A mine geological disaster dynamic early warning method based on space-air-ground multi-source data

Abstract

This invention discloses a dynamic early warning method for mine geological disasters based on multi-source data from air, space, and ground, belonging to the field of geological disaster monitoring and early warning technology. The method collects data from satellite remote sensing, BeiDou GNSS, UAV aerial surveying, and ground monitoring stations, completing standardized collection and traceability coding; it constructs a scenario-based dynamic weighted fusion model, assigning differentiated fusion weights to landslides, surface subsidence, and power transmission tower tilt; and outputs four-level risk levels and early warning durations for monitoring points through a multi-dimensional risk assessment model, triggering graded and targeted emergency responses and achieving identifiable push notifications of early warning information. This invention achieves dynamic fusion of multi-source data from air, space, and ground for precise early warning of mine disasters, improving accuracy by over 35% and reducing the missed detection rate to below 1%; the entire process is verifiable and traceable, solving the problem of obtaining evidence for technical infringement, adapting to existing mine monitoring systems, reducing implementation costs by 40%, and significantly improving early warning accuracy, timeliness, and implementability.

Description

Technical Field

[0001] This invention relates to the field of geological disaster monitoring and early warning technology, specifically a dynamic early warning method for mine geological disasters based on multi-source data from air, space, and ground, applicable to the accurate monitoring and graded early warning of geological disasters such as mine landslides, surface subsidence, and power transmission tower tilting. Background Technology

[0002] Mine geological hazard monitoring is a core aspect of mine safety production. Current technologies largely employ integrated space-air-ground technologies for monitoring geological hazards in mining areas, but these technologies still face numerous limitations in practical application, making it difficult to meet the high-precision and high-timeliness monitoring and early warning needs of mines.

[0003] 1. The data fusion method is singular, mostly adopting a fixed weight fusion mode of three types of data: satellite remote sensing, Beidou GNSS, and ground monitoring station data. The weights are not dynamically adjusted according to the monitoring focus of different types of mine disasters, resulting in low monitoring accuracy and easy to miss or misdetect disasters.

[0004] 2. The early warning results are vague, only outputting the overall disaster risk level of the mining area, and cannot achieve the dual positioning of precise monitoring points and specific warning duration. The mine site cannot carry out targeted emergency response work, resulting in low emergency response efficiency.

[0005] 3. The monitoring data coverage has blind spots, lacks localized and refined data supplemented by UAV aerial surveys, and macroscopic satellite remote sensing data is unable to capture small-scale, sudden geological deformation features, resulting in a lag in disaster early warning and response;

[0006] 4. The technical solution lacks unified verifiable features, and the format of the entire process of data collection, processing, and output is not standardized. This not only makes it impossible to achieve full-process traceability of mine monitoring data, but also makes it difficult to obtain evidence and lacks clear and objective basis for technical comparison when similar technologies infringe on rights.

[0007] While existing technologies attempt to improve the early warning effect of mine geological disasters through multi-source data fusion, none of them have proposed a dynamic weighted fusion scheme adapted to specific mine scenarios, nor have they formed a complete technical closed loop of "multi-source data collection - intelligent fusion analysis - precise hierarchical early warning - targeted emergency response". Furthermore, they have not designed verifiable and traceable feature identifiers in the technical solutions, and the protection measures for technical achievements are insufficient. There is an urgent need in this field for a dynamic early warning method for mine geological disasters that can solve the above-mentioned multiple pain points. Summary of the Invention

[0008] To address the shortcomings of existing technologies, this invention provides a dynamic early warning method for mine geological disasters based on multi-source data from air, space, and ground, which solves the technical problems of rigid data fusion, ambiguous early warning results, blind spots in monitoring, and difficulty in obtaining evidence of infringement in existing mine geological disaster monitoring and early warning systems.

