A mine deep exploration method, device and equipment based on air-ground fusion

Abstract

This application discloses an air-ground integrated method, apparatus, and equipment for deep mine exploration, relating to the field of mine safety monitoring. The method includes: conducting full-area aerial surveys of the target mining area to generate surface data, analyzing surface structure and alteration information, and delineating key exploration target areas; deploying muon imaging detection equipment in the key exploration target areas to collect flux data, reconstructing a three-dimensional density voxel model through joint transmission and scattering inversion, extracting deep anomaly parameters, and performing spatial registration and multi-source data fusion of the surface data and the three-dimensional density voxel model under a unified spatiotemporal reference; generating a risk classification map based on the fused air-ground-deep integrated data, establishing a routine air-ground collaborative monitoring cycle, and automatically issuing graded early warnings when monitoring data triggers preset thresholds. This application breaks down the air-ground-deep data barriers, achieving wide-area, accurate, low-cost, and full-cycle integrated monitoring and exploration, significantly improving the identification rate of hidden disaster-causing bodies and the mine safety early warning capability.

Description

Technical Field

[0001] This application relates to the field of mine safety monitoring, and in particular to a method, apparatus and equipment for deep mine exploration that integrates air and ground technologies. Background Technology

[0002] Currently, with the depletion of shallow mineral resources, global mining has fully entered the deep-mining stage (500-2000 meters). Deep mines have complex geological conditions, and hidden hazards such as goafs, faults, and water-bearing structures are the main causes of major safety accidents such as water inrushes, collapses, and landslides. At the same time, the exploration demand for strategic minerals such as lithium and rare earth elements associated with coal-bearing formations is increasingly urgent. However, existing mine monitoring and exploration technologies have the following main shortcomings: First, traditional drilling methods are "point" explorations, with single-hole costs ranging from hundreds of thousands to millions of yuan, and a verification success rate of less than 30%. Their detection range is limited to the area around the borehole, failing to comprehensively reflect the spatial distribution of goafs and ore bodies, resulting in a serious "one-hole view" problem and a high risk of missed or misjudged detections. Second, single geophysical exploration methods have obvious limitations. Traditional geophysical methods such as electrical resistivity tomography and seismic methods are easily affected by electromagnetic interference and topography in mines, have limited detection depth, and suffer from strong interpretability issues. Single UAV remote sensing technology can only acquire surface topography and image information, and cannot penetrate rock strata to detect deep anomalies. While single muon imaging technology can perform deep density inversion, it lacks the boundary constraints of surface geological information, leading to ambiguity in the interpretation of the anomaly's origin (whether it is a structure, ore body, or mined-out area). Third, the existing mine monitoring system is fragmented. Surface monitoring (manual inspection, GPS) and underground monitoring (borehole stress gauges, microseismic data) are independent, making data fusion difficult, hindering the construction of a unified geological model, and preventing comprehensive, multi-dimensional, coordinated early warning from the surface to the depths. Therefore, there is an urgent need for a method that can break down the "air-ground-depth" data barriers and achieve wide-area, accurate, low-cost, and full-cycle integrated monitoring and exploration. Summary of the Invention

[0003] The purpose of this application is to provide a method, apparatus and equipment for deep mine exploration that integrates air and ground, in order to solve the technical problems of limited exploration range, fragmentation of deep and surface information, high overall cost and low accuracy of identifying hidden disaster-causing bodies in the prior art.

[0004] To achieve the above objectives, this application provides the following solution: Firstly, this application provides an air-ground integrated method for deep mine exploration, including: Using drones equipped with lidar and / or multispectral sensors, a full-area aerial survey of the target mining area is conducted to generate surface data; the surface data includes: digital surface model, orthophoto, and multispectral / thermal infrared data. Based on the surface data, analyze the surface structure and alteration information, and delineate one or more key exploration target areas through spatial overlay and comprehensive evaluation of multi-source information. Mun imaging detection equipment was deployed in the key exploration target area to collect mun flux data penetrating the rock strata and preprocess the data. A three-dimensional density inversion was performed on the preprocessed muon flux data using a combined transmission imaging and scattering imaging method to reconstruct a three-dimensional density voxel model. Based on the three-dimensional density voxel model, a density threshold is set to perform binarization segmentation and three-dimensional connected component analysis to extract the spatial boundary and geometric parameters of deep anomalies. Using the extracted spatial boundary and geometric parameters as spatial constraint benchmarks, the surface data and the three-dimensional density voxel model are spatially registered and multi-source data fused under a unified spatiotemporal benchmark. Based on the geometric parameters, the risk weight factors of each anomaly are calculated, and an integrated air-ground-deep fusion model containing the spatial boundary, geometric parameters and risk weight factors is output. Based on the aforementioned integrated air-ground-deep fusion model, a risk classification map is generated to automatically identify high-risk anomaly areas; Establish a routine monitoring cycle and regularly conduct drone inspections and muon detection to monitor surface deformation sequences and underground anomaly evolution sequences; when monitoring data triggers preset thresholds, automatically issue graded warnings.

[0005] Optionally, based on the orthophotos, multispectral / thermal infrared data, and digital surface model, surface structure and alteration information are analyzed, and one or more key exploration target areas are delineated through spatial overlay and comprehensive evaluation of multi-source information, specifically including: Based on the orthophoto, the geometric distribution of linear and circular structures and the tonal anomalies caused by alteration are identified to obtain the first identification result; Based on the multispectral data, the information on iron staining, mudification or carbonate alteration is quantitatively extracted using band ratio or principal component analysis to obtain a second identification result. Based on the digital surface model, slope, aspect and curvature topographic factors are extracted to identify fault scarps, settlement funnels or landslide boundaries, and a third identification result is obtained. The first identification result, the second identification result, and the third identification result are spatially superimposed to delineate the key exploration target area.

