A mine off-site supervision over-limit over-mining dynamic inversion method and system

Abstract

The application relates to the technical field of mining administration, and discloses a mine off-site supervision cross-border over-mining dynamic inversion method and system, a whole-process automation technical system of "data fusion-constraint inversion-closed-loop supervision" is constructed, accurate dynamic inversion and off-site supervision of mine cross-border over-mining behavior in a large range such as a province are realized, compared with the prior art, by deeply coupling mining administration data and InSAR time sequence images, deformation information is directly associated with supervision rules, through a block parallel computing framework, global data processing can be efficiently completed, and the bottleneck of small-range point monitoring is overcome; meanwhile, under the improvement of three-dimensional mining administration constraints, the precision error of the inversion model is greatly reduced, and then, relying on the automation inversion comparison of the mining administration constraints, accurate cross-border over-mining behavior identification is realized, non-rule-breaking interference caused by legal mining and geological changes is effectively avoided, and the reliability of supervision is improved.

Description

Technical Field

[0001] This invention relates to the technical field of mining administration, specifically a dynamic inversion method and system for off-site monitoring of over-mining and boundary violations in mines. Background Technology

[0002] The global mining industry is accelerating its transformation towards remote and intelligent operations, and off-site supervision has become the mainstream international regulatory trend. Multiple factors have combined to provide a multi-faceted foundation for the research and application of off-site supervision technology in mining, including policy support, industry demand, and international reference. In particular, for large-scale mining scenarios such as provincial areas, building suitable off-site supervision technology has become an inevitable choice.

[0003] Existing mine monitoring and inversion technologies have significant shortcomings in data fusion, monitoring scope, inversion accuracy, and regulatory implementation. These common and individual deficiencies directly restrict the realization of large-scale off-site supervision of mines and accurate determination of over-mining and exceeding boundaries. Specifically, existing technologies have the following problems:

[0004] Insufficient data integration and lack of correlation with mining regulations mean that existing technologies have not achieved systematic integration of core mining management data with deformation monitoring data. They can only identify surface deformation or the existence of goaf areas, but cannot correlate the inversion results with regulatory rules. It is difficult to accurately determine whether mining activities fall under the compliance categories of "crossing the boundary" or "over-extraction". They can only output abnormal signals but cannot complete the characterization of violations, which is out of touch with the needs of mining law enforcement.

[0005] The monitoring scope is limited, the adaptability is insufficient, it is difficult to meet the regulatory needs of large-scale areas such as provinces, it has not solved the problem of balancing the efficiency and accuracy of large-scale data processing, it is difficult to adapt to the core requirements of full-domain non-site supervision, and there are obvious shortcomings in scenario adaptability.

[0006] The inversion accuracy is flawed, lacks support from mining regulations, and cannot distinguish between deformation differences caused by legal mining and over-mining. It is easy to misjudge settlement caused by legal mining as a violation signal, or to miss over-mining behavior due to similar deformation characteristics. The accuracy is insufficient to meet regulatory requirements.

[0007] The lack of effective implementation of supervision and the absence of a complete closed-loop supervision system necessitate reliance on manual secondary assessments. This prevents the full-process automation of "deformation inversion - rule verification - anomaly warning - dynamic updates," failing to meet the high-efficiency and automated requirements of off-site supervision by mining authorities. Summary of the Invention

[0008] The purpose of this invention is to provide a dynamic inversion method and system for off-site monitoring of over-extraction in mines to solve the problems mentioned in the background art.

[0009] To achieve the above objectives, the present invention provides the following technical solution:

[0010] A dynamic inversion method for off-site monitoring of over-extraction in mines, comprising:

[0011] Acquire multi-source information data on mineral administration within the supervised area and perform standardized data processing to establish a standardized mineral administration database. The multi-source information data includes mineral administration data, InSAR time-series images, and auxiliary geological data.

[0012] Based on the GIS spatial overlay algorithm, the mining administration data is mapped to the InSAR time series image to establish a spatiotemporal relationship, and dynamic weight optimization and allocation are performed through the random forest algorithm to achieve feature fusion and establish a fused feature field.

[0013] Based on the distribution of supervised regions, the fusion feature field is divided into grid blocks. The fusion feature fields of multiple sub-regions are output through multi-grid parallel processing and then stitched together to obtain the global fusion result.

[0014] A three-dimensional mineral policy constraint based on space, depth and reserves is constructed. The three-dimensional mineral policy constraint is assigned values ​​and the parameter inversion operation is performed on the fused feature field to identify and mark the over-extraction area.

[0015] The system excludes legitimate mining activities through fuzzy field matching and logical judgment, generates visual warnings for non-excluded illegal areas, and dynamically updates and monitors the process based on time-series iteration.

