A slope monitoring method and monitoring system based on Internet artificial intelligence

Abstract

The present invention discloses a slope monitoring method and a monitoring system based on Internet artificial intelligence, specifically relating to the technical field of slope monitoring, and solves the problems of slope terrain reshaping and internal stress field parameter drift caused by engineering activities. First, three-dimensional laser point cloud data of the slope is collected and terrain reshaping is judged by using a registration algorithm; then, a long short-term memory network is used to analyze the change of stress data to evaluate the potential instability risk inside the slope; then, the combined influence range is determined by combining the judgment; when the combined influence range exceeds the safety range, a random forest regression algorithm is used to analyze the shallow groundwater data to quantify the degree of deterioration of seepage stability; finally, the overall safety risk of the slope is evaluated according to the degree of deterioration, and it is judged whether to carry out an early warning response. Real-time data update and model self-adaptation are realized, and the accuracy of early warning is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of slope monitoring. More specifically, the present invention relates to a slope monitoring method and a monitoring system based on Internet artificial intelligence. Background Art

[0002] Existing slope monitoring methods mostly target natural slopes with long-term stability and less human interference, and the monitoring models are usually trained based on the initial landform and initial physical parameters. However, in actual engineering activities such as local excavation, anchoring, and support construction, etc., the local terrain and the internal stress state of the slope often change rapidly, resulting in a dislocation between the distribution of training data and the actual working conditions, and obvious drift and mismatch of model parameters, reducing the accuracy of monitoring and early warning.

[0003] In order to solve the above problems, a technical solution is provided now. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art, embodiments of the present invention provide a slope monitoring method and a monitoring system based on Internet artificial intelligence to solve the problems raised in the above background art.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A slope monitoring method based on Internet artificial intelligence, comprising the following steps:

[0007] Collect three-dimensional laser point cloud data of the slope, calculate the spatial geometric change area caused by local terrain reshaping based on the point cloud registration algorithm, and judge whether slope terrain reshaping occurs;

[0008] By analyzing the change of stress data in the stress drift area, use the long short-term memory neural network algorithm to evaluate the potential instability risk level inside the slope;

[0009] Based on the judgment result of whether slope terrain reshaping occurs and the potential instability risk level inside the slope, determine the combined influence range of slope terrain reshaping and slope stress drift;

[0010] When the combined influence range of slope terrain reshaping and slope stress drift exceeds the safety range, use the random forest regression algorithm to analyze the real-time monitoring data of shallow groundwater within the combined influence range, and evaluate the deterioration degree of internal seepage stability of the slope;

[0011] According to the deterioration degree of internal seepage stability of the slope, evaluate the overall safety risk of the slope, and judge whether the slope reaches the state that requires early warning response.

[0012] In a preferred embodiment, three-dimensional laser point cloud data of the slope is collected, and the spatial geometric change area caused by local terrain reshaping is calculated based on the point cloud registration algorithm to determine whether slope terrain reshaping has occurred. Specifically:

[0013] Collect the initial three-dimensional laser point cloud data of the slope;

[0014] Collect the current three-dimensional laser point cloud data of the slope in real time;

[0015] Based on the feature matching point cloud registration algorithm, calculate the deviation;

[0016] Extract the point cloud deviation and determine the abnormal area;

[0017] Analyze the spatial deviation magnitude of the abnormal area to determine whether terrain reshaping has occurred: compare the average deviation value with the critical deviation threshold. If there is at least one abnormal area such that the average deviation value is greater than the critical deviation threshold, it is determined that slope terrain reshaping has occurred in the abnormal area.

[0018] In a preferred embodiment, by analyzing the change of stress data in the stress drift area, the long short-term memory neural network algorithm is used to evaluate the potential instability risk level inside the slope. Specifically:

[0019] Collect the slope stress sensor data and construct continuous time-series stress data;

[0020] Perform filtering processing on the stress data to remove noise;

[0021] Use the trend analysis method to screen out the high-change data in the stress drift area;

[0022] Adopt the sliding window technique to construct local stress sequence sample data;

[0023] Construct a long short-term memory neural network model to extract the dynamic stress change characteristics;

[0024] Quantitatively evaluate the potential instability risk level inside the slope based on the model output.

[0025] In a preferred embodiment, based on the judgment result of whether slope terrain reshaping has occurred and the potential instability risk level inside the slope, determine the combined influence range of slope terrain reshaping and slope stress drift. Specifically:

[0026] Preset the combined influence threshold and compare the combined influence coefficient with the combined influence threshold:

[0027] When the combined influence coefficient is greater than or equal to the combined influence threshold, the combined influence range of slope terrain reshaping and slope stress drift exceeds the safe range;

[0028] When the combined influence coefficient is less than the combined influence threshold, the combined influence range of slope terrain reshaping and slope stress drift does not exceed the safe range.