[0009] To achieve the above-mentioned objectives, the present invention provides a dynamic early warning method for mine geological disasters based on multi-source data from air, space, and ground, comprising the following steps:

[0010] 1. Construction and Standardization of Multi-Source Data Acquisition Module

[0011] The system simultaneously acquires four types of air-space-ground data in the mine monitoring area: macroscopic surface deformation data from satellite remote sensing, real-time displacement data from BeiDou GNSS monitoring points, localized refined imagery data from UAV aerial surveys, and meteorological and geomechanical data from ground monitoring stations. All data types are connected to a unified data processing platform via standardized API interfaces to achieve real-time data synchronization. All raw and pre-processed data are uniformly converted to the JSON standard format, and each data point is assigned a unique traceability code. The coding rule is: Mine Number - Monitoring Type Code - Acquisition Timestamp - Monitoring Point Number. BeiDou GNSS monitoring points are deployed at intervals of 8-12 meters in key mine monitoring areas.

[0012] Among them, the UAV aerial survey data is acquired by inspection UAVs equipped with dual cameras of visible light and thermal imaging. The aerial survey path is dynamically planned according to the high-risk points of Beidou GNSS. The waypoints are set at 8-12 meters, and the aerial survey image resolution is 0.03-0.07 meters / pixel. The waypoint coordinate information is embedded in the traceability code of the aerial survey data. The preferred waypoints are set at 10 meters, and the preferred aerial survey image resolution is 0.05 meters / pixel.

[0013] 2. Construction of a Scenario-Based Dynamic Weight Fusion Model

[0014] The system pre-defines three core geological hazard types: mine landslides, surface subsidence, and power transmission tower tilting. Based on the monitoring focus of different hazard types, it automatically assigns fusion weights to four types of data sources. The weight allocation coefficients are embedded into the fusion model output results with fixed values. The comprehensive analysis data output by the fusion model retains the weight allocation coefficient identifiers and original data traceability codes, enabling the data processing end to be verifiable and comparable.

[0015] 3. Multi-dimensional disaster risk assessment and standardized output

[0016] The multi-source data after dynamic weighting is input into a preset disaster risk assessment model. Combined with the mining face recovery progress, historical geological disaster data, and real-time rainfall, the model outputs a four-level risk level and corresponding precise warning duration for each monitoring point. The risk level and warning duration have a unique correspondence. The risk assessment results are output in a standardized report format. The report includes six core fields: unique coordinates of the monitoring point, risk level identifier, precise warning duration, weight allocation coefficient, data traceability code, and assessment timestamp, so as to achieve standardization and verifiability of the assessment results.

[0017] 4. Tiered and targeted emergency response and identifiable push notifications

[0018] Based on the risk assessment results, a tiered response mechanism is triggered. Early warning information is pushed out in a targeted manner through three methods: cloud platform, mini-program, and SMS. The target audience and risk level are fixedly matched. All early warning information pushed through all channels includes six core contents: precise monitoring location, risk level, warning duration, emergency response suggestions, unique traceability code, and weight allocation coefficient. In addition, SMS and mini-program push information embeds fixed-format feature fields, with the field format being [Mine Geological Disaster Early Warning - Traceability Code: XXX - Weight Coefficient: XXX - Warning Duration: XXX], to achieve unique identification and comparability of the early warning response end.

[0019] The preferred weights for data source fusion in landslide disaster early warning are: 35%-45% for UAV aerial survey data, 25%-35% for BeiDou GNSS data, 15%-25% for meteorological data from ground monitoring stations, and 5%-15% for satellite remote sensing data. The preferred weights are 40%, 30%, 20%, and 10%.

[0020] The preferred weights for data source fusion in land subsidence disaster early warning are: BeiDou GNSS data weight 45%-55%, satellite remote sensing data weight 25%-35%, ground monitoring station geological data weight 10%-20%, and UAV aerial survey data weight 3%-8%, with preferred weights of 50%, 30%, 15%, and 5%.

[0021] Preferably, the data source fusion weights for power transmission tower tilt early warning are: BeiDou GNSS data weight 40%-50%, UAV aerial survey data weight 30%-40%, ground monitoring station geological data weight 10%-20%, and satellite remote sensing data weight 3%-8%, with preferred weights of 45%, 35%, 15%, and 5%.

[0022] The preferred and unique correspondence between the four risk levels and the precise warning duration is as follows: Level 1 extremely high risk, warning 4 hours in advance; Level 2 high risk, warning 1-3 days in advance; Level 3 medium risk, warning 7 days in advance; and Level 4 low risk, trend warning 15 days in advance.