[0016] As a further aspect of the present invention: the mineral administration data includes mineral ownership data, mineral right boundary vectors, depth vector data, and registered reserve data; the step of implementing feature fusion specifically includes:

[0017] Based on the GIS spatial overlay algorithm, the mining right boundary vector and InSAR time series image are matched pixel by pixel through intersection analysis and nearest neighbor interpolation to establish a one-to-one spatial mapping relationship between deformed pixels and mining right units.

[0018] A feature-level dynamic weight fusion model is constructed, and a random forest algorithm is used for 10-fold cross-validation to optimize multiple weight coefficients. These weight coefficients include mineral administration rule parameters, InSAR deformation values, and geological parameters. The sum of all weight coefficients is 1, and the mineral administration rule parameters are not less than 0.4. The dynamic weight fusion model is characterized as follows:

[0019] ,

[0020] in, Let be the InSAR deformation value at coordinates (x, y) at time t. These are parameters for mining regulations (quantified boundary distance and permissible depth threshold). For geological parameters, α, β, and γ are the corresponding weights (with values ​​ranging from 0 to 1, and α + β + γ = 1).

[0021] As a further embodiment of the present invention: the step of dividing the fusion feature field into grid blocks based on the distribution of supervised regions, outputting the fusion feature fields of multiple sub-regions through multi-grid parallel processing, and stitching them together to obtain the global fusion result specifically includes:

[0022] The fused feature field is divided into several independent sub-regions based on a preset unit boundary grid. Specifically, the preset unit boundary grid is 10KM*10KM.

[0023] Based on a parallel computing framework, dynamic weight fusion operations are performed synchronously on data from several sub-regions to obtain the fusion feature field of each sub-region.

[0024] Edge stitching and consistency verification are automatically performed on the fusion feature fields of each sub-region to obtain the global fusion result and ensure spatial continuity and uniformity.

[0025] As a further aspect of the present invention: the three-dimensional mineral constraints specifically include:

[0026] Spatial constraints are established using the mining right boundary vector as a hard constraint, and a boundary determination benchmark is set. When the inverted area exceeds the boundary by more than or equal to 5 meters, it is determined to be mining beyond the boundary.

[0027] Depth constraints are calculated by combining the mining depth registered with the mining administration with the regional average mining depth. The inversion depth threshold is represented as H threshold = registered mining depth - 50M. Mining below this value is judged as ultra-deep mining.

[0028] Reserve constraints are set by establishing a monthly mining limit based on the conversion relationship between deformation volume and extractable reserves. If the limit is exceeded, it is considered over-extraction. The conversion relationship is represented as: Reserves = Deformation volume × Rock density × Recovery rate.

[0029] As a further embodiment of the present invention: the step of performing parameter inversion calculation on the fused feature field to identify and mark the over-extraction area specifically includes:

[0030] Mining policy constraint equations are introduced and the traditional probability integral method is optimized to construct an objective function. The objective function is then solved using the gradient descent algorithm to invert the three-dimensional parameters of the goaf. The objective function is characterized as follows:

[0031] ,

[0032] in, The deformation value after fusion is P, where P represents the parameters of the goaf (length L, width W, depth H, and mining thickness M). Here, λ is the constraint function for mining regulations, and λ is the constraint weight (with a value of 0.6-0.8 to ensure the effectiveness of the constraint).

[0033] The three-dimensional parameters obtained by inversion are compared with the three-dimensional mining constraints. If any condition of the three-dimensional mining constraints is met, it is marked as an over-mining area.

[0034] This invention aims to provide a dynamic inversion system for off-site monitoring of over-extraction in mines, characterized by comprising:

[0035] The data acquisition module is used to acquire multi-source information data on mineral administration within the supervised area and perform standardized data processing to establish a standardized mineral administration database. The multi-source information data includes mineral administration data, InSAR time-series images, and auxiliary geological data.

[0036] The spatiotemporal correlation module is used to map the mining administration data to the InSAR time series image based on the GIS spatial overlay algorithm to establish spatiotemporal correlation, and to perform dynamic weight optimization and allocation through the random forest algorithm to achieve feature fusion and establish a fused feature field.

[0037] The gridding module is used to divide the fusion feature field into grid blocks based on the distribution of the supervised region. It outputs the fusion feature fields of multiple sub-regions through parallel processing of multiple grids and then stitches them together to obtain the global fusion result.

[0038] The boundary violation determination module is used to construct three-dimensional mineral policy constraints based on space, depth and reserves, assign values ​​to the three-dimensional mineral policy constraints and perform parameter inversion calculations on the fused feature field to identify and mark the boundary violation and over-mining areas.

[0039] The verification output module is used to exclude legal mining situations through fuzzy field matching and logical judgment, generate visual warnings for illegal areas that are not excluded, and dynamically update and monitor the process based on time-series iteration.

[0040] As a further aspect of the present invention: the mineral administration data includes mineral ownership data, mineral right boundary vectors, depth vector data, and registered reserve data; the spatiotemporal correlation module includes:

[0041] The spatial mapping unit is used to accurately match the mining right boundary vector with the InSAR time series image pixel by pixel based on the GIS spatial overlay algorithm through intersection analysis and nearest neighbor interpolation, and establish a one-to-one spatial mapping relationship between deformed pixels and mining right units.

