Land development information data intelligent management system based on multi-source data

Abstract

The application relates to the technical field of land development management, in particular to a land development information data intelligent management system based on multi-source data. The system comprises a land geological feature extraction module, a risk prediction module, a risk early warning model construction module and a management execution module; geological feature extraction between different regions is carried out based on multi-source data of a land development region, geological change structure instability risk prediction is carried out, underground water permeation change risk simulation is carried out, and underground water permeation change risk data is obtained; multi-target scene fitting is carried out according to the underground water permeation change risk data, a land development risk early warning model is constructed, and a land development risk early warning model is obtained; the land development risk early warning model is sent to a control center to execute development information data intelligent management. The optimization processing of the land development management technology makes the land development management technology more perfect.

Description

Technical Field

[0001] This invention relates to the field of land development management technology, and in particular to an intelligent management system for land development information data based on multi-source data. Background Technology

[0002] Multi-source data fusion can comprehensively reflect all dimensions of the land development process, improving the accuracy and timeliness of land development decisions. Through data integration and analysis, intelligent management systems can obtain real-time land use information, assess development potential, and analyze the socio-economic impacts of land use, thereby providing effective decision support for enterprises and other stakeholders. This system not only improves the transparency of land development but also promotes the rational allocation and sustainable use of land resources. Intelligent data processing technologies, such as artificial intelligence, big data analytics, and machine learning, can further optimize data mining and application. However, traditional intelligent management systems for land development information based on multi-source data suffer from inaccurate identification of geological risks during the development process, thus failing to provide accurate early warning management of land development risks. Summary of the Invention

[0003] Therefore, it is necessary to provide a land development information data intelligent management system based on multi-source data to solve at least one of the above-mentioned technical problems.

[0004] To achieve the above objectives, a land development information data intelligent management system based on multi-source data includes the following modules: The land geological feature extraction module is used to acquire multi-source data of land development areas; based on the multi-source data of land development areas, geological features are extracted between different areas to obtain land geological feature data between different areas; The risk prediction module is used to acquire land development planning data; based on the land development planning data, it performs geological change and structural instability risk prediction on land geological characteristic data between different regions, and obtains geological change and structural instability risk data; based on the geological change and structural instability risk data, it performs groundwater seepage change risk simulation, and obtains groundwater seepage change risk data. The risk warning model construction module is used to fit multi-objective scenarios based on groundwater seepage change risk data to obtain geological development multi-objective scenario risk fitting data; and to construct a land development risk warning model based on the geological development multi-objective scenario risk fitting data to obtain the land development risk warning model. The management execution module is used to send the land development risk early warning model to the control center to perform intelligent management of development information data.

[0005] Preferably, the land geological feature extraction module includes: Obtain multi-source data on land development areas; Multi-source data cleaning is performed on multi-source data of land development areas to obtain cleaned land development area data. Regional land topographic structure analysis was performed on the cleaned data of land development areas to obtain regional land topographic structure data. Based on multi-source data of land development areas, geological features of different regions are extracted from regional land topographic structure data to obtain land geological feature data of different regions.

[0006] Preferably, the risk prediction module includes: Obtain land development planning data; Based on land development planning data, theoretical topographic change demand analysis was conducted on land geological characteristic data of different regions to obtain theoretical topographic change data for land development. Based on the land development topographic theory change data, the geological geological feature data of different regions are used to predict the risk of geological change and structural instability, and the geological change and structural instability risk data are obtained. Based on the geological change and structural instability risk data, a groundwater seepage change risk simulation was conducted to obtain groundwater seepage change risk data.

[0007] Preferably, the prediction of geological change and structural instability risk based on land geological characteristic data of different regions using land development topographic theory change data includes: Based on the land development topography theory change data, the stratigraphic deformation structure analysis was carried out on the land geological feature data of different regions to obtain the stratigraphic deformation structure data of different regions. Sequence loose viscosity structure analysis of soil layers was performed on the stratigraphic deformation structure data of different regions to obtain the sequence loose viscosity structure data of soil layers. Based on the sequence loose viscosity structure data of soil layers, the internal friction angle loss under external force intervention is estimated to obtain the data of internal friction angle loss under external force intervention. The ultimate bearing capacity of the foundation soil layer is simulated and calculated based on the loss data of the internal friction angle caused by external force intervention, and the ultimate bearing capacity data of the foundation soil layer is obtained. Based on the ultimate bearing capacity data of foundation soil, the internal friction angle loss data of external force intervention, and the sequence loose viscosity structure data of soil layers, the risk of structural instability due to geological changes is predicted, and the risk data of structural instability due to geological changes is obtained.

[0008] Preferably, the estimation of the internal friction angle loss based on the sequence loose viscosity structure data of soil layers includes: Obtain initial data of excavation force; based on the soil layer sequence loose viscosity structure data, match the excavation force required for different soil layer structures to obtain the excavation force matching data required for the soil layer structure; Analysis of the average difference in cohesion among soil types was performed on the sequence loose viscosity structure data of soil layers to obtain the average difference in cohesion among soil types. Based on the external force matching data required for soil structure excavation, the cohesion difference of soil type is used to simulate the lateral pressure fluctuation of cohesion dilatation, and the cohesion dilatation lateral pressure fluctuation data is obtained. Nonlinear shear friction force analysis was performed on the lateral pressure fluctuation data of cohesive shear dilatation to obtain nonlinear shear friction force data; Based on the lateral pressure fluctuation data of cohesive shear dilatation and the nonlinear shear friction force data, the internal friction angle loss under external force intervention is estimated, and the internal friction angle loss data under external force intervention is obtained.

[0009] Preferably, the step of simulating groundwater seepage change risk based on geological change and structural instability risk data includes: Obtain groundwater distribution data in the development area; Geological layer variability analysis was performed on geological layer structural instability risk data to obtain geological layer variability structural data; The relative weak numerical intervals of geological layer variation structure data are evaluated to obtain the relative weak numerical intervals of geological layers. Based on the groundwater distribution data of the development area, the shear moment imbalance of the water flow is calculated in the numerical range of the relatively weak geological layer to obtain the shear moment imbalance data of the water flow. Based on the imbalance data of water flow shear pressure moment, the risk of groundwater seepage change was simulated, and the risk data of groundwater seepage change was obtained.

[0010] Preferably, the risk warning model construction module includes: Convolutional calculations are performed on groundwater seepage change risk data to obtain groundwater seepage change risk convolutional data. Multi-objective scenario fitting was performed based on groundwater seepage change risk convolution data and geological change structural instability risk data to obtain geological development multi-objective scenario risk fitting data. A land development risk early warning model was constructed based on risk fitting data from multiple objective scenarios of geological development.

[0011] Preferably, the multi-objective scenario fitting based on groundwater seepage change risk convolution data and geological change structural instability risk data includes the following steps: Based on the convolutional data of groundwater seepage change risk and the data of geological change and structural instability risk, the linkage factors of multi-objective risk factors are identified, and the linkage factors of multi-objective risk factors are obtained. The out-of-control risk factors are quantified by analyzing the linkage factors of multi-objective risk factors, and quantitative data of the linkage factors of out-of-control risks are obtained. Based on the quantitative data of the linkage factors of runaway risk, the antecedent risk mutation factors are identified, and the antecedent risk mutation factors are obtained. Based on the quantitative data of the antecedent risk mutation factors and the linkage factors of the runaway risk, multi-objective scenario fitting is performed to obtain multi-objective scenario risk fitting data for geological development.

[0012] Preferably, the construction of the land development risk early warning model based on multi-objective scenario risk fitting data for geological development includes: Normalize the risk fitting data of multi-objective scenarios in geological development to obtain normalized risk data of multi-objective scenarios. The risk normalization data for multi-objective scenarios is divided into training and testing sets to obtain the multi-objective scenario risk training set and the multi-objective scenario risk testing set. Random feature sampling is performed on the multi-target scenario risk training set to obtain a scenario risk random sampling training set; The initial land development risk early warning model is constructed by randomly sampling the training set of scenario risks based on the policy gradient algorithm. The initial land development risk early warning model was tested and optimized using a multi-objective scenario risk test set, resulting in a new land development risk early warning model.

