Air-space-ground collaborative evaluation model for dynamic pre-repair effect of coal mining collapse

Abstract

The invention relates to the technical field of coal mining collapse evaluation, and discloses an air-space-ground collaborative evaluation model for a dynamic pre-repair effect of coal mining collapse. The model comprises a data acquisition module, a space-time registration module, an index construction module, a dynamic simulation module and an effect feedback module. The data acquisition module acquires terrain, vegetation coverage and hydrological data of a coal mining collapse area through the space-air-ground multi-source monitoring equipment and divides the data into layers; the space-time registration module performs time synchronization and space coordinate unification on different levels of data, and constructs a space-time association relationship; the index construction module is used for extracting landform stability, ecological restoration and hydrological cycle indexes and calculating a comprehensive evaluation index; the dynamic simulation module simulates a pre-repairing scheme based on the index, predicts an index change trend and generates a dynamic simulation curve; and the effect feedback module is used for comparing the simulation curve with actual monitoring data, identifying deviation, adjusting model parameter weights and updating comprehensive evaluation indexes.

Description

Technical Field

[0001] This invention relates to the field of coal mining subsidence evaluation technology, specifically to a space-air-ground collaborative evaluation model for the dynamic pre-repair effect of coal mining subsidence. Background Technology

[0002] Surface subsidence caused by coal mining activities is a common geological and environmental problem in the process of mineral resource development. It not only changes the original topography but also significantly disturbs the regional ecosystem and hydrological cycle. With the large-scale mining of coal resources, the area of ​​coal mining subsidence zones continues to expand and the degree of subsidence gradually intensifies. As a result, problems such as waste of land resources, vegetation degradation, and water resource imbalance are becoming increasingly prominent, posing a severe challenge to regional sustainable development.

[0003] The evaluation of the restoration effect in coal mining subsidence areas often relies on a single monitoring method, such as obtaining local data through ground observation alone, or conducting large-scale macroscopic analysis based solely on remote sensing imagery. Single monitoring methods are insufficient to comprehensively capture the complex changes in subsidence areas. Ground observations are limited by spatial coverage, while remote sensing data suffers from deficiencies in detail accuracy and timeliness. Furthermore, existing evaluation methods have weak capabilities for integrating multi-source data; data from different sources and scales often cannot be effectively correlated due to inconsistencies in spatiotemporal references, thus affecting the accuracy and reliability of the evaluation results.

[0004] In terms of indicator selection, traditional evaluation systems often focus on single dimensions such as topographic stability or vegetation restoration, lacking a comprehensive consideration of the ecosystem and hydrological cycle, making it difficult to fully reflect the actual effects of restoration plans. Furthermore, existing evaluations are mostly static assessments, analyzing the restoration status only at a specific point in time, unable to predict and track dynamic changes during the restoration process, and difficult to adjust the evaluation model in real time based on actual monitoring data. This limits the guiding role of evaluation results in optimizing pre-restoration plans. Summary of the Invention

[0005] The purpose of this invention is to provide a space-air-ground collaborative evaluation model for the dynamic pre-repair effect of coal mining subsidence, so as to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, this invention provides a space-air-ground collaborative evaluation model for the dynamic pre-repair effect of coal mining subsidence, the model comprising: The data acquisition module is used to acquire topographic data, vegetation cover data and hydrological data of coal mining subsidence areas through multi-source monitoring equipment in air, space and ground, and divides the data acquisition level according to the data type; The spatiotemporal registration module is used to synchronize the time and unify the spatial coordinates of collected data at different levels, and to build spatiotemporal correlations between multi-source data. The indicator construction module is used to extract topographic stability indicators, ecological restoration indicators, and hydrological cycle indicators from spatiotemporal correlation data, and calculate a comprehensive evaluation index according to the indicator weights. The dynamic simulation module is used to perform numerical simulations of pre-repair schemes based on comprehensive evaluation indices, predict the trend of index changes at different repair stages, and generate dynamic simulation curves. The effect feedback module is used to compare the dynamic simulation curve with the actual monitoring data, identify simulation deviations, adjust the parameter weights of the evaluation model according to the deviation value, and update the comprehensive evaluation index.

[0007] Preferably, the data acquisition module can be implemented in the following ways: For any monitoring period in the coal mining subsidence area, deploy satellite remote sensing equipment, drone aerial photography equipment, and ground sensor networks; Regional-scale topographic elevation data and vegetation cover data are obtained through satellite remote sensing equipment; Data on the distribution of surface cracks and soil moisture content at the mesoscale were obtained using drone aerial photography equipment. Groundwater level and soil nutrient data at the point scale are acquired through a ground sensor network, and the data is divided into three data acquisition levels according to spatial resolution.

[0008] Preferably, the implementation methods for dividing data acquisition levels also include: Satellite remote sensing data, UAV aerial photography data, and ground sensor data are sorted according to timestamps to obtain a time-series data sequence; Spatial coordinate information is extracted from the time-series data sequence and matched with the reference coordinates in the regional geographic information system to obtain coordinate correction results; Based on the coordinate correction results, data from different sources are assigned to corresponding spatial grid cells to form a multi-level gridded data acquisition structure.

