Garden ecological slope protection stability monitoring system and method

Abstract

The invention provides a garden ecological slope protection stability monitoring system and method, and relates to the technical field of garden slope protection stability monitoring, and the method comprises the steps: dividing a preset monitoring range of a target garden ecological slope protection into a plurality of terrain structure grids, obtaining a key slope body pressure monitoring index and a key geological environment monitoring index of each topographic structure grid; time feature extraction is carried out on the key slope body pressure monitoring indexes and the key geological environment monitoring indexes of all the terrain structure grids, and corresponding dynamic change trend features are generated; spatial feature extraction is carried out on the dynamic change trend features of the key slope body pressure monitoring indexes and the key geological environment monitoring indexes of all the terrain structure grids, and a corresponding spatial-temporal feature sequence is generated; and constructing a slope protection stability early warning model based on a deep learning algorithm, inputting the spatial-temporal feature sequence of each topographic structure grid into the slope protection stability early warning model, and obtaining stability early warning information of the garden ecological slope protection.

Description

Technical Field

[0001] This invention relates to the field of landscape slope stability monitoring technology, specifically a landscape ecological slope stability monitoring system and method. Background Technology

[0002] As an important facility combining ecological protection and landscape functions, the stability of garden ecological slope protection directly affects the safety of the surrounding environment and ecological balance. Currently, the analysis of monitoring data for garden ecological slope protection is mostly limited to a single time or spatial dimension, failing to effectively integrate spatiotemporal characteristics and making it difficult to reflect the dynamic evolution of slope stability. Furthermore, it exhibits poor adaptability to complex slope environments, insufficient accuracy and reliability in early warning, and cannot provide timely and accurate stability warning information. This hinders the development and upgrading of ecological slope protection technology, failing to meet the actual needs of refined and intelligent monitoring of garden ecological slope protection.

[0003] Therefore, a system and method for monitoring the stability of ecological slope protection in gardens are provided. Summary of the Invention

[0004] In order to solve the above-mentioned technical problems, the purpose of this invention is to provide a system and method for monitoring the stability of garden ecological slope protection.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a method for monitoring the stability of garden ecological slope protection, the method comprising: Based on the pre-set monitoring range of the target garden ecological slope protection, combined with slope topographic structure data, historical slope pressure data and historical geological environment data, the pre-set monitoring range of the target garden ecological slope protection is divided into several topographic structure grids, and key slope pressure monitoring indicators and key geological environment monitoring indicators of each topographic structure grid are obtained. The time characteristics of key slope pressure monitoring indicators and key geological environment monitoring indicators of each terrain structure grid are extracted to generate corresponding dynamic change trend characteristics. Spatial features are extracted from the dynamic change trends of key slope pressure monitoring indicators and key geological environment monitoring indicators of each terrain structure grid to generate corresponding spatiotemporal feature sequences. A slope protection stability early warning model is constructed based on deep learning algorithms. The spatiotemporal feature sequences of various terrain structure grids are input into the slope protection stability early warning model to obtain stability early warning information of garden ecological slope protection.

[0006] Furthermore, the process of obtaining key slope pressure monitoring indicators and key geological environment monitoring indicators for each terrain structure grid includes: Based on the acquisition time of historical slope pressure data and historical geological environment data of each terrain structure grid, the historical slope pressure data and historical geological environment data of each terrain structure grid are sorted by time according to the corresponding acquisition time, thereby obtaining the time series sequence of slope pressure data and geological environment data of each terrain structure grid. Then, the slope pressure data time series and geological environment data time series corresponding to each terrain structure grid are extracted; Pre-set standard index threshold ranges corresponding to several slope pressure monitoring indicators; determine whether the time series of slope pressure index values ​​corresponding to each terrain structure grid is within the corresponding standard index threshold range, and thus obtain key slope pressure monitoring indicators. The time series of slope pressure index values ​​corresponding to the key slope pressure monitoring indicators of each terrain structure grid that exceed the corresponding standard index threshold range are marked as key time periods. Based on the Pearson correlation coefficient and according to the key time period, key geological environment monitoring indicators corresponding to each terrain structure grid are obtained.

[0007] Furthermore, the process of extracting the temporal characteristics of key slope pressure monitoring indicators and key geological environment monitoring indicators for each topographic structure grid includes: Construct a monitoring index prediction model; input the time series of slope pressure index values ​​corresponding to the key slope pressure monitoring indexes of each terrain structure grid in the current collection period and the time series of geological environment index values ​​corresponding to the key geological environment monitoring indexes into the trained monitoring index prediction model, thereby obtaining the numerical change trend characteristics of slope pressure monitoring indexes and geological environment monitoring indexes.

[0008] Furthermore, the process of constructing a predictive model for monitoring indicators includes: Time-series data related to key slope pressure monitoring indicators and key geological environment monitoring indicators in the slope stability risk early warning dataset were selected to form the basic dataset for model training. The basic dataset was preprocessed to obtain a standardized time-series dataset. The standardized time-series dataset was then divided into training set, validation set and test set, and the division process adopted stratified sampling. The main framework of the model is built based on a temporal convolutional neural network. The model is trained and its parameters are optimized. The relevant parameters for model training are set, the training set is input into the model for training, and the model performance is verified using a validation set. The model parameters are adjusted based on the validation set loss. After training, the model performance is evaluated using a test set. If the evaluation indicators meet the preset requirements, the model training is completed and the model parameters are saved. If the requirements are not met, the model structure parameters are adjusted or the dataset is reprocessed and trained again until the requirements are met, thus obtaining the corresponding monitoring indicator prediction model.