[0013] The beneficial effects of this invention lie in acquiring multi-source data (such as geological surveys, topographic maps, climate, soil types, etc.) of land development areas, and processing and analyzing the data of different areas to extract the geological characteristics of each area. This process helps identify the geological conditions of different areas, including important characteristics such as stratigraphic composition, soil and rock types, and groundwater distribution, providing accurate basic data support for subsequent risk assessment and planning design. Through the comprehensive processing of multi-source data, a more comprehensive understanding of the natural conditions of the development area can be achieved, improving the scientific nature and accuracy of land development. By acquiring land development planning data (such as development purpose, construction density, development methods, etc.), combined with the aforementioned geological characteristic data, the risk of geological change and structural instability in different areas can be predicted. By simulating changes in geological structure, the occurrence of structural instability phenomena is assessed, providing risk warnings for development projects. Subsequently, based on the geological change and structural instability risk data, further risk simulation of groundwater infiltration changes is conducted, thereby assessing the potential impact of groundwater flow and soil infiltration characteristics on the development area. This module effectively realizes risk assessment in the land development process, helping decision-makers identify and avoid potential risks during the planning stage. Based on groundwater seepage risk data, multi-objective scenario fitting technology is used to construct geological risk fitting data under different development scenarios. This fitting data reflects risk changes occurring under various development models and geological conditions. Next, a land development risk early warning model is constructed using this data to provide early warning and prediction of potential geological risks, groundwater changes, and their impact on land development. This module provides a comprehensive and accurate risk prediction model for land development projects, enabling developers to take corresponding preventative measures in advance, effectively reducing or avoiding the impact of sudden geological risks on the project. The land development risk early warning model is transmitted to the control center for intelligent management and real-time monitoring of development information data. Through the intelligent management platform, decision-makers can view risk early warning information for land development areas at any time, understand the dynamic risks such as geological changes and groundwater flow in different areas, and adjust development plans and countermeasures in real time. The management execution module ensures the safety and controllability of the land development process by providing data analysis results and risk early warning information, providing scientific and efficient management support for development projects, and promoting the sustainable development of land development and the rational use of resources. Therefore, this invention is an optimization of a traditional intelligent management system for land development information based on multi-source data. It solves the problem that the traditional intelligent management system for land development information based on multi-source data does not accurately identify geological risks in the development process, thus failing to conduct accurate land development risk early warning management. It improves the accuracy of geological risk identification in the development process, enhances the early warning capability for land development risks, and strengthens management capabilities. Attached Figure Description

[0014] Figure 1This is a schematic diagram of the module flow of an intelligent management system for land development information based on multi-source data; Figure 2 for Figure 1 A functional flowchart of the medium-risk prediction module; Figure 3 for Figure 1 A functional flowchart of the medium-risk early warning model construction module. Detailed Implementation

[0015] Please see Figures 1 to 3 A land development information data intelligent management system based on multi-source data includes the following modules: The land geological feature extraction module is used to acquire multi-source data of land development areas; based on the multi-source data of land development areas, geological features are extracted between different areas to obtain land geological feature data between different areas; The risk prediction module is used to acquire land development planning data; based on the land development planning data, it performs geological change and structural instability risk prediction on land geological characteristic data between different regions, and obtains geological change and structural instability risk data; based on the geological change and structural instability risk data, it performs groundwater seepage change risk simulation, and obtains groundwater seepage change risk data. The risk warning model construction module is used to fit multi-objective scenarios based on groundwater seepage change risk data to obtain geological development multi-objective scenario risk fitting data; and to construct a land development risk warning model based on the geological development multi-objective scenario risk fitting data to obtain the land development risk warning model. The management execution module is used to send the land development risk early warning model to the control center to perform intelligent management of development information data.

[0016] In this embodiment of the invention, reference Figure 1 The above is a schematic diagram of the module flow of a land development information data intelligent management system based on multi-source data according to the present invention. In this example, the land development information data intelligent management system based on multi-source data includes: S1: Land geological feature extraction module, used to acquire multi-source data of land development areas; based on the multi-source data of land development areas, geological features are extracted between different areas to obtain land geological feature data between different areas; In this embodiment of the invention, multi-source data of the land development area is acquired. The data sources include satellite remote sensing images, geological survey data, groundwater monitoring data, meteorological data, and regional historical geological disaster records. All of these data are processed in a standardized format to ensure that various types of data can be uniformly formatted for subsequent analysis. Data cleaning tools are used to denoise the raw data, removing unnecessary noise signals and outliers. Subsequently, the data is divided into regions. Based on Geographic Information System (GIS) technology, a spatial positioning-based zoning method is adopted to divide the land development area into several sub-regions according to topography, soil type, and hydrological conditions. The geological feature data of each sub-region is processed separately. The cleaned data is analyzed at multiple scales, from microscopic geological layers to macroscopic surface structure. Combined with multimodal data processing technology, geological feature data of different regions, such as stratigraphic structure, soil type, and surface morphology, are extracted. Finally, land geological feature data of each region is obtained. The data includes, but is not limited to, parameters such as soil viscosity, soil layer distribution, rock hardness, and groundwater flow characteristics. These data provide an important basis for subsequent geological risk assessment.

[0017] S2: Risk prediction module, used to acquire land development planning data; based on land development planning data, perform geological change and structural instability risk prediction on land geological characteristic data between different regions, and obtain geological change and structural instability risk data; perform groundwater seepage change risk simulation based on geological change and structural instability risk data, and obtain groundwater seepage change risk data. In this embodiment of the invention, when acquiring land development planning data, the land use planning layer of the development area is first imported into the data processing system. This layer typically includes information such as land development direction, building density, soil improvement measures, road and infrastructure layout. This planning data is cross-analyzed with specific development areas to deduce the development change requirements for each area, especially the changes in soil and surface. By analyzing the impact of land development on geological structure, and based on existing geological feature data, theoretical simulations of geological changes are performed for each area. Numerical simulation techniques, such as finite element analysis (FEM), are used to simulate the topographic changes in the developed area, obtaining data on the risk of geological structural instability during land development. Using this data, combined with groundwater monitoring data, groundwater infiltration changes are simulated through a hydrodynamic model. Considering the changes in hydrological conditions after land development, the infiltration changes of groundwater under different geological conditions are analyzed, and groundwater infiltration change risk data are generated to further determine the infiltration problems caused by groundwater, such as groundwater level rise or water flow shear effect.

[0018] S3: Risk warning model construction module, used to fit multi-objective scenarios based on groundwater seepage change risk data to obtain geological development multi-objective scenario risk fitting data; and to construct a land development risk warning model based on the geological development multi-objective scenario risk fitting data to obtain the land development risk warning model. In this embodiment of the invention, based on groundwater seepage change risk data, data normalization is first performed, and convolution operations are used to smooth the data to reduce noise interference. Then, multi-objective scenario fitting is performed, integrating risk data from different regions into the same model framework. Multi-objective optimization methods in machine learning algorithms, such as Support Vector Machine (SVM) or neural networks (ANN) in deep learning, are used to identify risk factors in different scenarios. Through these techniques, the various risk data are weighted and summed to form multi-objective risk fitting data, which reflects the impact of multiple variables on land development risks. Based on this multi-objective risk fitting data, a land development risk early warning model is constructed using the policy gradient algorithm or particle swarm optimization algorithm (PSO). The model identifies potential risks in the development process by weightedly evaluating risk factors under different scenarios, ultimately generating a land development risk early warning model to provide decision support for development management. During the model construction process, multiple iterations of optimization are performed on each objective data to ensure the model's accuracy and generalization ability, and to ensure its effectiveness under different development scenarios.

[0019] S4: Management Execution Module, used to send the land development risk early warning model to the control center to perform intelligent management of development information data.

[0020] In this embodiment of the invention, the constructed land development risk early warning model is sent to the control center. The control center monitors and schedules land development in real time through an intelligent data management system. First, the control center obtains real-time data of the land development site, such as soil moisture, groundwater level, and meteorological conditions, by accessing a remote monitoring system. This data is then matched with the early warning model. Based on the judgment of the early warning model, the risks generated during the development process are assessed in real time. If the early warning model identifies potential geological or hydrological risks, the control center immediately initiates an emergency response procedure and mobilizes on-site resources for processing. The control center automates the management of various information within the development area through an integrated intelligent management platform, such as automatically adjusting development plans and optimizing soil and hydrological condition improvement measures in real time. This intelligent management process can improve the efficiency of land development and reduce risks caused by uncontrollable factors such as geological changes. Finally, through big data analysis technology, the monitoring system regularly updates the early warning model, ensuring that the land development management process is always kept in an optimal state of risk control.

[0021] The land geological feature extraction module includes: Obtain multi-source data on land development areas; Multi-source data cleaning is performed on multi-source data of land development areas to obtain cleaned land development area data. Regional land topographic structure analysis was performed on the cleaned data of land development areas to obtain regional land topographic structure data. Based on multi-source data of land development areas, geological features of different regions are extracted from regional land topographic structure data to obtain land geological feature data of different regions.