[0009] Preferably, the methods for constructing spatiotemporal correlations of multi-source data also include: Outlier detection is performed on the spatiotemporally registered multi-source data to remove data samples that exceed a reasonable threshold. Analyze the time series correlation and spatial autocorrelation of data samples after removing those exceeding a reasonable threshold, and then associate and label data samples with correlations higher than a set threshold. Based on the results of the association labeling, a spatiotemporal association matrix of multi-source data is established, and the matrix elements represent the association strength of different data types in the same spatiotemporal unit.

[0010] Preferably, the implementation methods of the indicator construction module include: The evaluation criteria in the fields of terrain stability, ecological restoration and hydrological cycle are invoked to generate an initial index set, which represents the basic evaluation indicators without weight allocation. The Analytic Hierarchy Process (AHP) is used to calculate the weights of the initial indicator set and select indicators with weight values ​​higher than a set threshold as core evaluation indicators. The implementation steps of AHP include: constructing a hierarchical structure model with the target layer as the comprehensive evaluation index, the criterion layer as three types of indicators such as terrain stability, and the scheme layer as specific indicator items; constructing a judgment matrix through expert scoring; and performing consistency checks and correcting inconsistency matrices. By mapping the core evaluation indicators to the data items in the spatiotemporal correlation matrix, quantifiable topographic stability indicators, ecological restoration indicators, and hydrological cycle indicators are obtained.

[0011] Preferably, the implementation method of constructing a multi-level gridded data acquisition structure includes: taking the spatial grid unit in the gridded data as the basic unit, and combining the resolution information and data type labels of the data acquisition level, overlaying the multi-level data acquisition structure with the administrative division layer in the geographic information system; The implementation methods for overlaying with the administrative division layer include: checking the topological relationship between the boundaries of multi-level gridded data and the boundaries of township-level administrative divisions, classifying grid units that cross administrative divisions according to the principle of area proportion, and generating a grid index table with administrative division codes.

[0012] Preferably, the implementation methods of the dynamic simulation module include: The comprehensive evaluation index is input into the preset numerical simulation model, which includes a terrain evolution sub-model, a vegetation growth sub-model, and a hydrological cycle sub-model. Set the parameter variables for the pre-repair plan, including the amount of repair materials used, vegetation planting density, and drainage system layout; Run the numerical simulation model and output the index change curves under different combinations of parameter variables. Use the slope and inflection point of the curve as dynamic simulation feature values.

[0013] Preferably, the implementation methods of the effect feedback module include: The predicted values ​​of each evaluation indicator are extracted from the dynamic simulation curve, and the measured values ​​of the corresponding indicators are extracted from the actual monitoring data; the absolute error and relative error between the predicted value and the measured value are calculated, and the error value is compared with the preset allowable error range. If the error value exceeds the allowable error range, the weight parameters in the indicator construction module are corrected, and the comprehensive evaluation index is recalculated. The implementation of weight parameter correction includes: calculating the proportion of each indicator error value to the total error, allocating correction coefficients according to the proportion, linearly adjusting the weight parameters, recalculating the comprehensive evaluation index after correction, and verifying whether the error is within the allowable range.

[0014] The preferred method for updating the comprehensive evaluation index is as follows: Based on the corrected weight parameters output by the effect feedback module, adjust the weight ratios of the terrain stability index, ecological restoration index, and hydrological cycle index. The comprehensive evaluation index is recalculated according to the new weighting, and an updated evaluation report is generated. The updated evaluation report is compared with the historical evaluation report to record the trend of indicator changes and form a dynamic evaluation archive.

[0015] Preferably, the initialization parameters of the numerical simulation model include: original topographic data of the coal mining subsidence area, soil physical property parameters, vegetation species parameters and hydrogeological parameters, and the simulation duration covers the entire cycle of the pre-remediation scheme.

[0016] Compared with the prior art, the beneficial effects of the present invention are: The model acquires data through a multi-source monitoring system using air, space, and ground equipment, covering various aspects such as topography, vegetation cover, and hydrology in coal mining subsidence areas. Furthermore, it divides the data acquisition into different levels according to data types, making data acquisition more systematic and targeted, avoiding the limitations of a single data source, and ensuring that the basic information used for subsequent evaluations is more comprehensive.

[0017] The spatiotemporal registration module synchronizes the time and unifies the spatial coordinates of collected data at different levels, and constructs spatiotemporal correlations of multi-source data. This solves the problem that multi-source data is difficult to integrate due to inconsistent spatiotemporal benchmarks. It enables data from different sources and at different scales to be correlated with each other within the same spatiotemporal framework, providing a consistent data foundation for subsequent indicator extraction and comprehensive evaluation. This allows various data to be effectively combined to jointly serve the evaluation process.

[0018] The indicator construction module extracts topographic stability indicators, ecological restoration indicators, and hydrological cycle indicators from spatiotemporal correlation data. The selection of multi-dimensional indicators covers the key aspects of the restoration effect in coal mining subsidence areas, rather than being limited to a single dimension. Then, a comprehensive evaluation index is calculated through indicator weights, so that the evaluation results can comprehensively reflect the effects of the restoration plan at different levels, reflecting the comprehensiveness and completeness of the evaluation.