[0006] Optionally, the method of using a combined transmission imaging and scattering imaging inversion technique to perform three-dimensional density inversion on the preprocessed muon flux data and reconstruct a three-dimensional density voxel model specifically includes: Transmission imaging section: Based on the flux attenuation law of muons after passing through rock layers, the theoretical flux is calculated by forward modeling and iteratively compared with the measured flux to gradually infer the density distribution of the underground medium, which is used to delineate the boundary of large-scale density anomalies and obtain the transmission imaging inversion results. Scattering imaging section: By utilizing the angular change information of multiple scatterings between muons and high atomic number materials, uranium, lead or rare earth high-density heavy minerals are identified, and scattering imaging inversion results are obtained; The transmission imaging inversion results and the scattering imaging inversion results are jointly solved to output the three-dimensional density voxel model.

[0007] Optionally, during the joint solution of the transmission imaging inversion results and the scattering imaging inversion results, an iterative reconstruction algorithm is applied to discretize the two-dimensional projection data into a three-dimensional voxel grid for repeated iterative optimization until the forward modeling results converge to the measured data; and the density values ​​of the three-dimensional density voxel model are calibrated and corrected by combining the measured density values ​​of the regional borehole cores, and the density values ​​are mapped to specific lithology or geological body types.

[0008] Optionally, based on the three-dimensional density voxel model, a density threshold is set to perform binarization segmentation and three-dimensional connected component analysis to extract the spatial boundary and geometric parameters of the deep anomaly, specifically: Voxels with a density more than 15% lower than the average value of the surrounding rock were marked as low-density anomalies, and voxels with a density more than 10% higher than the average value of the surrounding rock were marked as high-density anomalies. A three-dimensional connected domain analysis algorithm was used to cluster and merge adjacent anomalous voxels with the same attributes, and the spatial boundary, volume and center coordinates of each independent anomaly were calculated as the geometric parameters.

[0009] Optionally, using the extracted spatial boundaries and geometric parameters as spatial constraint benchmarks, the digital surface model, orthophotos, and multispectral / thermal infrared data are spatially registered and multi-source data fused with the three-dimensional density voxel model under a unified spatiotemporal benchmark. Specifically, this includes: A unified spatiotemporal reference and coordinate system is constructed. Feature points that can be identified by both the surface data and the three-dimensional density voxel model are selected as control points. Coordinate transformation parameters are calculated through spatial affine transformation or polynomial fitting. The voxel center coordinates of the three-dimensional density voxel model are accurately registered to the plane coordinates and elevation reference of the surface data. Using the fine grid of the surface data as a reference, the three-dimensional density voxel model is resampled using Kriging interpolation or bilinear interpolation, so that the voxel coordinates of the three-dimensional density voxel model correspond one-to-one with the grid points of the surface data in space. At the same time, the surface data is aggregated and downsampled according to voxel size, so as to realize the collaborative analysis and fusion of multi-scale data under the same spatiotemporal reference.

[0010] Optionally, the specific implementation strategy for establishing a routine monitoring cycle is as follows: The drone inspection is repeated on a daily or weekly basis to obtain time-series data on surface deformation, and high-density data collection is carried out focusing on the key monitoring areas delineated by the spatial boundary. Muon detection is repeated monthly or quarterly to obtain time-series data on the evolution of subsurface anomalies, and the volume and center coordinates in the geometric parameters are updated in real time. Based on the temporal evolution trend of the geometric parameters, the dynamic evolution threshold is dynamically adjusted to achieve rolling refresh of the risk classification map.

[0011] Optionally, the risk classification map uses red, yellow, and green colors, with red representing high-risk areas, yellow representing medium-risk areas, and green representing low-risk areas. The classification warning method includes one or more combinations of on-site audible and visual alarms, SMS push notifications, and system pop-ups. The warning instruction carries the anomaly type, three-dimensional coordinates, and risk evolution trend generated by the spatial boundary and geometric parameters.

[0012] Secondly, this application provides an air-ground integrated deep mine exploration device, comprising: The surface data acquisition module is used to conduct full-area aerial surveys of the target mining area using a drone equipped with lidar and / or multispectral sensors to generate surface data; the surface data includes: digital surface model, orthophoto, and multispectral / thermal infrared data. The key exploration target area delineation module is used to analyze the surface structure and alteration information based on the surface data, and delineate one or more key exploration target areas through spatial overlay and comprehensive evaluation of multi-source information. The muon flux data acquisition module is used to deploy muon imaging detection equipment in the key exploration target area, collect muon flux data penetrating the rock layer and perform preprocessing. The 3D density inversion module is used to perform 3D density inversion on the preprocessed muon flux data using a joint inversion method of transmission imaging and scattering imaging, and reconstruct a 3D density voxel model. The spatial boundary and geometric parameter extraction module is used to perform binarization segmentation and three-dimensional connected component analysis based on the density threshold set by the three-dimensional density voxel model, and to extract the spatial boundary and geometric parameters of the deep anomaly. The air-ground-depth integrated fusion model construction module is used to perform spatial registration and multi-source data fusion of the surface data and the three-dimensional density voxel model under a unified spatiotemporal reference, using the extracted spatial boundary and geometric parameters as spatial constraint benchmarks, and to calculate the risk weight factor of each anomaly based on the geometric parameters, and output an air-ground-depth integrated fusion model containing the spatial boundary, geometric parameters and risk weight factor. The high-risk anomaly area identification module is used to generate a risk classification map based on the air-ground-deep integrated fusion model and automatically identify high-risk anomaly areas. The graded early warning module is used to establish a routine monitoring cycle, regularly perform UAV inspections and muon detection to monitor the surface deformation sequence and the evolution sequence of underground anomalies; when the monitoring data triggers a preset threshold, it automatically issues a graded early warning.

[0013] Thirdly, this application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the air-ground integrated deep mine exploration method described in any one of the above.

