Multi-source information data fusion method for mine area ground collapse detection

Abstract

The application relates to the technical field of data processing, in particular to a multi-source information data fusion method for mine area ground collapse detection. The method comprises the following steps: acquiring current static geological data, dynamic real-time data, historical collapse data and mining activity data of multiple areas in a mine area; comparing the static geological data and the historical collapse data of each area to determine the static fusion weight of each area; comparing the mining activity data of each area with the mining activity data of the multiple areas to determine the dynamic fusion weight of each area; and respectively performing fusion processing on the static geological data of each area and the dynamic real-time data of each area based on the static fusion weight of each area and the dynamic fusion weight of each area to obtain a ground collapse prediction result of each area. The method can improve the accuracy of ground collapse detection.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, in particular to a multi-source information data fusion method for mine area ground collapse detection. BACKGROUND

[0002] Mine area ground collapse is one of the main geological disasters in the process of mineral resource exploitation, which seriously threatens the safety of personnel life and production facilities in the mine area. In order to realize early identification and early warning of collapse risk, it is usually necessary to comprehensively utilize multi-source technical means such as geological exploration, remote sensing monitoring, sensor network, etc. to obtain multi-source data reflecting the geological conditions and surface deformation of the mine area. Through integration and analysis of multi-source data, the possibility of collapse occurrence is evaluated.

[0003] At present, when integrating and analyzing multi-source data, the data of different sources are usually standardized and feature-extracted uniformly, and then a fusion model is constructed for risk evaluation.

[0004] However, due to the difficulty in real-time acquisition of geological data in the process of exploitation, there is a time difference between the geological data and the dynamically collected surface deformation data. In the process of data fusion, the static data will gradually become distorted, which leads to the fact that the fusion result gradually deviates from the actual situation with the progress of exploitation, and thus the accuracy and reliability of the collapse detection result are reduced. SUMMARY

[0005] In order to solve the technical problem that in the process of data fusion, static data will gradually become distorted, which leads to the fact that the fusion result gradually deviates from the actual situation with the progress of exploitation, the purpose of the present application is to provide a multi-source information data fusion method for mine area ground collapse detection, and the technical scheme adopted is as follows:

[0006] The present application provides a multi-source information data fusion method for mine area ground collapse detection, which comprises: acquiring current static geological data, dynamic real-time data, historical collapse data and mining activity data of multiple regions in the mine area; comparing the static geological data and the historical collapse data of each region to determine the static fusion weight of each region; comparing the mining activity data of each region with the mining activity data of the multiple regions to determine the dynamic fusion weight of each region; and based on the static fusion weight of each region and the dynamic fusion weight of each region, respectively fusing the static geological data of each region and the dynamic real-time data of each region to obtain the ground collapse prediction result of each region.

[0007] Optionally, the method further comprises: acquiring static geological data of each coordinate point in the mine area; determining a ground collapse risk index of each coordinate point based on the static geological data of each coordinate point; clustering the coordinate points based on the ground collapse risk index to obtain multiple clustering clusters; and dividing the mine area into multiple regions based on the multiple clustering clusters.

[0008] Optionally, the static geological data includes a plurality of types of static data, and the determining of the static fusion weight of each region based on the comparison of the static geological data of each region with the historical collapse data comprises: determining an anomaly index of each type of static data in a first region based on a difference between each type of static data in the static geological data of the first region and the historical collapse data, the first region being any one of the plurality of regions; determining a collapse probability of the first region based on the anomaly indexes of the plurality of types of static data in the first region; and determining the static fusion weight of the first region based on the collapse probability of the first region, an area proportion of the first region, and the anomaly indexes of the plurality of types of static data in the first region, the area proportion of the first region being a proportion of an area of the first region in an area of the mining area.

[0009] Optionally, the determining of the collapse probability of the first region based on the anomaly indexes of the plurality of types of static data in the first region comprises: determining a type number of abnormal data in the static geological data, the abnormal data being static data with an anomaly index greater than an anomaly index threshold in the static geological data; and determining a ratio of the type number of the abnormal data to a total number of the plurality of types as the collapse probability of the first region.

[0010] Optionally, the mining activity data includes mining progress, mining scale, and mined duration, and the determining of the dynamic fusion weight of each region based on a difference between the mining activity data of each region and the mining activity data of the plurality of regions comprises: determining a mining progress index of a first region based on the mining progress and the mined duration of the first region; determining a mining scale index of the first region based on a ratio between the mining scale of the first region and an average mining scale of the plurality of regions; and determining the dynamic fusion weight of the first region based on the mining progress index of the first region and the mining scale index of the first region.