[0029] In a preferred embodiment, when the combined influence range of slope terrain reshaping and slope stress drift exceeds the safe range, the random forest regression algorithm is used to analyze the real-time monitoring data of shallow groundwater within the combined influence range to evaluate the deterioration degree of the internal seepage stability of the slope. Specifically:

[0030] Collect the real-time seepage data of shallow groundwater in the combined influence area;

[0031] Perform data correction on the real-time seepage data;

[0032] Construct a random forest regression model to train the relationship of groundwater seepage characteristics;

[0033] Predict the groundwater seepage data through the trained model;

[0034] Analyze the deterioration degree of slope seepage stability based on the predicted seepage data.

[0035] In a preferred embodiment, according to the deterioration degree of the internal seepage stability of the slope, evaluate the overall safety risk of the slope and determine whether the slope reaches the state that requires early warning response. Specifically:

[0036] Integrate the deterioration coefficient of seepage stability and calculate the overall safety risk probability of the slope;

[0037] Preset a safety critical threshold, compare the overall safety risk probability of the slope with the safety critical threshold, and determine whether it reaches the state that requires early warning response.

[0038] On the other hand, the present invention provides a slope monitoring system based on Internet artificial intelligence, including a terrain reshaping judgment unit, a risk level evaluation unit, an influence range determination unit, a deterioration degree evaluation unit, and a safety risk evaluation unit;

[0039] The terrain reshaping judgment unit collects the three-dimensional laser point cloud data of the slope, calculates the spatial geometric change area caused by local terrain reshaping based on the point cloud registration algorithm, and determines whether slope terrain reshaping occurs;

[0040] The risk level evaluation unit evaluates the potential instability risk level inside the slope by analyzing the stress data change in the stress drift area and using the long short-term memory neural network algorithm;

[0041] The influence range determination unit determines the combined influence range of slope terrain reshaping and slope stress drift based on the judgment result of whether slope terrain reshaping occurs and the potential instability risk level inside the slope;

[0042] When the combined influence range of slope terrain reshaping and slope stress drift exceeds the safe range, the deterioration degree evaluation unit uses the random forest regression algorithm to analyze the real-time monitoring data of shallow groundwater within the combined influence range, and evaluates the deterioration degree of the internal seepage stability of the slope;

[0043] The safety risk assessment unit evaluates the overall safety risk of the slope according to the deterioration degree of the internal seepage stability of the slope, and determines whether the slope reaches the state that requires early warning response.

[0044] The technical effects and advantages of a slope monitoring method and monitoring system based on Internet artificial intelligence according to the present invention:

[0045] By combining the three-dimensional laser point cloud registration algorithm and the long short-term memory network, the real-time and accurate identification of local terrain reshaping and stress parameter drift of the slope is realized, effectively solving the problems of parameter drift and model mismatch of the existing model in the scenario of severe human intervention; the random forest regression algorithm is introduced to analyze the deterioration degree of the seepage stability of shallow groundwater, improving the identification accuracy of the hidden risks inside the slope; by comparing the overall safety risk probability with the preset safety critical threshold in real time, it accurately determines whether the slope enters the early warning response state, improving the timeliness, accuracy and reliability of slope monitoring and early warning. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 It is a schematic diagram of a slope monitoring method based on Internet artificial intelligence according to the present invention;

[0047] Figure 2 It is a schematic diagram of the structure of a slope monitoring system based on Internet artificial intelligence according to the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0048] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention. Embodiment

[0049] Figure 1 A slope monitoring method based on Internet artificial intelligence according to the present invention is given, which includes the following steps:

[0050] Collect the three-dimensional laser point cloud data of the slope, calculate the spatial geometric change area caused by local terrain reshaping based on the point cloud registration algorithm, and determine whether slope terrain reshaping occurs;

[0051] By analyzing the stress data changes in the stress drift area, the long short-term memory neural network algorithm is used to evaluate the potential instability risk level inside the slope;

[0052] Based on the judgment result of whether slope terrain reshaping occurs and the potential instability risk level inside the slope, determine the combined influence range of slope terrain reshaping and slope stress drift;

[0053] When the combined influence range of slope terrain reshaping and slope stress drift exceeds the safe range, the random forest regression algorithm is used to analyze the real-time monitoring data of shallow groundwater within the combined influence range, and evaluate the deterioration degree of the internal seepage stability of the slope;

[0054] According to the deterioration degree of the internal seepage stability of the slope, evaluate the overall safety risk of the slope, and judge whether the slope reaches the state that requires early warning response.