[0042] The weight optimization unit is used to construct a feature-level dynamic weight fusion model. It employs a random forest algorithm for 10-fold cross-validation to optimize multiple weight coefficients. These weight coefficients include mineral administration rule parameters, InSAR deformation values, and geological parameters. The sum of all weight coefficients is 1, and the mineral administration rule parameters are not less than 0.4. The dynamic weight fusion model is characterized as follows:

[0043] ,

[0044] in, Let be the InSAR deformation value at coordinates (x, y) at time t. These are parameters for mining regulations (quantified boundary distance and permissible depth threshold). For geological parameters, α, β, and γ are the corresponding weights (with values ​​ranging from 0 to 1, and α + β + γ = 1).

[0045] As a further embodiment of the present invention: the meshing module includes:

[0046] The region segmentation unit is used to segment the fused feature field into several independent sub-regions based on a preset unit boundary grid. Specifically, the unit boundary grid is preset to be 10KM*10KM.

[0047] The parallel computing unit is used to synchronously perform dynamic weight fusion operations on data from several sub-regions based on a parallel computing framework, so as to obtain the fusion feature field of each sub-region;

[0048] The fusion verification unit is used to automatically perform edge stitching and consistency verification on the fusion feature field of each sub-region to obtain the global fusion result and ensure spatial continuity and uniformity.

[0049] As a further aspect of the present invention: the three-dimensional mineral constraints specifically include:

[0050] Spatial constraints are established using the mining right boundary vector as a hard constraint, and a boundary determination benchmark is set. When the inverted area exceeds the boundary by more than or equal to 5 meters, it is determined to be mining beyond the boundary.

[0051] Depth constraints are calculated by combining the mining depth registered with the mining administration with the regional average mining depth. The inversion depth threshold is represented as H threshold = registered mining depth - 50M. Mining below this value is judged as ultra-deep mining.

[0052] Reserve constraints are set by establishing a monthly mining limit based on the conversion relationship between deformation volume and extractable reserves. If the limit is exceeded, it is considered over-extraction. The conversion relationship is represented as: Reserves = Deformation volume × Rock density × Recovery rate.

[0053] As a further embodiment of the present invention: the boundary crossing determination module includes:

[0054] The inversion unit is used to introduce mining constraint equations and optimize the traditional probability integral method to construct an objective function. The objective function is then solved using a gradient descent algorithm to invert the three-dimensional parameters of the goaf. The objective function is characterized as follows:

[0055] ,

[0056] in, The deformation value after fusion is P, where P represents the parameters of the goaf (length L, width W, depth H, and mining thickness M). Here, λ is the constraint function for mining regulations, and λ is the constraint weight (with a value of 0.6-0.8 to ensure the effectiveness of the constraint).

[0057] The boundary violation determination unit is used to compare the three-dimensional parameters obtained by inversion with the three-dimensional mining constraints. If any condition of the three-dimensional mining constraints is met, it is marked as an over-mining area.

[0058] Compared with existing technologies, the beneficial effects of this invention are as follows: By constructing a fully automated technology system of "data fusion-constraint inversion-closed-loop supervision," it achieves accurate dynamic inversion and off-site supervision of mine boundary violations, over-deep mining, and over-extraction behaviors in large-scale scenarios such as provinces. Compared with existing technologies, by deeply coupling mining administration data with InSAR time-series images, a direct correlation between deformation information and regulatory rules is established. Through a block-parallel computing framework, it can efficiently complete full-domain data processing, overcoming the bottleneck of small-scale point monitoring. At the same time, under the improved three-dimensional mining administration constraints, the accuracy error of the inversion model is significantly reduced. Furthermore, relying on mining administration constraints for automated inversion comparison, accurate identification of boundary violations and over-extraction behaviors is achieved, effectively avoiding non-illegal interference caused by legal mining and geological changes, and improving the credibility of supervision. Attached Figure Description

[0059] Figure 1 This is a flowchart of a dynamic inversion method for off-site monitoring of over-extraction in mines.

[0060] Figure 2 This is a gridded flowchart of a dynamic inversion method for off-site monitoring of over-extraction in mines.

[0061] Figure 3 This is a block diagram of a dynamic inversion system for off-site monitoring of over-extraction in mines. Detailed Implementation

[0062] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0063] The specific implementation of the present invention will be described in detail below with reference to specific embodiments.

[0064] like Figure 1 As shown, an embodiment of the present invention provides a dynamic inversion method for off-site monitoring of over-extraction in mines, comprising the following steps:

[0065] S10, acquire multi-source information data on mineral administration within the supervised area and perform standardized data processing to establish a standardized mineral administration database. The multi-source information data includes mineral administration data, InSAR time-series images, and auxiliary geological data.