[0009] Furthermore, the process of extracting spatial features of the dynamic change trends of key slope pressure monitoring indicators and key geological environment monitoring indicators for each topographic structure grid includes: A unique grid spatial coordinate identifier is assigned to each terrain structure grid, the adjacency relationship and spatial distance parameters between grids are defined, and the grid spatial topology is constructed. The dynamic trend characteristics of key slope pressure monitoring indicators and key geological environment monitoring indicators for each terrain structure grid are obtained and bound to the corresponding grid spatial coordinate identifiers to form an associated dataset. If the slope pressure monitoring index and geological environment monitoring index of the terrain structure grid in the grid space topology do not exceed the corresponding threshold, they are used as the associated dataset of the corresponding terrain structure grid in the grid space topology. The graph attention network is used to learn the topology of the grid space, and the dynamic trend characteristics of the corresponding key slope pressure monitoring indicators and key geological environment monitoring indicators are input into the graph attention network. Take any terrain structure grid in the grid space topology as the target terrain structure grid and as the center grid, and denote the terrain structure grids adjacent to it as the neighborhood grids. By obtaining the collaborative influence coefficient of the neighboring grid on the central grid, the collaborative influence coefficient is used as the attention weight to assign weights to each neighboring grid. Based on feature aggregation technology, spatiotemporal feature sequences corresponding to the dynamic change trends of key slope pressure monitoring indicators and key geological environment monitoring indicators of each terrain structure grid are generated.

[0010] Furthermore, the process of obtaining the cooperative influence coefficient of the neighboring grid on the central grid includes: Key slope pressure monitoring indicators and key geological environment monitoring indicators of the central grid and each neighboring grid are obtained. Based on the key slope pressure monitoring indicators and key geological environment monitoring indicators of the central grid and each neighboring grid, the calculation formula of the synergistic influence coefficient of the neighboring grid on the central grid is obtained, and then the corresponding synergistic influence coefficient is obtained.

[0011] Furthermore, the process of constructing a slope protection stability early warning model based on deep learning algorithms includes: The study acquires the numerical trend characteristics and spatiotemporal characteristic sequences of slope pressure monitoring indicators for each terrain structure grid within the historical data acquisition period, as well as the numerical trend characteristics and spatiotemporal characteristic sequences of geological environment monitoring indicators for key geological environment monitoring indicators; and groups and labels these indicators as follows: It is a natural number; Will The data from the group is used as sample data, and Less than The natural numbers are used to obtain the mean of the sample data, which is denoted as the sample set; the data from the remaining historical collection periods are used as the test set; and a training sample set is formed based on the sample set and the test set. A standard prediction model is constructed based on deep learning algorithms. The training sample set is then input into the standard prediction model to train it until the loss function is stable. The model parameters are saved, and the standard prediction model is tested using a test set to output a slope stability early warning model.

[0012] A second aspect of the present invention also provides a garden ecological slope stability monitoring system, including a data acquisition module, a time feature extraction module, a spatial feature extraction module, and a slope stability prediction module; The data acquisition module is used to divide the preset monitoring range of the target garden ecological slope protection into several topographic structure grids based on the preset monitoring range of the target garden ecological slope protection, combined with slope topographic structure data, historical slope pressure data and historical geological environment data, and to obtain the key slope pressure monitoring indicators and key geological environment monitoring indicators of each topographic structure grid. The time feature extraction module is used to extract the time features of key slope pressure monitoring indicators and key geological environment monitoring indicators of each terrain structure grid, and generate corresponding dynamic change trend features. The spatial feature extraction module is used to extract the spatial features of the dynamic change trends of key slope pressure monitoring indicators and key geological environment monitoring indicators of each terrain structure grid, and generate corresponding spatiotemporal feature sequences. The slope stability prediction module constructs a slope stability early warning model based on deep learning algorithms. It inputs the spatiotemporal feature sequences of various terrain structure grids into the slope stability early warning model to obtain stability early warning information for garden ecological slopes.

[0013] Compared with existing technologies, the beneficial effects of this invention are as follows: The garden ecological slope stability monitoring method and system provided by this invention, by combining multi-source historical data to divide the terrain structure into grids and selectively screening key monitoring indicators, achieves full coverage of the monitoring range and precise focus of monitoring priorities, providing a high-quality data foundation for subsequent stability analysis; by using a monitoring indicator prediction model to extract dynamic change trend features in the time dimension and combining graph attention networks to mine spatial correlation features, a complete spatiotemporal feature sequence is generated, fully capturing the dynamic evolution law and spatial correlation characteristics of slope stability; the slope stability early warning model built based on deep learning algorithms effectively improves the accuracy and reliability of early warning information, and can promptly locate potential risk areas and output targeted early warning information. The entire method realizes comprehensive, refined, and intelligent monitoring of garden ecological slope stability, providing a scientific basis for slope maintenance and management, ensuring the safety of the surrounding environment and ecological balance. At the same time, the various modules of the system work together efficiently and are adaptable to garden ecological slopes with different terrain features, and have broad application value. Attached Figure Description

[0014] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0015] Figure 1 This is a schematic diagram illustrating the steps of a method for monitoring the stability of ecological slope protection in gardens.