[0055] Specifically, collect the three-dimensional laser point cloud data of the slope, calculate the spatial geometric change area caused by local terrain reshaping based on the point cloud registration algorithm, and judge whether slope terrain reshaping occurs, including:

[0056] Collect the initial three-dimensional laser point cloud data of the slope: Use a high-precision laser scanner to comprehensively scan the slope and collect the three-dimensional laser point cloud data at the initial moment. Define the initial point cloud data set as:

[0057] ; where represents the initial point cloud data set, that is, the set of all three-dimensional points on the slope surface collected by the laser scanner for the first time; represents the th three-dimensional point collected in the initial point cloud data set; represents the total number of three-dimensional points in the initial point cloud data set.

[0058] Each point is represented in Cartesian coordinate form as: ; where represents the coordinate value of the th point along the horizontal direction (X-axis) in the initial point cloud data collection; represents the coordinate value of the th point along the lateral direction (Y-axis) perpendicular to the X-axis in the initial point cloud data collection; represents the height or depth coordinate value of the th point along the vertical direction (Z-axis) in the initial point cloud data collection.

[0059] At the same time, in order to ensure the measurement accuracy, the error of the laser scanner is controlled within the allowable value Inside. The statistical outlier removal algorithm is used to filter the noise of the initial point cloud data, generating a set of preprocessed initial point cloud data .

[0060] Real-time acquisition of the current 3D laser point cloud data of the slope: When it is necessary to determine whether the terrain of the slope has been reshaped, the current 3D laser point cloud data of the slope is usually acquired in real time (currently) in the same area during a new monitoring period. Define the current point cloud data set as: ; where represents the current point cloud data set, that is, the set of all 3D points on the slope surface acquired in real time by the laser scanner during the new monitoring period; represents the th acquired 3D point in the current point cloud data set; represents the total number of 3D points in the current point cloud data set.

[0061] Similarly, each point is represented in Cartesian coordinate form as:

[0062] ; where represents the coordinate value of the th point along the horizontal direction (X-axis) in the initial point cloud data acquisition; represents the coordinate value of the th point along the lateral direction (Y-axis) perpendicular to the X-axis in the initial point cloud data acquisition; represents the height or depth coordinate value of the th point along the vertical direction (Z-axis) in the initial point cloud data acquisition.

[0063] The statistical outlier removal algorithm is also used to filter the noise of the current point cloud data set, generating a set of preprocessed current point cloud data .

[0064] Point cloud registration algorithm based on feature matching, calculating the deviation: In order to compare the terrain changes of the slope, a point cloud registration algorithm based on feature matching is used to spatially align the initial point cloud data set and the current point cloud data set . Define the rigid transformation matrix solved during the registration process as: ; where represents the rigid transformation matrix, which is used to map the current point cloud data to the coordinate system of the initial point cloud data through rotation and translation; is a 3×3 rotation matrix, describing the rotation relationship from the current point cloud data to the initial point cloud data; is a 3×1 translation vector, describing the translation relationship from the current point cloud data to the initial point cloud data.

[0065] The goal of the point cloud registration algorithm is to solve the optimal rigid transformation matrix so that each point in the initial point cloud data corresponds to the matching points in the current point cloud data to minimize the sum of the squares of the Euclidean distances, that is, to solve the minimum value of the following calculation formula:

[0066] ; where represents the set of allowed rigid transformations; represents the total number of successfully matched point pairs, that is, the number of point pairs that successfully establish a corresponding relationship between the initial point cloud data and the current point cloud data; represents the matching function, which defines how to select the points corresponding to the initial point cloud data from the current point cloud data ; represents the matching points in the current point cloud data corresponding to the initial point cloud data ; represents the Euclidean distance function; represents the coordinates of the matching points in the current point cloud data corresponding to the initial point cloud data after rigid transformation.

[0067] The point cloud registration algorithm adopts an iterative method and stops when the error between the current and previous iterations is reduced to a threshold or below.

[0068] Extract the point cloud deviation and determine the abnormal area: Define the deviation of each point calculated after registration as: ; where represents the spatial deviation value of the th matching point pair. For all spatial deviation values, by setting a predefined deviation threshold, mark the points with deviation values greater than the deviation threshold as abnormal points.

[0069] Adopt a clustering algorithm (such as the density-based DBSCAN algorithm) for grouping to group the abnormal points. Define the clustering result as: , ; where represents the th cluster, that is, the abnormal area, which contains all the abnormal points grouped into the same group by the clustering algorithm; represents the total number of identified abnormal areas.

