Non-coal mine area risk monitoring method and system based on gan and satellite image

Abstract

The application provides a non-coal mine area risk monitoring method and system based on GAN and satellite images, wherein the method comprises the following steps: accessing a satellite image data source, performing GAN expansion on satellite image data of a non-coal mine area to obtain target image data; performing image processing on the target image data to obtain an area monitoring data set; establishing a non-coal mine area database system according to the area monitoring data set; and performing real-time AI analysis according to the non-coal mine area database system to obtain a risk monitoring result of the non-coal mine area. The non-coal mine area risk monitoring method and system based on GAN and satellite images can realize real-time risk monitoring of the non-coal mine area through multi-source satellite data fusion, AI enhanced processing and intelligent analysis, and significantly improve the efficiency and response speed of non-coal mine risk monitoring.

Description

Technical Field

[0001] This invention relates to the field of intelligent safety monitoring technology, and in particular to a method and system for risk monitoring in non-coal mining areas based on GAN and satellite imagery. Background Technology

[0002] Non-coal mining areas mainly include open-pit mines, underground mines, and their corresponding supporting facilities. The risks involved involve multiple dimensions, including safety risks (such as landslides at spoil heaps) and environmental risks (such as water and soil pollution and vegetation destruction). In order to monitor and provide early warnings for these risks, it is necessary to combine multi-source data with intelligent analysis technology to identify and manage risks in advance in order to reduce losses.

[0003] Satellite imagery data is one of the aforementioned multi-source data sources. However, after acquiring satellite imagery data, existing technologies typically only perform routine image preprocessing before using it for subsequent risk factor extraction. If the image quality is low (e.g., cloud cover in the satellite imagery data), the efficiency of subsequent risk factor extraction will be greatly reduced. Furthermore, the efficiency and response speed of risk monitoring will also be reduced.

[0004] In view of this, there is an urgent need for risk monitoring methods and systems for non-coal mining areas based on GANs and satellite imagery, in order to at least address the above-mentioned shortcomings. Summary of the Invention

[0005] One of the objectives of this invention is to provide a method and system for risk monitoring of non-coal mining areas based on GAN and satellite imagery. By fusing multi-source satellite data, AI-enhanced processing, and intelligent analysis, the method enables real-time risk monitoring of non-coal mining areas, significantly improving the efficiency and response speed of risk monitoring in non-coal mines.

[0006] The risk monitoring method for non-coal mining areas based on GAN and satellite imagery provided in this invention includes:

[0007] By accessing satellite imagery data sources, GAN extension is performed on satellite imagery data of non-coal mining areas to obtain target imagery data.

[0008] Image processing is performed on the target image data to obtain a regional monitoring dataset;

[0009] Establish a non-coal mine regional database system based on regional monitoring datasets;

[0010] Based on the non-coal mining area database system, AI analysis is performed in real time to obtain risk monitoring results for non-coal mining areas.

[0011] Preferably, satellite imagery data is accessed, and GAN extension is performed on satellite imagery data of non-coal mining areas to obtain target imagery data, including:

[0012] Access to Gaofen-1B, C, D satellites and Gaofen-6E, F, G satellites to acquire satellite imagery data of non-coal mining areas;

[0013] Conditional GANs are used to supplement satellite imagery data to obtain target imagery data.

[0014] Preferably, image processing is performed on the target image data to obtain a regional monitoring dataset, including:

[0015] Enhancement processing is performed on the target image data;

[0016] After using Faster R-CNN target detection technology and U-Net model to perform target detection and segmentation on the enhanced target image data, parameter extraction and analysis are then performed to obtain a regional monitoring dataset.

[0017] Preferably, the target image data is enhanced, including:

[0018] Based on the target image data, obtain the input data, which includes: 1m resolution panchromatic image, 4m resolution multispectral image, DEM data, orbital parameters or RPC parameters;

[0019] Radiometric calibration, FLAASH atmospheric correction, and orthorectification are performed on multispectral images;

[0020] Orthorectification is performed on the panchromatic image; the orthorectification process combines DEM data and RPC parameters.

[0021] After both the multispectral and panchromatic images have been processed, they are fused and enhanced to obtain the pre-output enhanced image.