[0019] The dynamic simulation module uses a comprehensive evaluation index as a basis to perform numerical simulations on the pre-repair scheme, predict the trend of indicator changes at different repair stages and generate dynamic simulation curves. It can show the evolution of various indicators in advance during the repair process, allowing relevant personnel to understand the possible state of the repair scheme at different stages and helping to form a clear understanding of the overall direction of the scheme.

[0020] The effect feedback module compares the dynamic simulation curve with the actual monitoring data, identifies simulation deviations, adjusts the parameter weights of the evaluation model based on the deviation values, and updates the comprehensive evaluation index. This allows the model to be continuously optimized according to the actual situation, enhancing its adaptability and flexibility. The evaluation results can be dynamically adjusted as the actual data changes, making them more consistent with the real situation of the repair process.

[0021] The various modules work together to form a complete evaluation loop. From the comprehensiveness of data collection to the coordination of data processing, the comprehensiveness of indicator construction, the foresight of simulation prediction and the dynamism of model adjustment, the cooperation of multiple links makes the entire evaluation process more scientific and practical, and can more accurately reflect the actual effect of dynamic pre-repair of coal mining subsidence. Attached Figure Description

[0022] Figure 1 This is a time series diagram of the space-air-ground collaborative evaluation model for the dynamic pre-repair effect of coal mining subsidence as described in this invention; Figure 2 A diagram illustrating the working principle of the implementation method for dividing data acquisition levels; Figure 3 A diagram illustrating the working principle of constructing spatiotemporal correlations among multi-source data. Figure 4 This is a diagram illustrating the working principle of the dynamic simulation module. Detailed Implementation

[0023] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] Please see Figure 1 This invention provides a space-air-ground collaborative evaluation model for the dynamic pre-repair effect of coal mining subsidence. The model includes: a data acquisition module, a spatiotemporal registration module, an index construction module, a dynamic simulation module, and an effect feedback module. These modules work collaboratively to evaluate the dynamic pre-repair effect of coal mining subsidence. The specific implementation steps are as follows: The data acquisition module obtains topographic, vegetation cover, and hydrological data of the coal mining subsidence area through multi-source monitoring equipment (space, air, and ground), and divides the data acquisition levels according to data type. The spatiotemporal registration module synchronizes the time and unifies the spatial coordinates of the acquired data at different levels, constructing spatiotemporal correlations between the multi-source data. The index construction module extracts topographic stability, ecological restoration, and hydrological cycle indicators from the spatiotemporally correlated data, and calculates a comprehensive evaluation index according to the index weights. The dynamic simulation module uses the comprehensive evaluation index as a basis to numerically simulate the pre-remediation plan, predicts the indicator change trends at different remediation stages, and generates dynamic simulation curves. The effect feedback module compares the dynamic simulation curves with actual monitoring data, identifies simulation deviations, adjusts the parameter weights of the evaluation model based on the deviation values, and updates the comprehensive evaluation index.

[0025] Example 1: See Figure 2 In the implementation of the data acquisition module, satellite remote sensing equipment, UAV aerial photography equipment, and a ground sensor network are deployed for any monitoring period in the coal mining subsidence area. The satellite remote sensing equipment selects satellites with high spatial and temporal resolution to perform remote sensing imaging of the coal mining subsidence area at set time intervals (e.g., once every 7 days) within the monitoring period, acquiring regional-scale topographic elevation data and vegetation cover data. Topographic elevation data is acquired through a stereo mapping camera on the satellite and processed to generate a digital elevation model, which reflects the undulations and changes in the terrain within the area. Vegetation cover data is extracted by performing band calculations on the satellite remote sensing images and using the Normalized Difference Vegetation Index (NDVI) method. The NDVI value reflects the growth status and coverage of vegetation.

[0026] The aerial photography equipment used is a multi-rotor drone equipped with a high-resolution optical camera and a multispectral sensor. During the monitoring period, a mid-scale aerial photography operation is conducted every three days, depending on weather conditions and the actual situation of the subsidence area. The drone flies autonomously along a pre-set route covering the entire coal mining subsidence area, with the flight altitude set according to the required spatial resolution (e.g., 100 meters altitude corresponds to 5 meters resolution). Image data acquired by the high-resolution optical camera, after stitching and processing, allows for the extraction of surface crack distribution data, including crack location, length, width, and orientation. Image data acquired by the multispectral sensor is used to retrieve soil moisture content data. By analyzing the relationship between reflectivity in different wavelength bands and soil moisture content, the spatial distribution of soil moisture content within the area is obtained.

[0027] The ground-based sensor network consists of multiple monitoring nodes, distributed at a specific density (e.g., 5 nodes per square kilometer) within the coal mining subsidence area, with a focus on key locations such as areas with dense surface cracks, vegetation restoration zones, and hydrologically sensitive areas. Each monitoring node is equipped with a groundwater level sensor and a soil nutrient sensor. The groundwater level sensor is buried at a certain depth (e.g., 5 meters) through boreholes to collect groundwater level data in real time, with a data collection frequency of once per hour. The soil nutrient sensor is buried 20 centimeters below the surface to collect data on the content of nutrients such as nitrogen, phosphorus, and potassium in the soil, with a collection frequency of once per day. All data collected by the sensors is transmitted to a data center via a wireless transmission module, forming point-scale monitoring data.