[0066] S20, Based on the GIS spatial overlay algorithm, the mining administration data is mapped to the InSAR time series image to establish a spatiotemporal relationship, and dynamic weight optimization and allocation are performed through the random forest algorithm to achieve feature fusion and establish a fused feature field.

[0067] S30: Based on the distribution of supervised regions, the fusion feature field is divided into grid blocks. The fusion feature fields of multiple sub-regions are output through parallel processing of multiple grids and then stitched together to obtain the global fusion result.

[0068] S40, construct a three-dimensional mineral policy constraint based on space, depth and reserves, assign values ​​to the three-dimensional mineral policy constraint and perform parameter inversion calculation on the fused feature field to identify and mark the over-extraction area beyond the boundary;

[0069] S50 uses fuzzy field matching and logical judgment to exclude legal mining situations, generates visual warnings for illegal areas that are not excluded, and dynamically updates and monitors the process based on time-series iteration.

[0070] This embodiment presents a dynamic inversion method for off-site monitoring of mine over-extraction and boundary violations. By constructing a fully automated technical system of "data fusion - constraint inversion - closed-loop monitoring," it achieves accurate dynamic inversion and off-site monitoring of mine over-extraction and boundary violations in large-scale scenarios such as provincial areas. First, it collaboratively acquires large-scale mine administration data, InSAR time-series imagery, and auxiliary geological data. Using a three-in-one approach of "platform docking + satellite download + database retrieval," it automatically synchronizes real-time data from the provincial mine administration platform. It also batch downloads Sentinel-1 imagery through an open platform and retrieves structured data from the geological survey database to ensure data coverage. The system covers the entire province without blind spots, avoiding the limitations of point-based monitoring. It also standardizes and preprocesses multi-source heterogeneous data. Besides denoising operations commonly used in big data training, this standardization preprocessing includes: Standardization of mining administration data: Using the 2000 National Geodetic Coordinate System as a benchmark, coordinate transformation and topological correction are performed on mining right boundary vector data to remove duplicate and invalid boundaries; core fields such as permitted mining depth and registered reserves are extracted, quantified into numerical parameters, and a standardized mining administration database is constructed in Shapefile+MySQL format, supporting spatial queries; InSAR data temporal preprocessing: SBAS-InSAR technology is used to set spatiotemporal parameters. Baseline thresholds (spatial baseline ≤ 100m, temporal baseline ≤ 30 days) were used to register SAR images, generate interferograms, unwrap phases, and filter them. This corrected for incoherence issues caused by vegetation cover and atmospheric delay, generating monthly deformation rate maps and cumulative deformation fields. The output was 10m×10m spatial resolution raster data in GeoTIFF format. Lightweight geological data filtering was performed: redundant data with low correlation to surface deformation inversion were removed, retaining only core parameters such as rock hardness (quantized to 1-10 levels) and regional average mining depth. An auxiliary parameter set was constructed to reduce redundancy in large-scale data processing. Secondly, the mining rights vector data was integrated with the InSAR deformation raster. The system performs spatiotemporal correlation and feature-level deep fusion of grid data to establish a dynamic weighted fusion model. Next, it employs a block-parallel computing architecture to efficiently process massive fusion data at the provincial level. Then, it introduces mining right boundaries, permit depth, and registered reserves to construct three-dimensional mining administration hard constraints, driving an improved probabilistic integral method to achieve high-precision inversion of compliance-oriented goaf parameters. Finally, it automatically eliminates interference from legal mining through a rule-based automated verification algorithm, accurately identifies suspected over-mining areas, and generates an integrated visual report. Based on a monthly time-series iteration mechanism, it completes a closed-loop off-site supervision process covering the entire lifecycle from deformation monitoring, violation inversion, early warning output to dynamic updates.

[0071] In another preferred embodiment of the present invention, the mineral administration data includes mineral ownership data, mineral rights boundary vectors, depth vector data, and registered reserve data. The step of implementing feature fusion specifically includes:

[0072] Based on the GIS spatial overlay algorithm, the mining right boundary vector and InSAR time series image are matched pixel by pixel through intersection analysis and nearest neighbor interpolation to establish a one-to-one spatial mapping relationship between deformed pixels and mining right units.

[0073] A feature-level dynamic weight fusion model is constructed, and a random forest algorithm is used for 10-fold cross-validation to optimize multiple weight coefficients. These weight coefficients include mineral administration rule parameters, InSAR deformation values, and geological parameters. The sum of all weight coefficients is 1, and the mineral administration rule parameters are not less than 0.4. The dynamic weight fusion model is characterized as follows:

[0074] ,

[0075] in, Let be the InSAR deformation value at coordinates (x, y) at time t. These are parameters for mining regulations (quantified boundary distance and permissible depth threshold). For geological parameters, α, β, and γ are the corresponding weights (with values ​​ranging from 0 to 1, and α + β + γ = 1).