[0016] Figure 2 This is a schematic diagram of a module for a garden ecological slope protection stability monitoring system. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be described in detail below. Obviously, the described embodiments are merely some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other implementation methods obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0018] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0019] Example 1 like Figure 1 As shown, a method for monitoring the stability of ecological slope protection in gardens includes the following steps: Step S1: Based on the preset monitoring range of the target garden ecological slope protection, combined with slope topographic structure data, historical slope pressure data and historical geological environment data, the preset monitoring range of the target garden ecological slope protection is divided into several topographic structure grids, and the key slope pressure monitoring indicators and key geological environment monitoring indicators of each topographic structure grid are obtained. Step S2: Extract the time characteristics of key slope pressure monitoring indicators and key geological environment monitoring indicators for each terrain structure grid, and generate corresponding dynamic change trend characteristics. Step S3: Extract the spatial features of the dynamic change trends of key slope pressure monitoring indicators and key geological environment monitoring indicators for each terrain structure grid, and generate the corresponding spatiotemporal feature sequences. Step S4: Construct a slope protection stability early warning model based on deep learning algorithm, input the spatiotemporal feature sequence of each terrain structure grid into the slope protection stability early warning model, and obtain stability early warning information of garden ecological slope protection.

[0020] It should be further explained that, in the specific implementation process, based on the preset monitoring range of the target ecological slope protection, and combined with slope topographic structure data, historical slope pressure data, and historical geological environment data, the preset monitoring range of the target ecological slope protection is divided into several topographic structure grids, and the specific process of obtaining the slope pressure monitoring indicators and geological environment monitoring indicators for each topographic structure grid includes: Optionally, in this embodiment of the application, a slope stability risk early warning dataset is pre-constructed, including slope topographic structure data, historical slope pressure data, and historical geological environment data within the target garden ecological slope protection monitoring area. Slope topographic features are extracted based on the slope topographic structure data in the slope stability risk early warning dataset. The target garden ecological slope protection monitoring area is divided into several topographic structure grids based on the slope topographic features. The historical slope pressure data and historical geological environment data in the slope stability risk early warning dataset are divided into historical slope pressure data and historical geological environment data for each topographic structure grid. Based on historical slope pressure data and historical geological environment data of each terrain structure grid, slope pressure monitoring indicators and geological environment monitoring indicators of each terrain structure grid are obtained. Multiple data monitoring points are set up in each terrain structure grid. These monitoring points collect real-time values ​​of the slope pressure monitoring indicators and geological environment monitoring indicators corresponding to their respective terrain structure grids, and mark the collection time and set the collection period. The slope pressure monitoring indicators for each data monitoring point include both lateral and longitudinal slope pressure.

[0021] It should be noted that the slope pressure monitoring indicators include, but are not limited to, surface slope pressure monitoring indicators, deep slope pressure monitoring indicators, and pressure variation differences; the geological environment monitoring indicators include, but are not limited to, soil porosity, soil moisture content, and soil shear strength. Different geological environment monitoring indicators will produce corresponding pressure changes on the slope.

[0022] Furthermore, the significance of obtaining slope pressure monitoring indicators and geological environment monitoring indicators for each topographic structure grid lies in the following: Different slope topography and geomorphological characteristics lead to differences in slope pressure and geological environment across different regions. For example, in gentle slope areas, the soil layer thickness is usually more uniform, and the pressure distribution generated by the slope's own weight is relatively gentle. It is necessary to focus on monitoring the impact of slope surface pressure monitoring indicators and soil moisture content on pressure transmission in geological environment monitoring indicators, while also monitoring the deep slope pressure monitoring indicators and pressure variation differences in slope surface pressure monitoring indicators. In medium-slope areas, which are in the transitional slope stage, slope stability is easily affected. The impact of rain infiltration is a key factor. Monitoring of slope pressure indicators includes surface pressure, deep pressure, pressure variation, and soil porosity and moisture content in geological environment monitoring. Simultaneously, surface pressure indicators are also monitored. In steep slopes, the shear force along the slope surface is greater due to the slope's weight component, making local slippage more likely. Therefore, monitoring of surface pressure, deep pressure, pressure variation, and soil shear strength, porosity, and moisture content in geological environment monitoring is crucial for pressure transmission.

[0023] Meanwhile, considering the impact of seasonal changes on slope stability, different regions may need to focus on monitoring different key geological environments. For example, during the rainy season, rainwater can easily penetrate into the deep layers of the slope, so all regions need to strengthen the monitoring of soil moisture content. However, gentle slope areas need to pay extra attention to the increasing trend of surface pressure caused by rainwater soaking, while steep slope areas need to focus on monitoring the sudden changes in deep pressure caused by rainwater softening the soil layer. Taking all the above factors into account, the monitoring indicators for different regions may vary and be adjusted according to actual needs and circumstances.

