Slope safety coupling analysis method and device based on multi-source monitoring data

Abstract

The invention discloses a slope safety coupling analysis method and device based on multi-source monitoring data, and a server side method comprises the steps: obtaining and preprocessing the multi-source monitoring data of a to-be-monitored slope according to a preset period, and obtaining structure response data, environment driving data and visual appearance data; constructing a multi-dimensional feature matrix capable of representing the change of the slope to be monitored along with time through the structure response data, the environment driving data and the visual appearance data; inputting the multi-dimensional feature matrix into a pre-trained slope macroscopic behavior pattern recognition model, and outputting a current slope behavior pattern of the to-be-monitored slope; according to the current slope behavior mode and the multi-dimensional feature matrix, slope safety coupling analysis is carried out, and a coupling relation dynamic evolution result of the slope to be monitored is obtained. Therefore, by adopting the embodiment of the invention, the accuracy of the analysis result of the slope can be improved. Meanwhile, the complete deformation and damage process of the side slope can be accurately depicted, and then potential risks of the side slope can be early warned in time.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of slope safety monitoring, and in particular to a slope safety coupling analysis method and device based on multi-source monitoring data. BACKGROUND

[0002] In mountainous infrastructure construction, such as highways, railways, hydropower stations, etc., large-scale slope excavation and support engineering are inevitably involved. These slopes are subjected to self-weight stress, and are also long-term affected by complex factors such as rainfall infiltration, groundwater level fluctuation, weathering and erosion, earthquakes, etc., which are prone to induce geological disasters such as sliding and collapse, posing a serious threat to social public safety. Therefore, monitoring the relevant conditions of the slope for early prevention is a major demand to protect life and property safety.

[0003] In related technologies, slope monitoring mainly relies on manual inspection or a single type of sensor. Manual inspection is highly dependent on the personal experience of engineers, lacks quantitative scientific basis, and results in insufficient accuracy of analysis results. The sensor can only capture a single dimension of slope response, and cannot fully reflect the coupling effect of multiple physical fields (such as stress field, seepage field, temperature field), making it difficult to accurately depict the complete deformation and failure process of the slope, and resulting in the inability to timely warn potential risks.

[0004] In recent years, with the rapid progress of technologies such as Internet of Things (IoT), Global Navigation Satellite System (GNSS), Synthetic Aperture Radar Interferometry (InSAR), LiDAR, optical fiber sensing, unmanned aerial vehicle remote sensing, and artificial intelligence (AI) visual recognition, revolutionary opportunities have been brought to slope monitoring. Unprecedented multi-source heterogeneous data covering the whole space of "point-line-surface-body", the whole depth of "surface-internal", and the whole element of "structure-environment" can be obtained.

[0005] Therefore, how to effectively reveal the systematic evaluation model of the internal correlation and dynamic evolution among multiple factors through these information-abundant data streams is an urgent problem to be solved. SUMMARY

[0006] The embodiments of the present application provide a slope safety coupling analysis method and device based on multi-source monitoring data. In order to have a basic understanding of some aspects of the disclosed embodiments, a brief summary is given below. This summary is not a general review, nor is it intended to determine key / important components or delineate the scope of protection of these embodiments. Its only purpose is to present some concepts in a simple form as a prelude to the detailed description that follows.

[0007] In a first aspect, the embodiments of the present application provide a slope safety coupling analysis method based on multi-source monitoring data, applied to a server, and the method comprises: Obtaining and preprocessing multi-source monitoring data of the to-be-monitored slope according to a preset period, to obtain structural response data, environmental driving data and visual appearance data; Constructing a multi-dimensional feature matrix capable of representing the to-be-monitored slope changing over time through the structural response data, the environmental driving data and the visual appearance data; Inputting the multi-dimensional feature matrix into a pre-trained slope macro-behavior pattern recognition model to output a current slope behavior pattern of the to-be-monitored slope; Performing slope safety coupling analysis according to the current slope behavior pattern and the multi-dimensional feature matrix to obtain a coupling relationship dynamic evolution result of the to-be-monitored slope.

[0008] Optionally, collecting and preprocessing the multi-source monitoring data of the to-be-monitored slope to obtain the structural response data, the environmental driving data and the visual appearance data, comprising: Monitoring three-dimensional displacement data of the surface of the to-be-monitored slope through a GNSS device, and measuring opening and closing change data of cracks at a preset position of the to-be-monitored slope through a crack meter as structural monitoring source data; Recording rainfall intensity and cumulative rainfall data of the to-be-monitored slope through an automatic rain gauge, and monitoring change data of a seepage field inside the to-be-monitored slope through a pore water pressure gauge as environmental monitoring source data; Performing visual inspection on the slope surface of the to-be-monitored slope through a high-definition AI video ball machine to identify abnormal appearance change data of the slope surface of the to-be-monitored slope as visual monitoring source data; Performing data missing filling, abnormal data processing and data synchronization processing on the structural monitoring source data, the environmental monitoring source data and the visual monitoring source data to obtain the structural response data, the environmental driving data and the visual appearance data.

[0009] Optionally, performing slope safety coupling analysis according to the current slope behavior pattern and the multi-dimensional feature matrix to obtain a coupling relationship dynamic evolution result of the to-be-monitored slope, comprising: Selecting a plurality of monitoring feature parameters related to the current slope behavior pattern from the multi-dimensional feature matrix; Calculating similarity distances between different monitoring feature parameter time series according to the plurality of monitoring feature parameters to obtain a DTW similarity matrix; Quantifying the correlation strength between each monitoring feature parameter according to the DTW similarity matrix to obtain a grey correlation degree matrix; Taking the plurality of monitoring feature parameters as network nodes and taking the correlation strength in the grey correlation degree matrix as the weight of the edge to construct a slope state coupling network graph of the to-be-monitored slope; Determining the coupling relationship dynamic evolution result of the to-be-monitored slope based on the slope state coupling network graph.

[0010] Optionally, the current slope behavior pattern is one of the following: stable fluctuation type, environment-driven type, and structural deterioration and instability type. From the multidimensional feature matrix, several monitoring feature parameters related to the current slope behavior pattern are selected, including: Given that the current slope behavior pattern is stable-fluctuation, the first monitoring characteristic parameter related to low-frequency fluctuations is obtained from the multidimensional feature matrix. This first monitoring characteristic parameter includes at least the daily displacement increment, cumulative displacement, and short-term fluctuations in pore water pressure; or... Given that the current slope behavior pattern is environment-driven, a second monitoring characteristic parameter related to rainfall and pore water pressure is obtained from the multidimensional feature matrix. This second monitoring characteristic parameter includes at least hourly rainfall intensity, cumulative rainfall, and pore water pressure change rate; or... Given that the current slope behavior pattern is structural deterioration and instability, a third monitoring characteristic parameter related to the structural response is obtained from the multidimensional characteristic matrix. The third monitoring characteristic parameter includes at least displacement rate, crack opening and closing rate, and acceleration.