[0022] Use the pre-output enhanced image that has passed the quality check as the enhanced image.

[0023] Preferably, a non-coal mine regional database system is established based on the regional monitoring dataset, including:

[0024] The regional monitoring dataset is stored in a database system that combines ArcGIS FileGeodatabase, Personal Geodatabase, and file system, which is built using the ArcGeodatabase object relational database, to obtain a non-coal mine regional database system.

[0025] Preferably, based on the non-coal mining area database system, real-time AI analysis is performed to obtain risk monitoring results for non-coal mining areas, including:

[0026] Based on the vector change analysis method and the extreme learning machine algorithm, and using a database system of non-coal mining areas, AI analysis is performed in real time to obtain risk monitoring results for non-coal mining areas.

[0027] The risk monitoring method for non-coal mining areas based on GAN and satellite imagery provided in this invention also includes:

[0028] Analyze risk monitoring results to identify local risk areas;

[0029] Feature extraction is performed on local risk areas to obtain a regional feature set; the regional features include: regional satellite imagery data, risk type, and risk level.

[0030] Obtain the matching template for the deduction basis corresponding to the portfolio risk deduction scenario;

[0031] Based on the matching template and regional feature set of the deduction basis, the deduction basis corresponding to the combined risk is determined;

[0032] Based on the pre-set risk values ​​corresponding to the portfolio risks, from largest to smallest, and the deduction basis, the portfolio risks are deduced to obtain new risk areas;

[0033] The risk monitoring results will be updated based on newly added risk areas and risk combinations.

[0034] The risk monitoring method for non-coal mining areas based on GAN and satellite imagery provided in this invention also includes:

[0035] Identify the target risks associated with the portfolio risks being simulated;

[0036] If the risk levels of the target risks are all less than the preset risk level thresholds, and the risk value of the combined risks is greater than or equal to the preset risk value thresholds, obtain the first inference visualization model of the target risks and the second inference visualization model of the combined risks.

[0037] Obtain the preset simulation timeline;

[0038] The first simulation visualization models are set on the simulation timeline in ascending order of their corresponding risk levels. The earliest simulation visualization model starts its simulation at the starting time of the simulation timeline. The simulation duration of the first simulation visualization model is the first duration preset for the risk level corresponding to the first simulation visualization model. The timeline distance between the first simulation visualization models is less than the target axis distance.

[0039] Once the first simulation visualization model is fully set up, the second simulation visualization model is set up. The simulation duration of the second simulation visualization model is the second duration preset for the risk value corresponding to the second simulation visualization model. The time axis distance between the second simulation visualization model and the first simulation visualization model before it is less than the target axis distance.

[0040] The simulation is presented to administrators based on the established simulation timeline.

[0041] Preferred, the first inductive visualization model for obtaining target risk includes:

[0042] A basic simulation and visualization model for obtaining target risks;

[0043] The associated twin engine that determines the basis for the simulation of portfolio risks;

[0044] The associated twin engine is highlighted in the basic inference visualization model, and combined risk knowledge is searched according to the inference basis corresponding to the associated twin engine. The combined risk knowledge is marked in the preset range of the associated twin engine.

[0045] The non-coal mine area risk monitoring system based on GAN and satellite imagery provided in this embodiment of the invention includes:

[0046] The data access module is used to access satellite image data sources and perform GAN extension on satellite image data of non-coal mining areas to obtain target image data.

[0047] The image processing module is used to process the target image data to obtain the regional monitoring dataset;

[0048] The database construction module is used to establish a non-coal mine regional database system based on regional monitoring datasets.

[0049] The risk monitoring module is used to obtain risk monitoring results for non-coal mining areas in real time through AI analysis based on the non-coal mining area database system.

[0050] The beneficial effects of this invention are as follows:

[0051] This invention enables real-time risk monitoring of non-coal mining areas through multi-source satellite data fusion, AI-enhanced processing, and intelligent analysis, significantly improving the efficiency and response speed of risk monitoring in non-coal mines.

[0052] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in this application.