[0023] Preferably, the fixed matching relationship between the target of the early warning information and the risk level is as follows: Level 1 and Level 2 risks are pushed to the core team of mine safety management, and Level 3 and Level 4 risks are pushed to the person in charge of the mine area.

[0024] Beneficial effects

[0025] This invention provides a dynamic early warning method for mine geological disasters based on multi-source data from air, space, and ground. Compared with existing technologies, it has the following outstanding substantive features and significant progress:

[0026] 1. Significantly improves monitoring and early warning accuracy. For the first time, it integrates monitoring data from air, space, ground, and air, and designs a scenario-based dynamic weighted fusion model to accurately allocate data weights for different disaster types and set reasonable and effective ranges. Compared with the traditional fixed weighted fusion method, the early warning accuracy is improved by more than 35%, and the missed detection rate is reduced to less than 1%.

[0027] 2. To achieve more accurate early warning results, output three-dimensional early warning information of "monitoring point + risk level + early warning duration", solve the problem of vague early warning in existing technology, and enable targeted emergency response at the mine site, greatly improving emergency response efficiency;

[0028] 3. Achieve full-area monitoring without blind spots. By supplementing local refined data through UAV aerial surveys, it can make up for the monitoring blind spots of macroscopic data from satellite remote sensing, effectively capture small-scale and sudden geological deformations, and achieve full coverage of the mine monitoring area;

[0029] 4. It is highly practical and has low implementation costs. All data interfaces are standardized interfaces available on the market. Existing mine monitoring platforms can be directly connected and upgraded without the need for customized high-priced equipment, reducing implementation costs by 40%. Furthermore, the early warning response mechanism is compatible with existing mine safety management processes, without the need for additional adjustments to the organizational structure.

[0030] 5. Facilitate infringement evidence collection by designing verifiable, traceable, and unique technical features throughout the entire process of data collection, processing, evaluation, and push. These features are all objective and directly verifiable technical identifiers. In case of infringement, it is only necessary to compare whether the other party's technical solution / system output contains the above features to quickly determine the infringement, thus solving the industry pain points of difficulty in obtaining evidence and complexity in comparison of similar technologies.

[0031] 6. Strengthen the protection of technological achievements. The unique identifier and standardized output of the entire process form an "exclusive feature system" for the technical solution. Combined with the reasonable design of the numerical range, it is difficult for the other party to circumvent the infringement by simply modifying the surface parameters, which greatly enhances the protection of the patent. Attached Figure Description

[0032] Figure 1 This is a flowchart illustrating the overall process framework of the present invention.

[0033] Figure 2 This is a schematic diagram of the scenario-based dynamic weight fusion model structure of the present invention;

[0034] Figure 3 This is a schematic diagram of the hierarchical and targeted emergency response push process of the present invention.

[0035] The system includes: 1. Multi-source data acquisition module; 2. Unified data processing platform; 3. Scenario-based dynamic weight fusion model; 4. Disaster risk assessment model; 5. Tiered and targeted emergency response module; 6. Cloud platform push terminal; 7. Mini-program push terminal; 8. SMS push terminal; 9. Receiving terminal for the core team of mine safety management; and 10. Receiving terminal for the person in charge of the mine area. Detailed Implementation

[0036] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0037] Example 1: Dynamic early warning of mine landslide disasters

[0038] This embodiment takes the landslide disaster early warning of a coking coal mine in Shanxi Province as an example. The mine number is SXJH001. The slopes with high landslide incidence in this mine are the key monitoring areas. Ten Beidou GNSS monitoring stations and two ground meteorological and geological monitoring stations are set up every 10 meters. Aerial survey operations are carried out by inspection drones equipped with visible light + thermal imaging dual cameras.