[0070] Analyze the spatial deviation magnitude of the abnormal area and judge whether terrain reshaping has occurred: For each cluster in all the deviation values of the abnormal points are statistically analyzed to calculate the average deviation value: ; where represents the average deviation value of all points in the th abnormal area; Represents the spatial deviation value of the midpoint of the abnormal area ; Represents the number of midpoints in the abnormal area.

[0071] Compare the average deviation value with the critical deviation threshold: If there is at least one abnormal area such that the average deviation value is greater than the critical deviation threshold, it is determined that the abnormal area has undergone slope terrain reshaping.

[0072] To ensure the effectiveness of the abnormal area, it is also necessary to verify the spatial continuity and scale of the abnormal area. The boundary size of the abnormal area is used for judgment, and the minimum boundary size is defined as and the maximum boundary size is ; Among them, represents the minimum boundary size required for the abnormal area to have spatial coherence; represents the maximum boundary size required for the abnormal area to have spatial coherence.

[0073] Only when the boundary size of the abnormal area satisfies is the abnormal area considered valid, thus supporting the judgment of slope terrain reshaping. Among them represents the boundary size of the abnormal area.

[0074] Specifically, by analyzing the stress data changes in the stress drift area, a long short-term memory neural network algorithm is used to evaluate the potential instability risk level inside the slope, including:

[0075] Collect slope stress sensor data and construct continuous time-series stress data: Stress sensors are arranged inside the slope, and each sensor uploads the stress measurement value in real time through the Internet. Define the number of stress sensors as , the number of sampling points as . Define the originally collected stress data as: , ; . Among them, represents the stress measurement value of the th stress sensor at the th sampling point; represents the total number of stress sensors; represents the total number of sampling points.

[0076] Perform filtering processing on the stress data to remove noise interference: To improve the analysis accuracy, it is necessary to perform filtering operations on the original time-series stress data to remove high-frequency noise or outliers, and the filtered data is: ; Among them, represents the stress data obtained after filtering processing; A filtering function, which is used for noise suppression and smoothing processing, and algorithms such as Kalman filtering and wavelet threshold filtering can be adopted.

[0077] Use the trend analysis method to screen out the high-variation data in the stress drift area: After filtering, in order to identify whether there is local stress drift, trend analysis is required. Define a trend function to judge the direction and amplitude of stress change of the same stress sensor within a period of time. After applying it to the data of each stress sensor, ; where represents the th stress sensor at the th sampling point, and the larger the value, the stronger the stress change, and the smaller the value, the relatively gentle the stress change; represents the trend analysis function, which is used to calculate the change trend of the input data, and statistical indicators such as linear regression slope and moving average change rate are adopted; represents taking sampling points before and after the th sampling point as the center, and the stress data segment obtained after filtering composed of sampling points.

[0078] Set a high-variation threshold. If the stress change trend value is greater than the high-variation threshold, it means that the stress at the th sampling point has a significant change, and it can be considered that there is a sign of potential stress drift. At this time, the stress data at this point can be marked as high-variation data.

[0079] Adopt the sliding window technology to construct local stress sequence sample data: After screening out the high-variation data, it is necessary to analyze the local time series of the high-variation data to capture the dynamic evolution characteristics of stress. Introduce a sliding window , for each high-variation data, extract the stress data of a total of moments before and after it to construct a sequence sample. Define ; where represents the stress data sequence collected by the th stress sensor centered on the th sampling point; represents the length of the sliding window; respectively represent the stress data of taking sampling points forward and backward from the center sampling point in the th stress sensor.

[0080] Through the sliding window technology, the continuous change segments in the time dimension can be retained, providing time series context information for the training or prediction of the deep recurrent network.

[0081] Construct a long short-term memory neural network model to extract stress dynamic change features: To capture the long-term and short-term correlation features of slope stress in the time domain, a long short-term memory network structure is adopted. Define the data vector corresponding to each moment of the input stress data sequence as: ; where represents the stress data sequence input into the long short-term memory network; is the time index, and the range is from 0 to .

[0082] The long short-term memory network includes structures such as forget gates, input gates, and output gates, which are used to control the accumulation and forgetting of information in the time series, so as to better learn the hidden patterns of stress evolution over time. Define the output of the long short-term memory network as ; where represents the feature output of the th stress sensor at the th sampling point after being processed by the long short-term memory neural network. The output may be a scalar, vector, or high-dimensional feature representation, which is used to reflect the implicit pattern of stress dynamic change at this sampling point; represents the long short-term memory neural network model, which performs neural network operations on the input sequence. It internally includes gating mechanisms such as forget gates, input gates, and output gates, which are used to capture long-term and short-term dependencies in time series data.