[0011] Optionally, the mining activity data further includes mining planning, and the determining of the dynamic fusion weight of the first region based on the mining progress index of the first region and the mining scale index of the first region comprises: determining a mining activity change trend of the first region based on a change of the mining activity data of the first region within a historical time window; determining a dynamic influence of the first region based on the mining progress index, the mining scale index, and the mining activity change trend of the first region; and determining the dynamic fusion weight of the first region based on the dynamic influence of the first region.

[0012] Optionally, the determining the dynamic fusion weight of the first region based on the dynamic influence of the first region comprises: determining a terrain change intensity index of each region based on the static fusion weight of each region and the dynamic influence of each region; and determining the dynamic fusion weight of the first region based on the terrain change intensity index of the first region and the terrain change intensity index of the adjacent region of the first region.

[0013] Optionally, the determining the dynamic fusion weight of the first region based on the terrain change intensity index of the first region and the terrain change intensity index of the adjacent region of the first region comprises: determining a maximum terrain change intensity of the adjacent region when the terrain change intensity index of the first region is greater than the terrain change intensity index of all adjacent regions of the first region; determining a difference between the terrain change intensity index of the first region and the maximum terrain change intensity; and determining the dynamic fusion weight of the first region based on the difference, the terrain change intensity index of the first region, and an average of the terrain change intensity index of the adjacent region.

[0014] Optionally, the dynamic real-time data comprises a plurality of types of dynamic data, and the fusing the static geological data of each region and the dynamic real-time data of each region based on the static fusion weight of each region and the dynamic fusion weight of each region respectively to obtain the ground collapse prediction result of each region comprises: obtaining the dynamic real-time data of the mining area in a historical time period; dividing the historical time period into a plurality of time windows, the plurality of time windows comprising a current time window; determining an error probability of each type of dynamic data in each time window based on the difference between the plurality of types of dynamic data in the same pair of time windows; correcting the dynamic fusion weight of the first region based on the error probability of each type of dynamic data in the current time window to obtain the dynamic fusion weight of each type of dynamic data in the current dynamic real-time data of the first region; and fusing the static geological data of the first region and each type of dynamic data in the current dynamic real-time data of the first region based on the static fusion weight of the first region and the dynamic fusion weight of each type of dynamic data in the current dynamic real-time data of the first region respectively to obtain the ground collapse prediction result of each region.

[0015] Optionally, the dynamic fusion weight of each type of dynamic data in the current time window is modified based on the error probability of the first region to obtain the dynamic fusion weight of each type of dynamic data in the current dynamic real-time data of the first region, comprising: determining the error propagation influence coefficient of the first type of dynamic data in the current time window based on the error probability of the first type of dynamic data in the current time window and the data acquisition frequency, the first type being any one type included in the dynamic real-time data dynamic real-time data dynamic real-time data; and modifying the dynamic fusion weight of the first region based on the error propagation influence coefficient of the first type of dynamic data in the current time window to obtain the dynamic fusion weight of each type of dynamic data in the current dynamic real-time data of the first region.

[0016] The present application has the following beneficial effects:

[0017] In the present application, the mining area is divided into multiple regions and the weights are calculated respectively, realizing the differentiated processing of different regions and avoiding the limitations of unified analysis of the whole mining area. Then, the static fusion weight is determined by comparing the static geological data with the historical collapse data, and the dynamic fusion weight is determined by the relative difference of the mining activity data, so that the data fusion process can adaptively reflect the actual working condition changes of the mining area, and then based on the real-time changes of the data, the weights are adjusted in real time, different weights are given to the static geological data and the dynamic real-time data, and the fusion result can more accurately predict the ground collapse prediction result of each region, improving the accuracy of the ground collapse prediction result and the response ability to the mining environment changes. BRIEF DESCRIPTION OF DRAWINGS

[0018] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0019] Figure 1 A multi-source information data fusion method for mining area ground collapse detection provided by an embodiment of the present application;

[0020] Figure 2 Another multi-source information data fusion method for mining area ground collapse detection provided by an embodiment of the present application;

[0021] Figure 3 Another multi-source information data fusion method for mining area ground collapse detection provided by an embodiment of the present application;

[0022] Figure 4Another multi-source information data fusion method for mine area ground collapse detection provided by one embodiment of the present application. DETAILED DESCRIPTION

[0023] In order to further illustrate the technical means and effects adopted by the present application to achieve the predetermined object of the present application, the following describes in detail the specific embodiments, structure, features and effects of a multi-source information data fusion method for mine area ground collapse detection according to the present application, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0024] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs.