[0022] In this embodiment of the invention, when acquiring multi-source data for a land development area, geological survey data, satellite remote sensing images, meteorological data, surface hydrological data, and groundwater monitoring data are first obtained from different data sources. These data come from different remote sensing satellite systems, geological survey institutions, meteorological stations, and environmental monitoring stations. All data are stored in a standardized format during acquisition to ensure data consistency and comparability. Satellite remote sensing images are acquired by high-resolution remote sensing satellites, and the captured images are matched with geographic coordinates in a Geographic Information System (GIS). Geological survey data, including soil samples, rock strata data, and groundwater level monitoring data, are collected by a professional geological exploration team within the development area. Groundwater monitoring data is obtained by deploying groundwater monitoring wells and using sensors to continuously record changes in groundwater levels. Meteorological data, including precipitation, temperature, humidity, and wind speed, is obtained from nearby meteorological stations. After data cleaning and preprocessing, these data are stored in a unified database to ensure the timeliness and accuracy of the data. Furthermore, all data needs to be integrated with the GIS platform to facilitate subsequent spatial analysis and data mining. When performing multi-source data cleaning, the first step is to use data cleaning tools (such as the Pandas library in Python) to convert various types of multi-source data (such as remote sensing images, geological survey data, meteorological data, etc.) into a unified format to ensure that the data is synchronized in terms of time axis and spatial coordinates. Data with null or inconsistent values ​​are filled or deleted. Specifically, interpolation algorithms (such as Lagrange interpolation) are used to complete missing data, and extreme outliers or significantly biased values ​​are removed. Anomaly detection algorithms based on machine learning (such as the Isolation Forest algorithm) are used to automatically detect and eliminate noise signals. Data from different sources are standardized, and values ​​from different data sources are converted to a unified unit to ensure data compatibility. The cleaned data generates land development area cleaned data, which includes all processed geological data, remote sensing data, meteorological data, groundwater monitoring data, etc. In subsequent steps, this cleaned data will be used as the basis for more in-depth analysis.When conducting regional land topographic structure analysis on cleaned data of land development areas, the first step is to input the cleaned multi-source data into the spatial analysis module using a GIS platform. Digital Elevation Model (DEM) data is then used for topographic analysis, extracting features such as topographic relief, slope, and elevation changes. Slope analysis reveals the slope distribution within the region, highlighting its significant impact on geological stability; areas with steeper slopes pose a risk of geological disasters. Next, a hydrological analysis model, combined with topographic and hydrological data analysis, yields information on water flow distribution and groundwater flow paths within the region. This allows for the assessment of drainage characteristics and groundwater level changes in the land development area. Regional water flow analysis is conducted using hydrological models (such as the SWAT model) to analyze groundwater permeability and the resulting land subsidence. Combined with surface image data, image classification techniques are used to analyze soil type and surface vegetation distribution. Finally, the above analysis results are integrated to generate regional land topographic structure data. This data encompasses the region's topographic relief, slope, soil type, and hydrological characteristics, providing a data foundation for subsequent geological feature extraction and risk assessment. Based on multi-source data and regional land topography data of land development areas, when extracting geological features between different areas, the development area is first divided into several sub-regions with similar geological features using clustering analysis methods (such as K-means clustering algorithm). This process is based on the distribution characteristics of land topography data. The development area is divided into multiple sub-regions according to factors such as topography, soil type, and groundwater characteristics. On this basis, the data of each sub-region is analyzed in depth. The geological stability of each region is assessed through geomechanical models. The soil and rock layers of different regions are simulated using the finite element analysis (FEM) method to extract the geological feature data of each region, such as soil compressibility, rock compressive strength, and groundwater permeability. Through multiple regression analysis, the quantitative characteristics of the impact of regional geological differences on development are further extracted. Based on these data, land geological feature data between different regions are obtained, including important parameters such as geological stability, groundwater level change trend, and soil type of each sub-region. Finally, regional land geological feature data is generated to support subsequent risk prediction and management.

[0023] The risk prediction module includes: Obtain land development planning data; Based on land development planning data, theoretical topographic change demand analysis was conducted on land geological characteristic data of different regions to obtain theoretical topographic change data for land development. Based on the land development topographic theory change data, the geological geological feature data of different regions are used to predict the risk of geological change and structural instability, and the geological change and structural instability risk data are obtained. Based on the geological change and structural instability risk data, a groundwater seepage change risk simulation was conducted to obtain groundwater seepage change risk data.

[0024] As an example of the present invention, reference is made to Figure 2 As shown, Figure 1 A functional flowchart of the medium-risk prediction module is shown in this example. The functions of the risk prediction module include: S201: Obtain land development planning data; In this embodiment of the invention, the latest land development planning documents are obtained from publicly available websites. These planning documents include land use patterns, land development sequence, project distribution, land development density, and infrastructure construction plans required for development. This planning data is provided in electronic format and contains detailed information and spatial distribution of land parcels. To ensure the timeliness and accuracy of the data, it needs to be updated regularly. Spatial positioning of the planning data is performed using a Geographic Information System (GIS), and spatial visualization of the planning data is achieved using spatial analysis functions. Each land development project is marked with its geographical location and overlaid with land development area data in the GIS. Version control is implemented for planning data from different periods to ensure that the latest land development planning data is accurately applied. The obtained land development planning data is stored in a unified database, where the data is tagged for efficient association between development data and geological data, providing data support for subsequent risk prediction.

[0025] S202: Based on land development planning data, conduct theoretical topographic change demand analysis on land geological characteristic data of different regions to obtain theoretical change data of land development topography; In this embodiment of the invention, when conducting theoretical topographic change demand analysis on land geological feature data of different regions based on land development planning data, the topographic change requirements related to land development are first extracted from the land development planning data. These include development activities such as earthwork excavation, backfilling, and road construction. By analyzing the impact of these activities on the topographic structure, a topographic change analysis model is used to simulate each development area. In specific operation, combined with the geological feature data of the region, numerical simulation methods (such as finite element analysis) are used to model different development activities. The topographic change process is decomposed into different units for calculation, and the topographic changes of each unit are estimated. The analysis of how these topographic changes affect the geological structure, such as soil compaction, rock strata subsidence, and changes in groundwater flow paths, is performed to obtain theoretical topographic change data for each region. In this process, the numerical simulation software used (such as ABAQUS or PLAXIS) can accurately simulate the specific impact of different development methods on the land and dynamically predict the changed topography. Finally, theoretical topographic change data of land development is output. This data includes the estimated changes in regional topography, soil, and groundwater caused by different development activities, which can provide a theoretical basis for subsequent risk prediction.

[0026] S203: Based on the land development topographic theory change data, predict the geological change structural instability risk of land geological feature data between different regions, and obtain geological change structural instability risk data; In this embodiment of the invention, when predicting the risk of geological change structure instability, the key factors leading to geological changes, such as soil structure, groundwater level, and topographic slope, are first extracted by analyzing the theoretical data of land development and topographic changes. Modeling of land geological characteristics in different areas is then performed, and the geological change structure is simulated using the finite element analysis (FEA) method. Based on geological survey data, soil physicochemical properties, and groundwater flow patterns, a complex geological change model is constructed. This model considers factors such as soil subsidence, groundwater infiltration, and earthquakes that occur during development. Specifically, the spatial analysis tools of ArcGIS software are first used to extract the geological characteristics of the land development area, such as stratigraphic structure and soil compressibility. By establishing soil mechanics and hydrogeological models, the geological characteristics of different development areas are calculated and simulated. Then, a three-dimensional numerical simulation is performed using specialized geomechanical calculation software (such as PLAXIS) to analyze problems such as soil instability and groundwater level changes that occur during development, obtaining geological change structure instability risk data. This data can predict whether different areas will experience geological disaster risks such as landslides and subsidence.

[0027] In another embodiment, a CSV file of land development topographic theory change data is loaded using Python tools. The file content is parsed to extract the coordinates and topographic change demand values ​​for each cell. A CSV file of land geological characteristic data is also loaded to extract geological type, soil layer thickness, and lithological parameters. For example, if the geological type of a certain area is sandy clay, the soil layer thickness is 5 meters, the cohesion is 20 kPa, and the internal friction angle is 30°, the data is merged using Python's Pandas library, with cell coordinates as the merging condition, generating a merged dataset containing topographic change demand values ​​and geological characteristic values. The NumPy library is then used to perform stratigraphic analysis on the merged dataset. Deformation structure analysis is performed by calculating the stress change in the strata based on the topographic change demand value. The stress change calculation uses an elasticity model with parameters including Young's modulus of 50 MPa and Poisson's ratio of 0.3. The results are stratum deformation structure data; for example, the stress change in a cell is 15 kPa, and the deformation is 0.02 mm. Based on this stratum deformation structure data, sequence loose viscosity structure analysis is performed on the soil layers. The analysis method involves calculating loose viscosity based on soil layer thickness and lithological parameters. The results are also presented as sequence loose viscosity structure data; for example, the loose viscosity in a cell is 0.35. Based on this sequence loose viscosity structure data, the internal friction angle loss due to external force intervention is estimated. The estimation method involves obtaining initial excavation force data (e.g., 10 kN), calculating the required matching excavation force values ​​for different soil structures (e.g., 8 kN for sandy clay), calculating the average cohesion difference for each soil type (e.g., 5 kPa), simulating cohesive shear dilatation lateral pressure fluctuations based on the matching values ​​and average differences (e.g., lateral pressure fluctuation data, 3 kPa), performing nonlinear shear friction stress analysis based on the fluctuation data (e.g., nonlinear shear friction stress data, 12 kPa), and calculating the internal friction angle loss due to external force intervention (e.g., 5°) based on the fluctuation and stress data. The bearing capacity ultimate limit of the foundation soil layer is simulated using angular loss data based on Terzaghi bearing capacity theory. Parameters include an internal friction angular loss of 5° and cohesion of 20 kPa. The calculation result is the bearing capacity ultimate limit data of the foundation soil layer, for example, a bearing capacity ultimate limit of 150 kPa. The risk of structural instability due to geological changes is predicted by combining the bearing capacity ultimate limit data, internal friction angular loss data, and loose viscosity structure data. The prediction method is based on a weighted scoring model, with weighting coefficients of 0.4 for bearing capacity ultimate limit, 0.3 for internal friction angular loss, and 0.3 for loose viscosity. The calculation result is the risk data of structural instability due to geological changes, for example, a risk value of 0.65 for a certain cell. S204: Based on the geological change and structural instability risk data, groundwater seepage change risk simulation is performed to obtain groundwater seepage change risk data.