[0028] The satellite remote sensing data, UAV aerial photography data, and ground sensor data are sorted according to timestamps to obtain a time-series data sequence. Timestamps are accurate to the second to ensure data accuracy in the time dimension. Spatial coordinate information is extracted from the time-series data sequence. The spatial coordinates of the satellite remote sensing data and UAV aerial photography data are obtained through georeferenced imagery, while the spatial coordinates of the ground sensor data are obtained through GPS positioning at the time of sensor deployment. This spatial coordinate information is matched with the reference coordinates in the regional geographic information system (such as the National Geodetic Coordinate System 2000), and coordinate transformation algorithms (such as the seven-parameter transformation method) are used to perform coordinate correction on the data, obtaining the coordinate correction result, so that the spatial coordinates of all data are unified under the same reference.

[0029] Based on the coordinate correction results, data from different sources are assigned to corresponding spatial grid cells, forming a multi-level gridded data acquisition structure. The division of spatial grid cells is determined according to the spatial resolution of the data. The grid cell size for satellite remote sensing data is 30m × 30m, the grid cell size for UAV aerial photography data is 5m × 5m, and the grid cell size for ground sensor data is 1m × 1m. For satellite remote sensing data, the digital elevation model and vegetation cover data are divided into 30m × 30m grids, with each grid cell assigned a corresponding elevation and vegetation cover value. For UAV aerial photography data, the surface crack distribution data and soil moisture content data are divided into 5m × 5m grids, with each grid cell recording the presence, length, width, and soil moisture content of the cracks. For ground sensor data, groundwater level data and soil nutrient data are assigned to 1m × 1m grid cells, with each grid cell storing the sensor monitoring value at the corresponding location.

[0030] Example 2: See Figure 3When constructing the spatiotemporal correlation of multi-source data, outlier detection is first performed on the spatiotemporally registered multi-source data. Appropriate detection methods are used for different types of data. For topographic elevation data, a reasonable threshold is set based on the overall undulation of the regional topography; for example, data exceeding 20% ​​of the difference between the region's historical highest and lowest elevations are considered outliers. For vegetation cover data, since its theoretical range is between 0 and 1, data less than 0 or greater than 1 are directly considered outliers. For groundwater level data, a reasonable water level fluctuation range is set based on regional hydrogeological conditions; values ​​exceeding this range are considered outliers. For soil nutrient data, data significantly deviating from the conventional nutrient content range for regional soil types are marked as outliers. Detected outlier samples are manually verified and then removed to avoid interference from subsequent analysis.

[0031] After outlier removal, the temporal correlation and spatial autocorrelation of the remaining data samples are analyzed. Temporal correlation analysis examines the degree of association between different types of data at the same spatial location over time. For example, it analyzes the relationship between soil moisture content and vegetation cover within a grid cell during the monitoring period. This is achieved by calculating the correlation coefficients of different data types over time. Specifically, each data series is standardized to eliminate the influence of dimensions, and then the Pearson correlation coefficient is calculated. This coefficient ranges from -1 to 1, with positive values ​​indicating positive correlation and negative values ​​indicating negative correlation. A larger absolute value indicates a stronger correlation. Spatial autocorrelation analysis explores the degree of association of the same data type at different spatial locations. For example, it analyzes the distribution characteristics of surface crack distribution data between adjacent grid cells, using the Moran's index for quantification. The Moran's index also ranges from -1 to 1. A value greater than 0 indicates positive spatial correlation, meaning adjacent cells have similar attributes; a value less than 0 indicates negative spatial correlation, meaning adjacent cells have dissimilar attributes; and a value close to 0 indicates a random spatial distribution.

[0032] The time-series correlation coefficient and the Moran's index of spatial autocorrelation are compared with set thresholds. If the time-series correlation coefficient of two sets of data is greater than 0.6 and the spatial Moran's index is greater than 0.5, then the two sets of data are considered to have a strong correlation in the spatiotemporal dimensions and are labeled as such. The labeling includes the data type (e.g., topographic data and vegetation data), the corresponding timestamp (accurate to the time of collection), and specific spatial location information (e.g., grid cell number). The specific value of the correlation coefficient is also recorded as the basis for subsequent quantification of the correlation strength.

[0033] Based on the association labeling results, a spatiotemporal association matrix of multi-source data is constructed. The matrix is ​​constructed using data samples as the basic unit, with rows and columns corresponding to different data samples. Each data sample contains information such as data type, timestamp, and spatial coordinates. The element values ​​in the matrix represent the association strength of the corresponding row and column data samples within the same spatiotemporal unit. The association strength is obtained by combining the time series correlation coefficient and the Moran's index of spatial autocorrelation; for example, the average of the two is used as the quantification value of the association strength, ranging from 0 to 1, with values ​​closer to 1 indicating a stronger association. For datasets not labeled in the association labeling, their corresponding element values ​​in the matrix are set to 0, indicating a weak or no association. In this way, the complex spatiotemporal associations between multi-source data are presented intuitively in matrix form, forming a complete spatiotemporal association structure, providing a clear data association basis for subsequent extraction of evaluation indicators from the associated data. After the matrix is ​​constructed, a completeness check is performed to ensure that all data samples that have undergone outlier removal are included in the matrix, and that the matrix element values ​​accurately reflect the association between the data. If omissions or errors are found, they are promptly supplemented and corrected.