[0025] Ground collapse detection refers to monitoring multi-source, multi-dimensional data such as surface and underground deformation and geological environment changes through various technical means, so as to identify collapse hazards and assess risks. Through ground collapse detection in mine areas, ground collapse hazards can be identified in advance and warnings can be issued, which can save time for personnel evacuation and equipment prevention and control, and reduce mine area costs.

[0026] In the process of ground collapse detection in mine areas, it is usually necessary to analyze multi-source information data such as remote sensing data, topographic and geological data, ground detection data, and human activity data. In the existing method for ground collapse detection in mine areas, the data format, scale, etc. of different sources are first unified by standardizing the multi-source information data, and then the standardized multi-source information data is extracted and fused, and finally the risk of ground collapse is assessed according to the extracted feature information, so as to realize ground collapse detection in mine areas.

[0027] However, in the process of multi-source information fusion in mine areas, geological data cannot be obtained in real time during mine operation, which leads to the need for data fusion of static geological data and dynamic real-time surface topographic data during ground collapse detection in mine areas. This will cause the results of ground collapse detection to gradually lose accuracy as mine operations continue. At the same time, real-time data collection during mine operations will have a significant impact, so there will be some errors in the collected real-time multi-source information data. With the fusion of information data, data errors will accumulate and propagate, further increasing the error of ground collapse detection results.

[0028] In the process of ground collapse detection in a mining area, multi-source data of the mining area is usually combined for comprehensive analysis, including remote sensing data (used to identify changes in vegetation, changes in landforms, changes in surface subsidence, etc.), topographic and geological data (including regional geological maps, underlying lithology data, fault structure distribution, etc.), ground monitoring data (three-dimensional displacement data of ground points, underground soil stability, etc.), and human activity data (mining activity data, human activity data, etc.).

[0029] Among them, there are data with large time scale of change (topographic and geological data) and data with small time scale of change (human data, remote sensing data, etc.). In the process of data fusion, static data will gradually distort due to human activities, mining activities, etc., and static data cannot be obtained in real time in the process of mining area exploitation, which leads to the gradual increase of the error of the result of ground collapse detection in the mining area.

[0030] At the same time, due to the obvious influence of the mining area on the surrounding environment during the mining process, there will be errors in the real-time data collection of the mining area, and as the mining progresses and the scale expands, the errors will be gradually accumulated and propagated in the process of data fusion, thereby affecting the accuracy of ground collapse detection after multi-source information data fusion. Therefore, there is an urgent need for a multi-source information data fusion method for ground collapse detection in a mining area.

[0031] The specific scheme of the multi-source information data fusion method for ground collapse detection in a mining area provided by the present application will be described in detail below in combination with the drawings.

[0032] Please refer to Figure 1 , which shows the method flowchart of the multi-source information data fusion method for ground collapse detection in a mining area provided by an embodiment of the present application.

[0033] As Figure 1 shown, the multi-source information data fusion method for ground collapse detection in a mining area includes S101-S104.

[0034] S101, obtain current static geological data, dynamic real-time data, historical collapse data and mining activity data in multiple regions in the mining area.

[0035] It should be understood that the static geological data is a set of geological parameters that represent the basic geological conditions of the mining area and remain relatively stable in the short term; the dynamic real-time data is a set of time series data collected in real time by an automatic monitoring system, reflecting the dynamic changes of the ground surface and underground of the mining area.

[0036] It can be understood that, due to the different collection frequencies of static data and dynamic data, the data collection frequency of static data is lower, and therefore the static geological data and dynamic data updated last time can be determined as the current static geological data and dynamic real-time data.

[0037] Optionally, the static geological data includes multiple types of static data, such as fault distribution density, fold development degree, and other data for characterizing geological structure; surface elevation, slope, surface curvature, and other data for characterizing topography and geomorphology; groundwater depth, aquifer thickness, rock permeability coefficient, and other data for characterizing hydrogeology; rock cohesion, internal friction angle, uniaxial compressive strength, and other data for characterizing rock and soil mechanics, and the like.