[0011] Optionally, based on multiple monitoring feature parameters, the similarity distance between time series of different monitoring feature parameters is calculated to obtain the DTW similarity matrix, including: The monitoring values ​​of each monitoring characteristic parameter are obtained as a function of time, thus obtaining the time series data of each monitoring characteristic parameter. Initialize an empty similarity matrix with a size of n×n, where n is the number of multiple monitoring feature parameters. The rows and columns of the similarity matrix correspond to different monitoring feature parameters and are used to store the similarity distance between each pair of monitoring feature parameters. From multiple monitoring feature parameters, iterate through the monitoring feature parameter pairs that represent different detection feature parameter indices; Obtain the time series data of the monitoring feature parameter pairs from the time series data of each monitoring feature parameter; Based on the time series data of the monitored feature parameter pairs, the similarity distance of the monitored feature parameter pairs is calculated using a preset dynamic time warping algorithm; The similarity distances of the monitored feature parameter pairs are filled into the similarity matrix, and the step of traversing the monitored feature parameter pairs representing different detection feature parameter indices from multiple monitored feature parameters is continued until all multiple monitored feature parameters have been traversed, thus obtaining the DTW similarity matrix.

[0012] Optionally, based on the DTW similarity matrix, the correlation strength between each monitored feature parameter is quantified to obtain a gray correlation matrix, including: Initialize an empty initial correlation matrix. The size of the initial correlation matrix is ​​the same as that of the DTW similarity matrix. The rows and columns of the initial correlation matrix correspond to different monitoring feature parameters and are used to store the gray correlation between each pair of monitoring feature parameters. From multiple monitoring characteristic parameters, select one or more key monitoring characteristic parameters as reference parameters, and use the remaining monitoring characteristic parameters as comparison parameters; For each comparison sequence and reference sequence, calculate the difference sequence; Based on the difference sequence, calculate the correlation coefficient between each comparison sequence and the reference sequence; Calculate the average correlation coefficient to obtain the grey correlation degree between each comparison series and the reference series; The grey relational degree between each comparison sequence and the reference sequence is filled into the initial relational degree matrix to obtain the grey relational matrix; where, the difference sequence The calculation expression is:

[0013] in, It is an index for the time series. It is a reference sequence. It is the first A series of comparisons; Among them, the correlation coefficient The calculation expression is:

[0014] in, It is the resolution coefficient. It is the minimum value of the difference sequence. It is the maximum value of the difference sequence; where, gray relational degree The calculation expression is: in, It is the length of the time series.

[0015] Optionally, based on the slope state coupling network diagram, determine the dynamic evolution results of the coupling relationships of the slope to be monitored, including: Define the sliding time window based on the preset window width and preset sliding step size; Initialize the starting time point of the sliding time window; The current time window is determined by the starting time point, and the weights of all nodes and their edges within the current time window are extracted from the slope state coupling network diagram to obtain the coupling relationship subgraph within the current window. For each edge of the coupled subgraph within the current window, calculate the changing trend of the gray relation degree; Based on the changing trend of grey relational degree, the key characteristic parameters of the edges whose grey relational degree is enhanced or weakened and the changes in the coupling relationship on the edges are determined, and the dynamic evolution results of the coupling relationship of the slope to be monitored are obtained.

[0016] Optionally, the dynamic evolution results of the coupling relationship of the slope to be monitored include key characteristic parameters of the edges with enhanced or weakened grey relational degree and the changes in the coupling relationship on the edges; After obtaining the dynamic evolution results of the coupling relationship of the slope to be monitored, the following are also included: If the correlation strength between key characteristic parameters is greater than a preset strength threshold, it is determined that the slope to be monitored has potential risks; or, If the rate of change of the correlation strength between key characteristic parameters is greater than a preset rate threshold over a period of time, it is determined that there is an emergency risk event in the slope to be monitored. The system generates early warning information for the slope to be monitored and sends it to the client for early warning.

[0017] Optionally, a pre-trained slope macro-behavioral pattern recognition model is generated by following these steps: Collect and preprocess historical slope monitoring data, which includes historical structural response data, historical environmental driving data, and historical visual appearance data. Feature engineering is performed on preprocessed historical slope monitoring data to construct a historical multidimensional feature matrix that can characterize the changes of the target slope over time. Based on historical data and expert experience, the slope behavior patterns at each historical moment in the historical multidimensional feature matrix are labeled to obtain model training samples. Create a macroscopic behavior pattern recognition model for slopes; The training samples of the model are input into the slope macro-behavior pattern recognition model for machine learning, resulting in a pre-trained slope macro-behavior pattern recognition model.

[0018] Secondly, embodiments of this application provide a slope safety coupling analysis device based on multi-source monitoring data, the device comprising: The data preprocessing module is used to acquire and preprocess multi-source monitoring data of the slope to be monitored according to a preset cycle, so as to obtain structural response data, environmental driving data and visual appearance data. The feature matrix construction module is used to construct a multidimensional feature matrix that can characterize the changes of the monitored slope over time using structural response data, environmental driving data, and visual appearance data. The slope behavior pattern recognition module is used to input the multidimensional feature matrix into the pre-trained slope macro behavior pattern recognition model and output the current slope behavior pattern of the slope to be monitored. The slope safety coupling analysis module is used to perform slope safety coupling analysis based on the current slope behavior pattern and multidimensional feature matrix, and obtain the dynamic evolution results of the coupling relationship of the slope to be monitored.

[0019] The technical solutions provided in this application embodiment may include the following beneficial effects: In this embodiment, on the one hand, multi-dimensional data on structural response, environmental drivers, and visual appearance are acquired through multi-source monitoring equipment. Real-time acquisition and automated analysis of this multi-dimensional data can promptly capture changes in slope condition, thereby improving the accuracy of slope monitoring. On the other hand, a multi-dimensional feature matrix constructed from the multi-dimensional data, characterizing the changes of the monitored slope over time, can comprehensively reflect the dynamic response of the slope under the influence of multiple physical fields, accurately depicting the complete deformation and failure process of the slope. Simultaneously, based on the dynamic evolution results of coupling relationships, changes in slope condition can be monitored in real time, providing timely warnings of potential risks to the slope.

[0020] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0021] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0022] Figure 1 This is a schematic flowchart of a slope safety coupling analysis method based on multi-source monitoring data provided in an embodiment of this application; Figure 2 This is a schematic diagram of the overall area of ​​a slope to be monitored, provided in an embodiment of this application. Figure 3 This is a schematic diagram of a data overview result provided in an embodiment of this application; Figure 4 This is a schematic diagram of the interface of a behavior pattern recognition result provided in an embodiment of this application; Figure 5 This is a schematic diagram of the interface of a dynamic evolution result of a coupling relationship provided in an embodiment of this application; Figure 6 This is a schematic diagram of a warning interface received by a client, provided in an embodiment of this application. Figure 7 This is a schematic flowchart illustrating the process of generating dynamic evolution results of coupling relationships of a slope to be monitored, as provided in an embodiment of this application. Figure 8 This is a schematic block diagram of a slope safety coupling analysis process based on multi-source monitoring data provided in an embodiment of this application; Figure 9This is a flowchart illustrating a model training method for a slope macro-behavioral pattern recognition model provided in an embodiment of this application. Figure 10 This is a schematic diagram of the structure of a slope safety coupling analysis device based on multi-source monitoring data provided in an embodiment of this application; Figure 11 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0023] The following description and accompanying drawings fully illustrate specific embodiments of this application to enable those skilled in the art to practice them.

[0024] It should be understood that the described embodiments are merely some, not all, of the embodiments in this application. All other embodiments obtained by those skilled in the art based on the embodiments in this application without inventive effort are within the scope of protection of this application.

[0025] In the following description, when referring to the accompanying drawings, the same numbers in different drawings denote the same or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0026] In the description of this application, it should be understood that the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances. Furthermore, in the description of this application, unless otherwise stated, "multiple" refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship.