[0053] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0054] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0055] Figure 1 This is a schematic diagram of a risk monitoring method for non-coal mining areas based on GAN and satellite imagery in an embodiment of the present invention;

[0056] Figure 2 This is a schematic diagram of a risk monitoring system for non-coal mining areas based on GAN and satellite imagery, as described in an embodiment of the present invention. Detailed Implementation

[0057] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0058] This invention provides a method for risk monitoring in non-coal mining areas based on GAN and satellite imagery, such as... Figure 1 As shown, it includes:

[0059] By accessing satellite imagery data sources, GAN extension is performed on satellite imagery data of non-coal mining areas to obtain target imagery data.

[0060] This involves accessing satellite imagery data sources and performing GAN extensions on satellite imagery data from non-coal mining areas to obtain target imagery data, including:

[0061] Access to Gaofen-1B, C, D satellites and Gaofen-6E, F, G satellites to acquire satellite imagery data of non-coal mining areas;

[0062] Conditional GANs are used to supplement satellite imagery data to obtain target imagery data.

[0063] Image processing is performed on the target image data to obtain a regional monitoring dataset;

[0064] Image processing is performed on the target image data to obtain a regional monitoring dataset, including:

[0065] Enhancement processing is performed on the target image data;

[0066] After using Faster R-CNN target detection technology and U-Net model to perform target detection and segmentation on the enhanced target image data, parameter extraction and analysis are then performed to obtain a regional monitoring dataset.

[0067] Establish a non-coal mine regional database system based on regional monitoring datasets;

[0068] This includes establishing a non-coal mine regional database system based on regional monitoring datasets, including:

[0069] The regional monitoring dataset is stored in a database system that combines ArcGIS FileGeodatabase, Personal Geodatabase, and file system, which is built using the ArcGeodatabase object relational database, to obtain a non-coal mine regional database system;

[0070] Based on the non-coal mining area database system, AI analysis is performed in real time to obtain risk monitoring results for non-coal mining areas;

[0071] Among these measures, real-time AI analysis is conducted based on the non-coal mining area database system to obtain risk monitoring results for non-coal mining areas, including:

[0072] Based on the vector change analysis method and the extreme learning machine algorithm, and using a database system of non-coal mining areas, AI analysis is performed in real time to obtain risk monitoring results for non-coal mining areas.

[0073] The working principle and beneficial effects of the above technical solution are as follows:

[0074] By accessing Gaofen-1B, C, and D satellites and Gaofen-6E, F, and G satellites, satellite imagery data covering non-coal mining areas is acquired. For scenarios with missing data (e.g., images obscured by clouds), conditional GANs are used for supplementation. These conditional GANs generate new satellite images using a generative adversarial network (GAN) and additional conditional information (e.g., cloud removal), supplementing the existing satellite imagery data to obtain target imagery data. The target imagery data undergoes enhancement processing to improve image quality. Then, target detection is performed using Faster R-CNN to obtain target regions (e.g., detecting mining areas and spoil heaps in non-coal mining areas, outputting region bounding box coordinates). The target regions are segmented using a U-Net model, followed by parameter extraction and analysis (e.g., pixel-level segmentation of tailings dam images to extract key parameters such as rainfall, reservoir water level, seepage line, dam surface displacement, internal displacement, and deformation displacement), resulting in a regional monitoring dataset. This invention integrates ArcGIS Geodatabase, ArcGIS File Geodatabase, and Personal Geodatabase to form a hybrid storage architecture and store regional monitoring datasets, thus obtaining a database system for non-coal mining areas. Based on vector change analysis, it detects and identifies anomalous change areas (collapse, displacement) in the land surface, using these as input features for an Extreme Learning Machine (ELM). The ELM outputs the risk level of the changed areas. This invention utilizes multi-source satellite data fusion, AI-enhanced processing, and intelligent analysis to conduct real-time risk monitoring of non-coal mining areas, significantly improving the efficiency and response speed of risk monitoring in non-coal mining areas.

[0075] In one embodiment, enhancing the target image data includes:

[0076] Based on the target image data, obtain the input data, which includes: 1m resolution panchromatic image, 4m resolution multispectral image, DEM data, orbital parameters or RPC parameters;

[0077] Radiometric calibration, FLAASH atmospheric correction, and orthorectification are performed on multispectral images;

[0078] Orthorectification is performed on the panchromatic image; the orthorectification process combines DEM data and RPC parameters.