[0083] Quantitatively evaluate the potential instability risk level inside the slope based on the model output: After obtaining the feature output of the long short-term memory neural network for the stress sequence, define a risk assessment function to normalize the feature output, so as to quantitatively evaluate the potential instability risk level inside the slope. Its expression is: ; where represents the instability risk value of the th stress sensor at the th sampling point, and the value ranges from 0 to 1; represents the risk assessment function, which can be implemented based on a multi-layer perceptron, normalization method, or pre-trained rules.

[0084] Preset the first instability risk threshold and the second instability risk threshold, and the first instability risk threshold is less than the second instability risk threshold. According to the range of the instability risk value, different risk levels are divided:

[0085] When the instability risk value is less than the first instability risk threshold, the potential instability risk level inside the slope is divided into a safe state;

[0086] When the instability risk value is greater than or equal to the first instability risk threshold and less than the second instability risk threshold, the potential instability risk level inside the slope is divided into a low-risk level;

[0087] When the instability risk value is greater than or equal to the second instability risk threshold and less than 1, the potential instability risk level inside the slope is classified as high risk.

[0088] Specifically, based on the judgment result of whether slope terrain reshaping occurs and the potential instability risk level inside the slope, determine the combined influence range of slope terrain reshaping and slope stress drift, including:

[0089] Define the slope terrain reshaping indicator , when , it indicates that slope terrain reshaping occurs, and when , it indicates that slope terrain reshaping does not occur.

[0090] The potential instability risk levels inside the slope include the safe state, low risk, and high risk. Similarly, define the instability risk level indicator , when , the potential instability risk level inside the slope is in the safe state; when , the potential instability risk level inside the slope is low risk; when , the potential instability risk level inside the slope is high risk.

[0091] According to the slope terrain reshaping indicator and the instability risk level indicator, calculate the combined influence coefficient, and its calculation formula is: ; where represents the combined influence coefficient, which is used to determine the combined influence range of slope terrain reshaping and slope stress drift; and are the weight coefficients of the slope terrain reshaping indicator and the instability risk level indicator respectively, and their values are determined according to the actual engineering situation, so as to reflect the relative importance of slope geometric changes and stress changes in the overall influence; represents the slope terrain reshaping indicator; represents the instability risk level indicator.

[0092] Preset the combined influence threshold, and compare the combined influence coefficient with the combined influence threshold:

[0093] When the combined influence coefficient is greater than or equal to the combined influence threshold, it indicates that there are significant geometric terrain reshaping and relatively high potential instability risks in the slope area at the same time, indicating that the slope area is facing potential stability deterioration under the combined influence and is very likely to develop into an actual instability phenomenon; at this time, the combined influence range of slope terrain reshaping and slope stress drift exceeds the safe range;

[0094] When the combined influence coefficient is less than the combined influence threshold, it indicates that there may have been a certain degree of terrain reshaping or stress drift locally in the slope, but the combined influence formed by their synergistic effect has not reached the critical level; the overall slope area is generally within the controllable or safe range, and no immediate emergency intervention measures are required; at this time, the combined influence range of slope terrain reshaping and slope stress drift does not exceed the safe range.

[0095] The setting of the combined influence threshold is a key parameter determined based on historical monitoring data, on-site experiments, and expert evaluations, and is used to divide whether the combined effect of slope terrain reshaping and stress drift has reached a dangerous level.

[0096] Specifically, when the combined influence range of slope terrain reshaping and slope stress drift exceeds the safe range, the random forest regression algorithm is used to analyze the real-time monitoring data of shallow groundwater within the combined influence range, and evaluate the degree of deterioration of the internal seepage stability of the slope, including:

[0097] Collect real-time seepage data of shallow groundwater in the combined influence area: Install shallow groundwater monitoring devices in the determined combined influence area (i.e., the intersection of the slope area where slope terrain reshaping occurs and the stress drift area), and use wireless transmission technology to collect groundwater seepage data in real time. Assume that there are groundwater monitoring devices installed in this area, and the data collected by each device within a continuous time period constitutes a time series data set. Define the groundwater seepage rate collected by the th groundwater monitoring device at the th sampling point as: ; where, represents the groundwater seepage rate collected by the th groundwater monitoring device at the th sampling point; represents the total number of groundwater monitoring devices; represents the total number of sampling points for each groundwater monitoring device within the monitoring period.

[0098] Perform data correction on the real-time seepage data: Since there may be instrument drift, environmental interference, and data noise in the actual measurement process, it is necessary to perform data correction and preprocessing on the original seepage data. Introduce a data correction function to process each collected original seepage data, and the corrected seepage data is: ; where, represents the seepage data of the th groundwater monitoring device at the th sampling point after correction; represents the correction function, usually using an algorithm based on moving average or Kalman filter.