[0034] Example 3: Evaluation standards in the fields of terrain stability, ecological restoration, and hydrological cycle were applied. These standards cover nationally issued technical specifications for mine ecological restoration, industry-standard ecological environment evaluation indicator systems, and related indicator sets widely used in academic research. Based on these standards, basic evaluation indicators reflecting the pre-remediation effect of coal mining subsidence areas were identified, forming an initial indicator set. The initial indicator set includes unweighted indicators such as terrain slope, terrain undulation, surface crack density, vegetation coverage, vegetation growth rate, soil organic matter content, soil pH, groundwater level depth, groundwater level variation, soil moisture content, and surface runoff.

[0035] The Analytic Hierarchy Process (AHP) was used to calculate the weights of the initial indicator set. A hierarchical model was constructed, consisting of three levels: the target level, which is a comprehensive evaluation index used to reflect the overall pre-remediation effect; the criteria level, which includes three categories: topographic stability indicators, ecological restoration indicators, and hydrological cycle indicators, each measuring the remediation effect from different dimensions; and the scheme level, which contains specific indicators, such as topographic stability indicators including topographic slope, topographic relief, and surface crack density; ecological restoration indicators including vegetation coverage, vegetation growth rate, and soil organic matter content; and hydrological cycle indicators including groundwater depth, groundwater level variation, and soil moisture content. Experts from multiple fields were used to collect evaluations of the importance of each indicator in the scheme level relative to the criteria level. Each expert scored the indicators based on their relative importance, using a 1-9 scale, where 1 indicates equal importance, 3 indicates slightly more important, 5 indicates significant importance, 7 indicates strong importance, and 9 indicates extreme importance. 2, 4, 6, and 8 represent intermediate values ​​among these adjacent values. Based on the scores from all experts, the average value is calculated and a judgment matrix is ​​constructed. The elements in the judgment matrix represent the importance ratio of the corresponding row indicator to the corresponding column indicator.

[0036] The constructed judgment matrix is ​​subjected to a consistency test, and the consistency index CI and the average random consistency index RI are calculated. The formula for calculating CI is as follows: ; In the formula, To determine the largest eigenvalue of a matrix, The order of the judgment matrix is ​​used for evaluation. RI is the average random consistency index obtained through random simulation, and its value is determined according to the matrix order. For example, RI is 0.58 for an order of 3, and 0.90 for an order of 4. The consistency ratio CR is calculated as CR = CI / RI. If CR is less than 0.1, the judgment matrix is ​​considered to meet the consistency requirements. If CR is greater than or equal to 0.1, the judgment matrix needs to be corrected by adjusting the element values ​​and repeating the consistency check until CR is less than 0.1. Based on the judgment matrix that meets the consistency requirements, the weight value of each indicator is calculated. Specifically, this is obtained by solving the eigenvector of the judgment matrix. After normalization, the eigenvector becomes the weight of each indicator. Indicators with weight values ​​higher than a set threshold are selected as core evaluation indicators. The threshold is determined according to the actual evaluation needs; for example, the threshold can be set to 0.05, and indicators with weight values ​​lower than this threshold are excluded.

[0037] The core evaluation indicators are mapped to data items in the spatiotemporal correlation matrix. Based on the physical meaning and data characteristics of the indicators, the raw data is transformed and standardized. For example, for the terrain slope indicator, slope values ​​are extracted from terrain elevation data and converted into slope change rate to reflect the change of slope over time; for the surface crack density indicator, crack length per unit area is calculated based on surface crack distribution data and converted into crack closure degree to reflect the changing trend of cracks during the restoration process; for the vegetation cover indicator, vegetation cover data is converted into cover growth rate to reflect the speed of vegetation recovery; for the soil nutrient content indicator, soil nutrient data is standardized to eliminate dimensional differences between different nutrient types; for the groundwater level indicator, the rate of change of groundwater level is calculated to reflect the dynamic changes in water level; for the soil moisture content indicator, the average moisture content over a certain period is calculated to reflect the soil moisture status. Through the above processing, the raw data is transformed into quantifiable indicator values, ultimately yielding terrain stability indicators, ecological restoration indicators, and hydrological cycle indicators. These indicators can quantitatively reflect the pre-remediation effect of coal mining subsidence areas from different perspectives.

[0038] Example 4: See Figure 4 When constructing a multi-level gridded data acquisition structure, spatial grid units in the gridded data are used as the basic units. Combining the resolution information of the data acquisition levels and data type labels, the multi-level data acquisition structure is overlaid with the administrative division layer in the geographic information system (GIS). The resolution information of the data acquisition levels is determined according to the characteristics of different devices. The resolution of satellite remote sensing data is set to 30 meters, meaning each grid unit represents a 30m x 30m area on the ground; the resolution of UAV aerial photography data is set to 5 meters, corresponding to 5m x 5m grid units; and the resolution of ground sensor data is set to 1 meter, corresponding to 1m x 1m grid units. Data type labels include topography, vegetation, and hydrology. Topography labels correspond to topographic elevation data, surface crack distribution data, etc.; vegetation labels correspond to vegetation cover data, soil nutrient data, etc.; and hydrology labels correspond to groundwater level data, soil moisture content data, etc. The administrative division layer in the GIS contains township-level administrative boundary vector data. The boundary data is obtained from basic geographic information data released by the National Bureau of Surveying and Mapping and has clear boundary coordinates and administrative division codes.