[0028] In this embodiment of the invention, a Python tool is used to load a CSV file containing geological change and structural instability risk data. The file content is parsed, and the coordinates and risk values ​​of each cell are extracted. Groundwater distribution data for the development area is also loaded. The data format is NetCDF and includes groundwater level depth and permeability coefficient. For example, the groundwater level depth in a certain area is 2 meters, and the permeability coefficient is 1×10^-5. The data is analyzed using Python's Xarray library to parse NetCDF files at speeds of m / s, extracting groundwater distribution data. Geological layer variability analysis is then performed on the geological change and structural instability risk data. The analysis method involves calculating the coefficient of variation based on risk values. The results represent the geological layer variability structure data; for example, a cell might have a coefficient of variation of 0.25. Based on this variability structure data, a relative weak value interval assessment is conducted. The assessment method involves dividing intervals based on the coefficient of variation; for example, intervals with a coefficient of variation greater than 0.2 are considered relatively weak intervals. The assessment results represent the relative weak value intervals of the geological layer; for example, a certain area might have a weak interval of 0.2 to 0.3. Finally, based on the groundwater distribution data of the development area, a shear moment imbalance calculation is performed on the weak intervals. The calculation method involves calculating the shear moment based on the permeability coefficient and groundwater level depth; for example, a permeability coefficient of 1 × 10^-5... The groundwater level is 2 meters deep, and the calculated results are the shear pressure moment imbalance data of the water flow. For example, the imbalance value is 8 kPa. Based on the imbalance data, a groundwater seepage change risk simulation is performed using the finite difference method. The grid resolution is 5 meters × 5 meters, and the time step is 1 hour. The simulation results are groundwater seepage change risk data, for example, the risk value of a certain cell is 0.55. The prediction of geological change and structural instability risk based on land geological characteristic data of different regions using land development topographic theory change data includes: Based on the land development topography theory change data, the stratigraphic deformation structure analysis was carried out on the land geological feature data of different regions to obtain the stratigraphic deformation structure data of different regions. Sequence loose viscosity structure analysis of soil layers was performed on the stratigraphic deformation structure data of different regions to obtain the sequence loose viscosity structure data of soil layers. Based on the sequence loose viscosity structure data of soil layers, the internal friction angle loss under external force intervention is estimated to obtain the data of internal friction angle loss under external force intervention. The ultimate bearing capacity of the foundation soil layer is simulated and calculated based on the loss data of the internal friction angle caused by external force intervention, and the ultimate bearing capacity data of the foundation soil layer is obtained. Based on the ultimate bearing capacity data of foundation soil, the internal friction angle loss data of external force intervention, and the sequence loose viscosity structure data of soil layers, the risk of structural instability due to geological changes is predicted, and the risk data of structural instability due to geological changes is obtained.

[0029] In this embodiment of the invention, the coordinates and topographic change requirement values ​​of each cell are extracted, with a data resolution of 10 meters × 10 meters. A CSV file of land geological feature data is loaded, and geological type, soil layer thickness, and lithological parameters are extracted. The data resolution is 10 meters × 10 meters. Geological types include sandy clay and silty clay, soil layer thickness ranges from 2 meters to 8 meters, and lithological parameters include cohesion ranges from 15 kPa to 25 kPa and internal friction angle ranges from 25° to 35°. Data is merged using the Pandas library in Python, with cell coordinates as the merging condition, generating a merged dataset. The dataset contains topographic change requirement values ​​and geological feature values, with topographic change requirement values ​​ranging from 0.2 to 0.8. The NumPy library is used to process the merged dataset. A stratigraphic deformation structure analysis was conducted. The analysis method involved calculating stratigraphic stress changes based on topographic change demand values. An elasticity model was used for stress change calculation, with model parameters including a Young's modulus of 50 MPa and a Poisson's ratio of 0.3. The stress change in each cell was calculated, and the results are stratigraphic deformation structure data. The data includes the coordinates, stress change value, and deformation value of each cell. The stress change value ranges from 10 kPa to 20 kPa, and the deformation value ranges from 0.01 mm to 0.03 mm. The coordinates, stress change value, and deformation value of each cell were extracted. A CSV file of land geological feature data was loaded, and geological type, soil layer thickness, and lithological parameters were extracted. Geological types include sandy clay and silty clay, and soil layer thickness ranges from 2 meters to 8 meters. The parameters included cohesion ranging from 15 kPa to 25 kPa and internal friction angle ranging from 25° to 35°. Data was merged using the Pandas library in Python, with cell coordinates as the merging condition, generating a merged dataset containing stress variation values, deformation values, and geological characteristic values. The NumPy library was used to perform sequence loose viscosity structure analysis on the merged dataset. The analysis method involved calculating loose viscosity based on soil layer thickness and lithological parameters. The calculation process involved calculating the viscosity coefficient based on cohesion and internal friction angle, with the viscosity coefficient ranging from 0.2 to 0.5, and then calculating the looseness based on soil layer thickness, with the looseness ranging from 0.1 to 0.4. The results are the sequence loose viscosity structure data of the soil layers, containing the coordinates and loose viscosity of each cell. The loose viscosity values ​​range from 0.25 to 0.45. Sequence loose viscosity structure data of the soil layers is output. Based on this data, the coordinates and loose viscosity values ​​of each cell are extracted. The loose viscosity values ​​range from 0.25 to 0.45. Initial excavation force data is loaded in CSV format, containing excavation force values ​​ranging from 5 kN to 15 kN. Data is merged using the Pandas library in Python, with cell coordinates as the merging condition, generating a merged dataset containing both loose viscosity and excavation force values. The NumPy library is used to match the required excavation force for the soil structure in the merged dataset. The matching method is based on the loose viscosity values, with matching values ​​ranging from 6 kN to 12 kN.The mean deviation of cohesion for different soil types was calculated. The mean deviation was calculated based on the standard deviation within the cohesion range of 15 kPa to 25 kPa, and the mean deviation ranged from 3 kPa to 5 kPa. Lateral pressure fluctuations due to cohesion dilatation were simulated based on the matching values ​​and the mean deviation. The simulation method was finite element analysis with a mesh resolution of 5 m × 5 m. The simulation results were lateral pressure fluctuation data, with fluctuation values ​​ranging from 2 kPa to 4 kPa. Nonlinear shear friction stress analysis was performed on the lateral pressure fluctuation data. The analysis method was based on shear stress to calculate the stress values, with stress values ​​ranging from 10 kPa to 15 kPa. The internal friction angle loss due to external force intervention was calculated based on the fluctuation data and the stress data. The calculation method was based on the stress values ​​and fluctuation values ​​to calculate the loss values, with loss values ​​ranging from 3° to 7°. The data on the internal friction angle loss due to external force intervention were output. The ultimate bearing capacity simulation of foundation soil layers was performed based on the data of internal friction angle loss caused by external forces. A CSV file containing the data was loaded using Python, its contents were parsed, and the coordinates and internal friction angle loss values ​​of each cell were extracted. The internal friction angle loss values ​​ranged from 3° to 7°. A CSV file containing land geological characteristic data was also loaded, and geological type, soil layer thickness, and lithological parameters were extracted. Geological types included sandy clay and silty clay, with soil layer thicknesses ranging from 2 meters to 8 meters. Lithological parameters included cohesion ranging from 15 kPa to 25 kPa and internal friction angles ranging from 25° to 35°. Data was merged using Python's Pandas library, with cell coordinates as the merging condition, generating a merged dataset containing internal friction angle loss values ​​and geological characteristic values. The ultimate bearing capacity simulation of the foundation soil layers was performed using the NumPy library based on the Terzaghi bearing capacity theory. The calculation parameters included internal friction angle loss values, cohesion values, and soil layer thickness values. The calculation process involved adjusting the internal friction angle loss values... The internal friction angle was adjusted to range from 18° to 32°. The ultimate bearing capacity was calculated by combining the cohesion and soil layer thickness values, with the ultimate bearing capacity ranging from 120 kPa to 180 kPa. The ultimate bearing capacity data of the foundation soil layers was output in CSV format. An example of predicting structural instability risk due to geological changes based on the ultimate bearing capacity data of the foundation soil layers, the internal friction angle loss data due to external forces, and the sequence loose viscosity structure data of the soil layers is presented: A Python tool is used to load the CSV file of the ultimate bearing capacity data of the foundation soil layers and parse the text... The data was processed as follows: The coordinates and ultimate bearing capacity of each cell were extracted (range: 120 kPa to 180 kPa). A CSV file containing data on the internal friction angle loss due to external force was loaded, and the internal friction angle loss values ​​(range: 3° to 7°) were extracted. A CSV file containing data on the loose viscosity structure of soil layers was also loaded, and the loose viscosity values ​​(range: 0.25 to 0.45) were extracted. Finally, the data was merged using Python's Pandas library, with cell coordinates as the merging condition, to generate a merged dataset.The dataset contains ultimate bearing capacity values, internal friction angle loss values, and loose viscosity values. The NumPy library is used to predict the structural instability risk due to geological changes from the merged dataset. The prediction method is based on a weighted scoring model, with weighting coefficients of 0.4 for ultimate bearing capacity, 0.3 for internal friction angle loss, and 0.3 for loose viscosity. The risk value for each cell is calculated, ranging from 0.3 to 0.7. The output is the structural instability risk data due to geological changes, in CSV format.