[0099] The purpose of the calibration process is to eliminate the interference of instrument bias and outliers and obtain more stable and accurate groundwater seepage rate data. After data calibration, the seepage data can be used as reliable input for model training and prediction.

[0100] Construct a random forest regression model to train the groundwater seepage characteristic relationship: Use the calibrated seepage data to construct a random forest regression model to establish the characteristic relationship between groundwater seepage and slope infiltration status. First, construct a training sample set. Set the feature vector and the target variable to be respectively: , ; where, represents the feature vector constructed by the th groundwater monitoring device at the th sampling point, including the calibration data of the current and the previous sampling points; is the length of the feature vector; represents the corresponding target vector, that is, the seepage data predicted for the next sampling point (i.e., the sampling point).

[0101] After constructing the random forest regression model, the goal of model training is to minimize the prediction error, and its expression is: ; where, represents the set of parameters of all trees in the random forest regression model; represents the seepage data predicted by the random forest regression model based on the model parameters for the input feature vector ; represents the set of all adjustable parameters in the random forest regression model. These parameters include the thresholds for dividing nodes in each tree, the selected features, the predicted values of leaf nodes, etc. During the model training process, the parameters will be continuously optimized so that the model can more accurately map the input features to the target predicted values, thereby minimizing the prediction error.

[0102] After the random forest regression model training is completed, the model can fully capture the non-linear relationship and multi-moment correlation characteristics of groundwater seepage data and provide data support for prediction.

[0103] Predict groundwater seepage data through the trained model: Use the trained random forest regression model to predict the groundwater seepage in the joint influence area. For the th groundwater monitoring device at the th sampling point, use the feature vector to calculate the predicted value:

[0104] ; among them, represents the predicted seepage data of the th groundwater monitoring device at the th sampling point; represents the function mapping of the random forest regression model.

[0105] The prediction results can reveal the changes in the groundwater seepage path and the changing trend of the seepage velocity. By comparing the predicted data with the historical data, the dynamic evolution of groundwater movement can be judged.

[0106] Based on the predicted seepage data analysis, evaluate the deterioration degree of slope seepage stability: According to the predicted groundwater seepage data, use trend analysis and statistical methods to evaluate the deterioration degree of internal seepage stability of the slope. Set a seepage stability deterioration evaluation function to calculate the prediction results: ; among them, represents the seepage stability deterioration coefficient, reflecting the impact of seepage data changes on slope stability; represents the seepage stability deterioration evaluation function, which is used to map the predicted seepage data into the seepage stability deterioration coefficient, designed based on statistical indicators (such as standard deviation, change rate, cumulative change amount, etc.) or combined with expert experience rules, and is used to quantitatively reflect the relationship between groundwater seepage changes and slope seepage stability.

[0107] The larger the seepage stability deterioration coefficient, the more abnormal the internal moisture penetration of the slope is, and the more significant the influence of the water flow erosion on the rock and soil mass structure is. A larger seepage stability deterioration coefficient usually indicates that local saturation and shear strength decline may have occurred inside the slope, which in turn increases the risk of landslide instability, and engineering reinforcement, drainage and other measures need to be taken to ensure the overall safety of the slope.

[0108] Specifically, according to the deterioration degree of internal seepage stability of the slope, evaluate the overall safety risk of the slope, and judge whether the slope reaches the state that requires early warning response, including:

[0109] Integrate the seepage stability deterioration coefficient and calculate the overall safety risk probability of the slope: Integrate the seepage stability deterioration coefficients obtained by each groundwater monitoring device in the joint influence area: , ; among them, represents the total number of groundwater monitoring devices.

[0110] Through weighted average, integrate the seepage stability deterioration coefficient into the overall safety risk probability of the slope, and its calculation formula is: ; among them, represents the overall safety risk probability of the slope, and its value range is usually normalized to 0 to 1; represents the The weight factor corresponding to the slope area monitored by a groundwater monitoring device is determined based on the area of the region, the credibility of the monitoring data, or other actual engineering parameters, ensuring that the impact of each slope area on the overall risk assessment is proportionally distributed.

[0111] A preset safety critical threshold is compared with the overall safety risk probability of the slope to determine whether it reaches the state that requires early warning response:

[0112] When the overall safety risk probability of the slope is greater than or equal to the safety critical threshold, it indicates that the overall safety risk of the slope is a high risk, and the overall structural stability is severely threatened, reaching the state that requires early warning response; the management department needs to immediately activate the early warning response mechanism, take emergency measures such as engineering reinforcement, drainage diversion, and local support, and strengthen on-site inspections and monitoring;

[0113] When the overall safety risk probability of the slope is greater than or equal to the safety critical threshold, it indicates that the current overall state of the slope is relatively safe, the overall safety risk of the slope is a low risk, and it does not reach the state that requires early warning response; at this time, the routine monitoring work can be maintained, data review can be carried out regularly according to the normal monitoring cycle, the change trend of the slope hydrological conditions can be recorded, and the deterioration of the local seepage stability can be continuously monitored.