[0039] The boundaries of multi-level gridded data are topologically checked against township-level administrative division boundaries using a topology check tool in Geographic Information System (GIS) software. The check includes whether grid cells are entirely located within a single township administrative region, whether they span two or more township administrative regions, and whether there is overlap or gap between grid cell boundaries and administrative division boundaries. For grid cells entirely located within a single township administrative region, they are directly assigned to that township. For grid cells spanning multiple administrative divisions, their assignment is determined by calculating the area percentage of the grid cell within each of the relevant township administrative regions. For example, a 5m x 5m grid cell is partially located in township A and partially in township B. Using a measurement tool, the area of ​​this grid cell in township A is calculated to be 18 square meters, and its area in township B is 7 square meters, representing 72% and 28% of the area respectively. Therefore, this grid cell is assigned to township A. After assigning all grid cells, a grid index table with administrative division codes is generated. The index table is stored in a database table format and includes fields such as the unique identifier of the grid cell, the administrative division code of the township to which it belongs, the data resolution, the data type label, and the coordinates of the top-left and bottom-right corners of the grid cell, facilitating subsequent data querying and management.

[0040] When the dynamic simulation module is running, the comprehensive evaluation index is input into the preset numerical simulation model. This model consists of a terrain evolution sub-model, a vegetation growth sub-model, and a hydrological cycle sub-model. The sub-models are coupled through a data interface. That is, the output of the terrain evolution sub-model is used as the input parameter of the vegetation growth sub-model and the hydrological cycle sub-model. The output of the vegetation growth sub-model and the hydrological cycle sub-model will also have a feedback effect on the terrain evolution sub-model, forming a dynamic feedback mechanism. The initial parameters of the numerical simulation model were obtained through preliminary data collection and laboratory analysis. The original topographic data of the coal mining subsidence area was obtained using a high-precision topographic elevation map of the area before the subsidence occurred, derived from historical survey data and satellite remote sensing imagery. Soil physical property parameters, including soil texture (such as sand, silt, and clay content), soil bulk density, and porosity, were determined through laboratory analysis of collected soil samples. Vegetation species parameters, including the species' growth cycle, suitable soil moisture content range, and salt tolerance, were obtained by consulting relevant botanical literature based on the selected vegetation type in the remediation plan. Hydrogeological parameters, including aquifer thickness, permeability coefficient, and specific yield, were obtained from geological survey reports and pumping test data. The simulation duration was set to the entire lifecycle of the pre-remediation plan. For example, if the pre-remediation plan is planned for 5 years, the simulation duration is set to 5 years, with a time step of 1 month, meaning simulation results are output monthly.

[0041] When setting the parameters for the pre-remediation scheme, the amount of remediation material used includes the thickness of the soil layer (e.g., different gradients such as 10 cm, 20 cm, 30 cm), and the amount of organic fertilizer applied (e.g., 1 kg, 2 kg, 3 kg per square meter); the planting density includes the spacing between trees (e.g., 2 m × 2 m, 3 m × 3 m), the planting density of shrubs (e.g., 3 plants per square meter, 5 plants per square meter), and the sowing amount of herbaceous plants (e.g., 20 g, 30 g per square meter); the layout of the drainage system includes the direction of the drainage ditches (e.g., along the contour lines, perpendicular to the contour lines), the pipe diameter (e.g., 300 mm, 500 mm), and the spacing between drainage ditches (e.g., 50 m, 100 m). When running the numerical simulation model, multiple simulation schemes are generated by adjusting the combinations of parameter variable values, with each scheme corresponding to a set of parameter variable values. During model operation, real-time values ​​of terrain stability, ecological restoration, and hydrological cycle indicators are output. Based on these values, indicator change curves are plotted, with the horizontal axis representing the simulation time (months) and the vertical axis representing the quantified indicator value. Feature extraction is performed on the generated dynamic simulation curves, using the curve slope as a quantified representation of the indicator's rate of change. For example, a curve segment with a positive slope and a large absolute value indicates that the indicator is growing rapidly within the corresponding time period. Inflection points on the curves correspond to the points in time when the indicator's trend changes; for instance, an inflection point on the vegetation cover curve may correspond to the point in time when vegetation growth transitions from a slow phase to a rapid phase.

[0042] Example 5: During the operation of the effect feedback module, predicted values ​​of each evaluation indicator at different time points are extracted from the dynamic simulation curve. The selection of time points is consistent with the actual data collection time, for example, predicted values ​​are extracted once a month on the 1st, covering specific values ​​of topographic stability indicators (such as topographic slope change rate and surface crack closure), ecological restoration indicators (such as vegetation coverage growth rate and soil nutrient content), and hydrological cycle indicators (such as groundwater level change rate and average soil moisture content). At the same time, the measured values ​​of the corresponding indicators at the same time points are extracted from the actual monitoring data. The measured values ​​are obtained from satellite remote sensing data, UAV aerial photography data, and ground sensor data continuously acquired by the data acquisition module, after spatiotemporal registration and data processing.