[0038] Optionally, the data can be obtained through geological exploration, topographic mapping, drilling sampling, and laboratory testing, one-time or periodically.

[0039] Optionally, the dynamic real-time data includes multiple types of dynamic data, such as surface subsidence rate, surface temperature anomaly, vegetation index change data, surface inclination angle, soil displacement rate, pore water pressure, surface three-dimensional displacement rate, displacement direction angle, and the like.

[0040] Optionally, the surface subsidence rate, surface temperature anomaly, and vegetation index change data can be obtained through satellite or unmanned aerial vehicle remote sensing technology; the surface inclination angle, soil displacement rate, and pore water pressure can be obtained through geological sensors arranged in the mining area; the surface three-dimensional displacement rate and displacement direction angle can be obtained through a global navigation satellite system monitoring station.

[0041] It should be understood that the historical collapse data refers to a collection of detailed records of collapse events that have occurred within the mining area, serving as a reference benchmark for risk assessment.

[0042] Optionally, the historical collapse data includes multiple historical collapse cases, and each historical collapse case includes static geological data (such as the static geological data listed above) at the time of the collapse, historical collapse pit location coordinates, impact range, collapse area, depth and volume, geological conditions at the time of the collapse, and mining activities.

[0043] Optionally, the data can be obtained from historical collapse event archives.

[0044] It should be understood that the mining activity data refers to a collection of dynamic parameters reflecting the state and intensity of mining operations in the mining area.

[0045] Optionally, the mining activity data can include mining progress, mining scale, mining process, mining plan, and mining duration. Specifically, the mining progress is used to represent the proportion of the area that has been mined in the total area, the advancing speed of the mining working face, and the remaining mining life; the mining scale is used to represent the daily mining amount, the monthly mining amount, and the mining intensity distribution; the mining process data is used to represent the mining method, the support parameter, and the blasting operation frequency; the mining plan is used to represent the planned mining area, the expected mining amount, and the mining time sequence arrangement; and the mining duration is used to represent the duration of the mining.

[0046] Optionally, human and economic data and meteorological data of each region can also be obtained. The human and economic data refers to a set of parameters related to the human activities and the social and economic development conditions around the mining area, and specifically can include the population distribution around the region, the building density, the traffic load, the location of the infrastructure (such as roads and pipelines), and the like; and the meteorological data can include the precipitation, the precipitation intensity, the air temperature, and the like.

[0047] In an optional implementation, the above multi-source information data (i.e., static geological data, dynamic real-time data, historical subsidence data, mining activity data, human and economic data, and meteorological data) for ground subsidence detection can be obtained by the mining enterprises, the geological departments, the remote sensing service providers, and the meteorological monitoring stations in the mining area.

[0048] In an optional implementation, after the above multi-source information data is obtained, data preprocessing operations can be performed on the data.

[0049] Optionally, the multi-source information data can be converted in format. Specifically, all the data can be converted into a fusion required format, for example, the vector data such as the geological structure distribution map and the topographic map can be uniformly converted into the ESRI Shapefile format (.shp), and it is ensured that the projection coordinate system is consistent with the coordinate system of the mining area. The raster data (such as remote sensing images) is converted into the GeoTIFF format, and the geographic reference information is embedded.

[0050] Optionally, the multi-source information data can be denoised and corrected. Specifically, for the time series data such as the GNSS deformation data and the underground stress monitoring data, the data points deviating from the mean value by more than 3 times the standard deviation are removed, and the missing values are filled by the linear interpolation method; and for the optical remote sensing images, the data of the cloud pollution area is removed.

[0051] Optionally, the multi-source information data can be time and space aligned. All the dynamic data are synchronized according to a uniform time step (such as an hourly level), and the spatial data are unified in the coordinate system.

[0052] Optionally, the mining activity data, human and economic data, and meteorological data are classified as auxiliary data, and the data after preprocessing is stored according to the classification of “static data-dynamic data-auxiliary data-historical collapse data”.

[0053] In an optional implementation, before S101 is performed, the mining area can be divided into multiple regions. Specifically, static geological data of each coordinate point in the mining area can be obtained; a ground collapse risk index of each coordinate point is determined based on the static geological data of each coordinate point; the coordinate points are clustered based on the ground collapse risk index to obtain multiple clustering clusters; and the mining area is divided into multiple regions based on the multiple clustering clusters.