[0114] The setting of the safety critical threshold is based on a large amount of historical monitoring data, on-site test results, and numerical simulation analysis. A risk probability value is determined by statistical methods and expert experience, which reflects the stable limit state of the slope under various working conditions and takes into account factors such as geological conditions, rainfall influence, and engineering activities to ensure that it can truly reflect the overall stability of the slope. Embodiment

[0115] The difference between Embodiment 2 and Embodiment 1 of the present invention is that this embodiment introduces a slope monitoring system based on Internet artificial intelligence.

[0116] Figure 2 The structural schematic diagram of a slope monitoring system based on Internet artificial intelligence according to the present invention is given. A slope monitoring system based on Internet artificial intelligence includes a terrain reshaping judgment unit, a risk level assessment unit, an influence range determination unit, a deterioration degree assessment unit, and a safety risk assessment unit;

[0117] The terrain reshaping judgment unit collects the three-dimensional laser point cloud data of the slope, calculates the spatial geometric change area caused by local terrain reshaping based on the point cloud registration algorithm, and judges whether slope terrain reshaping occurs;

[0118] The risk level assessment unit evaluates the potential internal instability risk level of the slope by analyzing the stress data change in the stress drift area and using the long short-term memory neural network algorithm;

[0119] The influence range determination unit determines the combined influence range of slope terrain reshaping and slope stress drift based on the judgment result of whether slope terrain reshaping occurs and the potential instability risk level inside the slope.

[0120] When the combined influence range of slope terrain reshaping and slope stress drift exceeds the safety range, the deterioration degree evaluation unit uses the random forest regression algorithm to analyze the real-time monitoring data of shallow groundwater within the combined influence range, and evaluates the deterioration degree of the internal seepage stability of the slope.

[0121] The safety risk assessment unit evaluates the overall safety risk of the slope according to the deterioration degree of the internal seepage stability of the slope, and judges whether the slope reaches the state that requires early warning response.

[0122] The above formulas are all dimensionless and take their numerical values for calculation. The formulas are obtained by collecting a large amount of data for software simulation to get a formula closest to the real situation. The preset parameters and threshold selection in the formulas are set by those skilled in the art according to the actual situation.

[0123] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions described in the embodiments of the present application are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center by wired (such as infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains one or more collections of available media. The available medium can be a magnetic medium (such as a floppy disk, hard disk, magnetic tape), an optical medium (such as a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0124] Those of ordinary skill in the art can realize that the modules and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0125] Those skilled in the art can clearly understand that for the convenience and conciseness of description, the specific working processes of the systems, devices, and modules described above can refer to the corresponding processes in the foregoing method embodiments, and will not be elaborated herein.

[0126] In several embodiments provided in this application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the modules is only a logical function division, and there can be other division methods in actual implementation. For example, multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces, and the indirect couplings or communication connections of the devices or modules can be in electrical, mechanical, or other forms.

[0127] The modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical modules. They can be located in one place or distributed to multiple network modules. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0128] In addition, the functional modules in each embodiment of this application can be integrated into a processing module, or each module can exist physically alone, or two or more modules can be integrated into one module.

[0129] When the above-mentioned functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of this application. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs that can store program codes.

[0130] As described above, the above is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art within the technical scope disclosed in this application can easily think of changes or substitutions, which should all be covered by the protection scope of this application. Therefore, the protection scope of this application should be subject to the protection scope of the claims.

[0131] Finally: The above is only the preferred embodiment of the present invention and is not used to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A slope monitoring method based on Internet artificial intelligence, characterized in that: The steps include: Collect three-dimensional laser point cloud data of the slope, calculate the spatial geometric change area caused by local terrain reshaping based on the point cloud registration algorithm, and determine whether slope terrain reshaping has occurred; By analyzing the stress data changes in the stress drift area, the long short-term memory neural network algorithm is used to evaluate the potential instability risk level inside the slope. Based on the judgment result of whether the slope topography remodeling occurs and the potential instability risk level inside the slope, determine the combined impact range of the slope topography remodeling and the slope stress drift; When the combined influence range of slope topography remodeling and slope stress drift exceeds the safety range, the random forest regression algorithm is used to analyze the real-time monitoring data of shallow groundwater within the combined influence range to evaluate the degree of degradation of the internal seepage stability of the slope. Based on the degree of degradation of the internal seepage stability of the slope, the overall safety risk of the slope is assessed, and it is determined whether the slope has reached a state requiring an early warning response.