[0024] It should be further explained that, in the specific implementation process, the process of obtaining key slope pressure monitoring indicators and key geological environment monitoring indicators for each terrain structure grid based on historical slope pressure data and historical geological environment data includes: Optionally, in this embodiment of the application, based on the acquisition time of the historical slope pressure data and historical geological environment data of each terrain structure grid, the historical slope pressure data and historical geological environment data of each terrain structure grid are sorted by time according to the corresponding acquisition time, thereby obtaining the time series sequence of slope pressure data and geological environment data of each terrain structure grid. Extract several slope pressure monitoring indicators and their corresponding slope pressure monitoring indicator value time series from the slope pressure data time series corresponding to each terrain structure grid, and denot them as slope pressure indicator value time series. Extract several geological environment monitoring indicators and their corresponding index value time series from the geological environment data time series corresponding to each topographic structure grid, and denot them as the geological environment index value time series. Preset a standard index threshold range for several slope pressure monitoring indicators; determine whether the time series of slope pressure index values ​​for each terrain structure grid is within the corresponding standard index threshold range. Slope pressure monitoring indicators that exceed the corresponding standard indicator threshold range are designated as key slope pressure monitoring indicators. The time series of slope pressure index values ​​corresponding to the key slope pressure monitoring indicators of each terrain structure grid that exceed the corresponding standard index threshold range are marked as key time periods. Based on the Pearson correlation coefficient, the autocorrelation coefficient between the time series of geological environmental monitoring indicators corresponding to each topographic structure grid during the critical time period and the time series of geological environmental monitoring indicators corresponding to the non-critical time period is obtained. An autocorrelation coefficient threshold is preset, and geological environmental monitoring indicators whose autocorrelation coefficient between the time series of geological environmental monitoring indicators corresponding to the critical time period and the time series of geological environmental monitoring indicators corresponding to the non-critical time period is greater than or equal to the autocorrelation coefficient threshold are designated as key geological environmental monitoring indicators.

[0025] It should be further explained that the formula for calculating the autocorrelation coefficient between the time series of geological environmental monitoring indicators corresponding to each topographic structure grid during the critical time period and the time series of geological environmental indicators corresponding to the non-critical time period is as follows: ;in, Indicates the first In the first terrain structure grid The autocorrelation coefficient of geological environment monitoring indicators; Indicates the first Class III geological environment monitoring indicators during non-critical time periods The index value corresponding to the given time; Indicates the first The average value of the corresponding indicator for geological environment monitoring indicators during non-critical time periods; This represents the total number of moments within a non-critical time period. Indicates the first Geological environment monitoring indicators during key time periods The index value corresponding to the given time; Indicates the first The average value of the corresponding indicator for geological environment monitoring indicators during key time periods; This indicates the total number of moments within the key time period.

[0026] It should be further explained that, in the specific implementation process, the extraction of time characteristics of key slope pressure monitoring indicators and key geological environment monitoring indicators for each terrain structure grid, and the generation of dynamic change trend characteristics of key slope pressure monitoring indicators and key geological environment monitoring indicators, specifically includes: It should be noted that the dynamic trend characteristics of the corresponding key slope pressure monitoring indicators and key geological environment monitoring indicators include the numerical trend characteristics of the slope pressure monitoring indicators and the numerical trend characteristics of the geological environment monitoring indicators.

[0027] Optionally, in this embodiment of the application, a monitoring index prediction model is constructed; the time series sequences of slope pressure index values ​​corresponding to the key slope pressure monitoring indexes of each terrain structure grid in the current collection period and the time series sequences of geological environment index values ​​corresponding to the key geological environment monitoring indexes are input into the trained monitoring index prediction model; the trend features of index values ​​change corresponding to the key slope pressure monitoring indexes of each terrain structure grid in the current collection period are obtained according to the output layer of the monitoring index prediction model, and are recorded as the trend features of index values ​​change of slope pressure monitoring indexes and the trend features of index values ​​change corresponding to the key geological environment monitoring indexes, and are recorded as the trend features of index values ​​change of geological environment monitoring indexes.

[0028] It should be noted that the monitoring indicator prediction model is constructed based on the slope stability risk early warning dataset and a temporal convolutional neural network. Since the number of monitoring indicators for the corresponding key slope pressure monitoring indicators and key geological environment monitoring indicators is related to the actual situation, the number of trend features for the numerical change trends of the corresponding slope pressure monitoring indicators and geological environment monitoring indicators also varies accordingly.

[0029] To further clarify, the slope pressure monitoring index numerical trend characteristics represent a numerical matrix composed of the predicted values ​​of key slope pressure monitoring indicators corresponding to the topographic structure grid; the geological environment monitoring index numerical trend characteristics represent a numerical matrix composed of the predicted values ​​of key geological environment monitoring indicators corresponding to the topographic structure grid. The data collection period, data collection time, and corresponding predicted index values ​​for each topographic structure grid must correspond one-to-one.

[0030] It should be further explained that, in the specific implementation process, the process of constructing the monitoring indicator prediction model includes: Optionally, in this embodiment, a pre-constructed slope stability risk early warning dataset is used as the basis. Time-series data related to key slope pressure monitoring indicators and key geological environment monitoring indicators are selected to form the basic dataset for model training. The basic dataset is cleaned by using the IQR outlier detection method to remove outliers generated during the collection process; linear interpolation is used to supplement missing values ​​generated during the collection process; then the time-series data is standardized to obtain a standardized time-series dataset; finally, the standardized time-series dataset is divided into a training set, a validation set, and a test set in a 7:2:1 ratio. The division process uses stratified sampling to ensure that the time-series distribution characteristics of different monitoring indicators within each dataset are consistent.