[0027] To address the existing technical problems, this application provides a slope safety coupling analysis method and apparatus based on multi-source monitoring data, thereby resolving the issues mentioned above. In the embodiments of this application, on the one hand, multi-dimensional data on structural response, environmental driving forces, and visual appearance are acquired through multi-source monitoring equipment. Real-time acquisition and automated analysis of this multi-dimensional data can promptly capture changes in slope condition, thus improving the accuracy of slope monitoring. On the other hand, a multi-dimensional feature matrix constructed from the multi-dimensional data, characterizing the changes of the monitored slope over time, can comprehensively reflect the dynamic response of the slope under the action of multiple physical fields, accurately depicting the complete deformation and failure process of the slope. Simultaneously, based on the dynamic evolution results of the coupling relationship, changes in slope condition can be monitored in real time, providing timely warnings of potential risks to the slope. The following detailed description uses exemplary embodiments.

[0028] The following will be combined with the appendix Figure 1 - Appendix Figure 9 This application provides a detailed description of the slope safety coupling analysis method based on multi-source monitoring data provided in its embodiments. This method can be implemented using a computer program and can run on a slope safety coupling analysis device based on multi-source monitoring data and the von Neumann architecture. This computer program can be integrated into applications or run as a standalone tool application.

[0029] Please see Figure 1 This document provides a flowchart illustrating a slope safety coupling analysis method based on multi-source monitoring data, applicable to the server side. For example... Figure 1 As shown, the method in this application embodiment includes the following steps: S101, the server acquires and preprocesses multi-source monitoring data of the slope to be monitored according to a preset cycle to obtain structural response data, environmental driving data and visual appearance data; In the slope monitoring system, the server is responsible for receiving data from sensors and monitoring equipment, processing and analyzing the data, and outputting results. The preset cycle is a pre-defined time interval used to periodically acquire and process monitoring data. The length of the cycle can be adjusted according to monitoring needs and slope stability requirements. The slope to be monitored is the area being coupled and analyzed by the server, for example... Figure 2As shown. Multi-source monitoring data refers to data from different monitoring devices and sensors, covering various physical responses and environmental factors of the slope. Structural response data refers to data reflecting the deformation and stress state of the slope structure, which can be obtained through devices such as displacement sensors (e.g., GNSS) and crack gauges. Environmentally driven data refers to data on external environmental factors affecting slope stability, such as rainfall and pore water pressure, which can be obtained through devices such as rain gauges and pore water pressure gauges. Visual appearance data is visual information about the slope surface obtained through visual devices (e.g., high-definition cameras, AI video PTZ cameras), such as abnormal appearance changes like cracks, water seepage, and spalling.

[0030] In some embodiments of this application, the specific process of collecting and preprocessing multi-source monitoring data of the slope to be monitored to obtain structural response data, environmental driving data, and visual appearance data includes: monitoring the three-dimensional displacement data of the surface of the slope to be monitored using GNSS equipment, and measuring the opening and closing changes of cracks at preset locations on the slope to be monitored using a crack gauge, as structural monitoring source data; recording the rainfall intensity and cumulative rainfall data of the slope to be monitored using an automated rain gauge, and monitoring the changes in the seepage field inside the slope to be monitored using a pore water pressure gauge, as environmental monitoring source data; visually inspecting the slope surface of the slope to be monitored using a high-definition AI video PTZ camera to identify abnormal appearance changes on the slope surface of the slope to be monitored, as visual monitoring source data; and performing data missing filling, abnormal data processing, and data synchronization processing on the structural monitoring source data, environmental monitoring source data, and visual monitoring source data to obtain structural response data, environmental driving data, and visual appearance data.

[0031] Specifically, in data missing filling, abnormal data processing, and data synchronization, spline interpolation is used to fill short-term data missing data; the 3σ criterion or isolated forest algorithm is used to identify and process abnormal jump points in the data stream; and resampling technology is used to unify sensor data of different frequencies to the same time base.

[0032] For example, targeting Figure 2 For the slope, a total of 38 sets of multimodal monitoring equipment were deployed on site, and three core monitoring profiles were set up along the route, with the following specific composition: (1) Surface deformation monitoring network: consisting of 9 GNSS monitoring points and 5 crack gauges. The GNSS monitoring points are deployed along the top, middle and bottom of each profile to capture the macroscopic deformation framework; the crack gauges are precisely deployed in the early tension crack development zone at the rear edge of the top of the slope. Among them, GNSS monitoring point G5, located in the key stress zone of the main profile (K0+650), is the core of this event analysis.

[0033] (2) Internal seepage pressure monitoring: Six sets of pore water pressure gauges were installed along the main profile borehole to accurately monitor the dynamic response of water pressure on the potential sliding zone. Among them, sensor P3 was installed near the red clay-bedrock interface below point G5 and was the key probe for capturing changes in deep seepage.

[0034] (3) Environmental and visual perception system: It consists of one set of automated rain gauges and two high-definition AI video PTZ cameras. The rain gauges provide the baseline rainfall data for the entire field, while the AI ​​video terminals perform all-weather automated visual inspection and appearance change recognition of the entire slope.

[0035] For example, a multidimensional feature matrix for a given hour might contain the following data: Structural response data: displacement increment is 0.5 mm, crack opening and closing change is 0.2 mm.

[0036] Environmental driving data: rainfall intensity was 10 mm / h, and pore water pressure was 10 kPa.

[0037] Visual appearance data: The area of ​​water seepage from the crack is 0.1 m².

[0038] S102, the server constructs a multi-dimensional feature matrix that can characterize the changes of the monitored slope over time by using structural response data, environmental driving data and visual appearance data; Among them, the multidimensional feature matrix is ​​a matrix containing multiple feature parameters. The data of each feature parameter changing over time are organized into a matrix form to characterize the dynamic state of the slope.

[0039] In some embodiments of this application, feature parameters, such as displacement increment, displacement rate, and crack opening / closing rate, are extracted from structural response data. Feature parameters, such as hourly rainfall intensity, cumulative rainfall, and pore water pressure change rate, are extracted from environmental driving data. Feature parameters, such as crack seepage area and spalling frequency, are extracted from visual appearance data. The extracted feature parameters are normalized to eliminate dimensional differences and ensure that the value range of all feature parameters is uniform. The normalized feature parameters are arranged in chronological order to form a multidimensional feature matrix. Each row of the matrix represents a time point, and each column represents a feature parameter.

[0040] Specifically, this application constructs a multi-layered monitoring feature parameter system for different types of data: (1) Structural response parameters: Displacement parameters include daily displacement increment and cumulative displacement, which reflect the current activity level and long-term deformation trend of the slope.

[0041] Rate parameters: The core is the displacement rate and its changing trend (i.e., acceleration). A continuous increase in rate, especially exhibiting a nonlinear acceleration trend, is a strong precursor signal of slope instability.

[0042] (2) Environment-related driver parameters: Rainfall parameters include hourly rainfall intensity, cumulative rainfall, and preceding effective rainfall. These indicators directly quantify the infiltration and recharge intensity of rainfall on the slope.

[14] .

[0043] Hydrological parameters: The core parameter is pore water pressure. This indicator is a crucial bridge connecting rainfall input and changes in the internal stress state of the slope.

[0044] (3) Visual appearance parameters: To transform unstructured visual information into analytical evidence, this application introduces a deep learning-based YOLOv5 object detection model. Trained on slope image samples, this model can assist in identifying and locating abnormal seepage marks, small rockfalls, and other apparent changes on the slope surface.

[0045] For example, when the preset period is 30 days, the overview of the slope data over 30 days is as follows: Figure 3 As shown.