[0039] Multi-source data acquisition and standardized processing: Surface deformation data of the mining area at a scale of 1:10,000 was obtained by connecting to a commercial satellite data platform. The monitoring type code is W, the acquisition timestamp is 20240926100000, the monitoring point number is 001, and a unique traceability code is generated: SXJH001-W-20240926100000-001. The data was converted to JSON standard format. One Beidou mini monitoring station was deployed every 10 meters on the landslide-prone slopes of the mine to acquire horizontal / vertical displacement data of the monitoring points. The sampling frequency was 5 minutes / time. The traceability code for monitoring station No. 1 is: SXJH001-B-20240927174000-001. The data was converted to JSON standard format. Aerial survey routes were dynamically planned based on high-risk BeiDou GNSS locations, with waypoints deployed every 10 meters. The aerial survey resolution was set to 0.05 meters per pixel, with an update frequency of 2 hours per update. The aerial survey data traceability code is: SXJH001-U-20240927180000-001, embedding the waypoint coordinates X=4023151.11 and Y=12715661. Ground monitoring stations were deployed at the bottom of the slope to acquire real-time rainfall and soil moisture content data, with a sampling frequency of 1 minute per update. The traceability code is: SXJH001-D-20240927190000-001. The data was converted to JSON standard format. All data was connected to the mine's existing cloud monitoring platform through a standardized interface program to achieve unified data format and real-time synchronization. All raw and pre-processed data retained a unique traceability code.

[0040] Scenario-based dynamic weighted fusion model construction: For landslide disasters, a preset landslide weight allocation scheme is activated, and the optimal weights are adopted: 40% for UAV aerial survey data, 30% for Beidou GNSS data, 20% for meteorological data from ground monitoring stations, and 10% for satellite remote sensing data. The four types of data are feature extracted and weighted fused through a weighted fusion model. In the comprehensive deformation data output by the fusion model, a fixed weight allocation coefficient identifier "40-30-20-10" is embedded, and the unique traceability code of all original data is retained to realize the traceability and verifiability of the data processing process.

[0041] Multi-dimensional disaster risk assessment and standardized output: The integrated deformation data was input into the risk assessment model. Combined with the recent mining progress of the mine face (daily mining volume of 5 meters) and real-time rainfall (cumulative rainfall of 80 mm over 3 consecutive days), the horizontal displacement of monitoring point No. 1 on the slope was calculated to be 342.6 mm, with a displacement rate of 23 mm / day, which was determined to be a Level 1 risk (extremely high), and a warning was issued 4 hours in advance (warning time: 2024-09-27 21:40). The risk assessment results were output in a standardized report format, and the report fully included six core fields: unique coordinates of the monitoring point (X=4023151.11, Y=12715661), risk level identifier (Level 1), precise warning duration (4 hours), weight allocation coefficient (40-30-20-10), data traceability code (SXJH001-B-20240927174000-001), and assessment timestamp (20240927174000).

[0042] Tiered and targeted emergency response and identifiable push notifications: The system automatically triggers the first-level risk response mechanism and pushes mini-program messages and SMS alerts to the core team of mine safety management according to fixed matching relationships. All push information contains six core contents and embeds fixed-format feature fields. Example of SMS alert content:

Mine Geological Disaster Warning - Traceability Code: SXJH001-B-20240927174000-001-Weight Coefficient: 40-30-20-10-Warning Duration: 4 hours

[0043] Example 2: Dynamic early warning of tilting power transmission towers in mines

[0044] This embodiment takes the early warning of tilting power transmission towers in a coal mine in Shaanxi Province as an example. The mine number is SXMK002. There are 20 power transmission towers in the mining area, which are the core facilities for power supply in the mining area. Ten key towers are selected and Beidou GNSS monitoring points are set up every 10 meters. Inspection drones are used to carry out geological aerial surveys of the towers and surrounding areas.

[0045] Multi-source data acquisition and standardization processing: Surface deformation data of the mining area was acquired, updated once a day. Monitoring type code: W; acquisition timestamp: 20241010080000; monitoring point number: 005; unique traceability code: SXMK002-W-20241010080000-005. Data was converted to JSON standard format. Beidou mini monitoring stations were deployed at the foundations of power transmission towers, one monitoring point per tower, with a sampling frequency of 5 minutes per station. The traceability code for monitoring station No. 5 is: SXMK002-B-20241010140000-005. Data was converted to JSON standard format. The system dynamically plans aerial survey routes based on BeiDou GNSS monitoring data, deploying waypoints every 10 meters with a resolution of 0.05 meters per pixel and an update frequency of 3 hours per update. The aerial survey data traceability code is SXMK002-U-20241010150000-005, embedding waypoint coordinates X=3856214.32 and Y=11987652.45. Ground monitoring stations are deployed around the tower to acquire geomechanical and environmental data, sampling at a frequency of 1 minute per update. The traceability code is SXMK002-D-20241010160000-005, and the data is converted to JSON standard format. All data is connected to the mine cloud monitoring platform through a standardized interface program, uniformly formatted as JSON and retaining a unique traceability code, achieving real-time data synchronization and traceability.