[0031] The main framework of the model is built based on a temporal convolutional neural network, which includes an input layer, a temporal convolutional layer, a residual connection layer, a pooling layer, and an output layer. The input dimension of the input layer is determined by the length of a single time-series data point (i.e., the number of consecutive data points for a single monitoring indicator) and the number of monitoring indicators, and is set as [batch_size, seq_len, feature_num], where batch_size is the batch size, seq_len is the length of the time-series data, and feature_num is the total number of key monitoring indicators. The temporal convolutional layer adopts a causal convolutional structure to avoid interference from future data on the current prediction results, and the kernel size is set to [missing value]. (Adaptively adjustable based on the length of the time series data), the number of convolutional kernels is gradually increased from 64 to 256, extracting time series features at different scales through multi-layer convolution stacking; each time convolutional layer is followed by a normalization layer and a ReLU activation function to reduce the gradient vanishing problem and improve the model convergence speed; the residual connection layer is used to connect adjacent convolutional layers, superimposing the output of the previous layer with the output of the current layer before passing it to the next layer, ensuring the effective transmission of deep features; the pooling layer uses average pooling to reduce the dimensionality of the time series features extracted by convolution, retaining the core trend features; the output layer uses a fully connected layer, with the output dimension consistent with the length of the time series data in the input layer, corresponding to the output of the predicted values ​​of the monitoring indicators at each time step, forming a feature matrix of indicator value change trends.

[0032] The model was trained and its parameters optimized. Training parameters were set, with the Adam optimizer chosen and the initial learning rate set to 0.001. A learning rate decay strategy was employed (i.e., the learning rate was halved if the validation set loss did not decrease for five consecutive iterations). The mean squared error loss function was used to measure the deviation between the model's predicted and actual values. The maximum number of iterations was set to 100, and an early stopping mechanism was introduced: training stopped if the validation set loss did not decrease for ten consecutive iterations to prevent overfitting. The training set was input into the model for training. After each iteration, the model performance was validated using the validation set, and the model parameters were adjusted based on the validation set loss. After training, the model performance was evaluated using the test set, with evaluation metrics including mean absolute error, root mean square error, and coefficient of determination. If the evaluation metrics met the preset requirements, the model training was complete, and the model parameters were saved. If the requirements were not met, the model structure parameters (e.g., kernel size, number of convolutional layers) were adjusted, or the dataset preprocessing was repeated, and training continued until the requirements were met.

[0033] The trained monitoring indicator prediction model was adapted to the landscape ecological slope stability monitoring process. To address the differences in the number of monitoring indicators for different terrain structure grids, the model dynamically adjusted the feature_num parameter of the input layer. Three sets of real-time monitoring data from different seasons and terrain types were selected and input into the model for prediction verification. The trend characteristics of the indicator values ​​output by the model were compared with the trend characteristics of the actual monitoring data to verify the adaptability and prediction accuracy of the model in different scenarios. This ensured that the model could accurately output the trend characteristics of slope pressure monitoring indicators and geological environment monitoring indicators that met the actual needs.

[0034] It should be further explained that, in the specific implementation process, the extraction of spatial features from the dynamic change trends of key slope pressure monitoring indicators and key geological environment monitoring indicators of each terrain structure grid, and the generation of corresponding spatiotemporal feature sequences, includes the following: Optionally, in this embodiment of the application, a unique grid spatial coordinate identifier is assigned to each terrain structure grid to clarify the adjacency relationship and spatial distance parameters between grids, thereby constructing a grid spatial topology; The dynamic trend characteristics of key slope pressure monitoring indicators and key geological environment monitoring indicators for each terrain structure grid are obtained and bound to the corresponding grid spatial coordinate identifiers to form an associated dataset. It should be noted that the adjacency relationship includes horizontal adjacency and vertical elevation difference adjacency.

[0035] If the slope pressure monitoring index and geological environment monitoring index of the terrain structure grid in the grid spatial topology do not exceed the corresponding threshold, the real-time collected slope pressure monitoring index and geological environment monitoring index will be used as the associated dataset of the corresponding terrain structure grid in the grid spatial topology to ensure the continuity of spatial data within the monitoring range.

[0036] The graph attention network is used to learn the topology of the grid space. The dynamic trend features of key slope pressure monitoring indicators and key geological environment monitoring indicators in the associated dataset of each terrain structure grid in the grid space topology are input into the graph attention network. Take any terrain structure grid in the grid space topology as the target terrain structure grid and as the center grid, and denote the terrain structure grids adjacent to it as the neighborhood grids. By obtaining the collaborative influence coefficient of the neighboring grid on the central grid, the collaborative influence coefficient is used as the attention weight to assign weights to each neighboring grid. Based on feature aggregation technology, spatiotemporal feature sequences corresponding to the dynamic change trends of key slope pressure monitoring indicators and key geological environment monitoring indicators of each terrain structure grid are generated. The spatiotemporal feature sequences include the spatiotemporal feature sequences of slope pressure and the spatiotemporal feature sequences of geological environment.