[0076] In this embodiment, addressing the core shortcomings of existing technologies such as "insufficient data fusion and lack of correlation with mining regulations," a data correspondence relationship is established through spatial registration, and a dynamic weighted fusion model is constructed to achieve deep coupling between mining regulatory rules and InSAR deformation data. Based on the GIS spatial overlay analysis algorithm, the mining right boundary vector data and InSAR deformation raster data are matched pixel by pixel. The "intersection analysis + nearest neighbor interpolation" method is used to establish a one-to-one mapping relationship between "each deformation pixel and the corresponding mining right unit," clarifying the mining right subject and permitted scope of the deformation area and eliminating correlation errors caused by spatial misalignment.

[0077] like Figure 2 As shown, in another preferred embodiment of the present invention, the step of dividing the fusion feature field into grid blocks based on the distribution of supervised regions, outputting the fusion feature fields of multiple sub-regions through multi-grid parallel processing, and stitching them together to obtain the global fusion result specifically includes:

[0078] S31, the fused feature field is divided into regions based on a preset unit boundary grid to obtain several independent sub-regions. Specifically, the preset unit boundary grid is 10KM*10KM.

[0079] S32, based on a parallel computing framework, synchronously performs dynamic weight fusion operations on data from several sub-regions to obtain the fusion feature field of each sub-region;

[0080] S33 automatically performs edge stitching and consistency verification on the fusion feature field of each sub-region to obtain the global fusion result and ensure spatial continuity and uniformity.

[0081] In this embodiment, the problem of balancing efficiency and accuracy in large-scale data processing is solved by block parallel processing. The provincial data is divided into sub-regions of 10km×10km. A GPU parallel computing framework is used to perform fusion operations synchronously on each sub-region. After processing, edge stitching and consistency verification are performed to ensure that the fusion results across the entire region are seamless and have uniform accuracy, thus compressing the provincial data processing time to within 24 hours.

[0082] In another preferred embodiment of the present invention, the three-dimensional mineral constraints specifically include:

[0083] Spatial constraints are established using the mining right boundary vector as a hard constraint, and a boundary determination benchmark is set. When the inverted area exceeds the boundary by more than or equal to 5 meters, it is determined to be mining beyond the boundary.

[0084] Depth constraints are calculated by combining the mining depth registered with the mining administration with the regional average mining depth. The inversion depth threshold is represented as H threshold = registered mining depth - 50M. Mining below this value is judged as ultra-deep mining.

[0085] Reserve constraints are set by establishing a monthly mining limit based on the conversion relationship between deformation volume and extractable reserves. If the limit is exceeded, it is considered over-extraction. The conversion relationship is represented as: Reserves = Deformation volume × Rock density × Recovery rate.

[0086] Furthermore, the step of performing parameter inversion calculations on the fused feature field to identify and mark the over-extraction areas specifically includes:

[0087] Mining policy constraint equations are introduced and the traditional probability integral method is optimized to construct an objective function. The objective function is then solved using the gradient descent algorithm to invert the three-dimensional parameters of the goaf. The objective function is characterized as follows:

[0088] ,

[0089] in, The deformation value after fusion is P, where P represents the parameters of the goaf (length L, width W, depth H, and mining thickness M). Here, λ is the constraint function for mining regulations, and λ is the constraint weight (with a value of 0.6-0.8 to ensure the effectiveness of the constraint).

[0090] The three-dimensional parameters obtained by inversion are compared with the three-dimensional mining constraints. If any condition of the three-dimensional mining constraints is met, it is marked as an over-mining area.

[0091] In this embodiment, in response to the shortcomings of "low inversion accuracy, lack of mining regulations constraints, and inability to distinguish deformation causes", the traditional probability integral method is improved by using mining regulations as hard constraints, realizing the transformation from "simple deformation inversion" to "compliance-oriented inversion", accurately locating over-mining behavior, while ensuring the accuracy of inversion over a wide range.

[0092] Furthermore, step S50 addresses the shortcomings of existing technologies, such as "lack of regulatory implementation and no complete supervisory loop." Based on a technical chain of data verification, result output, and dynamic iteration, it constructs a fully automated closed loop of "inversion-verification-early warning-update." Through algorithmic rule verification and data time-series iteration, it avoids reliance on secondary manual assessment, realizing the transformation of technical monitoring into compliance judgment and adapting to the automation needs of large-scale off-site supervision. By calling standardized real-time mining administration ledger data, it constructs a compliance verification algorithm model, using mining rights change records and temporary mining permits... Fuzzy matching and logical judgment of fields such as mining rights boundary fine-tuning and emergency mining are used to automatically exclude legitimate adjustment situations, output a dataset of suspected violations and verification logs, and ensure that the verification process is free of human intervention. Based on timed task scheduling technology, a monthly fixed update cycle is set to automatically trigger the batch download of Sentinel-1 image data and the synchronization of mining administration ledger data. The core algorithm modules of steps 1-3 are called to realize the automatic recalculation of the whole process, complete the time-series iteration of inversion results and early warning information, and build a full life cycle data recording system for violation areas from identification and early warning to rectification tracking.