[0030] The estimation of the internal friction angle loss based on the sequence loose viscosity structure data of soil layers includes: Obtain initial data of excavation force; based on the soil layer sequence loose viscosity structure data, match the excavation force required for different soil layer structures to obtain the excavation force matching data required for the soil layer structure; Analysis of the average difference in cohesion among soil types was performed on the sequence loose viscosity structure data of soil layers to obtain the average difference in cohesion among soil types. Based on the external force matching data required for soil structure excavation, the cohesion difference of soil type is used to simulate the lateral pressure fluctuation of cohesion dilatation, and the cohesion dilatation lateral pressure fluctuation data is obtained. Nonlinear shear friction force analysis was performed on the lateral pressure fluctuation data of cohesive shear dilatation to obtain nonlinear shear friction force data; Based on the lateral pressure fluctuation data of cohesive shear dilatation and the nonlinear shear friction force data, the internal friction angle loss under external force intervention is estimated, and the internal friction angle loss data under external force intervention is obtained.

[0031] In this embodiment of the invention, the initial data of the excavating external force is obtained through the actual operation records of engineering machinery equipment (such as excavators), including that the direction of the excavating external force is vertically downward and the intensity is 100 kN / m. 2The excavation force was applied for 10 seconds at a depth of 5 meters, covering an area of ​​10 square meters. This data was stored in tabular form and recorded and transmitted in real-time through a data acquisition system to ensure accuracy and completeness. A CSV file containing the loose viscosity structure data of the soil stratification was loaded using Python. The file content was parsed, and the coordinates and loose viscosity values ​​of each cell were extracted. The loose viscosity values ​​ranged from 0.25 to 0.45, and the data resolution was 10 meters × 10 meters. A CSV file containing the initial excavation force data was also loaded, and the timestamps and excavation force values ​​were extracted. The excavation force values ​​ranged from 5 kN to 15 kN. The data was merged using Python's Pandas library, with cell coordinates as the merging condition, generating a merged dataset containing both loose viscosity and excavation force values. The NumPy library was used to calculate the required excavation force for the soil stratification structure from the merged dataset. The matching method is based on the loose viscosity value to calculate the matching value. The calculation process involves classifying the soil layer type based on the loose viscosity value. The soil layer type includes sandy clay and silty clay. The loose viscosity range of sandy clay is 0.25 to 0.35, and the loose viscosity range of silty clay is 0.36 to 0.45. The matching value of sandy clay is 0.8 times the excavation force value, and the matching value of silty clay is 1.2 times the excavation force value. The calculation result is the matching data of the excavation force required for the soil structure. The data includes the coordinates of each cell and the matching value. The matching value ranges from 4kN to 18kN. The output is the matching data of the excavation force required for the soil structure in CSV format. This paper uses Python to load a CSV file containing soil sequence loose viscosity structure data, parses the file content, and extracts the coordinates and loose viscosity values ​​for each cell (ranging from 0.25 to 0.45). It also loads a CSV file containing land geological characteristics data, extracting the geological type and cohesion values ​​(including sandy clay and silty clay, with cohesion values ​​ranging from 15 kPa to 25 kPa). Finally, it uses Python's Pandas library to merge the data based on cell coordinates, generating a merged dataset containing both loose viscosity and cohesion values. The NumPy library is then used to further refine the dataset. The dataset was merged to perform mean variance analysis of soil type cohesion. The analysis method was to calculate the standard deviation based on the cohesion value. The calculation process was based on geological type grouping. The cohesion range of sandy clay was 15 kPa to 20 kPa, and the cohesion range of silty clay was 21 kPa to 25 kPa. The standard deviation of each group was calculated, with the standard deviation of sandy clay being 3 kPa and the standard deviation of silty clay being 2 kPa. The calculation result is the mean variance of soil type cohesion. The data includes the coordinates of each cell and the mean variance value, with the mean variance value ranging from 2 kPa to 3 kPa. The output is the mean variance data of soil type cohesion in CSV format.A CSV file containing the required excavation force matching data for the soil structure was loaded using Python tools. The file content was parsed, and the coordinates and matching values ​​of each cell were extracted. The matching values ​​ranged from 4kN to 18kN. A CSV file containing the average difference in cohesion for soil types was also loaded, and the average difference values ​​were extracted. The average difference values ​​ranged from 2kPa to 3kPa. The data was then merged using Python's Pandas library, with cell coordinates as the merging condition, generating a merged dataset containing both matching values ​​and average difference values. Finally, the NumPy library was used to perform cohesive shear dilatation lateral compression analysis on the merged dataset. Force fluctuation simulation was performed using finite element analysis with a mesh resolution of 5m × 5m. Simulation parameters included matching values ​​and mean difference values. The calculation process involved calculating shear stress based on matching values, ranging from 8kPa to 15kPa, calculating lateral pressure based on mean difference values, ranging from 5kPa to 10kPa, and calculating fluctuation values, ranging from 2kPa to 4kPa. The results were calculated as lateral pressure fluctuation data for viscoelastic shear dilatation, including the coordinates and fluctuation values ​​of each cell. The output data was in CSV format. A CSV file containing lateral pressure fluctuation data of cohesive shear dilatation was loaded using Python. The file content was parsed, and the coordinates and fluctuation values ​​of each cell were extracted. The fluctuation values ​​ranged from 2 kPa to 4 kPa. A CSV file containing land geological feature data was also loaded, and the geological type and internal friction angle value were extracted. The geological types included sandy clay and silty clay, and the internal friction angle value ranged from 25° to 35°. The data was merged using the Pandas library in Python, with the merging condition being cell coordinates. A merged dataset containing fluctuation values ​​and internal friction angle values ​​was generated. The NumPy library was used to perform nonlinear shear friction force analysis on the merged dataset. The analysis method was to calculate the force value based on shear stress and internal friction angle. The calculation process was to calculate the shear stress based on the fluctuation values, with the shear stress ranged from 8 kPa to 15 kPa, and to calculate the friction force based on the internal friction angle value, with the friction force ranged from 10 kPa to 18 kPa. The calculation results were nonlinear shear friction force data, containing the coordinates and force value of each cell, with the force value ranged from 10 kPa to 18 kPa. The nonlinear shear friction force data was output in CSV format.A CSV file containing lateral pressure fluctuation data of viscoelastic shear dilatation was loaded using Python. The file content was parsed, and the coordinates and fluctuation values ​​of each cell were extracted. The fluctuation values ​​ranged from 2 kPa to 4 kPa. A CSV file containing nonlinear shear friction force data was also loaded, and the force values ​​ranged from 10 kPa to 18 kPa. The data was merged using the Pandas library in Python, with cell coordinates as the merging condition, generating a merged dataset containing both fluctuation and force values. The NumPy library was used to estimate the internal friction angle loss due to external force intervention on the merged dataset. The estimation method was to calculate the loss value based on the fluctuation and force values. The calculation process involved calculating the shear stress change based on the fluctuation values, with the shear stress change ranged from 3 kPa to 5 kPa, and then calculating the internal friction angle loss based on the force values, with the internal friction angle loss ranged from 3° to 7°. The calculation results were the data on the internal friction angle loss due to external force intervention, containing the coordinates and loss value of each cell. The data was output as CSV.

[0032] The simulation of groundwater seepage change risk based on geological change and structural instability risk data includes: Obtain groundwater distribution data in the development area; Geological layer variability analysis was performed on geological layer structural instability risk data to obtain geological layer variability structural data; The relative weak numerical intervals of geological layer variation structure data are evaluated to obtain the relative weak numerical intervals of geological layers. Based on the groundwater distribution data of the development area, the shear moment imbalance of the water flow is calculated in the numerical range of the relatively weak geological layer to obtain the shear moment imbalance data of the water flow. Based on the imbalance data of water flow shear pressure moment, the risk of groundwater seepage change was simulated, and the risk data of groundwater seepage change was obtained.