[0079] After both the multispectral and panchromatic images have been processed, they are fused and enhanced to obtain the pre-output enhanced image.

[0080] Use the pre-output enhanced image that has passed the quality check as the enhanced image.

[0081] The working principle and beneficial effects of the above technical solution are as follows:

[0082] This invention fuses and enhances panchromatic and multispectral full-view images from Gaofen series satellites to generate images with both high resolution and multispectral characteristics. Standardized processing is carried out from radiometric calibration to orthorectification and then to fusion enhancement, which improves the efficiency of automated process for single-view image processing. Differential image analysis further enhances the detection accuracy of change areas.

[0083] In one embodiment, it also includes:

[0084] Analyze risk monitoring results to identify local risk areas;

[0085] Feature extraction is performed on local risk areas to obtain a regional feature set; the regional features include: regional satellite imagery data, risk type, and risk level.

[0086] Obtain the matching template for the deduction basis corresponding to the portfolio risk deduction scenario;

[0087] Based on the matching template and regional feature set of the deduction basis, the deduction basis corresponding to the combined risk is determined;

[0088] Based on the pre-set risk values ​​corresponding to the portfolio risks, from largest to smallest, and the deduction basis, the portfolio risks are deduced to obtain new risk areas;

[0089] The risk monitoring results will be updated based on newly added risk areas and risk combinations.

[0090] The working principle and beneficial effects of the above technical solution are as follows:

[0091] Risk monitoring results may contain multiple risk types. Besides single risks (such as slope instability, mine water accumulation, and goaf collapse), different risks may constitute combined risks (for example, when slope instability and mine water accumulation coexist, slope stability is further reduced, easily leading to slope collapse; simultaneously, the water accumulation in the mine pit increases, potentially causing a mine flooding accident). Local risk areas are the non-coal mine areas with risks monitored in satellite imagery. Combined risk extrapolation scenarios involve extrapolating from the individual risks in a combination to obtain further possible risk outcomes. For example, extrapolating the risk range and scale when slope instability and mine water accumulation coexist. The extrapolation basis matching template is a template used to match the basis for the corresponding combined risk extrapolation, such as the distance between the slope instability area and the mine water accumulation area, and their respective risk scales. Based on the matching template of the inference basis, the inference basis corresponding to the combined risks is selected from the set of regional features. Then, according to the pre-set risk values ​​of the combined risks in descending order, the combined risks are inferred based on the inference basis to obtain new risk areas. These new risk areas are those that cannot be directly obtained when monitoring single risks corresponding to combined risks. Based on the new risk areas and combined risks, the risk monitoring results are updated, improving the comprehensiveness of risk monitoring.

[0092] In one embodiment, combined risk is deduced based on the order of preset risk values ​​corresponding to the combined risks from largest to smallest and the deduction basis to obtain newly added risk areas, including:

[0093] Identify the target risks associated with the portfolio risks being simulated;

[0094] If the risk levels of the target risks are all less than the preset risk level thresholds, and the risk value of the combined risks is greater than or equal to the preset risk value thresholds, obtain the first inference visualization model of the target risks and the second inference visualization model of the combined risks.

[0095] Obtain the preset simulation timeline;

[0096] The first simulation visualization models are set on the simulation timeline in ascending order of their corresponding risk levels. The earliest simulation visualization model starts its simulation at the starting time of the simulation timeline. The simulation duration of the first simulation visualization model is the first duration preset for the risk level corresponding to the first simulation visualization model. The timeline distance between the first simulation visualization models is less than the target axis distance.

[0097] Once the first simulation visualization model is fully set up, the second simulation visualization model is set up. The simulation duration of the second simulation visualization model is the second duration preset for the risk value corresponding to the second simulation visualization model. The time axis distance between the second simulation visualization model and the first simulation visualization model before it is less than the target axis distance.

[0098] The simulation is presented to administrators based on the established simulation timeline.