[0093] like Figure 3 As shown, the present invention also provides a dynamic inversion system for off-site monitoring of over-extraction in mines, comprising:

[0094] The data acquisition module 100 is used to acquire multi-source information data on mineral administration within the supervised area and perform standardized data processing to establish a standardized mineral administration database. The multi-source information data includes mineral administration data, InSAR time-series images, and auxiliary geological data.

[0095] The spatiotemporal correlation module 200 is used to map the mining administration data to the InSAR time series image based on the GIS spatial overlay algorithm to establish spatiotemporal correlation, and to perform dynamic weight optimization and allocation through the random forest algorithm to achieve feature fusion and establish a fused feature field.

[0096] The gridding module 300 is used to divide the fusion feature field into grid blocks based on the distribution of the supervised region. It outputs the fusion feature fields of multiple sub-regions through parallel processing of multiple grids and stitches them together to obtain the global fusion result.

[0097] The boundary violation determination module 400 is used to construct three-dimensional mineral policy constraints based on space, depth and reserves, assign values ​​to the three-dimensional mineral policy constraints and perform parameter inversion calculations on the fused feature field to identify and mark the boundary violation and over-mining areas.

[0098] The verification output module 500 is used to exclude legal mining situations through fuzzy field matching and logical judgment, generate visual warnings for illegal areas that are not excluded, and perform dynamic updates and closed-loop monitoring of the process based on time-series iteration.

[0099] In another preferred embodiment of the present invention, the mineral administration data includes mineral ownership data, mineral rights boundary vectors, depth vector data, and registered reserve data, and the spatiotemporal correlation module includes:

[0100] The spatial mapping unit is used to accurately match the mining right boundary vector with the InSAR time series image pixel by pixel based on the GIS spatial overlay algorithm through intersection analysis and nearest neighbor interpolation, and establish a one-to-one spatial mapping relationship between deformed pixels and mining right units.

[0101] The weight optimization unit is used to construct a feature-level dynamic weight fusion model. It employs a random forest algorithm for 10-fold cross-validation to optimize multiple weight coefficients. These weight coefficients include mineral administration rule parameters, InSAR deformation values, and geological parameters. The sum of all weight coefficients is 1, and the mineral administration rule parameters are not less than 0.4. The dynamic weight fusion model is characterized as follows:

[0102] ,

[0103] in, Let be the InSAR deformation value at coordinates (x, y) at time t. These are parameters for mining regulations (quantified boundary distance and permissible depth threshold). For geological parameters, α, β, and γ are the corresponding weights (with values ​​ranging from 0 to 1, and α + β + γ = 1).

[0104] In another preferred embodiment of the present invention, the meshing module includes:

[0105] The region segmentation unit is used to segment the fused feature field into several independent sub-regions based on a preset unit boundary grid. Specifically, the unit boundary grid is preset to be 10KM*10KM.

[0106] The parallel computing unit is used to synchronously perform dynamic weight fusion operations on data from several sub-regions based on a parallel computing framework, so as to obtain the fusion feature field of each sub-region;

[0107] The fusion verification unit is used to automatically perform edge stitching and consistency verification on the fusion feature field of each sub-region to obtain the global fusion result and ensure spatial continuity and uniformity.

[0108] In another preferred embodiment of the present invention, the three-dimensional mineral constraints specifically include:

[0109] Spatial constraints are established using the mining right boundary vector as a hard constraint, and a boundary determination benchmark is set. When the inverted area exceeds the boundary by more than or equal to 5 meters, it is determined to be mining beyond the boundary.

[0110] Depth constraints are calculated by combining the mining depth registered with the mining administration with the regional average mining depth. The inversion depth threshold is represented as H threshold = registered mining depth - 50M. Mining below this value is judged as ultra-deep mining.

[0111] Reserve constraints are set by establishing a monthly mining limit based on the conversion relationship between deformation volume and extractable reserves. If the limit is exceeded, it is considered over-extraction. The conversion relationship is represented as: Reserves = Deformation volume × Rock density × Recovery rate.

[0112] In another preferred embodiment of the present invention, the boundary crossing determination module includes:

[0113] The inversion unit is used to introduce mining constraint equations and optimize the traditional probability integral method to construct an objective function. The objective function is then solved using a gradient descent algorithm to invert the three-dimensional parameters of the goaf. The objective function is characterized as follows:

[0114] ,

[0115] in, The deformation value after fusion is P, where P represents the parameters of the goaf (length L, width W, depth H, and mining thickness M). Here, λ is the constraint function for mining regulations, and λ is the constraint weight (with a value of 0.6-0.8 to ensure the effectiveness of the constraint).