[0014] According to the specific embodiments provided in this application, this application has the following technical effects: This application provides a method, apparatus, and equipment for deep mine exploration that integrates air and ground exploration, which has the following significant advantages: Full-area three-dimensional detection, eliminating blind spots: The organic integration of UAV wide-area surface monitoring (area coverage) and muon imaging deep volume detection (volume coverage) completely breaks the limitations of traditional drilling "point" detection, and realizes full-area, three-dimensional and blind-spot-free monitoring of the mine "above ground + underground", greatly reducing the missed detection rate.

[0015] Significantly improves detection accuracy and reliability: By using a centimeter-level surface model to constrain subsurface property inversion and cross-validating with multi-source data, the identification accuracy of deep anomalies is improved from the traditional meter level to the centimeter level. Simultaneously, the correlation analysis between the surface and subsurface effectively suppresses the ambiguity of geophysical inversion, significantly improving the accuracy of orebody delineation and goaf identification.

[0016] Significantly reducing overall exploration costs and time: A new model combining "drone surveys + detailed surveys + drilling verification" replaces the traditional "blind drilling" approach. Field verification has shown that this can reduce overall exploration costs by over 50% and shorten the exploration cycle by over 70%.

[0017] Green and non-destructive, in line with policy guidance: Drone inspection and Muzi imaging are both non-invasive, radiation-free, and low-energy-consumption green technologies that solve the problem of the great damage to the ecological environment caused by traditional drilling and geophysical exploration, and fully comply with the national requirements for "green exploration" and "intelligent mine" construction.

[0018] Achieving full-cycle intelligent early warning: A complete closed loop has been constructed from early exploration to mid-term monitoring and then to later services. Through normalized and automated monitoring and intelligent early warning, the mine safety management model has been upgraded from "passive rescue" to "proactive prevention". Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 A schematic diagram of the overall process of an integrated air-ground deep mine monitoring and exploration method provided in an embodiment of this application; Figure 2 A schematic flowchart illustrating an air-ground integrated deep mine exploration method provided in an embodiment of this application; Figure 3 A schematic diagram comparing the effects of the air-ground integrated deep mine exploration method of the present invention with traditional drilling methods, provided as an embodiment of this application; Figure 4 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0021] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0022] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0023] See Figure 1This system illustrates an integrated air-ground architecture for deep ore body exploration, comprised of three core units: a UAV remote sensing module, a muon imaging detection module, and a data fusion and early warning platform. These units form a complete closed-loop operating system through data flow and feedback mechanisms. On the left is the UAV remote sensing module (1), which uses a UAV equipped with high-precision airborne remote sensing equipment to perform large-scale, non-contact scanning of the mining area. This allows for the acquisition of aerial imagery, topographic elevation information, and multispectral or lidar data, enabling functions such as 3D modeling of the surface terrain, extraction of structural lines, and identification of vegetation and landforms. This module can quickly grasp the surface environmental characteristics of the mining area, providing accurate surface boundary conditions and spatial coordinate benchmarks for subsequent deep exploration. In the middle is the muon imaging detection module (2), which receives high-energy muon particles generated by cosmic rays by placing a muon detector in a ground borehole. Because muons have extremely strong penetrating power, they can penetrate thousands of meters of rock layers from bottom to top. The attenuation of muons varies significantly between different media, allowing for the inversion of underground rock density structure based on the detected particle flux and angular distribution. The system can achieve deep rock strata penetration, ore body density imaging, ore body occurrence morphology identification, 3D structure reconstruction, and ore body volume and reserve estimation, with a detection depth of over 2000 meters. The lower part of the image shows the 3D density distribution of the target ore body deep underground in the form of a 3D model, clearly reflecting the spatial distribution characteristics of the ore body. On the right is the 3-data fusion and early warning platform, responsible for the unified management and comprehensive analysis of UAV remote sensing data and muon imaging data. The platform has functions such as data management, 3D visualization, data fusion, ore body modeling, reserve estimation, report generation, and system management. The platform ultimately outputs a 3D ore body distribution map, spatial location distribution, reserve estimation results, and exploration results report, and can further generate intelligent early warning information. The bottom of the image also shows the complete information flow: muon particle paths propagate upwards from underground, surface remote sensing data is transmitted to the central module, deep imaging results are aggregated to the platform, and the platform can then issue tasks and feedback results to the front end, forming a collaborative working mechanism of "surface perception—deep exploration—intelligent analysis—decision feedback". The numbers 4, 5, and 6 in the diagram correspond to the surface, underground rock strata, and deep target ore bodies, respectively. This emphasizes that the system can simultaneously cover the entire process of exploration from the surface to the deep underground, and is a new type of deep mineral exploration technology system that integrates high precision, low disturbance, and intelligence.

[0024] Specifically, in one exemplary embodiment, such as Figure 2 As shown, a ground-air integrated method for deep mine exploration is provided. This method is executed by computer equipment, specifically by a terminal or server alone, or by both a terminal and a server. In this embodiment, the method includes the following steps: Step 101: Use a drone equipped with lidar and / or multispectral sensors to conduct a full-area aerial survey of the target mining area to generate surface data; the surface data includes: digital surface model, orthophotos and multispectral / thermal infrared data.

[0025] Step 102: Based on the surface data, analyze the surface structure and alteration information, and delineate one or more key exploration target areas through spatial overlay of multi-source information and comprehensive evaluation.

[0026] Specifically, drones equipped with lidar and / or multispectral sensors are used to conduct full-area aerial surveys of the target mining area. When lidar is used, the sensor emits high-frequency laser pulses towards the ground. By accurately measuring the time difference between the emission and return of each pulse (combined with real-time position and attitude information obtained from GPS and inertial measurement units), the precise three-dimensional coordinates of the ground surface and all objects on the ground (such as vegetation and buildings) are calculated, thereby directly generating high-density three-dimensional point cloud data. After filtering, classification, and rasterization, a digital surface model containing the height of all ground features is finally formed. When using an optical camera for photogrammetry, the drone collects highly overlapping sequence images along a predetermined route. The Structure for Motion Restoration (SfM) algorithm in computer vision is used to automatically match corresponding image points on multiple images. Combined with aerial triangulation, the three-dimensional spatial coordinates of each image point are calculated, thereby generating a dense three-dimensional point cloud and interpolating it into a regular grid digital surface model.