[0046] Scenario-based dynamic weighted fusion model construction: For power transmission tower tilt disasters, a preset tower tilt weight allocation scheme is activated, and the optimal weights are adopted: Beidou GNSS data 45%, UAV aerial survey data 35%, ground monitoring station geological data 15%, and satellite remote sensing data 5%. The four types of data are feature extracted and weighted fused through a weighted fusion model. In the comprehensive analysis data output by the fusion model, a fixed weight allocation coefficient identifier "45-35-15-5" is embedded, and the unique traceability code of all original data is retained.

[0047] Multi-dimensional disaster risk assessment and standardized output: The integrated analysis data was input into the risk assessment model. Combined with the mining progress of the working face (daily mining volume of 4 meters) and the geological settlement data around the transmission tower, the horizontal displacement of the No. 5 transmission tower foundation was calculated to be 126.8 mm and the tilt angle was 3.2°, which was judged as a level 2 risk (high), and a warning was issued 2 days in advance (warning time 2024-10-12 14:00). The risk assessment results were output in a standardized report format. The report fully includes six core fields: unique coordinates of the monitoring point (X=3856214.32, Y=11987652.45), risk level identifier (level 2), precise warning duration (2 days), weight allocation coefficient (45-35-15-5), data traceability code (SXMK002-B-20241010140000-005), and assessment timestamp (20241010160000).

[0048] Tiered and targeted emergency response and identifiable push notifications: The system automatically triggers the secondary risk response mechanism and pushes mini-program messages and SMS alerts to the core team of mine safety management according to fixed matching relationships. The push information embeds fixed-format feature fields. Example of SMS warning content:

Mine Geological Disaster Warning - Traceability Code: SXMK002-B-20241010140000-005-Weight Coefficient: 45-35-15-5-Warning Duration: 2 days

[0049] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A dynamic early warning method for mine geological disasters based on multi-source data from air, space, and ground, characterized in that, Includes the following steps: S1: A multi-source data acquisition module is built to simultaneously acquire macroscopic data on surface deformation from satellite remote sensing of the mine monitoring area, real-time displacement data from BeiDou GNSS monitoring points, local fine-grained image data from UAV aerial surveys, and meteorological and geomechanical data from ground monitoring stations. All types of data are connected to a unified data processing platform through standardized API interfaces. All raw and pre-processed data are uniformly converted into JSON standard format, and each data is assigned a unique traceability code. The coding rule is: mine number - monitoring type code - acquisition timestamp - monitoring point number. BeiDou GNSS monitoring points are deployed at 8-12 meters in key monitoring areas of the mine. Among them, UAV aerial survey data is acquired by inspection UAVs equipped with visible light + thermal imaging dual cameras. The aerial survey path is dynamically planned according to the high-risk points of BeiDou GNSS, the waypoints are deployed at 8-12 meters, and the aerial survey image resolution is 0.03-0.07 meters / pixel. The waypoint coordinate information is embedded in the traceability code of the aerial survey data. S2: Construct a scenario-based dynamic weight fusion model, pre-set three core geological hazard types: mine landslide, surface subsidence, and power transmission tower tilt. Automatically assign fusion weights to four types of data sources for different hazard types. The weight allocation coefficients are embedded into the output results of the fusion model with fixed values. The comprehensive analysis data output by the fusion model retains the weight allocation coefficient identifier and the original data traceability code. S3: Input the multi-source data after dynamic weight fusion into the disaster risk assessment model. Combine the mining progress of the working face, historical geological disaster data, and real-time rainfall to output the four-level risk level and corresponding precise warning duration for each monitoring point. The risk level and the warning duration have a unique correspondence. The risk assessment results are output in a standardized report format. The report includes six core fields: unique coordinates of the monitoring point, risk level identifier, precise warning duration, weight allocation coefficient, data traceability code, and assessment timestamp. S4: Based on the risk assessment results, a tiered response mechanism is triggered. Warning information is pushed to users via cloud platform, mini-program, and SMS. The target audience and risk level are matched in a fixed way. All warning information pushed through all channels includes six core contents: precise monitoring location, risk level, warning duration, emergency response suggestions, unique traceability code, and weight allocation coefficient. In addition, SMS and mini-program push information embeds fixed-format feature fields, with the field format being: [Mine Geological Disaster Warning - Traceability Code: XXX - Weight Coefficient: XXX - Warning Duration: XXX].