[0046] S103, the server inputs the multi-dimensional feature matrix into the pre-trained slope macro behavior pattern recognition model and outputs the current slope behavior pattern of the slope to be monitored. The pre-trained slope macro-behavior pattern recognition model is a machine learning or deep learning model trained on historical data, capable of identifying the macro-behavior patterns of slopes under different conditions. Based on the input multi-dimensional feature matrix, the model outputs the current behavior pattern of the slope (e.g., stable fluctuation type, environment-driven type, structural deterioration and instability type, etc.). The current slope behavior pattern represents the slope's behavior state at the current point in time, such as stable fluctuation type, environment-driven type, structural deterioration and instability type, etc.

[0047] In some embodiments of this application, a pre-trained slope macroscopic behavior pattern recognition model is loaded. A multi-dimensional feature matrix is ​​input into the model, which processes the input data according to the trained parameters and outputs the current behavior pattern of the slope. The current slope behavior pattern output by the model can be a classification label (e.g., "stable fluctuation type," "environment-driven type," "structural degradation and instability type") or a probability distribution (e.g., "stable fluctuation type" probability is 0.8, "environment-driven type" probability is 0.2). The UI interface displaying the current slope behavior pattern in a given scenario is as follows: Figure 4 As shown.

[0048] In some embodiments of this application, a pre-trained slope macro-behavior pattern recognition model can be generated according to the following steps: collecting and preprocessing historical slope monitoring data, including historical structural response data, historical environmental driving data, and historical visual appearance data; performing feature engineering on the preprocessed historical slope monitoring data to construct a historical multidimensional feature matrix that can characterize the changes of the target slope over time; labeling the slope behavior pattern at each historical moment in the historical multidimensional feature matrix based on historical data and expert experience to obtain model training samples; creating a slope macro-behavior pattern recognition model; and inputting the model training samples into the slope macro-behavior pattern recognition model for machine learning to obtain the pre-trained slope macro-behavior pattern recognition model.

[0049] For example, when the above multidimensional feature matrix is ​​input into the model, the model will process the input data according to the trained parameters, and the model's output will be: Stable fluctuation type: probability is 0.1; Environment-driven: Probability is 0.7; Structural deterioration and instability type: probability is 0.2.

[0050] At this point, the model's final output of the current slope behavior pattern is "environment-driven" because this pattern has the highest probability (0.7).

[0051] S104, the server performs slope safety coupling analysis based on the current slope behavior pattern and multidimensional feature matrix to obtain the dynamic evolution results of the coupling relationship of the slope to be monitored.

[0052] In some embodiments of this application, the specific process of performing slope safety coupling analysis based on the current slope behavior pattern and multidimensional feature matrix to obtain the dynamic evolution result of the coupling relationship of the slope to be monitored includes: selecting multiple monitoring feature parameters related to the current slope behavior pattern from the multidimensional feature matrix; calculating the similarity distance between the time series of different monitoring feature parameters based on the multiple monitoring feature parameters to obtain the DTW similarity matrix; quantifying the correlation strength between each monitoring feature parameter based on the DTW similarity matrix to obtain the gray correlation matrix; constructing a slope state coupling network diagram of the slope to be monitored by using multiple monitoring feature parameters as network nodes and the correlation strength in the gray correlation matrix as the edge weights; and determining the dynamic evolution result of the coupling relationship of the slope to be monitored based on the slope state coupling network diagram.

[0053] Among them, the current slope behavior pattern is one of the following: stable fluctuation type, environment-driven type, and structural deterioration and instability type.

[0054] In some embodiments of this application, the specific process of selecting multiple monitoring feature parameters related to the current slope behavior mode from the multidimensional feature matrix includes: when the current slope behavior mode is stable fluctuation type, obtaining a first monitoring feature parameter related to low-frequency fluctuation from the multidimensional feature matrix, the first monitoring feature parameter including at least daily displacement increment, cumulative displacement, and short-term fluctuation of pore water pressure; or, when the current slope behavior mode is environment-driven type, obtaining a second monitoring feature parameter related to rainfall and pore water pressure from the multidimensional feature matrix, the second monitoring feature parameter including at least hourly rainfall intensity, cumulative rainfall, and pore water pressure change rate; or, when the current slope behavior mode is structural deterioration and instability type, obtaining a third monitoring feature parameter related to structural response from the multidimensional feature matrix, the third monitoring feature parameter including at least displacement rate, crack opening and closing rate, and acceleration.

[0055] In some embodiments of this application, the specific process of calculating the similarity distance between time series of different monitoring feature parameters based on multiple monitoring feature parameters to obtain the DTW similarity matrix includes: obtaining the monitoring value of each monitoring feature parameter changing over time to obtain the time series data of each monitoring feature parameter; initializing an empty similarity matrix with a size of n×n, where n is the number of multiple monitoring feature parameters, and the rows and columns of the similarity matrix correspond to different monitoring feature parameters, used to store the similarity distance between pairs of monitoring feature parameters; traversing the monitoring feature parameter pairs representing different detection feature parameter indices from multiple monitoring feature parameters; obtaining the time series data of monitoring feature parameter pairs from the time series data of each monitoring feature parameter; calculating the similarity distance of the monitoring feature parameter pairs based on the time series data of the monitoring feature parameter pairs and combining a preset dynamic time warping algorithm; filling the similarity distance of the monitoring feature parameter pairs into the similarity matrix, and continuing to execute the step of traversing the monitoring feature parameter pairs representing different detection feature parameter indices from multiple monitoring feature parameters until all multiple monitoring feature parameters have been traversed, thus obtaining the DTW similarity matrix.

[0056] In some embodiments of this application, the specific process of quantifying the correlation strength between various monitoring feature parameters and obtaining a gray correlation matrix based on the DTW similarity matrix includes: initializing an empty initial correlation matrix, the size of which is the same as the size of the DTW similarity matrix, with rows and columns of the initial correlation matrix corresponding to different monitoring feature parameters, used to store the gray correlation between pairs of monitoring feature parameters; selecting one or more key monitoring feature parameters as reference sequences from multiple monitoring feature parameters, and using the remaining monitoring feature parameters as comparison sequences; calculating the difference sequence for each comparison sequence and reference sequence; calculating the correlation coefficient between each comparison sequence and reference sequence based on the difference sequence; calculating the average of the correlation coefficients to obtain the gray correlation between each comparison sequence and reference sequence; filling the gray correlation between each comparison sequence and reference sequence into the initial correlation matrix to obtain the gray correlation matrix; wherein, the difference sequence... The calculation expression is:

[0057] in, It is an index for the time series. It is a reference sequence. It is the first A series of comparisons; Among them, the correlation coefficient The calculation expression is:

[0058] in, It is the resolution coefficient. It is the minimum value of the difference sequence. It is the maximum value of the difference sequence; where, gray relational degree The calculation expression is: in, It is the length of the time series.

[0059] In some embodiments of this application, the specific process of determining the dynamic evolution result of the coupling relationship of the slope to be monitored based on the slope state coupling network diagram includes: defining a sliding time window based on a preset window width and a preset sliding step size; initializing the starting time point of the sliding time window; determining the current time window based on the starting time point, and extracting the weights of all nodes and their edges within the current time window from the slope state coupling network diagram to obtain a coupling relationship subgraph within the current window; calculating the changing trend of gray relation degree for each edge of the coupling relationship subgraph within the current window; and determining the edges with enhanced or weakened gray relation degree and the key feature parameters of the changes in coupling relationship on these edges based on the changing trend of gray relation degree to obtain the dynamic evolution result of the coupling relationship of the slope to be monitored. This dynamic evolution result of the coupling relationship is, for example... Figure 5As shown.