[0027] Then, a drone equipped with an optical camera is used for high-overlap aerial photography, and the position and attitude of the images are recorded with the help of GPS and IMU. Then, aerial triangulation is performed using the structure-reconstruction-motion algorithm to solve the camera parameters and sparse point cloud. Finally, the original images are differentially corrected pixel by pixel using a digital surface model to eliminate the geometric distortion caused by terrain undulation and camera tilt. Then, multiple corrected images are mosaicked and color-matched to generate an orthophoto map with a uniform scale that can be directly measured.

[0028] Then, based on orthophotos, multispectral / thermal infrared data, and combined with a digital surface model, the surface structure and alteration information were analyzed. The specific analysis method is as follows: 1. Based on orthophotos: Visual interpretation identifies linear and ring structures (such as fault traces and joint zones) and tonal anomalies caused by alteration (such as limonite mineralization appearing yellowish-brown). This is the most direct method for identifying structures and alterations.

[0029] 2. Based on multispectral / thermal infrared data: Using methods such as band ratio (e.g., red light / near infrared) or principal component analysis, quantitative information on alteration such as iron staining, mudification, and carbonation is extracted. This is the core technology for identifying alteration.

[0030] 3. Based on digital surface model: mainly used to assist in the analysis of structural morphology. By extracting topographic factors such as slope, aspect, and curvature, it identifies three-dimensional surface structural features such as fault scarps, settlement funnels, and landslide boundaries.

[0031] Based on digital surface models, anomalies in landform such as subsidence funnels and fault scarps are identified. Orthophotos are used to interpret the geometric distribution of linear and ring-shaped structures. Multispectral data is used to extract alteration anomalies such as iron staining and mudstone formation. This multi-source information is then spatially overlaid and comprehensively evaluated to ultimately delineate one or more key exploration target areas within the target region. The accuracy of this full-domain aerial survey reaches the centimeter level. The sensors carried by the UAV are selected from one or more combinations of lidar, multispectral cameras, and thermal infrared cameras.

[0032] Step 103: Deploy muon imaging detection equipment in the key exploration target area to collect muon flux data penetrating the rock strata and perform preprocessing.

[0033] Step 104: The three-dimensional density inversion of the preprocessed muon flux data is performed using a combined transmission imaging and scattering imaging inversion method to reconstruct a three-dimensional density voxel model.

[0034] Step 105: Based on the three-dimensional density voxel model, set a density threshold to perform binarization segmentation and three-dimensional connected component analysis to extract the spatial boundary and geometric parameters of the deep anomaly.

[0035] Step 106: Using the extracted spatial boundary and geometric parameters as spatial constraint benchmarks, perform spatial registration and multi-source data fusion of the surface data and the three-dimensional density voxel model under a unified spatiotemporal benchmark, and calculate the risk weight factor of each anomaly based on the geometric parameters, and output an integrated air-ground-deep fusion model containing the spatial boundary, geometric parameters and risk weight factor.

[0036] Specifically, muon imaging detection equipment is deployed in the key exploration target area to conduct passive deep exploration and collect muon flux data penetrating the rock strata. Modeling is then performed. First, the raw muon flux data collected by the muon imaging detection equipment deployed on the surface of the key exploration target area is preprocessed, including cosmic ray background subtraction, detector response correction and noise filtering, in order to improve the data signal-to-noise ratio.

[0037] Then, a joint inversion method combining transmission imaging and scattering imaging is employed: the transmission imaging part, based on the flux attenuation law of muons after passing through rock strata, calculates the theoretical flux through forward modeling and iteratively compares it with the measured flux to gradually infer the density distribution of the underground medium, mainly used to delineate large-scale density anomalies such as goaf areas and orebody boundaries; the scattering imaging part utilizes the angular change information of multiple scatterings between muons and high atomic number materials to identify high-density heavy minerals such as uranium, lead, and rare earth elements. In the inversion calculation, inversion algorithms or classical iterative reconstruction algorithms (such as MLEM and SART) are applied to discretize the two-dimensional projection data into a three-dimensional voxel grid. Through repeated iterative optimization, the forward modeling results of the inversion model continuously approach the measured data until they converge to the optimal solution.

[0038] Finally, a three-dimensional density voxel model with clear spatial coordinates is reconstructed. Through post-processing operations such as filtering, smoothing, and artifact removal, three-dimensional density slice maps or isosurface maps are generated to visually display the spatial morphology and density differences of deep anomalies (such as low-density goafs, high-density aquifers, or ore bodies). When necessary, the inversion results are calibrated and corrected by combining measured density values ​​from regional borehole cores, converting the density values ​​into specific lithological or geological body types.

[0039] Based on the 3D density voxel model obtained from inversion and reconstruction, the model is binarized and segmented by setting density thresholds. For example, voxels with a density more than 15% lower than the average value of the surrounding rock are marked as low-density anomalies (corresponding to goafs or aquifers), and voxels with a density more than 10% higher than the average value of the surrounding rock are marked as high-density anomalies (corresponding to ore bodies or tight rock masses). Then, a 3D connected component analysis algorithm is used to cluster and merge adjacent anomalous voxels with the same attributes, extracting independent anomalous body clusters, and calculating the spatial boundaries, volume, center coordinates, and other geometric parameters of each anomalous body. Next, by generating 3D density slice maps at different depth levels, the continuity and stratigraphic distribution of the anomalous bodies in the vertical direction are observed. At the same time, by drawing density isosurface models, the 3D contours of the anomalous bodies are presented in a visual form, intuitively showing their spatial distribution morphology (such as lenticular goafs, layered veins, or columnar caving zones). Finally, the genesis is determined by combining regional geological background knowledge—for example, low-density anomalies with continuous layered distribution may be coal-bearing strata or goaf areas; low-density anomalies with isolated masses may be karst cavities; high-density anomalies with clear boundaries may be metallic ore bodies; and high-density anomalies that coincide with the spatial location of faults may be aquifers or dense dikes. Through the above steps, the spatial morphology and density differences of various deep anomalies can be accurately identified from the three-dimensional density model. These deep anomalies include one or more of the following: ore-bearing strata, goaf areas, aquifers, and concealed faults.