2. The dynamic early warning method for mine geological disasters based on multi-source data from air, space, and ground as described in claim 1, is characterized in that, In step 1), the preferred layout of UAV aerial survey waypoints is every 10 meters, and the preferred resolution of aerial survey images is 0.05 meters per pixel.

3. The dynamic early warning method for mine geological disasters based on multi-source data from air, space, and ground as described in claim 1, is characterized in that... In step 2), the data source fusion weights for landslide disaster early warning are as follows: UAV aerial survey data weight 35%-45%, BeiDou GNSS data weight 25%-35%, ground monitoring station meteorological data weight 15%-25%, and satellite remote sensing data weight 5%-15%.

4. The dynamic early warning method for mine geological disasters based on multi-source data from air, space, and ground as described in claim 1, characterized in that, In step 2), the data source fusion weights for the land subsidence disaster early warning are as follows: BeiDou GNSS data weight 45%-55%, satellite remote sensing data weight 25%-35%, ground monitoring station geological data weight 10%-20%, and UAV aerial survey data weight 3%-8%.

5. The dynamic early warning method for mine geological disasters based on multi-source data from air, space, and ground as described in claim 1, characterized in that, In step 2), the data source fusion weights for the early warning of power transmission tower tilt are as follows: BeiDou GNSS data weight 40%-50%, UAV aerial survey data weight 30%-40%, ground monitoring station geological data weight 10%-20%, and satellite remote sensing data weight 3%-8%.

6. The dynamic early warning method for mine geological disasters based on multi-source data from air, space, and ground as described in claim 3, is characterized in that... The optimal weighting for data source fusion in landslide disaster early warning is: 40% UAV aerial survey data, 30% BeiDou GNSS data, 20% meteorological data from ground monitoring stations, and 10% satellite remote sensing data.

7. The dynamic early warning method for mine geological disasters based on multi-source data from air, space, and ground as described in claim 4, is characterized in that... The optimal weighting for data source fusion in the early warning of land subsidence disasters is as follows: 50% BeiDou GNSS data, 30% satellite remote sensing data, 15% geological data from ground monitoring stations, and 5% UAV aerial survey data.

8. The dynamic early warning method for mine geological disasters based on multi-source data from air, space, and ground as described in claim 5, is characterized in that... The optimal data source fusion weight for power transmission tower tilt early warning is: 45% BeiDou GNSS data, 35% UAV aerial survey data, 15% ground monitoring station geological data, and 5% satellite remote sensing data.

9. The dynamic early warning method for mine geological disasters based on multi-source data from air, space, and ground as described in claim 1, characterized in that, In step 3), the unique correspondence between the four risk levels and the precise warning duration is as follows: Level 1, extremely high risk, provides a warning 4 hours in advance; Level 2, high risk, provides a warning 1-3 days in advance; Level 3, medium risk, provides a warning 7 days in advance; and Level 4, low risk, provides a trend warning 15 days in advance.

10. The dynamic early warning method for mine geological disasters based on multi-source data from air, space, and ground as described in claim 1, characterized in that, In step 4), the fixed matching relationship between the target of the early warning information and the risk level is as follows: Level 1 and Level 2 risks are pushed to the core team of mine safety management, and Level 3 and Level 4 risks are pushed to the person in charge of the mine area.

Images