2. The slope monitoring method based on Internet artificial intelligence according to claim 1 is characterized in that: Collect the three-dimensional laser point cloud data of the slope, calculate the spatial geometric change area caused by local terrain reshaping based on the point cloud registration algorithm, and determine whether the slope terrain reshaping occurs. Specifically: Collect initial 3D laser point cloud data of slope; Real-time collection of current 3D laser point cloud data of the slope; Calculate the deviation based on the feature matching point cloud registration algorithm; Extract point cloud deviations and determine abnormal areas; Analyze the magnitude of spatial deviation in abnormal areas to determine whether topographic remodeling has occurred: compare the average deviation value with the critical deviation threshold. If there is at least one abnormal area such that the average deviation value is greater than the critical deviation threshold, it is determined that slope topographic remodeling has occurred in the abnormal area.

3. The slope monitoring method based on Internet artificial intelligence according to claim 2 is characterized in that: By analyzing the stress data changes in the stress drift area, the long short-term memory neural network algorithm is used to evaluate the potential instability risk level inside the slope, specifically: Collect slope stress sensor data and construct continuous time series stress data; Filter the stress data to remove noise; Use trend analysis method to filter out high-variation data in the stress drift area; The sliding window technique is used to construct the local stress series sample data; Construct a long short-term memory neural network model to extract the dynamic change characteristics of stress; The potential instability risk level inside the slope is quantitatively assessed based on the model output.

4. The slope monitoring method based on Internet artificial intelligence according to claim 3 is characterized in that: Based on the judgment result of whether the slope topography remodeling occurs and the potential instability risk level inside the slope, the combined influence range of the slope topography remodeling and the slope stress drift is determined, specifically: Preset the joint impact threshold and compare the joint impact coefficient with the joint impact threshold: When the combined influence coefficient is greater than or equal to the combined influence threshold, the combined influence range of slope topography remodeling and slope stress drift exceeds the safety range; When the combined influence coefficient is less than the combined influence threshold, the combined influence range of slope topography remodeling and slope stress drift does not exceed the safety range.

5. The slope monitoring method based on Internet artificial intelligence according to claim 4 is characterized in that: When the combined influence range of slope topography remodeling and slope stress drift exceeds the safety range, the random forest regression algorithm is used to analyze the real-time monitoring data of shallow groundwater within the combined influence range to evaluate the degree of degradation of the internal seepage stability of the slope, specifically: Collect real-time seepage data of shallow groundwater in the joint impact area; Carry out data correction on real-time seepage data; Construct a random forest regression model to train the groundwater seepage characteristic relationship; Predict groundwater seepage data by training models; The degree of slope seepage stability degradation is analyzed based on the predicted seepage data.

6. The slope monitoring method based on Internet artificial intelligence according to claim 5 is characterized in that: According to the degree of degradation of the seepage stability inside the slope, the overall safety risk of the slope is assessed, and it is determined whether the slope has reached a state that requires an early warning response, specifically: Integrate the seepage stability degradation coefficient and calculate the overall safety risk probability of the slope; A safety critical threshold is preset, and the overall safety risk probability of the slope is compared with the safety critical threshold to determine whether a state requiring an early warning response has been reached.

7. A slope monitoring system based on Internet artificial intelligence, used to implement a slope monitoring method based on Internet artificial intelligence as described in any one of claims 1 to 6, characterized in that: It includes terrain reshaping judgment unit, risk level assessment unit, impact range determination unit, degradation degree assessment unit and safety risk assessment unit; The terrain reshaping judgment unit collects the three-dimensional laser point cloud data of the slope, calculates the spatial geometric change area caused by the local terrain reshaping based on the point cloud registration algorithm, and judges whether the slope terrain reshaping occurs; The risk level assessment unit analyzes the stress data changes in the stress drift area and uses the long short-term memory neural network algorithm to assess the potential instability risk level inside the slope; The impact range determination unit determines the combined impact range of slope topography remodeling and slope stress drift based on the judgment result of whether slope topography remodeling occurs and the potential instability risk level inside the slope; When the combined influence range of slope topography remodeling and slope stress drift exceeds the safety range, the deterioration degree assessment unit uses the random forest regression algorithm to analyze the real-time monitoring data of shallow groundwater within the combined influence range to assess the degree of degradation of the internal seepage stability of the slope; The safety risk assessment unit evaluates the overall safety risk of the slope based on the degree of degradation of the internal seepage stability of the slope, and determines whether the slope has reached a state requiring an early warning response.

Images