[0040] The muon imaging detection employs a combined inversion method of transmission imaging and scattering imaging. Transmission imaging is used to delineate the boundaries of the ore body and the extent of the goaf, while scattering imaging is used to identify high atomic number materials. Then, the surface data obtained in step S101 and the three-dimensional density voxel model obtained in step S104 are spatially registered and fused under a unified spatiotemporal reference.

[0041] Specifically, it includes the following sub-steps: Construct a unified spatiotemporal reference and coordinate system; Surface data is obtained by conducting full-area aerial surveys of the target mining area using drones equipped with lidar or multispectral sensors. Lidar emits laser pulses and receives echoes, combining them with GPS data to directly generate a digital surface model and 3D point cloud. Multispectral cameras acquire images, which are then stitched and corrected to generate orthophotos. Surface structure and alteration information are further analyzed, and the data is ultimately stored in raster or vector format. Subsurface data is obtained by deploying muon imaging detection equipment in key exploration target areas to continuously collect muon flux data penetrating rock strata. Then, a combined transmission and scattering imaging inversion method, combined with inversion algorithms or iterative reconstruction algorithms, is used to calculate 3D density, reconstructing a subsurface density model represented in the form of a 3D voxel grid, i.e., a 3D density voxel model. Each voxel contains spatial coordinates and a density value. All data is unified to the same coordinate system and projection system. Then, identifiable feature points shared by both the surface and subsurface are selected as control points. Coordinate transformation parameters are calculated through spatial affine transformation or polynomial fitting, thereby accurately registering the voxel center coordinates of the subsurface density model to the planar coordinates and elevation datum of the surface data. Scale matching addresses the spatial resolution differences between the surface data and the subsurface density model. Specifically, using the fine grid of the surface data as a reference, Kriging interpolation or bilinear interpolation methods are used to resample the subsurface density model, ensuring a one-to-one spatial correspondence between subsurface voxel coordinates and surface grid points. Simultaneously, the surface data is aggregated and downsampled according to the subsurface voxel size, ultimately achieving collaborative analysis and fusion of multi-scale data under the same spatiotemporal reference.

[0042] The causal relationship between surface anomalies and subsurface anomalies is linked by a multi-source data fusion algorithm.

[0043] Step 107: Generate a risk classification map based on the air-ground-deep integrated fusion model and automatically identify high-risk anomaly areas.

[0044] Based on the integrated air-ground-deep fusion model, a risk grading map is generated to automatically identify high-risk anomaly areas. The risk grading map is a three-color risk grading map with red representing high-risk areas, yellow representing medium-risk areas, and green representing low-risk areas.

[0045] Step 108: Establish a routine monitoring cycle and regularly conduct UAV inspections and muon detection to monitor the surface deformation sequence and the evolution sequence of underground anomalies; when the monitoring data triggers a preset threshold, an automatic graded warning will be issued.

[0046] Specifically, drone inspections are repeated daily or weekly to monitor surface deformation; muon detection is repeated monthly or quarterly to monitor the evolution of underground anomalies.

[0047] When monitored data triggers a preset threshold, a tiered warning will be automatically issued. The tiered warning methods include one or more combinations of on-site alarms, SMS push notifications, and system pop-ups.

[0048] Based on the same inventive concept, this application also provides an air-ground integrated deep mine exploration device for implementing the air-ground integrated deep mine exploration method described above. The solution provided by this device is similar to the solution described in the above method. Therefore, the specific limitations of one or more air-ground integrated deep mine exploration device embodiments provided below can be found in the limitations of the air-ground integrated deep mine exploration method described above, and will not be repeated here.

[0049] In one exemplary embodiment, an air-ground integrated deep mine exploration device is provided, comprising: The surface data acquisition module is used to conduct full-area aerial surveys of the target mining area using a drone equipped with lidar and / or multispectral sensors to generate surface data; the surface data includes: digital surface model, orthophoto, and multispectral / thermal infrared data. The key exploration target area delineation module is used to analyze the surface structure and alteration information based on the surface data, and delineate one or more key exploration target areas through spatial overlay and comprehensive evaluation of multi-source information. The muon flux data acquisition module is used to deploy muon imaging detection equipment in the key exploration target area, collect muon flux data penetrating the rock layer and perform preprocessing. The 3D density inversion module is used to perform 3D density inversion on the preprocessed muon flux data using a joint inversion method of transmission imaging and scattering imaging, and reconstruct a 3D density voxel model. The spatial boundary and geometric parameter extraction module is used to perform binarization segmentation and three-dimensional connected component analysis based on the density threshold set by the three-dimensional density voxel model, and to extract the spatial boundary and geometric parameters of the deep anomaly. The air-ground-depth integrated fusion model construction module is used to perform spatial registration and multi-source data fusion of the surface data and the three-dimensional density voxel model under a unified spatiotemporal reference, using the extracted spatial boundary and geometric parameters as spatial constraint benchmarks, and to calculate the risk weight factor of each anomaly based on the geometric parameters, and output an air-ground-depth integrated fusion model containing the spatial boundary, geometric parameters and risk weight factor. The high-risk anomaly area identification module is used to generate a risk classification map based on the air-ground-deep integrated fusion model and automatically identify high-risk anomaly areas. The graded early warning module is used to establish a routine monitoring cycle, regularly perform UAV inspections and muon detection to monitor the surface deformation sequence and the evolution sequence of underground anomalies; when the monitoring data triggers a preset threshold, it automatically issues a graded early warning.