[0099] The working principle and beneficial effects of the above technical solution are as follows:

[0100] The target risk associated with the combined risks being simulated is the individual risk corresponding to the combined risks being simulated. If the risk levels of the target risks are all below the preset risk level thresholds, and the risk values ​​of the combined risks are greater than or equal to the preset risk value thresholds, it indicates that individual risks are unlikely to attract the attention of management, but combined risks can synergistically bring unexpected harm. Therefore, a corresponding visual model is presented to management.

[0101] The corresponding visualization models include: a first-stage visualization model for target risk and a second-stage visualization model for combined risk. The preset simulation timeline is a timeline used to indicate the simulation order and duration of the visualization models; the visualization model set at the earlier timeline point is presented first.

[0102] The first visualization model is set on the simulation timeline in ascending order of risk level. The simulation duration of the first visualization model is the preset duration for the corresponding risk level. For example, for risk level 1, the first duration is 30 seconds; for risk level 2, the first duration is 60 seconds. Simulating in ascending order of risk level allows managers to gradually discover the impact of these individual risks, improving the smoothness of transitioning to the second visualization model where the risk value of the simulated combined risk is greater than or equal to the preset risk value threshold. The timeline distance between the visualization models is constrained to be less than the target axis distance. The target axis distance is the minimum first duration (e.g., 20 seconds) corresponding to the axis distance of the simulation timeline. This allows managers to quickly view the simulation results of target risks and combined risks, and to promptly establish the connection between individual target risks and combined risks, so as to intervene in the management of target risks in advance and avoid greater losses.

[0103] In one embodiment, obtaining a first inductive visualization model of the target risk includes:

[0104] A basic simulation and visualization model for obtaining target risks;

[0105] The associated twin engine that determines the basis for the simulation of portfolio risks;

[0106] The associated twin engine is highlighted in the basic inference visualization model, and combined risk knowledge is searched according to the inference basis corresponding to the associated twin engine. The combined risk knowledge is marked in the preset range of the associated twin engine.

[0107] The working principle and beneficial effects of the above technical solution are as follows:

[0108] The basic visualization model for target risk projection is a visualization model constructed based on the regional feature set of the local risk area corresponding to the target risk, used to project the target risk. The associated twin engine for projecting the combined risk is a twin component in the visualization model related to the projected combined risk. For example, if the projection basis is the distance between the slope instability area and the mine water accumulation area, and their respective risk scales, then the corresponding associated twin engine is: slope component, mine component, and local ground surface component between the slope instability area and the mine water accumulation area. The associated twin engine is highlighted in the basic simulation visualization model (e.g., highlighted annotation). This allows managers to intuitively understand how target risks are combined and correlated when performing simulations of a single target risk. Furthermore, based on the simulation basis corresponding to the associated twin engine, combined risk knowledge is searched. For example, if the associated twin engine is a component of the ground in a local area between waterlogged areas in a mine, and the simulation basis is the distance between the slope instability area and the waterlogged area in the mine, then the combined risk knowledge is the knowledge obtained from big data about the positional relationship between the slope instability area and the waterlogged area that leads to a greater risk combination. This combined risk knowledge is marked within a preset range of the associated twin engine, such as a blank area within 5 centimeters next to the associated twin engine. This further helps managers pre-learn combined risk knowledge when performing simulations of a single target risk, facilitating a global understanding when the second simulation visualization model is presented.

[0109] This invention provides a risk monitoring system for non-coal mining areas based on GAN and satellite imagery, such as... Figure 2 As shown, it includes:

[0110] Data access module 1 is used to access satellite image data sources, perform GAN extension on satellite image data of non-coal mining areas, and obtain target image data;

[0111] Image processing module 2 is used to process the target image data to obtain a regional monitoring dataset;

[0112] Database construction module 3 is used to establish a non-coal mine regional database system based on the regional monitoring dataset;

[0113] Risk monitoring module 4 is used to obtain risk monitoring results for non-coal mining areas in real time through AI analysis based on the non-coal mining area database system.

[0114] The risk monitoring system for non-coal mining areas based on GAN and satellite imagery also performs the following operations:

[0115] Analyze risk monitoring results to identify local risk areas;

[0116] Feature extraction is performed on local risk areas to obtain a regional feature set; the regional features include: regional satellite imagery data, risk type, and risk level.

