as weight adjustment factors to perform multi-objective optimization by minimizing the false alarm rate and the missed alarm rate, using the weight adjustment factors to weight the objective function, and constructing the fitness function of the firefly algorithm through the inverse of the weighted objective function; Initialize the firefly population according to the warning threshold constraint range, use the fitness function to calculate the brightness of each firefly, introduce an adaptive step size that decreases with the iteration coefficient and a random perturbation vector to update the firefly position, so that the firefly moves to a brighter neighbor; After iteration until convergence, the firefly with the highest brightness is selected as the current dynamic warning threshold.

7. The slope deformation monitoring and dynamic early warning method based on multi-sensor data according to claim 1 is characterized in that: The optimized warning threshold and deformation prediction results are used to classify the risk levels and generate warning information, specifically: Obtain deformation prediction results for a preset time period in the future, compare the deformation prediction results with the current warning threshold, and generate warning information of different risk levels based on the comparison results; When the warning information exceeds the preset risk level, multimodal verification is triggered to detect the consistency between the deformation prediction result and the measured displacement trend, and similarity calculation is performed based on the current environmental parameter characteristics to retrieve historical landslide events with similarity that meets the requirements; When the consistency verification and the number of historical landslide events meet the preset multimodal verification standards, early warning information corresponding to the risk level is generated.

8. A slope deformation monitoring and dynamic early warning system based on multi-sensor data, implementing the slope deformation monitoring and dynamic early warning method based on multi-sensor data according to any one of claims 1 to 7, the system comprising: Data acquisition and preprocessing module, graph structure modeling and intelligent analysis module, federated learning and model optimization module, dynamic early warning and decision support module; The data acquisition and preprocessing module uses pre-deployed multiple types of sensors to collect multi-source sensor data, performs spatiotemporal registration and feature extraction on the multi-source sensor data, and obtains pre-processed multi-source sensor data; The graph structure modeling and intelligent analysis module constructs graph structure data based on sensor distribution and preprocessed multi-source sensor data, uses a spatiotemporal graph convolutional network to model the graph structure data, builds a slope deformation monitoring model, and realizes deformation prediction; The federated learning and model optimization module introduces a clustered federated learning strategy to jointly train slope deformation monitoring models at multiple sites, thereby improving model generalization capabilities and protecting data privacy. The dynamic warning and decision support module extracts key influencing factors of landslide disasters, introduces an improved firefly algorithm based on the key influencing factors to dynamically optimize the warning threshold, uses the optimized warning threshold and deformation prediction results to divide the risk level, and generates warning information.

Images