[0050] Below, taking the monitoring of goaf and water hazards in a certain underground mine in northern Shaanxi and the Shanglin coal-bearing lithium mine in Guangxi as examples, the above-mentioned method of this invention is applied to strategic mineral exploration: Monitoring of mined-out areas and water hazards in a well mine in northern Shaanxi This embodiment takes a coal mine in northern Shaanxi, China, with a mining depth of 500-800 meters and a risk of water inrush as an example, and applies the method of the present invention to monitor the goaf and water hazards.

[0051] S1: Preliminary reconnaissance and target area delineation In May 2025, a DJI M300 industrial-grade drone equipped with a LiDAR sensor was used to conduct a full-area aerial survey of a 15-square-kilometer target mining area. The flight altitude was set at 120 meters, with a forward overlap of 80% and a lateral overlap of 60%. After acquiring point cloud data, it underwent denoising, filtering, and classification processing to generate a 1:500 scale digital surface model and a 3D point cloud model of the mining area.

[0052] By interpreting the digital surface model, a significant surface subsidence funnel area was discovered in the southwest of the mining area, covering approximately 0.5 square kilometers, with a maximum subsidence of 15 centimeters. Combined with surface alteration information from multispectral image analysis, this subsidence funnel area and its surrounding region were designated as the key exploration target area T1.

[0053] S2: Mid-term refined monitoring and data fusion Three portable muon imaging detectors were deployed in a triangular array above the target area T1, with a spacing of 50 meters between the detectors. Continuous detection was conducted for 45 days, collecting data on the cosmic ray muon flux that penetrated the rock layers.

[0054] After data acquisition, a joint inversion method combining transmission and scattering imaging was used to process the data. Transmission imaging was used to invert the underground density distribution, while scattering imaging was used to identify high atomic number materials. A self-developed "seed" inversion algorithm was then applied for three-dimensional density reconstruction.

[0055] The inversion results clearly show that within a depth range of 350-450 meters underground, there exists a low-density anomaly (15% lower density than the surrounding rock), measuring 200 meters long, 30 meters high, and 80 meters wide, whose morphological characteristics match those of known goaf areas. Simultaneously, at a depth of approximately 500 meters below this goaf area, a high-density anomaly was identified, which, combined with regional geological data, is determined to be a water-rich layer.

[0056] The digital land surface model obtained in step S1 is spatially registered with the subsurface 3D density model obtained in step S2. The CGCS2000 coordinate system is used uniformly, and multi-scale data fusion is achieved through feature point matching and kriging interpolation.

[0057] Fusion analysis revealed that the surface subsidence funnel (center coordinates X: 364500, Y: 412300) has a good spatial correlation in the vertical direction with the underground low-density goaf (center coordinates X: 364510, Y: 412280, depth -400 meters) and high-density aquifer (depth -520 meters). It was determined that there is a hydraulic connection between the goaf and the deep aquifer, and the risk level of collapse and water inrush is high.

[0058] S3: After-sales service and full-cycle early warning The fused data is imported into the data fusion and early warning platform. The platform automatically generates a red, yellow, and green risk classification map, marking target area T1 as a red high-risk area.

[0059] Establish a routine monitoring cycle: conduct drone inspections every two months to monitor changes in surface subsidence; conduct muon detection every six months to monitor the evolution trend of underground mining voids and aquifers.

[0060] When monitoring data triggers a preset threshold, the platform automatically issues a tiered warning. In this embodiment, the surface subsidence monitoring data for the third month shows that the subsidence has increased to 18 centimeters, exceeding the preset threshold. The platform automatically sends SMS warnings and system pop-up warnings to the mine safety management department.

[0061] Based on this, the mine took preventative measures to reinforce the goaf and reduce water pressure. Subsequent drilling verification showed that the goaf and water inrush were accurately exposed at the predicted location, validating the accuracy of the invention and successfully preventing potential water inrush accidents.

[0062] Statistics after one year of continuous operation show that, compared with traditional monitoring methods, the accuracy rate of identifying goaf and aquifers has reached over 85%; the rate of collapse and water inrush accidents has decreased by over 80%; the incidence of major safety accidents has decreased by over 60%; annual disaster prevention and control costs have been saved by approximately 10 million yuan; resource recovery rate has increased by 6%, resulting in an annual increase in revenue of approximately 20 million yuan; and the overall monitoring cost has been reduced by over 55%.

[0063] Exploration of coal-bearing lithium deposits in Shanglin, Guangxi This embodiment takes a coal-bearing lithium mine in Shanglin, Guangxi as an example to demonstrate the application of the method of the present invention for strategic mineral exploration.

[0064] S1: Preliminary reconnaissance and target area delineation From January to March 2026, a Pegasus V10 drone equipped with a multispectral camera and a thermal infrared camera was used to conduct aerial surveys of a 15-square-kilometer area in the mining area. By interpreting surface alteration information through multispectral data and combining it with thermal infrared data to identify temperature anomaly zones, three prospecting target areas were initially delineated.

[0065] S2: Mid-term refined monitoring and data fusion Mun imaging detectors were deployed in three target areas and continuously probed for two months, covering depths of 100-500 meters. Through mun imaging inversion, lithium-bearing anomalies with relatively small density differences from the surrounding rocks were identified.

[0066] By fusing surface alteration information with underground density anomalies, it was found that the surface alteration zone and the underground lithium-bearing strata are highly consistent in spatial location, further verifying the existence of lithium-bearing strata.

[0067] Post-verification Based on the fusion results, the drilling layout was optimized, and subsequent borehole verification showed that the accuracy of ore body delineation exceeded 85%. Compared with traditional drilling methods, exploration costs were reduced by more than 50%, and the exploration cycle was shortened by 70%.