[0117] Obtain the matching template for the deduction basis corresponding to the portfolio risk deduction scenario;

[0118] Based on the matching template and regional feature set of the deduction basis, the deduction basis corresponding to the combined risk is determined;

[0119] Based on the pre-set risk values ​​corresponding to the portfolio risks, from largest to smallest, and the deduction basis, the portfolio risks are deduced to obtain new risk areas;

[0120] The risk monitoring results were updated based on the newly added risk areas and risk combinations.

[0121] Identify the target risks associated with the portfolio risks being simulated;

[0122] If the risk levels of the target risks are all less than the preset risk level thresholds, and the risk value of the combined risks is greater than or equal to the preset risk value thresholds, obtain the first inference visualization model of the target risks and the second inference visualization model of the combined risks.

[0123] Obtain the preset simulation timeline;

[0124] The first simulation visualization models are set on the simulation timeline in ascending order of their corresponding risk levels. The earliest simulation visualization model starts its simulation at the starting time of the simulation timeline. The simulation duration of the first simulation visualization model is the first duration preset for the risk level corresponding to the first simulation visualization model. The timeline distance between the first simulation visualization models is less than the target axis distance.

[0125] Once the first simulation visualization model is fully set up, the second simulation visualization model is set up. The simulation duration of the second simulation visualization model is the second duration preset for the risk value corresponding to the second simulation visualization model. The time axis distance between the second simulation visualization model and the first simulation visualization model before it is less than the target axis distance.

[0126] The simulation is displayed to management personnel based on the established simulation timeline;

[0127] The first inductive visualization model for obtaining target risk includes:

[0128] A basic simulation and visualization model for obtaining target risks;

[0129] The associated twin engine that determines the basis for the simulation of portfolio risks;

[0130] The associated twin engine is highlighted in the basic inference visualization model, and combined risk knowledge is searched according to the inference basis corresponding to the associated twin engine. The combined risk knowledge is marked in the preset range of the associated twin engine.

[0131] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A risk monitoring method for non-coal mining areas based on GAN and satellite imagery, characterized in that, include: By accessing satellite imagery data sources, GAN extension is performed on satellite imagery data of non-coal mining areas to obtain target imagery data. Image processing is performed on the target image data to obtain a regional monitoring dataset; Establish a non-coal mine regional database system based on regional monitoring datasets; Based on the non-coal mining area database system, AI analysis is performed in real time to obtain risk monitoring results for non-coal mining areas; Analyze risk monitoring results to identify local risk areas; Feature extraction is performed on local risk areas to obtain regional feature sets; Regional characteristics include: regional satellite imagery data, risk type, and risk level; Obtain the matching template for the deduction basis corresponding to the portfolio risk deduction scenario; Based on the matching template and regional feature set of the deduction basis, the deduction basis corresponding to the combined risk is determined; Based on the pre-set risk values ​​corresponding to the portfolio risks, from largest to smallest, and the deduction basis, the portfolio risks are deduced to obtain new risk areas; The risk monitoring results will be updated based on newly added risk areas and risk combinations.

2. The risk monitoring method for non-coal mining areas based on GAN and satellite imagery as described in claim 1, characterized in that, By accessing satellite imagery data sources, GAN extension is applied to satellite imagery data of non-coal mining areas to obtain target imagery data, including: Access to Gaofen-1B, C, D satellites and Gaofen-6E, F, G satellites to acquire satellite imagery data of non-coal mining areas; Conditional GANs are used to supplement satellite imagery data to obtain target imagery data.

3. The risk monitoring method for non-coal mining areas based on GAN and satellite imagery as described in claim 1, characterized in that, Image processing is performed on the target image data to obtain a regional monitoring dataset, including: Enhancement processing is performed on the target image data; After using Faster R-CNN target detection technology and U-Net model to perform target detection and segmentation on the enhanced target image data, parameter extraction and analysis are then performed to obtain a regional monitoring dataset.