[0043] Calculate the absolute and relative errors between the predicted and measured values. The absolute error is the absolute value of the difference between the predicted and measured values, and the relative error is the ratio of the absolute error to the measured value, rounded to two decimal places. Compare the calculated error values ​​with the preset allowable error ranges. The allowable error ranges are determined based on the characteristics of the indicator and the required evaluation accuracy. For example, the allowable absolute error range for the rate of change of terrain slope is ±0.02° / month, the allowable relative error range for the growth rate of vegetation cover is ±10%, and the allowable absolute error range for the rate of change of groundwater level is ±0.05 meters / month.

[0044] If the error value exceeds the allowable error range, the weight parameter correction process in the indicator construction module is initiated. The proportion of each indicator's error value to the total error is calculated. The total error is the sum of the absolute errors of all indicators, and the error proportion of each indicator is the ratio of its absolute error to the total error, rounded to two decimal places. Correction coefficients are allocated according to the error proportion; the higher the error proportion, the larger the corresponding correction coefficient, with a total correction coefficient of 1. The weight parameters are linearly adjusted based on the correction coefficients. For example, if an indicator's original weight is 0.15, its error proportion is 20%, and its assigned correction coefficient is 0.2, then the adjusted weight is the original weight plus (correction coefficient × adjustment magnitude). The adjustment magnitude is determined based on the degree to which the error exceeds the allowable range; for example, if the error exceeds the allowable range by 10%, the adjustment magnitude is 0.02. The sum of all adjusted indicator weights must remain at 1. If the adjusted sum is not equal to 1, normalization is performed to ensure the rationality of the weight allocation. After correction, the comprehensive evaluation index is recalculated, and the new predicted value is extracted again and compared with the measured value. This correction process is repeated until the error values ​​of all indicators are within the allowable error range.

[0045] When updating the comprehensive evaluation index, the weight percentages of the topographic stability, ecological restoration, and hydrological cycle indicators are redefined based on the corrected weight parameters output by the effect feedback module. For example, if the weight percentage of the topographic stability indicator was originally 30%, it is adjusted to 35% after correction; the weight percentage of the ecological restoration indicator is adjusted from 40% to 38%; and the weight percentage of the hydrological cycle indicator is adjusted from 30% to 27%. According to the new weight percentages, the quantified value of each indicator at the corresponding time point is multiplied by its weight, and the products are summed to obtain the new comprehensive evaluation index value. An updated evaluation report is generated, which includes the evaluation period, the comprehensive evaluation index at each time point, the weight and quantified value of each indicator, and indicator change curves, presented in a combination of charts and text.

[0046] The updated evaluation report is compared with historical evaluation reports, which include the previous and earlier evaluation results. The comparison covers changes in the comprehensive evaluation index, adjustments to the weights of each indicator, and specific trends in the values ​​of topographic stability, ecological restoration, and hydrological cycle indicators. Changes observed during the comparison process are recorded, such as upward or downward trends in the comprehensive evaluation index, continuous increases or decreases in the weights of certain indicators, and fluctuations in the values ​​of individual indicators. These records are compiled into a dynamic evaluation archive, stored electronically and archived chronologically. The archive includes the original data, calculation process, evaluation report, and comparative analysis results from each evaluation, facilitating the tracking of the evolution of the evaluation model and the long-term changes in the pre-remediation effects.

[0047] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0048] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A space-air-ground collaborative evaluation model for the dynamic pre-repair effect of coal mining subsidence, characterized in that, include: The data acquisition module is used to acquire topographic data, vegetation cover data and hydrological data of coal mining subsidence areas through multi-source monitoring equipment in air, space and ground, and divides the data acquisition level according to the data type; The spatiotemporal registration module is used to synchronize the time and unify the spatial coordinates of collected data at different levels, and to build spatiotemporal correlations between multi-source data. The indicator construction module is used to extract topographic stability indicators, ecological restoration indicators, and hydrological cycle indicators from spatiotemporal correlation data, and calculate a comprehensive evaluation index according to the indicator weights. The dynamic simulation module is used to perform numerical simulations of pre-repair schemes based on comprehensive evaluation indices, predict the trend of index changes at different repair stages, and generate dynamic simulation curves. The effect feedback module is used to compare the dynamic simulation curve with the actual monitoring data, identify simulation deviations, adjust the parameter weights of the evaluation model according to the deviation value, and update the comprehensive evaluation index.

2. The space-air-ground collaborative evaluation model for the dynamic pre-repair effect of coal mining subsidence according to claim 1, characterized in that, The data acquisition module can be implemented in the following ways: For any monitoring period in the coal mining subsidence area, deploy satellite remote sensing equipment, drone aerial photography equipment, and ground sensor networks; Regional-scale topographic elevation data and vegetation cover data are obtained through satellite remote sensing equipment; Data on the distribution of surface cracks and soil moisture content at the mesoscale were obtained using drone aerial photography equipment. Groundwater level and soil nutrient data at the point scale are acquired through a ground sensor network, and the data is divided into three data acquisition levels according to spatial resolution.