[0068] Below, the integrated air-ground deep mineral exploration method of the present invention is compared with traditional drilling methods. See below. Figure 3This figure compares the differences between the "invention (integrated air-ground deep prospecting method)" and the "traditional drilling method" in terms of exploration process, result form, and technical and economic indicators, highlighting the significant advantages of the new method in terms of coverage, accuracy, detection depth, cost, and cycle. The left side of the figure shows the scheme of the invention. The upper part first shows the surface remote sensing results (UAV remote sensing), which quickly acquires the topography and anomaly information of the mining area through a color elevation model, achieving continuous coverage of the entire area. Then, the deep exploration results (münden imaging) are shown, which use three-dimensional volume data to reflect the density anomaly distribution within the range of 0-2000 meters underground. The color from blue to red indicates that the density difference is from low to high, which can intuitively locate potential mineralization centers. The lower part shows the three-dimensional model of the ore body (fusion result), which integrates the surface remote sensing and underground density imaging results to construct a complete spatial model of the ore body, and expresses it in a hierarchical manner according to different confidence levels such as "inferred ore body - control ore body - presumed ore body". The key indicators of this method are further quantified below the figure: the exploration coverage achieves "full area coverage," no longer limited to the vicinity of the borehole; the ore body positioning accuracy can reach 10-20 meters; the detection depth exceeds 2000 meters; the overall exploration cost can be reduced by 50-70%; and the exploration cycle can be shortened by more than 60%. The "Advantages Summary" at the bottom points out its outstanding features such as full area coverage, high-precision imaging, deep detection, low cost and high efficiency, and green and non-destructive operation. In contrast, the traditional drilling method shown on the right side of the figure mainly relies on manual surface surveys and discrete borehole layout. Geological survey results only provide two-dimensional geological map information, and the borehole layout requires a large number of drilling rigs for point-like construction. The actual drilling results can only be locally verified at a few borehole locations, and the interpretation of the formed ore body relies on interpolation inference, which has significant uncertainty. Its indicators show that: the exploration coverage is limited to the point area around the borehole; the ore body positioning accuracy is usually 50-100 meters; the detection depth generally does not exceed 1500 meters and is significantly limited by the drilling rig capacity; the drilling cost is high; and the exploration cycle usually takes several months to several years. The "Summary of Shortcomings" at the bottom points out that traditional methods suffer from problems such as small coverage, low accuracy, limited depth, high cost, long cycle time, and significant destructiveness. This comparison chart clearly demonstrates that the integrated air-ground deep mineral exploration method overcomes the technical bottleneck of traditional drilling's "point verification and local speculation," enabling comprehensive detection and quantitative evaluation of deep ore bodies in a non-destructive, rapid, and high-precision manner. This provides a more efficient, economical, and intelligent new technological approach for deep resource exploration.

[0069] In one exemplary embodiment, a computer device is provided, which may be a server or a terminal, and its internal structure diagram may be as follows. Figure 4As shown, the computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media. The database stores integrated air-ground mining deep exploration data. The I / O interfaces allow the processor to exchange information with external devices. The communication interface allows communication with external terminals via a network connection. When the computer program is executed by the processor, it implements an integrated air-ground mining deep exploration method.

[0070] Those skilled in the art will understand that Figure 4 The structures shown are merely block diagrams of some structures related to the present application and do not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than shown in the figures, or combine certain components, or have different component arrangements. In an exemplary embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0071] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0072] 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 computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0073] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0074] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0075] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for deep mine exploration integrating air and ground operations, characterized in that, The air-ground integrated deep mine exploration method includes: Using drones equipped with lidar and / or multispectral sensors, a full-area aerial survey of the target mining area is conducted to generate surface data; the surface data includes: digital surface model, orthophoto, and multispectral / thermal infrared data. Based on the surface data, analyze the surface structure and alteration information, and delineate one or more key exploration target areas through spatial overlay and comprehensive evaluation of multi-source information. Mun imaging detection equipment was deployed in the key exploration target area to collect mun flux data penetrating the rock strata and preprocess the data. A three-dimensional density inversion was performed on the preprocessed muon flux data using a combined transmission imaging and scattering imaging method to reconstruct a three-dimensional density voxel model. Based on the three-dimensional density voxel model, a density threshold is set to perform binarization segmentation and three-dimensional connected component analysis to extract the spatial boundary and geometric parameters of deep anomalies. Using the extracted spatial boundary and geometric parameters as spatial constraint benchmarks, the surface data and the three-dimensional density voxel model are spatially registered and multi-source data fused under a unified spatiotemporal benchmark. Based on the geometric parameters, the risk weight factors of each anomaly are calculated, and an integrated air-ground-deep fusion model containing the spatial boundary, geometric parameters and risk weight factors is output. Based on the aforementioned integrated air-ground-deep fusion model, a risk classification map is generated to automatically identify high-risk anomaly areas; Establish a routine monitoring cycle and regularly conduct drone inspections and muon detection to monitor surface deformation sequences and underground anomaly evolution sequences; when monitoring data triggers preset thresholds, automatically issue graded warnings.

2. The air-ground integrated deep mine exploration method according to claim 1, characterized in that, Based on the orthophotos, multispectral / thermal infrared data, and digital surface models, surface structure and alteration information are analyzed. One or more key exploration target areas are delineated through spatial overlay and comprehensive evaluation of multi-source information. Specifically, these include: Based on the orthophoto, the geometric distribution of linear and circular structures and the tonal anomalies caused by alteration are identified to obtain the first identification result; Based on the multispectral data, the information on iron staining, mudification or carbonate alteration is quantitatively extracted using band ratio or principal component analysis to obtain a second identification result. Based on the digital surface model, slope, aspect and curvature topographic factors are extracted to identify fault scarps, settlement funnels or landslide boundaries, and a third identification result is obtained. The first identification result, the second identification result, and the third identification result are spatially superimposed to delineate the key exploration target area.