[0060] Among them, the dynamic evolution results of the coupling relationship of the slope to be monitored include key characteristic parameters of the edges with enhanced or weakened grey relational degree and the changes in the coupling relationship on the edges.

[0061] In some embodiments of this application, after obtaining the dynamic evolution results of the coupling relationship of the slope to be monitored, the specific process of issuing an early warning includes: determining that the slope to be monitored has a potential risk when the correlation strength between key characteristic parameters is greater than a preset strength threshold; or determining that the slope to be monitored has an emergency risk event when the rate of change of the correlation strength between key characteristic parameters over a period of time is greater than a preset rate threshold; generating early warning information for the slope to be monitored and sending it to the client for early warning. The early warning interface received by the client is, for example... Figure 6 As shown.

[0062] For example Figure 7 As shown, Figure 7 This application provides a schematic flowchart of the process for generating dynamic evolution results of coupling relationships on a monitored slope. The multidimensional feature matrix contains various data types collected from slope monitoring, such as structural response data, environmentally driven data, and visual appearance data. These data are integrated into a matrix to comprehensively describe the slope's state. The multidimensional feature matrix is ​​input into a pre-trained model capable of identifying macroscopic behavior patterns of the slope. The model outputs the current slope behavior pattern, which may be one of stable fluctuation, environmentally driven, or structurally deteriorating and unstable. Based on the identified slope behavior pattern, feature parameters related to that pattern are selected from the multidimensional feature matrix. These parameters will be used for subsequent coupling analysis. The similarity distance between the selected feature parameter time series is calculated using the Dynamic Time Warping (DTW) algorithm. DTW is an effective method for measuring the similarity between two time series of potentially different lengths. Based on the DTW similarity matrix, the correlation strength between each monitored feature parameter is quantified to obtain a grey relational matrix. Selected feature parameters are used as network nodes, and the association strength in the grey relational degree matrix is ​​used as the edge weights to construct a coupling network diagram of slope state. A sliding time window technique is used to extract data within a specific time period from the coupling network diagram to analyze the changes in slope state over time. The dynamic evolution results of the slope coupling relationship are obtained, such as... Figure 5 As shown, users can click "Play Evolution" to view the dynamic analysis results for the day.

[0063] For example Figure 8 As shown, Figure 8This application provides a schematic flowchart of a slope safety coupling analysis process based on multi-source monitoring data. Multi-source monitoring data is sent to a server. On the server, a preprocessing module performs preliminary processing on the raw monitoring data. The preprocessed data is integrated into a multi-dimensional feature matrix, which contains multiple feature parameters. The data for each parameter changing over time is organized into a matrix to characterize the dynamic state of the slope. The multi-dimensional feature matrix is ​​input into a pre-trained model that can identify the macroscopic behavior patterns of the slope, such as stable fluctuation, environment-driven, or structural deterioration and instability. The model outputs the current slope behavior pattern, which is a classification of the current slope state and used to guide subsequent safety coupling analysis. Based on the current slope behavior pattern and the multi-dimensional feature matrix, slope safety coupling analysis is performed to obtain the dynamic evolution results of the slope coupling relationship. Based on the dynamic evolution results of the coupling relationship, a risk warning module assesses the slope stability and generates warning information when necessary. The warning information is finally sent to the client, which can be a monitoring center, an engineer's equipment, or other relevant parties, who then take appropriate measures based on the warning information.

[0064] In this embodiment, on the one hand, multi-dimensional data on structural response, environmental drivers, and visual appearance are acquired through multi-source monitoring equipment. Real-time acquisition and automated analysis of this multi-dimensional data can promptly capture changes in slope condition, thereby improving the accuracy of slope monitoring. On the other hand, a multi-dimensional feature matrix constructed from the multi-dimensional data, characterizing the changes of the monitored slope over time, can comprehensively reflect the dynamic response of the slope under the influence of multiple physical fields, accurately depicting the complete deformation and failure process of the slope. Simultaneously, based on the dynamic evolution results of coupling relationships, changes in slope condition can be monitored in real time, providing timely warnings of potential risks to the slope.

[0065] Please see Figure 9 This is a flowchart illustrating a model training method for a slope macro-behavioral pattern recognition model, as provided in this application embodiment. Figure 9 As shown, the method in this application embodiment may include the following steps: S201, collect and preprocess historical slope monitoring data, which includes historical structural response data, historical environmental driving data and historical visual appearance data; S202 uses preprocessed historical slope monitoring data for feature engineering to construct a historical multidimensional feature matrix that can characterize the changes of the target slope over time. S203, based on historical data and expert experience, the slope behavior patterns at each historical moment in the historical multidimensional feature matrix are labeled to obtain model training samples; S204, Create a macroscopic behavior pattern recognition model for slopes; S205. Input the model training samples into the slope macro-behavior pattern recognition model for machine learning to obtain the pre-trained slope macro-behavior pattern recognition model.

[0066] In some embodiments of this application, the model training samples are input into the slope macro-behavior pattern recognition model, and a model loss value is output. When the model loss value reaches its minimum, a pre-trained slope macro-behavior pattern recognition model is obtained. Alternatively, if the model loss value has not reached its minimum, the model parameters are adjusted, and the step of inputting the model training samples into the slope macro-behavior pattern recognition model for machine learning can continue until the model loss value reaches its minimum.

[0067] In this embodiment, on the one hand, multi-dimensional data on structural response, environmental drivers, and visual appearance are acquired through multi-source monitoring equipment. Real-time acquisition and automated analysis of this multi-dimensional data can promptly capture changes in slope condition, thereby improving the accuracy of slope monitoring. On the other hand, a multi-dimensional feature matrix constructed from the multi-dimensional data, characterizing the changes of the monitored slope over time, can comprehensively reflect the dynamic response of the slope under the influence of multiple physical fields, accurately depicting the complete deformation and failure process of the slope. Simultaneously, based on the dynamic evolution results of coupling relationships, changes in slope condition can be monitored in real time, providing timely warnings of potential risks to the slope.

[0068] The following are embodiments of the apparatus described in this application, which can be used to execute the embodiments of the method described in this application. For details not disclosed in the apparatus embodiments of this application, please refer to the embodiments of the method described in this application.

[0069] Please see Figure 10 This illustration shows a schematic diagram of a slope safety coupling analysis device based on multi-source monitoring data, provided in an exemplary embodiment of this application. This slope safety coupling analysis device based on multi-source monitoring data can be implemented as all or part of an electronic device through software, hardware, or a combination of both. The device 1 includes a data preprocessing module 10, a feature matrix construction module 20, a slope behavior pattern recognition module 30, and a slope safety coupling analysis module 40.

[0070] The data preprocessing module 10 is used to acquire and preprocess multi-source monitoring data of the slope to be monitored according to a preset cycle to obtain structural response data, environmental driving data and visual appearance data. The feature matrix construction module 20 is used to construct a multidimensional feature matrix that can characterize the changes of the monitored slope over time using structural response data, environmental driving data, and visual appearance data. The slope behavior pattern recognition module 30 is used to input the multidimensional feature matrix into the pre-trained slope macro behavior pattern recognition model and output the current slope behavior pattern of the slope to be monitored. The slope safety coupling analysis module 40 is used to perform slope safety coupling analysis based on the current slope behavior pattern and multidimensional feature matrix, and obtain the dynamic evolution results of the coupling relationship of the slope to be monitored.