[0116] The boundary violation determination unit is used to compare the three-dimensional parameters obtained by inversion with the three-dimensional mining constraints. If any condition of the three-dimensional mining constraints is met, it is marked as an over-mining area.

[0117] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0118] Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the disclosure in the specification and embodiments. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The specification and embodiments are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the claims.

[0119] It should be understood that this disclosure is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this disclosure is limited only by the appended claims.

Claims

1. A dynamic inversion method for off-site monitoring of over-mining in mines, characterized in that, Include: Acquire multi-source information data on mineral administration within the supervised area and perform standardized data processing to establish a standardized mineral administration database. The multi-source information data includes mineral administration data, InSAR time-series images, and auxiliary geological data. Based on the GIS spatial overlay algorithm, the mining administration data is mapped to the InSAR time series image to establish a spatiotemporal relationship, and dynamic weight optimization and allocation are performed through the random forest algorithm to achieve feature fusion and establish a fused feature field. Based on the distribution of supervised regions, the fusion feature field is divided into grid blocks. The fusion feature fields of multiple sub-regions are output through multi-grid parallel processing and then stitched together to obtain the global fusion result. A three-dimensional mineral policy constraint based on space, depth and reserves is constructed. The three-dimensional mineral policy constraint is assigned values ​​and the parameter inversion operation is performed on the fused feature field to identify and mark the over-extraction area. The system excludes legitimate mining activities through fuzzy field matching and logical judgment, generates visual warnings for non-excluded illegal areas, and dynamically updates and monitors the process based on time-series iteration.

2. The dynamic inversion method for off-site monitoring of over-mining in mines according to claim 1, characterized in that, The mineral administration data includes mineral ownership data, mineral right boundary vectors, depth vector data, and registered reserve data. The steps for achieving feature fusion specifically include: Based on the GIS spatial overlay algorithm, the mining right boundary vector and InSAR time series image are matched pixel by pixel through intersection analysis and nearest neighbor interpolation to establish a one-to-one spatial mapping relationship between deformed pixels and mining right units. A feature-level dynamic weight fusion model is constructed, and a random forest algorithm is used for 10-fold cross-validation to optimize multiple weight coefficients. These weight coefficients include mineral administration rule parameters, InSAR deformation values, and geological parameters. The sum of all weight coefficients is 1, and the mineral administration rule parameters are not less than 0.

4. The dynamic weight fusion model is characterized as follows: ， in, Let be the InSAR deformation value at coordinates (x, y) at time t. These are parameters for mining regulations (quantified boundary distance and permissible depth threshold). For geological parameters, α, β, and γ are the corresponding weights (with values ​​ranging from 0 to 1, and α + β + γ = 1).

3. The dynamic inversion method for off-site monitoring of over-mining in mines according to claim 2, characterized in that, The steps of dividing the fusion feature field into grid blocks based on the distribution of supervised regions, outputting the fusion feature fields of multiple sub-regions through multi-grid parallel processing, and stitching them together to obtain the global fusion result specifically include: The fused feature field is divided into several independent sub-regions based on a preset unit boundary grid. Specifically, the preset unit boundary grid is 10KM*10KM. Based on a parallel computing framework, dynamic weight fusion operations are performed synchronously on data from several sub-regions to obtain the fusion feature field of each sub-region. Edge stitching and consistency verification are automatically performed on the fusion feature fields of each sub-region to obtain the global fusion result and ensure spatial continuity and uniformity.

4. The dynamic inversion method for off-site monitoring of over-mining in mines according to claim 1, characterized in that, The three-dimensional mineral resource constraints specifically include: Spatial constraints are established using the mining right boundary vector as a hard constraint, and a boundary determination benchmark is set. When the inverted area exceeds the boundary by more than or equal to 5 meters, it is determined to be mining beyond the boundary. Depth constraints are calculated by combining the mining depth registered with the mining administration with the regional average mining depth. The inversion depth threshold is represented as H threshold = registered mining depth - 50M. Mining below this value is judged as ultra-deep mining. Reserve constraints are set by establishing a monthly mining limit based on the conversion relationship between deformation volume and extractable reserves. If the limit is exceeded, it is considered over-extraction. The conversion relationship is represented as: Reserves = Deformation volume × Rock density × Recovery rate.

5. The dynamic inversion method for off-site monitoring of over-mining in mines according to claim 4, characterized in that, The steps of performing parameter inversion calculations on the fused feature field to identify and mark over-extraction areas specifically include: Mining policy constraint equations are introduced and the traditional probability integral method is optimized to construct an objective function. The objective function is then solved using the gradient descent algorithm to invert the three-dimensional parameters of the goaf. The objective function is characterized as follows: ， in, The deformation value after fusion is P, where P represents the parameters of the goaf (length L, width W, depth H, and mining thickness M). Here, λ is the constraint function for mining regulations, and λ is the constraint weight (with a value of 0.6-0.8 to ensure the effectiveness of the constraint). The three-dimensional parameters obtained by inversion are compared with the three-dimensional mining constraints. If any condition of the three-dimensional mining constraints is met, it is marked as an over-mining area.