3. The space-air-ground collaborative evaluation model for the dynamic pre-repair effect of coal mining subsidence according to claim 2, characterized in that, Other ways to implement data acquisition hierarchy division include: Satellite remote sensing data, UAV aerial photography data, and ground sensor data are sorted according to timestamps to obtain a time-series data sequence; Spatial coordinate information is extracted from the time-series data sequence and matched with the reference coordinates in the regional geographic information system to obtain coordinate correction results; Based on the coordinate correction results, data from different sources are assigned to corresponding spatial grid cells to form a multi-level gridded data acquisition structure.

4. The space-air-ground collaborative evaluation model for the dynamic pre-repair effect of coal mining subsidence according to claim 1, characterized in that, Other methods for constructing spatiotemporal correlations of multi-source data include: Outlier detection is performed on the spatiotemporally registered multi-source data to remove data samples that exceed a reasonable threshold. Analyze the time series correlation and spatial autocorrelation of data samples after removing those exceeding a reasonable threshold, and then associate and label data samples with correlations higher than a set threshold. Based on the results of the association labeling, a spatiotemporal association matrix of multi-source data is established, and the matrix elements represent the association strength of different data types in the same spatiotemporal unit.

5. The space-air-ground collaborative evaluation model for the dynamic pre-repair effect of coal mining subsidence according to claim 1, characterized in that, The implementation methods for the indicator construction module include: The evaluation criteria in the fields of terrain stability, ecological restoration and hydrological cycle are invoked to generate an initial index set, which represents the basic evaluation indicators without weight allocation. The Analytic Hierarchy Process (AHP) is used to calculate the weights of the initial indicator set and select indicators with weight values ​​higher than a set threshold as core evaluation indicators. The implementation steps of AHP include: constructing a hierarchical structure model with the target layer as the comprehensive evaluation index, the criterion layer as three types of indicators such as terrain stability, and the scheme layer as specific indicator items; constructing a judgment matrix through expert scoring; and performing consistency checks and correcting inconsistency matrices. By mapping the core evaluation indicators to the data items in the spatiotemporal correlation matrix, quantifiable topographic stability indicators, ecological restoration indicators, and hydrological cycle indicators are obtained.

6. The space-air-ground collaborative evaluation model for the dynamic pre-repair effect of coal mining subsidence according to claim 3, characterized in that, The implementation methods for constructing a multi-level gridded data acquisition structure include: using spatial grid units in the gridded data as basic units, combining the resolution information and data type labels of the data acquisition level, and overlaying the multi-level data acquisition structure with the administrative division layer in the geographic information system; The implementation methods for overlaying with the administrative division layer include: checking the topological relationship between the boundaries of multi-level gridded data and the boundaries of township-level administrative divisions, classifying grid units that cross administrative divisions according to the principle of area proportion, and generating a grid index table with administrative division codes.

7. The space-air-ground collaborative evaluation model for the dynamic pre-repair effect of coal mining subsidence according to claim 1, characterized in that, The implementation methods of the dynamic simulation module include: The comprehensive evaluation index is input into the preset numerical simulation model, which includes a terrain evolution sub-model, a vegetation growth sub-model, and a hydrological cycle sub-model. Set the parameter variables for the pre-repair plan, including the amount of repair materials used, vegetation planting density, and drainage system layout; Run the numerical simulation model and output the index change curves under different combinations of parameter variables. Use the slope and inflection point of the curve as dynamic simulation feature values.

8. The space-air-ground collaborative evaluation model for the dynamic pre-repair effect of coal mining subsidence according to claim 7, characterized in that, The implementation methods for the effect feedback module include: The predicted values ​​of each evaluation indicator are extracted from the dynamic simulation curve, and the measured values ​​of the corresponding indicators are extracted from the actual monitoring data; the absolute error and relative error between the predicted value and the measured value are calculated, and the error value is compared with the preset allowable error range. If the error value exceeds the allowable error range, the weight parameters in the indicator construction module are corrected, and the comprehensive evaluation index is recalculated. The implementation of weight parameter correction includes: calculating the proportion of each indicator error value to the total error, allocating correction coefficients according to the proportion, linearly adjusting the weight parameters, recalculating the comprehensive evaluation index after correction, and verifying whether the error is within the allowable range.

9. The space-air-ground collaborative evaluation model for the dynamic pre-repair effect of coal mining subsidence according to claim 1, characterized in that, The method for updating the comprehensive evaluation index is as follows: Based on the corrected weight parameters output by the effect feedback module, adjust the weight ratios of the terrain stability index, ecological restoration index, and hydrological cycle index. The comprehensive evaluation index is recalculated according to the new weighting, and an updated evaluation report is generated. The updated evaluation report is compared with the historical evaluation report to record the trend of indicator changes and form a dynamic evaluation archive.

10. The space-air-ground collaborative evaluation model for the dynamic pre-repair effect of coal mining subsidence according to claim 7, characterized in that, The initialization parameters of the numerical simulation model include: original topographic data of the coal mining subsidence area, soil physical property parameters, vegetation species parameters and hydrogeological parameters, and the simulation duration covers the entire cycle of the pre-remediation plan.

Images