[0071] It should be noted that the slope safety coupling analysis device based on multi-source monitoring data provided in the above embodiments is only illustrated by the division of the above functional modules when executing the slope safety coupling analysis method based on multi-source monitoring data. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the slope safety coupling analysis device based on multi-source monitoring data and the slope safety coupling analysis method embodiment based on multi-source monitoring data provided in the above embodiments belong to the same concept, and the implementation process is detailed in the method embodiment, which will not be repeated here.

[0072] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0073] In this embodiment, on the one hand, multi-dimensional data on structural response, environmental drivers, and visual appearance are acquired through multi-source monitoring equipment. Real-time acquisition and automated analysis of this multi-dimensional data can promptly capture changes in slope condition, thereby improving the accuracy of slope monitoring. On the other hand, a multi-dimensional feature matrix constructed from the multi-dimensional data, characterizing the changes of the monitored slope over time, can comprehensively reflect the dynamic response of the slope under the influence of multiple physical fields, accurately depicting the complete deformation and failure process of the slope. Simultaneously, based on the dynamic evolution results of coupling relationships, changes in slope condition can be monitored in real time, providing timely warnings of potential risks to the slope.

[0074] This application also provides a computer-readable medium having program instructions stored thereon, which, when executed by a processor, implement the slope safety coupling analysis method based on multi-source monitoring data provided in the above-described method embodiments.

[0075] This application also provides a computer program product containing instructions that, when run on a computer, cause the computer to execute the slope safety coupling analysis method based on multi-source monitoring data described in the above-described method embodiments.

[0076] Please see Figure 11 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 11 As shown, the electronic device 1000 may include: at least one processor 1001, at least one network interface 1004, a user interface 1003, a memory 1005, and at least one communication bus 1002.

[0077] The communication bus 1002 is used to realize the connection and communication between these components.

[0078] The user interface 1003 may include a display screen and a camera. Optionally, the user interface 1003 may also include a standard wired interface and a wireless interface.

[0079] The network interface 1004 may optionally include a standard wired interface or a wireless interface (such as a Wi-Fi interface).

[0080] The processor 1001 may include one or more processing cores. The processor 1001 connects to various parts within the electronic device 1000 using various interfaces and lines. It executes various functions and processes data by running or executing instructions, programs, code sets, or instruction sets stored in the memory 1005, and by calling data stored in the memory 1005. Optionally, the processor 1001 may be implemented using at least one hardware form selected from Digital Signal Processing (DSP), Field-Programmable Gate Array (FPGA), and Programmable Logic Array (PLA). The processor 1001 may integrate one or more of the following: a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), and a modem. The CPU primarily handles the operating system, user interface, and applications; the GPU is responsible for rendering and drawing the content to be displayed on the screen; and the modem handles wireless communication. It is understood that the modem may also be implemented as a separate chip, without being integrated into the processor 1001.

[0081] The memory 1005 may include random access memory (RAM) or read-only memory. Optionally, the memory 1005 may include a non-transitory computer-readable storage medium. The memory 1005 can be used to store instructions, programs, code, code sets, or instruction sets. The memory 1005 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as touch function, sound playback function, image playback function, etc.), instructions for implementing the above-described method embodiments, etc.; the data storage area may store data involved in the above-described method embodiments, etc. Optionally, the memory 1005 may also be at least one storage system located remotely from the aforementioned processor 1001. Figure 11 As shown, the memory 1005, which serves as a computer storage medium, may include an operating system, a network communication module, a user interface module, and a slope safety coupling analysis application based on multi-source monitoring data.

[0082] exist Figure 11 In the illustrated electronic device 1000, the user interface 1003 is mainly used to provide an input interface for the user and to acquire user input data; while the processor 1001 can be used to call the slope safety coupling analysis application based on multi-source monitoring data stored in the memory 1005, and specifically perform the following operations: According to the preset cycle, multi-source monitoring data of the slope to be monitored are acquired and preprocessed to obtain structural response data, environmental driving data and visual appearance data. A multidimensional feature matrix that can characterize the changes of the monitored slope over time is constructed by using structural response data, environmental driving data, and visual appearance data. The multidimensional feature matrix is ​​input into a pre-trained slope macro-behavior pattern recognition model, which outputs the current slope behavior pattern of the slope to be monitored. Based on the current slope behavior pattern and multidimensional feature matrix, slope safety coupling analysis is performed to obtain the dynamic evolution results of the coupling relationship of the slope to be monitored.

[0083] In one embodiment, when the processor 1001 collects and preprocesses multi-source monitoring data of the slope to be monitored to obtain structural response data, environmental driving data, and visual appearance data, it specifically performs the following operations: The three-dimensional displacement data of the surface of the slope to be monitored is monitored by GNSS equipment, and the opening and closing changes of cracks at preset locations of the slope to be monitored are measured by crack gauges, which serve as source data for structural monitoring. The rainfall intensity and cumulative rainfall data of the slope to be monitored are recorded by an automated rain gauge, and the changes in the seepage field inside the slope to be monitored are monitored by a pore water pressure gauge, which serves as the source data for environmental monitoring. A high-definition AI video PTZ camera is used to visually inspect the slope surface of the slope to be monitored, so as to identify abnormal appearance changes on the slope surface of the slope to be monitored, and use them as visual monitoring source data. Data missing data filling, abnormal data processing, and data synchronization processing are performed on structural monitoring source data, environmental monitoring source data, and visual monitoring source data to obtain structural response data, environmental driving data, and visual appearance data.

[0084] In one embodiment, when the processor 1001 performs slope safety coupling analysis based on the current slope behavior pattern and multidimensional feature matrix to obtain the dynamic evolution results of the coupling relationship of the slope to be monitored, it specifically performs the following operations: From the multidimensional feature matrix, multiple monitoring feature parameters related to the current slope behavior pattern are selected; Based on multiple monitoring feature parameters, the similarity distance between time series of different monitoring feature parameters is calculated to obtain the DTW similarity matrix; Based on the DTW similarity matrix, the correlation strength between each monitoring feature parameter is quantified to obtain the gray correlation matrix; Multiple monitoring feature parameters are used as network nodes, and the correlation strength in the gray relational degree matrix is ​​used as the weight of the edge to construct a slope state coupling network diagram of the slope to be monitored. Based on the slope state coupling network diagram, the dynamic evolution of the coupling relationship of the slope to be monitored is determined.

[0085] In one embodiment, when the processor 1001 performs the following operations to filter out multiple monitoring feature parameters related to the current slope behavior pattern from the multidimensional feature matrix: Given that the current slope behavior pattern is stable-fluctuation, the first monitoring characteristic parameter related to low-frequency fluctuations is obtained from the multidimensional feature matrix. This first monitoring characteristic parameter includes at least the daily displacement increment, cumulative displacement, and short-term fluctuations in pore water pressure; or... Given that the current slope behavior pattern is environment-driven, a second monitoring characteristic parameter related to rainfall and pore water pressure is obtained from the multidimensional feature matrix. This second monitoring characteristic parameter includes at least hourly rainfall intensity, cumulative rainfall, and pore water pressure change rate; or... Given that the current slope behavior pattern is structural deterioration and instability, a third monitoring characteristic parameter related to the structural response is obtained from the multidimensional characteristic matrix. The third monitoring characteristic parameter includes at least displacement rate, crack opening and closing rate, and acceleration.