[0037] It should be noted that the calculation formula for the spatiotemporal characteristic sequence of slope pressure is as follows: ;in, Indicates the first The spatiotemporal characteristic sequence corresponding to the dynamic change trend of key slope pressure monitoring indicators of each terrain structure grid, namely the spatiotemporal characteristic sequence of slope pressure. Indicates the activation function; No. The dynamic trend characteristics of key slope pressure monitoring indicators of a terrain structure grid, namely, the numerical trend characteristics of slope pressure monitoring indicators. Indicates the relationship with the first A set of terrain feature grids adjacent to a terrain structure grid; Indicates the synergistic impact coefficient; This represents the feature transformation parameter matrix.

[0038] The calculation formula for the spatiotemporal characteristic sequence of the geological environment is as follows: ;in, Indicates the first The spatiotemporal characteristic sequence corresponding to the dynamic change trend of key geological environment monitoring indicators of each topographic structure grid, i.e., the spatiotemporal characteristic sequence of the geological environment; Indicates the activation function; No. The dynamic trend characteristics of key geological environment monitoring indicators of each topographic structure grid, namely, the numerical trend characteristics of geological environment monitoring indicators. Indicates the relationship with the first A set of terrain feature grids adjacent to a terrain structure grid; Indicates the synergistic impact coefficient; This represents the feature transformation parameter matrix.

[0039] Further explanation of the numerical trend characteristics of the slope pressure monitoring index. ; Indicates the first The terrain structure grid in the first A matrix of slope pressure monitoring index values ​​corresponding to key slope pressure monitoring indicators at specific times. The trend characteristics of the numerical changes of the geological environment monitoring indicators. ; Indicates the first The terrain structure grid in the first A matrix of geological environment monitoring index values ​​corresponding to key geological environment monitoring indicators at any given time.

[0040] It should be further explained that, in the specific implementation process, the process of obtaining the cooperative influence coefficient of the neighboring grid on the central grid includes: Optionally, in this embodiment, key slope pressure monitoring indicators and key geological environment monitoring indicators of the central grid and each neighboring grid are obtained. Based on these indicators, the formula for calculating the synergistic influence coefficient of the neighboring grids on the central grid is as follows: ;in, Indicates the first The terrain structure grid and the first Cooperative influence coefficient among terrain structure grids; Indicates the conversion factor; , Indicates the weighting factor; Indicates the first The terrain structure grid and the first The total number of key slope pressure monitoring indicators of different types in each terrain structure grid; Indicates the first The terrain structure grid and the first The number of key slope pressure monitoring indicators of the same type in each terrain structure grid; Indicates the first The terrain structure grid and the first The total number of different types of key geological environment monitoring indicators in each topographic structure grid; Indicates the first The terrain structure grid and the first The number of key geological environment monitoring indicators of the same type in each topographic structure grid.

[0041] It should be further explained that, in the specific implementation process, the process of constructing a slope stability early warning model based on deep learning algorithms includes: Optionally, in this application embodiment, the numerical change trend characteristics of the slope pressure monitoring indicators corresponding to the key slope pressure monitoring indicators of each terrain structure grid during the historical collection period and the corresponding slope pressure spatiotemporal feature sequence, as well as the numerical change trend characteristics of the geological environment monitoring indicators corresponding to the key geological environment monitoring indicators and the corresponding geological environment spatiotemporal feature sequence are obtained. The numerical trend characteristics of key slope pressure monitoring indicators and their corresponding spatiotemporal characteristic sequences for each topographic structure grid within the historical data collection period, as well as the numerical trend characteristics of key geological environment monitoring indicators and their corresponding spatiotemporal characteristic sequences, are grouped and labeled as follows: It is a natural number; Will The sample data includes the numerical trend characteristics of key slope pressure monitoring indicators and their corresponding spatiotemporal characteristic sequences for each topographic structure grid within a historical data collection period, as well as the numerical trend characteristics of key geological environment monitoring indicators and their corresponding spatiotemporal characteristic sequences. Less than The natural numbers are used to obtain the mean of the sample data, which is denoted as the sample set; the numerical change trend characteristics of the slope pressure monitoring indicators corresponding to the key slope pressure monitoring indicators of each terrain structure grid in the remaining several groups of historical collection periods, as well as the corresponding spatiotemporal characteristic sequences of slope pressure, and the numerical change trend characteristics of the geological environment monitoring indicators corresponding to the key geological environment monitoring indicators, as well as the corresponding spatiotemporal characteristic sequences of the geological environment, are used as the test set; a training sample set is formed based on the sample set and the test set. A standard prediction model is constructed based on deep learning algorithms. The training sample set is then input into the standard prediction model to train it until the loss function is stable. The model parameters are saved, and the standard prediction model is tested using a test set until it meets the preset requirements. Finally, a slope stability early warning model is output. The slope pressure monitoring index numerical change trend characteristics and corresponding slope pressure spatiotemporal feature sequences for each topographic structure grid within the current acquisition period, as well as the geological environment monitoring index numerical change trend characteristics and corresponding geological environment spatiotemporal feature sequences for each topographic structure grid, are input into the slope protection stability early warning model to output the risk occurrence probability prediction for each topographic structure grid within the current acquisition period.