4. The risk monitoring method for non-coal mining areas based on GAN and satellite imagery as described in claim 3, characterized in that, Enhancement processing of the target image data includes: Based on the target image data, obtain the input data, which includes: 1m resolution panchromatic image, 4m resolution multispectral image, DEM data, orbital parameters or RPC parameters; Radiometric calibration, FLAASH atmospheric correction, and orthorectification are performed on multispectral images; Orthorectification is performed on the panchromatic image; the orthorectification process combines DEM data and RPC parameters. After both the multispectral and panchromatic images have been processed, they are fused and enhanced to obtain the pre-output enhanced image. Use the pre-output enhanced image that has passed the quality check as the enhanced image.

5. The risk monitoring method for non-coal mining areas based on GAN and satellite imagery as described in claim 1, characterized in that, A non-coal mine regional database system was established based on regional monitoring datasets, including: The regional monitoring dataset is stored in a database system that combines ArcGIS FileGeodatabase, Personal Geodatabase, and file system, which is built using the ArcGeodatabase object relational database, to obtain a non-coal mine regional database system.

6. The risk monitoring method for non-coal mining areas based on GAN and satellite imagery as described in claim 1, characterized in that, Based on the non-coal mining area database system, real-time AI analysis is used to obtain risk monitoring results for non-coal mining areas, including: Based on the vector change analysis method and the extreme learning machine algorithm, and using a database system of non-coal mining areas, AI analysis is performed in real time to obtain risk monitoring results for non-coal mining areas.

7. The risk monitoring method for non-coal mining areas based on GAN and satellite imagery as described in claim 1, characterized in that, Also includes: Identify the target risks associated with the portfolio risks being simulated; If the risk levels of the target risks are all less than the preset risk level thresholds, and the risk value of the combined risks is greater than or equal to the preset risk value thresholds, obtain the first inference visualization model of the target risks and the second inference visualization model of the combined risks. Obtain the preset simulation timeline; The first simulation visualization models are set on the simulation timeline in ascending order of their corresponding risk levels. The earliest simulation visualization model starts its simulation at the starting time of the simulation timeline. The simulation duration of the first simulation visualization model is the first duration preset for the risk level corresponding to the first simulation visualization model. The timeline distance between the first simulation visualization models is less than the target axis distance. Once the first simulation visualization model is fully set up, the second simulation visualization model is set up. The simulation duration of the second simulation visualization model is the second duration preset for the risk value corresponding to the second simulation visualization model. The time axis distance between the second simulation visualization model and the first simulation visualization model before it is less than the target axis distance. The simulation is presented to administrators based on the established simulation timeline.

8. The risk monitoring method for non-coal mining areas based on GAN and satellite imagery as described in claim 7, characterized in that, The first inductive visualization model for obtaining target risk includes: A basic simulation and visualization model for obtaining target risks; The associated twin engine that determines the basis for the simulation of portfolio risks; The associated twin engine is highlighted in the basic inference visualization model, and combined risk knowledge is searched according to the inference basis corresponding to the associated twin engine. The combined risk knowledge is marked in the preset range of the associated twin engine.

9. A risk monitoring system for non-coal mining areas based on GAN and satellite imagery, characterized in that, include: The data access module is used to access satellite image data sources and perform GAN extension on satellite image data of non-coal mining areas to obtain target image data. The image processing module is used to process the target image data to obtain the regional monitoring dataset; The database construction module is used to establish a non-coal mine regional database system based on regional monitoring datasets. The risk monitoring module is used to perform real-time AI analysis based on the non-coal mining area database system to obtain risk monitoring results for non-coal mining areas; The risk monitoring system for non-coal mining areas based on GAN and satellite imagery also performs the following operations: Analyze risk monitoring results to identify local risk areas; Feature extraction is performed on local risk areas to obtain regional feature sets; Regional characteristics include: regional satellite imagery data, risk type, and risk level; Obtain the matching template for the deduction basis corresponding to the portfolio risk deduction scenario; Based on the matching template and regional feature set of the deduction basis, the deduction basis corresponding to the combined risk is determined; Based on the pre-set risk values ​​corresponding to the portfolio risks, from largest to smallest, and the deduction basis, the portfolio risks are deduced to obtain new risk areas; The risk monitoring results will be updated based on newly added risk areas and risk combinations.

Images