3. The air-ground integrated deep mine exploration method according to claim 1, characterized in that, The method employs a combined transmission and scattering imaging inversion technique to perform three-dimensional density inversion on the preprocessed muon flux data, reconstructing a three-dimensional density voxel model. Specifically, this includes: Transmission imaging section: Based on the flux attenuation law of muons after passing through rock layers, the theoretical flux is calculated by forward modeling and iteratively compared with the measured flux to gradually infer the density distribution of the underground medium, which is used to delineate the boundary of large-scale density anomalies and obtain the transmission imaging inversion results. Scattering imaging section: By utilizing the angular change information of multiple scatterings between muons and high atomic number materials, uranium, lead or rare earth high-density heavy minerals are identified, and scattering imaging inversion results are obtained; The transmission imaging inversion results and the scattering imaging inversion results are jointly solved to output the three-dimensional density voxel model.

4. The air-ground integrated deep mine exploration method according to claim 3, characterized in that, In the process of jointly solving the transmission imaging inversion results and the scattering imaging inversion results, an iterative reconstruction algorithm is applied to discretize the two-dimensional projection data into a three-dimensional voxel grid for repeated iterative optimization until the forward modeling results converge to the measured data; and the density values ​​of the three-dimensional density voxel model are calibrated and corrected by combining the measured density values ​​of regional borehole cores, and the density values ​​are mapped to specific lithology or geological body types.

5. The air-ground integrated deep mine exploration method according to claim 1, characterized in that, Based on the aforementioned three-dimensional density voxel model, a density threshold is set for binarization segmentation and three-dimensional connected component analysis to extract the spatial boundary and geometric parameters of deep anomalies. Specifically: Voxels with a density more than 15% lower than the average value of the surrounding rock were marked as low-density anomalies, and voxels with a density more than 10% higher than the average value of the surrounding rock were marked as high-density anomalies. A three-dimensional connected domain analysis algorithm was used to cluster and merge adjacent anomalous voxels with the same attributes, and the spatial boundary, volume and center coordinates of each independent anomaly were calculated as the geometric parameters.

6. The air-ground integrated deep mine exploration method according to claim 1, characterized in that, Using the extracted spatial boundaries and geometric parameters as spatial constraint benchmarks, the digital land surface model, orthophotos, and multispectral / thermal infrared data are spatially registered and multi-source data fused with the three-dimensional density voxel model under a unified spatiotemporal benchmark. Specifically, this includes: A unified spatiotemporal reference and coordinate system is constructed. Feature points that can be identified by both the surface data and the three-dimensional density voxel model are selected as control points. Coordinate transformation parameters are calculated through spatial affine transformation or polynomial fitting. The voxel center coordinates of the three-dimensional density voxel model are accurately registered to the plane coordinates and elevation reference of the surface data. Using the fine grid of the surface data as a reference, the three-dimensional density voxel model is resampled using Kriging interpolation or bilinear interpolation, so that the voxel coordinates of the three-dimensional density voxel model correspond one-to-one with the grid points of the surface data in space. At the same time, the surface data is aggregated and downsampled according to voxel size, so as to realize the collaborative analysis and fusion of multi-scale data under the same spatiotemporal reference.

7. The air-ground integrated deep mine exploration method according to claim 1, characterized in that, The specific implementation strategy for establishing a routine monitoring cycle is as follows: The drone inspection is repeated on a daily or weekly basis to obtain time-series data on surface deformation, and high-density data collection is carried out focusing on the key monitoring areas delineated by the spatial boundary. Muon detection is repeated monthly or quarterly to obtain time-series data on the evolution of subsurface anomalies, and the volume and center coordinates in the geometric parameters are updated in real time. Based on the temporal evolution trend of the geometric parameters, the dynamic evolution threshold is dynamically adjusted to achieve rolling refresh of the risk classification map.

8. The air-ground integrated deep mine exploration method according to claim 1, characterized in that, The risk grading map uses red, yellow, and green colors for identification, with red representing high-risk areas, yellow representing medium-risk areas, and green representing low-risk areas. The graded early warning methods include one or more combinations of on-site audible and visual alarms, SMS push notifications, and system pop-ups. The early warning instructions carry the anomaly type, three-dimensional coordinates, and risk evolution trend generated by the spatial boundary and geometric parameters.

9. A deep mine exploration device integrating air and ground capabilities, characterized in that, The air-ground integrated deep mine exploration device includes: The surface data acquisition module is used to conduct full-area aerial surveys of the target mining area using a drone equipped with lidar and / or multispectral sensors to generate surface data; the surface data includes: digital surface model, orthophoto, and multispectral / thermal infrared data. The key exploration target area delineation module is used to analyze the surface structure and alteration information based on the surface data, and delineate one or more key exploration target areas through spatial overlay and comprehensive evaluation of multi-source information. The muon flux data acquisition module is used to deploy muon imaging detection equipment in the key exploration target area, collect muon flux data penetrating the rock layer, and perform preprocessing. The 3D density inversion module is used to perform 3D density inversion on the preprocessed muon flux data using a joint inversion method of transmission imaging and scattering imaging, and reconstruct a 3D density voxel model. The spatial boundary and geometric parameter extraction module is used to perform binarization segmentation and three-dimensional connected component analysis based on the density threshold set by the three-dimensional density voxel model, and to extract the spatial boundary and geometric parameters of the deep anomaly. The air-ground-depth integrated fusion model construction module is used to perform spatial registration and multi-source data fusion of the surface data and the three-dimensional density voxel model under a unified spatiotemporal reference, using the extracted spatial boundary and geometric parameters as spatial constraint benchmarks, and to calculate the risk weight factor of each anomaly based on the geometric parameters, and output an air-ground-depth integrated fusion model containing the spatial boundary, geometric parameters and risk weight factor. The high-risk anomaly area identification module is used to generate a risk classification map based on the air-ground-deep integrated fusion model and automatically identify high-risk anomaly areas. The graded early warning module is used to establish a routine monitoring cycle, regularly perform UAV inspections and muon detection to monitor the surface deformation sequence and the evolution sequence of underground anomalies; when the monitoring data triggers a preset threshold, it automatically issues a graded early warning.

10. A computer device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the air-ground integrated deep mine exploration method according to any one of claims 1-8.

Images