[0033] In this embodiment of the invention, hydrogeological exploration technology, including borehole exploration and groundwater monitoring well data, is used to obtain groundwater distribution data in the development area. This data includes a groundwater depth of 5 meters, an aquifer thickness of 10 meters, a permeability coefficient of 0.001 m / s, a northeast-southeast flow direction, and a flow velocity of 0.1 m / d. This data is stored in a spatial database and visualized using a Geographic Information System (GIS) to ensure clear visibility of the spatial distribution characteristics. A CSV file containing geological change and structural instability risk data is loaded using Python. The file content is parsed, and the coordinates and risk values ​​of each cell are extracted. The risk values ​​range from 0.3 to 0.7, and the data resolution is 10 meters × 10 meters. A CSV file containing land geological feature data is also loaded, extracting the geological type and soil layer thickness. Geological types include sandy clay and silty clay, and soil layer thickness ranges from 2 meters to 8 meters. Data is merged using Python's Pandas library, with cell coordinates as the merging condition, generating a merged dataset containing risk values ​​and geological feature values. NumPy is then used to perform the merge. The y library performs geological layer variability analysis on a merged dataset. The analysis method involves calculating the coefficient of variation based on risk values. The calculation process is based on geological type grouping. The risk values ​​for sandy clay range from 0.3 to 0.5, and for silty clay from 0.5 to 0.7. The coefficient of variation for each group is calculated, with 0.15 for sandy clay and 0.25 for silty clay. The results are geological layer variability structure data, including the coordinates of each cell and the coefficient of variation, ranging from 0.15 to 0.25. The output is geological layer variability structure data in CSV format. The file is then parsed. The content involves extracting the coordinates and coefficient of variation (COP) of each cell, with the COP ranging from 0.15 to 0.25. A CSV file of land geological feature data is loaded, and the geological type and soil layer thickness are extracted. Geological types include sandy clay and silty clay, and soil layer thickness ranges from 2 to 8 meters. Data is merged using Python's Pandas library, with cell coordinates as the merging condition, generating a merged dataset containing COP and geological feature values. The merged dataset is then evaluated for weak numerical intervals using the NumPy library, with the evaluation method based on COP interval division and the division rule being the COP. The intervals with a coefficient of variation greater than 0.2 are considered relatively weak intervals, while the intervals with a coefficient of variation less than or equal to 0.2 are considered stable intervals. The calculation results are the numerical intervals of relatively weak geological layers. The data includes the coordinates of each cell and the interval type, which includes weak intervals and stable intervals. The coefficient of variation for weak intervals ranges from 0.21 to 0.25, and the coefficient of variation for stable intervals ranges from 0.15 to 0.20. The output is the numerical interval data of relatively weak geological layers in CSV format. The file content is parsed to extract the timestamp, groundwater level depth, and permeability coefficient. The groundwater level depth ranges from 1.5 meters to 3 meters.A CSV file containing data on relatively weak geological layers was loaded, with a permeability coefficient ranging from 1×10^-6 m / s to 5×10^-5 m / s. The coordinates and interval type of each cell were extracted, including weak and stable intervals. Data was merged using Python's Xarray and Pandas libraries, with cell coordinates as the merging condition, generating a merged dataset containing groundwater level depth, permeability coefficient, and interval type. The NumPy library was used to calculate the shear moment imbalance of the merged dataset. The calculation method involved calculating the shear moment based on the permeability coefficient and groundwater level depth. The calculation process involved calculating the water flow velocity based on the permeability coefficient, with the velocity ranging from 0.001 m / s to 0.005 m / s, combined with groundwater level measurement... Shear moment was calculated, ranging from 5 kPa to 10 kPa. For weaker sections, moment imbalance values ​​were calculated, ranging from 6 kPa to 8 kPa. The results are water flow shear-pressure moment imbalance data, including the coordinates and imbalance value of each cell. The output is water flow shear-pressure moment imbalance data in CSV format. The file content is parsed to extract the coordinates and imbalance value of each cell, ranging from 6 kPa to 8 kPa. A NetCDF file of groundwater distribution data for the development area is loaded to extract groundwater level depth and permeability coefficient, ranging from 1.5 meters to 3.5 meters, and permeability coefficient ranging from 1 × 10⁻⁶ m / s to 5 × 10⁻⁵. The data was merged using Python's Pandas and Xarray libraries, with cell coordinates as the merging condition. This generated a merged dataset containing imbalance values, groundwater level depth, and permeability coefficients. The NumPy library was used to simulate groundwater seepage risk on this merged dataset using the finite difference method. The grid resolution was 5m × 5m, and the time step was 1 hour. Simulation parameters included imbalance values, permeability coefficients, and groundwater level depth. The calculation process involved calculating the seepage change rate based on the imbalance values, ranging from 0.0001m / s to 0.0005m / s. The risk value was then calculated based on the groundwater level depth, ranging from 0.4 to 0.6. The results are groundwater seepage risk data, including the coordinates and risk value of each cell. The output data is in CSV format.

[0034] The risk warning model construction module includes: Convolutional calculations are performed on groundwater seepage change risk data to obtain groundwater seepage change risk convolutional data. Multi-objective scenario fitting was performed based on groundwater seepage change risk convolution data and geological change structural instability risk data to obtain geological development multi-objective scenario risk fitting data. A land development risk early warning model was constructed based on risk fitting data from multiple objective scenarios of geological development.

[0035] As an example of the present invention, reference is made to Figure 3 As shown, Figure 1 A functional flowchart of the risk warning model construction module is shown. In this example, the functions of the risk warning model construction module include: S301: Perform convolution calculations on groundwater seepage change risk data to obtain groundwater seepage change risk convolution data; In this embodiment of the invention, a CSV file of groundwater seepage change risk data is loaded using Python tools. The file content is parsed to extract the coordinates and risk value of each cell. The risk value ranges from 0.4 to 0.6, and the data resolution is 5 meters × 5 meters. The risk data is converted into a two-dimensional array using Python's NumPy library. The array dimension is 1000 × 1000, and the array elements are risk values. Convolution calculation is performed using the convolution function of the SciPy library. The convolution kernel is a 3 × 3 mean filter kernel with a kernel element value of 1 / 9. The convolution calculation process involves sliding the convolution kernel to cover the two-dimensional array, calculating the weighted average of each sliding window, and the sliding step size is 1. The calculation result is the groundwater seepage change risk convolution data, which includes the coordinates and convolution value of each cell. The convolution value ranges from 0.35 to 0.55. The groundwater seepage change risk convolution data is output in CSV format.

[0036] S302: Multi-objective scenario fitting is performed based on groundwater seepage change risk convolution data and geological change structural instability risk data to obtain geological development multi-objective scenario risk fitting data; In this embodiment of the invention, a CSV file containing convolutional data on groundwater seepage change risk is loaded using Python tools. The file content is parsed, and the coordinates and convolutional values ​​of each cell are extracted. The convolutional values ​​range from 0.35 to 0.55. A CSV file containing geological change and structural instability risk data is also loaded, and the coordinates and risk values ​​of each cell are extracted. The risk values ​​range from 0.3 to 0.7. The data resolution is 10 meters × 10 meters. The data is merged using Python's Pandas library, with cell coordinates as the merging condition, generating a merged dataset containing convolutional values ​​and risk values. The NumPy library is used to identify multi-objective risk factor linkage factors in the merged dataset. The identification method is based on correlation analysis to calculate linkage factors, with the linkage factors ranging from 0 to 1. The risk factors of runaway risk are quantified for the linkage factor from 0.2 to 0.8. The quantification method is to calculate the quantified value based on the linkage factor value, and the quantified value range is 0.1 to 0.5. Based on the quantified value, the antecedent risk mutation factors are identified. The identification method is to determine the mutation factors based on the interval where the quantified value is greater than 0.3, and the mutation factors range from 0.31 to 0.5. Based on the mutation factors and quantified values, multi-objective scenario fitting is performed. The fitting method is to calculate the fitted value based on the weighted average. The weighting coefficient is 0.6 for mutation factors and 0.4 for quantified values. The calculation result is the geological development multi-objective scenario risk fitting data. The data includes the coordinates of each cell and the fitted value. The fitted value range is 0.25 to 0.65. The geological development multi-objective scenario risk fitting data is output in CSV format.

[0037] S303: Based on the risk fitting data of multi-objective geological development scenarios, a land development risk early warning model is constructed, and the land development risk early warning model is obtained.

[0038] In this embodiment of the invention, a CSV file containing multi-objective scenario risk fitting data for geological development is loaded using Python tools. The file content is parsed, and the coordinates and fitted values ​​of each cell are extracted. The fitted values ​​range from 0.25 to 0.65. The fitted data is then normalized using Python's NumPy library using a min-max normalization method, with a normalization range of 0 to 1. The normalized result is multi-objective scenario risk normalized data, containing the coordinates and normalized values ​​of each cell, with normalized values ​​ranging from 0.1 to 0.9. The Pandas library is used to train the normalized data. The training set and test set were divided with a ratio of 70% and 30% respectively. The training set contained 7000 samples, and the test set contained 3000 samples. Random feature sampling was performed on the training set, using a random number generator to extract 50% of the features. The sampling result was a randomized training set for scenario risks. An initial land development risk early warning model was constructed based on the policy gradient algorithm on the sampled training set. The algorithm parameters included a learning rate of 0.001, 1000 iterations, a three-layer neural network structure with 64 hidden layer nodes, and ReLU (Rectified Linear Unit) activation function. The constructed result was the initial land development risk early warning model. The initial model was tested and optimized using the test set. The optimization method was to adjust the model parameters based on gradient descent, with the optimization objective being to minimize the mean squared error (MSE), which ranged from 0.01 to 0.05. The optimized result was the land development risk early warning model.