[0042] Based on the risk probability prediction corresponding to each terrain structure grid, stability early warning information for garden ecological slope protection is obtained; the stability early warning information includes stability level assessment, location of potential slippage risk areas, and risk probability prediction. To further explain, in constructing the slope protection stability early warning model, this invention selects a convolutional neural network algorithm suitable for time series analysis and uses the MES loss function. The prepared training set is input into the selected convolutional neural network algorithm to begin training. During training, the weights are continuously updated through the backpropagation algorithm, causing the loss function to gradually decrease until a stable state is reached. Overfitting is avoided based on the early stopping technique, and various parameters of the multi-source data prediction model are optimized based on grid search. These parameters include, but are not limited to, the learning rate, batch size, and regularization coefficient.

[0043] Once the standard prediction model has been trained and its parameters adjusted, a final evaluation is performed using a test set to obtain the evaluation results of the slope protection stability early warning model. These evaluation results include, but are not limited to, classification metrics such as accuracy, recall, and F1 score. Based on the evaluation results on the test set, it is determined whether the standard prediction model has met the expected standards. If the requirements are met, the model parameters are saved and deployment is prepared; if the requirements are not met, it is necessary to return to a previous stage to re-examine issues such as data quality, model structure, or training strategy.

[0044] Example 2 like Figure 2 As shown, a garden ecological slope stability monitoring system is provided, which includes, but is not limited to, a data acquisition module, a time feature extraction module, a spatial feature extraction module, and a slope stability prediction module. The data acquisition module is used to divide the preset monitoring range of the target garden ecological slope protection into several topographic structure grids based on the preset monitoring range of the target garden ecological slope protection, combined with slope topographic structure data, historical slope pressure data and historical geological environment data, and to obtain the key slope pressure monitoring indicators and key geological environment monitoring indicators of each topographic structure grid. The time feature extraction module is used to extract the time features of key slope pressure monitoring indicators and key geological environment monitoring indicators of each terrain structure grid, and generate corresponding dynamic change trend features. The spatial feature extraction module is used to extract the spatial features of the dynamic change trends of key slope pressure monitoring indicators and key geological environment monitoring indicators of each terrain structure grid, and generate corresponding spatiotemporal feature sequences. The slope stability prediction module constructs a slope stability early warning model based on deep learning algorithms. It inputs the spatiotemporal feature sequences of various terrain structure grids into the slope stability early warning model to obtain stability early warning information for garden ecological slopes.

[0045] Optionally, in this embodiment, those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing the hardware related to the terminal device. The program can be stored in a computer-readable storage medium, which may include: flash drive, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0046] 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.

[0047] If the integrated units in the above embodiments are implemented as software functional units and sold or used as independent products, they can be stored in the aforementioned 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 all or part of the 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 to cause one or more electronic devices to execute all or part of the steps of the methods described in the various embodiments of this application.

[0048] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0049] In the several embodiments provided in this application, it should be understood that the disclosed application can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection of units or modules may be electrical or other forms.

[0050] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0051] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0052] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for monitoring the stability of ecological slope protection in gardens, characterized in that, The method includes: Based on the pre-set monitoring range of the target garden ecological slope protection, combined with slope topographic structure data, historical slope pressure data and historical geological environment data, the pre-set monitoring range of the target garden ecological slope protection is divided into several topographic structure grids, and key slope pressure monitoring indicators and key geological environment monitoring indicators of each topographic structure grid are obtained. The time characteristics of key slope pressure monitoring indicators and key geological environment monitoring indicators of each terrain structure grid are extracted to generate corresponding dynamic change trend characteristics. Spatial features are extracted from the dynamic change trends of key slope pressure monitoring indicators and key geological environment monitoring indicators of each terrain structure grid to generate corresponding spatiotemporal feature sequences. A slope protection stability early warning model is constructed based on deep learning algorithms. The spatiotemporal feature sequences of various terrain structure grids are input into the slope protection stability early warning model to obtain stability early warning information of garden ecological slope protection.

2. The method for monitoring the stability of garden ecological slope protection according to claim 1, characterized in that, The process of obtaining key slope pressure monitoring indicators and key geological environment monitoring indicators for each terrain structure grid includes: Based on the acquisition time of historical slope pressure data and historical geological environment data of each terrain structure grid, the historical slope pressure data and historical geological environment data of each terrain structure grid are sorted by time according to the corresponding acquisition time, thereby obtaining the time series sequence of slope pressure data and geological environment data of each terrain structure grid. Then, the slope pressure data time series and geological environment data time series corresponding to each terrain structure grid are extracted; Pre-set standard index threshold ranges corresponding to several slope pressure monitoring indicators; determine whether the time series of slope pressure index values ​​corresponding to each terrain structure grid is within the corresponding standard index threshold range, and thus obtain key slope pressure monitoring indicators. The time series of slope pressure index values ​​corresponding to the key slope pressure monitoring indicators of each terrain structure grid that exceed the corresponding standard index threshold range are marked as key time periods. Based on the Pearson correlation coefficient and according to the key time period, key geological environment monitoring indicators corresponding to each terrain structure grid are obtained.

3. The method for monitoring the stability of garden ecological slope protection according to claim 2, characterized in that, The process of extracting the temporal characteristics of key slope pressure monitoring indicators and key geological environment monitoring indicators for each terrain structure grid includes: Construct a monitoring index prediction model; input the time series of slope pressure index values ​​corresponding to the key slope pressure monitoring indexes of each terrain structure grid in the current collection period and the time series of geological environment index values ​​corresponding to the key geological environment monitoring indexes into the trained monitoring index prediction model, thereby obtaining the numerical change trend characteristics of slope pressure monitoring indexes and geological environment monitoring indexes.