[0086] In one embodiment, when the processor 1001 calculates the similarity distance between time series of different monitoring feature parameters based on multiple monitoring feature parameters to obtain the DTW similarity matrix, it specifically performs the following operations: The monitoring values ​​of each monitoring characteristic parameter are obtained as a function of time, thus obtaining the time series data of each monitoring characteristic parameter. Initialize an empty similarity matrix with a size of n×n, where n is the number of multiple monitoring feature parameters. The rows and columns of the similarity matrix correspond to different monitoring feature parameters and are used to store the similarity distance between each pair of monitoring feature parameters. From multiple monitoring feature parameters, iterate through the monitoring feature parameter pairs that represent different detection feature parameter indices; Obtain the time series data of the monitoring feature parameter pairs from the time series data of each monitoring feature parameter; Based on the time series data of the monitored feature parameter pairs, the similarity distance of the monitored feature parameter pairs is calculated using a preset dynamic time warping algorithm; The similarity distances of the monitored feature parameter pairs are filled into the similarity matrix, and the step of traversing the monitored feature parameter pairs representing different detection feature parameter indices from multiple monitored feature parameters is continued until all multiple monitored feature parameters have been traversed, thus obtaining the DTW similarity matrix.

[0087] In one embodiment, when the processor 1001 executes the operation of quantifying the correlation strength between various monitored feature parameters based on the DTW similarity matrix to obtain a gray correlation matrix, it specifically performs the following operations: Initialize an empty initial correlation matrix. The size of the initial correlation matrix is ​​the same as that of the DTW similarity matrix. The rows and columns of the initial correlation matrix correspond to different monitoring feature parameters and are used to store the gray correlation between each pair of monitoring feature parameters. From multiple monitoring characteristic parameters, select one or more key monitoring characteristic parameters as reference parameters, and use the remaining monitoring characteristic parameters as comparison parameters; For each comparison sequence and reference sequence, calculate the difference sequence; Based on the difference sequence, calculate the correlation coefficient between each comparison sequence and the reference sequence; Calculate the average correlation coefficient to obtain the grey correlation degree between each comparison series and the reference series; The grey relational degree between each comparison sequence and the reference sequence is filled into the initial relational degree matrix to obtain the grey relational matrix; where, the difference sequence The calculation expression is:

[0088] in, It is an index for the time series. It is a reference sequence. It is the first A series of comparisons; Among them, the correlation coefficient The calculation expression is:

[0089] in, It is the resolution coefficient. It is the minimum value of the difference sequence. It is the maximum value of the difference sequence; where, gray relational degree The calculation expression is: in, It is the length of the time series.

[0090] In one embodiment, when the processor 1001 executes the dynamic evolution result of the coupling relationship of the slope to be monitored based on the slope state coupling network diagram, it specifically performs the following operations: Define the sliding time window based on the preset window width and preset sliding step size; Initialize the starting time point of the sliding time window; The current time window is determined by the starting time point, and the weights of all nodes and their edges within the current time window are extracted from the slope state coupling network diagram to obtain the coupling relationship subgraph within the current window. For each edge of the coupled subgraph within the current window, calculate the changing trend of the gray relation degree; Based on the changing trend of grey relational degree, the key characteristic parameters of the edges whose grey relational degree is enhanced or weakened and the changes in the coupling relationship on the edges are determined, and the dynamic evolution results of the coupling relationship of the slope to be monitored are obtained.

[0091] In one embodiment, after obtaining the dynamic evolution results of the coupling relationship of the slope to be monitored, the processor 1001 also performs the following operations: If the correlation strength between key characteristic parameters is greater than a preset strength threshold, it is determined that the slope to be monitored has potential risks; or, If the rate of change of the correlation strength between key characteristic parameters is greater than a preset rate threshold over a period of time, it is determined that there is an emergency risk event in the slope to be monitored. The system generates early warning information for the slope to be monitored and sends it to the client for early warning.

[0092] In one embodiment, when the processor 1001 executes the generation of a pre-trained slope macroscopic behavior pattern recognition model, it specifically performs the following operations: Collect and preprocess historical slope monitoring data, which includes historical structural response data, historical environmental driving data, and historical visual appearance data. Feature engineering is performed on preprocessed historical slope monitoring data to construct a historical multidimensional feature matrix that can characterize the changes of the target slope over time. Based on historical data and expert experience, the slope behavior patterns at each historical moment in the historical multidimensional feature matrix are labeled to obtain model training samples. Create a macroscopic behavior pattern recognition model for slopes; The training samples of the model are input into the slope macro-behavior pattern recognition model for machine learning, resulting in a pre-trained slope macro-behavior pattern recognition model.

[0093] In this embodiment, on the one hand, multi-dimensional data on structural response, environmental drivers, and visual appearance are acquired through multi-source monitoring equipment. Real-time acquisition and automated analysis of this multi-dimensional data can promptly capture changes in slope condition, thereby improving the accuracy of slope monitoring. On the other hand, a multi-dimensional feature matrix constructed from the multi-dimensional data, characterizing the changes of the monitored slope over time, can comprehensively reflect the dynamic response of the slope under the influence of multiple physical fields, accurately depicting the complete deformation and failure process of the slope. Simultaneously, based on the dynamic evolution results of coupling relationships, changes in slope condition can be monitored in real time, providing timely warnings of potential risks to the slope.

[0094] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program for slope safety coupling analysis based on multi-source monitoring data can be stored in a computer-readable storage medium. When executed, the program can include the processes of the embodiments of the above methods. The storage medium for the program for slope safety coupling analysis based on multi-source monitoring data can be a magnetic disk, optical disk, read-only memory, or random access memory, etc.

[0095] The above-disclosed embodiments are merely preferred embodiments of this application and should not be construed as limiting the scope of this application. Therefore, any equivalent variations made in accordance with the claims of this application shall still fall within the scope of this application.

Claims

1. A slope safety coupling analysis method based on multi-source monitoring data, characterized in that, Applied to the server side, the method includes: According to the preset cycle, multi-source monitoring data of the slope to be monitored are acquired and preprocessed to obtain structural response data, environmental driving data and visual appearance data. Using the structural response data, environmental driving data, and visual appearance data, a multidimensional feature matrix is ​​constructed that can characterize the changes of the monitored slope over time. The multidimensional feature matrix is ​​input into a pre-trained slope macro-behavior pattern recognition model, which outputs the current slope behavior pattern of the slope to be monitored. Based on the current slope behavior pattern and the multidimensional feature matrix, slope safety coupling analysis is performed to obtain the dynamic evolution results of the coupling relationship of the slope to be monitored.

2. The method according to claim 1, characterized in that, Multi-source monitoring data of the slope to be monitored were collected and preprocessed to obtain structural response data, environmental driving data, and visual appearance data, including: The three-dimensional displacement data of the surface of the slope to be monitored is monitored by GNSS equipment, and the opening and closing changes of cracks at a preset location of the slope to be monitored are measured by crack gauge as source data for structural monitoring. The rainfall intensity and cumulative rainfall data of the slope to be monitored are recorded by an automated rain gauge, and the changes in the seepage field inside the slope to be monitored are monitored by a pore water pressure gauge, which serves as the source data for environmental monitoring. The slope surface of the slope to be monitored is visually inspected using a high-definition AI video PTZ camera to identify abnormal appearance changes on the slope surface of the slope to be monitored, which serve as the source data for visual monitoring. Data missing data filling, abnormal data processing, and data synchronization processing are performed on the structural monitoring source data, the environmental monitoring source data, and the visual monitoring source data to obtain structural response data, environmental driving data, and visual appearance data.