[0039] The multi-objective scenario fitting based on groundwater seepage change risk convolution data and geological change structural instability risk data includes the following steps: Based on the convolutional data of groundwater seepage change risk and the data of geological change and structural instability risk, the linkage factors of multi-objective risk factors are identified, and the linkage factors of multi-objective risk factors are obtained. The out-of-control risk factors are quantified by analyzing the linkage factors of multi-objective risk factors, and quantitative data of the linkage factors of out-of-control risks are obtained. Based on the quantitative data of the linkage factors of runaway risk, the antecedent risk mutation factors are identified, and the antecedent risk mutation factors are obtained. Based on the quantitative data of the antecedent risk mutation factors and the linkage factors of the runaway risk, multi-objective scenario fitting is performed to obtain multi-objective scenario risk fitting data for geological development.

[0040] In this embodiment of the invention, the Apriori algorithm is used to analyze convolutional data on groundwater seepage change risk and geological change / structural instability risk. Input data includes a groundwater seepage change risk feature value of 0.8 and a geological change / structural instability risk feature value of 0.6. By setting a minimum support of 0.5 and a minimum confidence of 0.7, the linkage relationship between multiple risk factors is identified. For example, in a certain area, the identification results show that the linkage factor between groundwater seepage change risk and geological change / structural instability risk is 0.75, and the linkage area is mainly concentrated in the central location of the development area. A risk quantification model is used to quantify the linkage factor of multiple risk factors. Input data includes a linkage factor of 0.75. By setting risk quantification standards as high risk greater than 0.8, medium risk 0.5-0.8, and low risk less than 0.5, the quantified value of the runaway risk factor is calculated. For example, in a certain area, the quantification results show that the runaway risk value is 0.7, the risk level is medium risk, and the quantified area is mainly concentrated at the edge of the development area. Mutation detection algorithms (such as CUSUM, Cumulative Sum Control) are used. Charts (cumulative and control charts) are used to analyze the quantitative data of factors linking the risk of runaway risk. The input data includes a runaway risk value of 0.7. By setting a mutation threshold of 0.6, antecedent risk mutation factors are identified. For example, in a certain area, the identification results show that the antecedent risk mutation factor is a groundwater seepage path deviation of more than 5 meters. The mutation area is mainly concentrated in the center of the development area. A multi-objective optimization algorithm (such as NSGA-II, Non-dominated Sorting Genetic Algorithm II) is used to fit the quantitative data of antecedent risk mutation factors and runaway risk linkage factors. The input data includes an antecedent risk mutation factor of groundwater seepage path deviation of more than 5 meters and a runaway risk linkage factor quantitative value of 0.7. By setting a weight coefficient of 0.5 and the objective function of minimizing the risk value, the risk fitting value under the multi-objective scenario is calculated. For example, in a certain area, the fitting result shows that the risk fitting value of the multi-objective scenario is 0.65, and the fitting area is mainly concentrated at the edge of the development area.

[0041] In another embodiment, a CSV file containing groundwater seepage change risk convolutional data is loaded using Python tools. The file content is parsed, and the coordinates and convolutional values ​​of each cell are extracted. The convolutional values ​​range from 0.35 to 0.55, and the data resolution is 5 meters × 5 meters. A CSV file containing geological change and structural instability risk data is also loaded, and the coordinates and risk values ​​of each cell are extracted. The risk values ​​range from 0.3 to 0.7, and the data resolution is 10 meters × 10 meters. The data is merged using Python's Pandas library, with cell coordinates as the merging condition, generating a merged dataset containing convolutional and risk values. The NumPy library is used to identify multi-objective risk factor linkage factors in the merged dataset. The identification method is based on Pearson correlation analysis to calculate linkage factors. The calculation process involves calculating the correlation coefficient between the convolutional and risk values, with the correlation coefficient ranging from 0.2 to 0.8. The calculation result is the multi-objective risk factor linkage factor, containing the coordinates and linkage factor values ​​of each cell, with linkage factor values ​​ranging from 0.25 to 0.75. The multi-objective risk factor linkage factor data is output in CSV format. Load a CSV file containing multi-objective risk factor linkage factor data, parse the file content, and extract the coordinates and linkage factor values ​​for each cell. The linkage factor values ​​range from 0.25 to 0.75. Load a CSV file containing geological change and structural instability risk data, and extract the risk values, which range from 0.3 to 0.7. Merge the data using Python's Pandas library, using cell coordinates as the merging condition, to generate a merged dataset containing linkage factor values ​​and risk values. Use the NumPy library to quantify the runaway risk factors in the merged dataset, using a quantification method based on linkage factor values ​​and risk... The risk value is calculated by dividing the value into intervals based on the linkage factor value. The intervals range from 0.25 to 0.5 and from 0.51 to 0.75. The quantified value for the interval 0.25 to 0.5 is 0.5 times the linkage factor value, and the quantified value for the interval 0.51 to 0.75 is 0.8 times the linkage factor value. The quantified value is adjusted in conjunction with the risk value, with an adjustment range of 0.1 to 0.5. The calculation result is the quantified data of the linkage factors of runaway risk. The data includes the coordinates and quantified value of each cell, with the quantified value ranging from 0.15 to 0.45. The output is the quantified data of the linkage factors of runaway risk in CSV format.This paper uses Python to load a CSV file containing quantitative data on runaway risk linkage factors, parses the file content, and extracts the coordinates and quantitative values ​​of each cell (ranging from 0.15 to 0.45). It also loads a CSV file containing convolutional data on groundwater seepage change risk and extracts the convolutional values ​​(ranging from 0.35 to 0.55). The data is then merged using Python's Pandas library, with cell coordinates as the merging condition, generating a merged dataset containing both quantitative and convolutional values. NumPy is then used to identify antecedent risk mutation factors in the merged dataset. The identification method involves calculating mutation factors based on both quantitative and convolutional values. The calculation process involves determining mutation intervals based on quantified values ​​greater than 0.3 (ranging from 0.31 to 0.45), adjusting the mutation factors based on the convolutional values ​​(ranging from 0.31 to 0.5), and obtaining the antecedent risk mutation factors. This data includes the coordinates and mutation factor values ​​of each cell (ranging from 0.32 to 0.48). The antecedent risk mutation factor data is then output in CSV format. Load a CSV file containing data on preceding risk mutation factors, parse the file content, and extract the coordinates and mutation factor values ​​for each cell. The mutation factor values ​​range from 0.32 to 0.48. Load a CSV file containing quantified data on runaway risk linkage factors, and extract the quantified values, which range from 0.15 to 0.45. Merge the data using the Pandas library in Python, with cell coordinates as the merging condition, to generate a merged dataset containing mutation factor values ​​and quantified values. Use the NumPy library to perform multi-objective scenario fitting on the merged dataset. The fitting method is based on a weighted average to calculate the fitted value, with the weighting coefficients being 0.6 for mutation factor values ​​and 0.4 for quantified values. The calculation process involves calculating a weighted average based on the mutation factor values ​​and quantified values, with the fitted value ranging from 0.25 to 0.65. The calculation result is the multi-objective scenario risk fitting data for geological development, containing the coordinates and fitted values ​​for each cell. Output the multi-objective scenario risk fitting data for geological development.

[0042] The construction of the land development risk early warning model based on multi-objective scenario risk fitting data for geological development includes: Normalize the risk fitting data of multi-objective scenarios in geological development to obtain normalized risk data of multi-objective scenarios. The risk normalization data for multi-objective scenarios is divided into training and testing sets to obtain the multi-objective scenario risk training set and the multi-objective scenario risk testing set. Random feature sampling is performed on the multi-target scenario risk training set to obtain a scenario risk random sampling training set; The initial land development risk early warning model is constructed by randomly sampling the training set of scenario risks based on the policy gradient algorithm. The initial land development risk early warning model was tested and optimized using a multi-objective scenario risk test set, resulting in a new land development risk early warning model.