4. The method for monitoring the stability of garden ecological slope protection according to claim 3, characterized in that, The process of building a predictive model for monitoring indicators includes: Time-series data related to key slope pressure monitoring indicators and key geological environment monitoring indicators in the slope stability risk early warning dataset were selected to form the basic dataset for model training. The basic dataset was preprocessed to obtain a standardized time-series dataset. The standardized time-series dataset was then divided into training set, validation set and test set, and the division process adopted stratified sampling. The main framework of the model is built based on a temporal convolutional neural network. The model is trained and its parameters are optimized. The relevant parameters for model training are set, the training set is input into the model for training, and the model performance is verified using a validation set. The model parameters are adjusted based on the validation set loss. After training, the model performance is evaluated using a test set. If the evaluation indicators meet the preset requirements, the model training is completed and the model parameters are saved. If the requirements are not met, the model structure parameters are adjusted or the dataset is reprocessed and trained again until the requirements are met, thus obtaining the corresponding monitoring indicator prediction model.

5. The method for monitoring the stability of garden ecological slope protection according to claim 4, characterized in that, The process of extracting spatial features of the dynamic change trends of key slope pressure monitoring indicators and key geological environment monitoring indicators for each terrain structure grid includes: A unique grid spatial coordinate identifier is assigned to each terrain structure grid, the adjacency relationship and spatial distance parameters between grids are defined, and the grid spatial topology is constructed. The dynamic trend characteristics of key slope pressure monitoring indicators and key geological environment monitoring indicators for each terrain structure grid are obtained and bound to the corresponding grid spatial coordinate identifiers to form an associated dataset. If the slope pressure monitoring index and geological environment monitoring index of the terrain structure grid in the grid space topology do not exceed the corresponding threshold, they are used as the associated dataset of the corresponding terrain structure grid in the grid space topology. The graph attention network is used to learn the topology of the grid space, and the dynamic trend characteristics of the corresponding key slope pressure monitoring indicators and key geological environment monitoring indicators are input into the graph attention network. Take any terrain structure grid in the grid space topology as the target terrain structure grid and as the center grid, and denote the terrain structure grids adjacent to it as the neighborhood grids. By obtaining the collaborative influence coefficient of the neighboring grid on the central grid, the collaborative influence coefficient is used as the attention weight to assign weights to each neighboring grid. Based on feature aggregation technology, spatiotemporal feature sequences corresponding to the dynamic change trends of key slope pressure monitoring indicators and key geological environment monitoring indicators of each terrain structure grid are generated.

6. The method for monitoring the stability of garden ecological slope protection according to claim 5, characterized in that, The process of obtaining the cooperative influence coefficient of the neighboring grid on the central grid includes: Key slope pressure monitoring indicators and key geological environment monitoring indicators of the central grid and each neighboring grid are obtained. Based on the key slope pressure monitoring indicators and key geological environment monitoring indicators of the central grid and each neighboring grid, the calculation formula of the synergistic influence coefficient of the neighboring grid on the central grid is obtained, and then the corresponding synergistic influence coefficient is obtained.

7. The method for monitoring the stability of garden ecological slope protection according to claim 6, characterized in that, The process of constructing a slope protection stability early warning model based on deep learning algorithms includes: The study acquires the numerical trend characteristics and spatiotemporal characteristic sequences of slope pressure monitoring indicators for each terrain structure grid within the historical data acquisition period, as well as the numerical trend characteristics and spatiotemporal characteristic sequences of geological environment monitoring indicators for key geological environment monitoring indicators; and groups and labels these indicators as follows: It is a natural number; Will The data from the group is used as sample data, and Less than The natural numbers are used to obtain the mean of the sample data, which is denoted as the sample set; the data from the remaining historical collection periods are used as the test set; and a training sample set is formed based on the sample set and the test set. A standard prediction model is constructed based on deep learning algorithms. The training sample set is then input into the standard prediction model to train it until the loss function is stable. The model parameters are saved, and the standard prediction model is tested using a test set to output a slope stability early warning model.

8. A landscape ecological slope protection stability monitoring system, implementing the landscape ecological slope protection stability monitoring method according to any one of claims 1 to 7, characterized in that, include: The system includes a data acquisition module, a time feature extraction module, a spatial feature extraction module, and a slope stability prediction module. The data acquisition module is used to divide the preset monitoring range of the target garden ecological slope protection into several topographic structure grids based on the preset monitoring range of the target garden ecological slope protection, combined with slope topographic structure data, historical slope pressure data and historical geological environment data, and to obtain the key slope pressure monitoring indicators and key geological environment monitoring indicators of each topographic structure grid. The time feature extraction module is used to extract the time features of key slope pressure monitoring indicators and key geological environment monitoring indicators of each terrain structure grid, and generate corresponding dynamic change trend features. The spatial feature extraction module is used to extract the spatial features of the dynamic change trends of key slope pressure monitoring indicators and key geological environment monitoring indicators of each terrain structure grid, and generate corresponding spatiotemporal feature sequences. The slope stability prediction module constructs a slope stability early warning model based on deep learning algorithms. It inputs the spatiotemporal feature sequences of various terrain structure grids into the slope stability early warning model to obtain stability early warning information for garden ecological slopes.

Images