6. A dynamic inversion system for off-site monitoring of over-extraction in mines, characterized in that, Include: The data acquisition module is used to acquire multi-source information data on mineral administration within the supervised area and perform standardized data processing to establish a standardized mineral administration database. The multi-source information data includes mineral administration data, InSAR time-series images, and auxiliary geological data. The spatiotemporal correlation module is used to map the mining administration data to the InSAR time series image based on the GIS spatial overlay algorithm to establish spatiotemporal correlation, and to perform dynamic weight optimization and allocation through the random forest algorithm to achieve feature fusion and establish a fused feature field. The gridding module is used to divide the fusion feature field into grid blocks based on the distribution of the supervised region. It outputs the fusion feature fields of multiple sub-regions through parallel processing of multiple grids and then stitches them together to obtain the global fusion result. The boundary violation determination module is used to construct three-dimensional mineral policy constraints based on space, depth and reserves, assign values ​​to the three-dimensional mineral policy constraints and perform parameter inversion calculations on the fused feature field to identify and mark the boundary violation and over-mining areas. The verification output module is used to exclude legal mining situations through fuzzy field matching and logical judgment, generate visual warnings for illegal areas that are not excluded, and dynamically update and monitor the process based on time-series iteration.

7. A dynamic inversion system for off-site monitoring of over-extraction in mines according to claim 6, characterized in that, The mineral administration data includes mineral ownership data, mineral right boundary vectors, depth vector data, and registered reserve data. The spatiotemporal correlation module includes: The spatial mapping unit is used to accurately match the mining right boundary vector with the InSAR time series image pixel by pixel based on the GIS spatial overlay algorithm through intersection analysis and nearest neighbor interpolation, and establish a one-to-one spatial mapping relationship between deformed pixels and mining right units. The weight optimization unit is used to construct a feature-level dynamic weight fusion model. It employs a random forest algorithm for 10-fold cross-validation to optimize multiple weight coefficients. These weight coefficients include mineral administration rule parameters, InSAR deformation values, and geological parameters. The sum of all weight coefficients is 1, and the mineral administration rule parameters are not less than 0.

4. The dynamic weight fusion model is characterized as follows: ， in, Let be the InSAR deformation value at coordinates (x, y) at time t. These are parameters for mining regulations (quantified boundary distance and permissible depth threshold). For geological parameters, α, β, and γ are the corresponding weights (with values ​​ranging from 0 to 1, and α + β + γ = 1).

8. A dynamic inversion system for off-site monitoring of over-mining in mines according to claim 7, characterized in that, The meshing module includes: The region segmentation unit is used to segment the fused feature field into several independent sub-regions based on a preset unit boundary grid. Specifically, the unit boundary grid is preset to be 10KM*10KM. The parallel computing unit is used to synchronously perform dynamic weight fusion operations on data from several sub-regions based on a parallel computing framework, so as to obtain the fusion feature field of each sub-region; The fusion verification unit is used to automatically perform edge stitching and consistency verification on the fusion feature field of each sub-region to obtain the global fusion result and ensure spatial continuity and uniformity.

9. A dynamic inversion system for off-site monitoring of over-extraction in mines according to claim 6, characterized in that, The three-dimensional mineral resource constraints specifically include: Spatial constraints are established using the mining right boundary vector as a hard constraint, and a boundary determination benchmark is set. When the inverted area exceeds the boundary by more than or equal to 5 meters, it is determined to be mining beyond the boundary. Depth constraints are calculated by combining the mining depth registered with the mining administration with the regional average mining depth. The inversion depth threshold is represented as H threshold = registered mining depth - 50M. Mining below this value is judged as ultra-deep mining. Reserve constraints are set by establishing a monthly mining limit based on the conversion relationship between deformation volume and extractable reserves. If the limit is exceeded, it is considered over-extraction. The conversion relationship is represented as: Reserves = Deformation volume × Rock density × Recovery rate.

10. A dynamic inversion system for off-site monitoring of over-extraction in mines according to claim 9, characterized in that, The boundary crossing determination module includes: The inversion unit is used to introduce mining constraint equations and optimize the traditional probability integral method to construct an objective function. The objective function is then solved using a gradient descent algorithm to invert the three-dimensional parameters of the goaf. The objective function is characterized as follows: ， in, The deformation value after fusion is P, where P represents the parameters of the goaf (length L, width W, depth H, and mining thickness M). Here, λ is the constraint function for mining regulations, and λ is the constraint weight (with a value of 0.6-0.8 to ensure the effectiveness of the constraint). The boundary violation determination unit is used to compare the three-dimensional parameters obtained by inversion with the three-dimensional mining constraints. If any condition of the three-dimensional mining constraints is met, it is marked as an over-mining area.

Images