[0043] In this embodiment of the invention, a CSV file containing risk fitting data for multi-objective scenarios in geological development is loaded. The file content is parsed, and the coordinates and fitted values ​​of each cell are extracted. The fitted values ​​range from 0.25 to 0.65, and the data resolution is 5 meters × 5 meters. The fitted data is normalized using Python's NumPy library. The normalization method is based on minimum-maximum standardization. The calculation process involves calculating the normalized value based on the minimum fitted value of 0.25 and the maximum fitted value of 0.65. The normalized value ranges from 0 to 1. The calculation result is the normalized risk data for the multi-objective scenario. This dataset contains the coordinates and normalized value of each cell, with the normalized value ranging from 0.1 to 0.9. The output is normalized risk data for multi-objective scenarios in CSV format. The CSV file containing this data is loaded, its contents are parsed, and the coordinates and normalized value of each cell are extracted. The normalized value ranges from 0.1 to 0.9, and the total sample size is 10,000. The normalized data is then split into training and test sets using Python's Pandas library. The splitting method is based on random partitioning, with the training set comprising 70% and the test set 30%. The training set contains 7... The test set contains 3000 samples, and the coordinates and normalized values ​​of each sample are extracted. The normalized values ​​range from 0.1 to 0.9. The total number of samples is 7000, and the feature dimension is 10, including geological type, soil layer thickness, permeability coefficient, etc. Random feature sampling is performed on the training set using Python's NumPy library. The sampling method is to extract 50% of the features based on a random number generator with a random number seed of 42. The sampling result is a random sampling training set for scene risk, containing 7000 samples with a feature dimension of 5. The sampled features include geological features. The model outputs a random sampling training set of scenario risks, including features such as type and penetration coefficient. A CSV file containing this training set is loaded using Python tools. The file content is parsed to extract the coordinates, sampling feature values, and normalized values ​​for each sample. The total number of samples is 7000, and the feature dimension is 5. An initial land development risk early warning model is constructed using Python's TensorFlow library. The model structure is a three-layer neural network with 5 input layer nodes, 64 hidden layer nodes, and 1 output layer node. The activation function is ReLU (Rectified Linear Unit). The model is trained based on a policy gradient algorithm with parameters including a learning rate of 0.001, a discount factor of 0.95, and 1000 iterations. The training process calculates the loss based on the sampling feature values ​​and normalized values, with the loss ranging from 0.01 to 0.05. Update model parameters, constructing the initial land development risk early warning model. Load the CSV file of the multi-objective scenario risk test set, parse the file content, and extract the coordinates, feature values, and normalized values ​​of each sample. The total number of samples is 3000, and the feature dimension is 10. Load the initial land development risk early warning model (model path: D:\LandPlanningData\InitialModel). Test the initial model using Python's TensorFlow library. The testing process involves calculating predicted values ​​based on the feature values ​​of the test set, with predicted values ​​ranging from 0.15 to 0.85. Calculate the mean squared error between the predicted values ​​and the normalized values, with a mean squared error ranging from 0.01 to 0.05. Optimize the model using the gradient descent algorithm, with optimization parameters including a learning rate of 0.0005 and 500 iterations. The optimization process involves adjusting model parameters based on the mean squared error. The optimized result is the land development risk early warning model.

[0044] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

Claims

1. A land development information data intelligent management system based on multi-source data, characterized in that, Includes the following steps: The land geological feature extraction module is used to acquire multi-source data of land development areas; Geological features of different regions are extracted based on multi-source data of land development areas to obtain land geological feature data of different regions. The risk prediction module is used to obtain land development planning data; Based on land development planning data, geological change and structural instability risk prediction is carried out on land geological characteristic data of different regions to obtain geological change and structural instability risk data. Based on the geological change and structural instability risk data, groundwater seepage change risk simulation was performed to obtain groundwater seepage change risk data; The risk warning model construction module is used to fit multi-objective scenarios based on groundwater seepage change risk data to obtain multi-objective scenario risk fitting data for geological development. A land development risk early warning model was constructed based on risk fitting data from multiple objective scenarios of geological development, and the land development risk early warning model was obtained. The management execution module is used to send the land development risk early warning model to the control center to perform intelligent management of development information data.

2. The intelligent management system for land development information based on multi-source data according to claim 1, characterized in that, The land geological feature extraction module includes: Obtain multi-source data on land development areas; Multi-source data cleaning is performed on multi-source data of land development areas to obtain cleaned land development area data. Regional land topographic structure analysis was performed on the cleaned data of land development areas to obtain regional land topographic structure data. Based on multi-source data of land development areas, geological features of different regions are extracted from regional land topographic structure data to obtain land geological feature data of different regions.

3. The intelligent management system for land development information based on multi-source data according to claim 1, characterized in that, The risk prediction module includes: Obtain land development planning data; Based on land development planning data, theoretical topographic change demand analysis was conducted on land geological characteristic data of different regions to obtain theoretical topographic change data for land development. Based on the land development topographic theory change data, the geological geological feature data of different regions are used to predict the risk of geological change and structural instability, and the geological change and structural instability risk data are obtained. Based on the geological change and structural instability risk data, a groundwater seepage change risk simulation was conducted to obtain groundwater seepage change risk data.

4. The intelligent management system for land development information based on multi-source data according to claim 3, characterized in that, The prediction of geological change and structural instability risk based on land geological characteristic data of different regions using land development topographic theory change data includes: Based on the land development topography theory change data, the stratigraphic deformation structure analysis was carried out on the land geological feature data of different regions to obtain the stratigraphic deformation structure data of different regions. Sequence loose viscosity structure analysis of soil layers was performed on the stratigraphic deformation structure data of different regions to obtain the sequence loose viscosity structure data of soil layers. Based on the sequence loose viscosity structure data of soil layers, the internal friction angle loss under external force intervention is estimated to obtain the data of internal friction angle loss under external force intervention. The ultimate bearing capacity of the foundation soil layer is simulated and calculated based on the loss data of the internal friction angle caused by external force intervention, and the ultimate bearing capacity data of the foundation soil layer is obtained. Based on the ultimate bearing capacity data of foundation soil, the internal friction angle loss data of external force intervention, and the sequence loose viscosity structure data of soil layers, the risk of structural instability due to geological changes is predicted, and the risk data of structural instability due to geological changes is obtained.

5. The intelligent management system for land development information based on multi-source data according to claim 4, characterized in that, The estimation of the internal friction angle loss based on the sequence loose viscosity structure data of soil layers includes: Obtain initial data of excavation force; based on the soil layer sequence loose viscosity structure data, match the excavation force required for different soil layer structures to obtain the excavation force matching data required for the soil layer structure; Analysis of the average difference in cohesion among soil types was performed on the sequence loose viscosity structure data of soil layers to obtain the average difference in cohesion among soil types. Based on the external force matching data required for soil structure excavation, the cohesion difference of soil type is used to simulate the lateral pressure fluctuation of cohesion dilatation, and the cohesion dilatation lateral pressure fluctuation data is obtained. Nonlinear shear friction force analysis was performed on the lateral pressure fluctuation data of cohesive shear dilatation to obtain nonlinear shear friction force data; Based on the lateral pressure fluctuation data of cohesive shear dilatation and the nonlinear shear friction force data, the internal friction angle loss under external force intervention is estimated, and the internal friction angle loss data under external force intervention is obtained.

6. The intelligent management system for land development information based on multi-source data according to claim 3, characterized in that, The simulation of groundwater seepage change risk based on geological change and structural instability risk data includes: Obtain groundwater distribution data in the development area; Geological layer variability analysis was performed on geological layer structural instability risk data to obtain geological layer variability structural data; The relative weak numerical intervals of geological layer variation structure data are evaluated to obtain the relative weak numerical intervals of geological layers. Based on the groundwater distribution data of the development area, the shear moment imbalance of the water flow is calculated in the numerical range of the relatively weak geological layer to obtain the shear moment imbalance data of the water flow. Based on the imbalance data of water flow shear pressure moment, the risk of groundwater seepage change was simulated, and the risk data of groundwater seepage change was obtained.

7. The intelligent management system for land development information based on multi-source data according to claim 1, characterized in that, The risk warning model construction module includes: Convolutional calculations are performed on groundwater seepage change risk data to obtain groundwater seepage change risk convolutional data. Multi-objective scenario fitting was performed based on groundwater seepage change risk convolution data and geological change structural instability risk data to obtain geological development multi-objective scenario risk fitting data. A land development risk early warning model was constructed based on risk fitting data from multiple objective scenarios of geological development.

8. The intelligent management system for land development information based on multi-source data according to claim 7, characterized in that, The multi-objective scenario fitting based on groundwater seepage change risk convolution data and geological change structural instability risk data includes the following steps: Based on the convolutional data of groundwater seepage change risk and the data of geological change and structural instability risk, the linkage factors of multi-objective risk factors are identified, and the linkage factors of multi-objective risk factors are obtained. The out-of-control risk factors are quantified by analyzing the linkage factors of multi-objective risk factors, and quantitative data of the linkage factors of out-of-control risks are obtained. Based on the quantitative data of the linkage factors of runaway risk, the antecedent risk mutation factors are identified, and the antecedent risk mutation factors are obtained. Based on the quantitative data of the antecedent risk mutation factors and the linkage factors of the runaway risk, multi-objective scenario fitting is performed to obtain multi-objective scenario risk fitting data for geological development.

9. The intelligent management system for land development information based on multi-source data according to claim 7, characterized in that, The construction of the land development risk early warning model based on multi-objective scenario risk fitting data for geological development includes: Normalize the risk fitting data of multi-objective scenarios in geological development to obtain normalized risk data of multi-objective scenarios. The risk normalization data for multi-objective scenarios is divided into training and testing sets to obtain the multi-objective scenario risk training set and the multi-objective scenario risk testing set. Random feature sampling is performed on the multi-target scenario risk training set to obtain a scenario risk random sampling training set; The initial land development risk early warning model is constructed by randomly sampling the training set of scenario risks based on the policy gradient algorithm. The initial land development risk early warning model was tested and optimized using a multi-objective scenario risk test set, resulting in a new land development risk early warning model.

Images