3. The method according to claim 1, characterized in that, The step of performing slope safety coupling analysis based on the current slope behavior pattern and the multidimensional feature matrix to obtain the dynamic evolution results of the coupling relationship of the slope to be monitored includes: From the multidimensional feature matrix, multiple monitoring feature parameters related to the current slope behavior pattern are selected; Based on the multiple monitoring feature parameters, the similarity distance between time series of different monitoring feature parameters is calculated to obtain the DTW similarity matrix; Based on the DTW similarity matrix, the correlation strength between each monitored feature parameter is quantified to obtain the gray correlation matrix; Using the multiple monitoring feature parameters as network nodes and the correlation strength in the gray correlation matrix as the edge weights, a slope state coupling network diagram of the slope to be monitored is constructed. Based on the slope state coupling network diagram, the dynamic evolution of the coupling relationship of the slope to be monitored is determined.

4. The method according to claim 3, characterized in that, The current slope behavior pattern is one of the following: stable fluctuation type, environment-driven type, and structural deterioration and instability type. The step of filtering out multiple monitoring feature parameters related to the current slope behavior pattern from the multidimensional feature matrix includes: When the current slope behavior pattern is the stable fluctuation type, a first monitoring feature parameter related to low-frequency fluctuations is obtained from the multi-dimensional feature matrix. This first monitoring feature parameter includes at least the daily displacement increment, cumulative displacement, and short-term fluctuations in pore water pressure; or... When the current slope behavior pattern is environment-driven, a second monitoring feature parameter related to rainfall and pore water pressure is obtained from the multidimensional feature matrix. This second monitoring feature parameter includes at least hourly rainfall intensity, cumulative rainfall, and pore water pressure change rate; or... When the current slope behavior mode is the structural deterioration and instability type, a third monitoring feature parameter related to the structural response is obtained from the multidimensional feature matrix. The third monitoring feature parameter includes at least displacement rate, crack opening and closing rate, and acceleration.

5. The method according to claim 3, characterized in that, The step of calculating the similarity distance between time series of different monitoring feature parameters based on the multiple monitoring feature parameters to obtain the DTW similarity matrix includes: The monitoring value of each monitoring feature parameter changes over time, thus obtaining the time series data of each monitoring feature parameter; Initialize an empty similarity matrix of size n×n, where n is the number of the plurality of monitoring feature parameters. The rows and columns of the similarity matrix correspond to different monitoring feature parameters and are used to store the similarity distance between each pair of monitoring feature parameters. From the plurality of monitoring feature parameters, traverse the monitoring feature parameter pairs that represent different detection feature parameter indices; Obtain the time series data of the monitoring feature parameter pair from the time series data of each monitoring feature parameter; Based on the time series data of the monitored feature parameter pairs, the similarity distance of the monitored feature parameter pairs is calculated using a preset dynamic time warping algorithm; The similarity distance of the monitoring feature parameter pairs is filled into the similarity matrix, and the step of traversing the monitoring feature parameter pairs representing different detection feature parameter indices from the multiple monitoring feature parameters is continued until all the multiple monitoring feature parameters have been traversed, and the DTW similarity matrix is ​​obtained.

6. The method according to claim 3, characterized in that, The step of quantifying the correlation strength between each monitored feature parameter based on the DTW similarity matrix to obtain a gray correlation matrix includes: Initialize an empty initial correlation matrix, the size of which is the same as the size of the DTW similarity matrix. The rows and columns of the initial correlation matrix correspond to different monitoring feature parameters, and are used to store the gray correlation between each pair of monitoring feature parameters. From the plurality of monitoring feature parameters, one or more key monitoring feature parameters are selected as reference parameters, and the remaining monitoring feature parameters are used as comparison parameters. For each comparison sequence and the reference sequence, calculate the difference sequence; Based on the difference sequence, calculate the correlation coefficient between each comparison sequence and the reference sequence; Calculate the average value of the correlation coefficients to obtain the grey correlation degree between each comparison sequence and the reference sequence; The grey relational degree between each comparison sequence and the reference sequence is filled into the initial relational degree matrix to obtain the grey relational matrix; wherein, the difference sequence The calculation expression is: in, It is an index for time series data. It is a reference sequence. It is the first A series of comparison sequences; Wherein, the correlation coefficient The calculation expression is: in, It is the resolution coefficient. It is the minimum value of the difference sequence. It is the maximum value of the difference sequence; where the gray relational degree is... The calculation expression is: in, It is the length of the time series.

7. The method according to claim 3, characterized in that, The determination of the dynamic evolution of the coupling relationship of the slope to be monitored based on the slope state coupling network diagram includes: Define the sliding time window based on the preset window width and preset sliding step size; Initialize the starting time point of the sliding time window; The current time window is determined based on the starting time point, and the weights of all nodes and their edges within the current time window are extracted from the slope state coupling network diagram to obtain the coupling relationship subgraph within the current window. For each edge of the coupling subgraph within the current window, calculate the changing trend of the gray relation degree; Based on the changing trend of the gray relation degree, the key characteristic parameters of the edges whose gray relation degree is enhanced or weakened and the changes in the coupling relationship on the edges are determined, and the dynamic evolution results of the coupling relationship of the slope to be monitored are obtained.

8. The method according to claim 1, characterized in that, The dynamic evolution results of the coupling relationship of the slope to be monitored include key characteristic parameters of the edges with enhanced or weakened grey relational degree and the changes in the coupling relationship on the edges. After obtaining the dynamic evolution results of the coupling relationship of the slope to be monitored, the method further includes: If the correlation strength between the key characteristic parameters is greater than a preset strength threshold, it is determined that the slope to be monitored has a potential risk; or, If the rate of change of the correlation strength between the key characteristic parameters is greater than a preset rate threshold over a period of time, it is determined that there is an emergency risk event in the slope to be monitored. The system generates early warning information for the slope to be monitored and sends it to the client for early warning.

9. The method according to claim 1, characterized in that, Generate a pre-trained slope macro-behavioral pattern recognition model according to the following steps: Collect and preprocess historical slope monitoring data, which includes historical structural response data, historical environmental driving data, and historical visual appearance data; Feature engineering is performed on preprocessed historical slope monitoring data to construct a historical multidimensional feature matrix that can characterize the changes of the target slope over time. Based on historical data and expert experience, the slope behavior patterns at each historical moment in the historical multidimensional feature matrix are labeled to obtain model training samples; Create a macroscopic behavior pattern recognition model for slopes; The training samples of the model are input into the slope macro-behavior pattern recognition model for machine learning to obtain the pre-trained slope macro-behavior pattern recognition model.

10. A slope safety coupling analysis device based on multi-source monitoring data, characterized in that, The device includes: The data preprocessing module is used to acquire and preprocess multi-source monitoring data of the slope to be monitored according to a preset cycle, so as to obtain structural response data, environmental driving data and visual appearance data. The feature matrix construction module is used to construct a multidimensional feature matrix that can characterize the changes of the monitored slope over time using the structural response data, environmental driving data, and visual appearance data. The slope behavior pattern recognition module is used to input the multidimensional feature matrix into a pre-trained slope macro behavior pattern recognition model and output the current slope behavior pattern of the slope to be monitored. The slope safety coupling analysis module is used to perform slope safety coupling analysis based on the current slope behavior pattern and the multidimensional feature matrix, and obtain the dynamic evolution results of the coupling relationship of the slope to be monitored.

Images