[0054] Optionally, the boundary of the mining area can be first delineated, and a spatial coordinate system covering the entire mining area, such as the CGCS2000 coordinate system, can be established. The accuracy of the spatial coordinate system can be the maximum accuracy that can be collected during data collection. Then, the static geological data of each coordinate point in the entire mining area can be collected.

[0055] Optionally, the ground collapse risk index of each coordinate point can be calculated by a multi-factor weighted evaluation model. Specifically, each type of static data in the static geological data is normalized to determine an influence factor of each type of static data, and the weights of the influence factors are determined based on the analytic hierarchy process. Then, the data of each influence factor of each coordinate point is weighted and summed based on the weights to obtain the ground collapse risk index of each coordinate point.

[0056] Optionally, the influence factor can include the distance of the coordinate point to the nearest fault, the rock type code, the surface slope value, the surface curvature value, the groundwater depth, and the distance to the historical collapse point.

[0057] Optionally, the ground collapse risk index of each coordinate point can also be evaluated in combination with the mining activity data, human and economic data, and meteorological data. The influence factor can also include the distance of the coordinate point to the mining working face, the daily mining amount, the precipitation, the building density near the coordinate point, and the traffic load.

[0058] It should be understood that the human and economic data can be analyzed by analyzing the population density and infrastructure distribution, and the high-sensitivity areas that can cause significant losses due to ground collapse can be identified, so as to give priority consideration in the warning level division and emergency resource allocation. The meteorological data can be used to analyze the influence of external environmental factors on the geological stability, especially the disturbance effect of heavy rainfall on the groundwater level and soil strength, and the deterioration effect of freeze-thaw cycle on the mechanical properties of rock-soil body. These two types of auxiliary data, in combination with the static geological data and dynamic real-time data, constitute a complete multi-source information evaluation system for ground collapse in the mining area, which significantly improves the forward-looking of risk identification and the scientificity of prevention and control decision-making.

[0059] Optionally, the K-means clustering algorithm can be used to cluster the coordinate points, and in the clustering process, the ground collapse risk index of each coordinate point is taken as a feature input, and through iterative calculation, the coordinate points with similar risk levels are classified into the same cluster. The coordinate points in each cluster of the obtained multiple clusters have similar geological risk characteristics, while the different clusters show obvious differences in risk levels, thereby realizing the risk classification of the mining area.

[0060] It should be understood that in this K-means clustering algorithm, The size of the value is used to represent the number of divided regions.

[0061] Optionally, in the K-means clustering algorithm, the number of clusters Satisfies the following formula:

[0062]

[0063] Wherein, The region number parameter is represented by N, The total number of coordinate points in the coordinate system of the mining area is represented by N, The preset region size parameter is represented by N, which represents the average number of coordinate points contained in a region in the mining area.

[0064] Optionally, the It can be set to 20.

[0065] It should be understood that one cluster corresponds to one region. After obtaining multiple clusters, the coordinate points belonging to the same cluster and spatially adjacent can be connected to form a continuous region boundary, and multiple regions are obtained.

[0066] Optionally, geometric feature calculation can also be performed on each region, including region area, perimeter, shape index and other parameters.

[0067] Optionally, after dividing the multiple regions, the multi-source information data of each region can be collected respectively.

[0068] Optionally, the multi-source information data of the entire mining area can also be collected, and after dividing the multiple regions, the multi-source information data of each region is divided from the multi-source information of the entire mining area.

[0069] It can be understood that by clustering through the ground collapse risk index, the coordinate points with similar geological properties are merged into the same region, so that the geological conditions inside each region are relatively uniform, and the risk characteristics between regions are obviously distinguished. The ground collapse index detection can be performed in a regional manner, which improves the pertinence and accuracy of the detection.

[0070] S102, compare the static geological data of each region with the historical collapse data to determine the static fusion weight of each region.

[0071] It can be understood that when the current geological conditions (such as rock thickness, soil density) of a region are less different (i.e. more similar) to the geological conditions in the historical collapse data, it means that the geological conditions of the region are closer to the state that has led to collapse, and at this time, a higher static fusion weight can be given to the static data of the region, so that the monitoring data of the region is given more attention in subsequent data fusion, so as to identify the region with similar collapse hidden dangers in advance.

[0072] In an optional implementation, the same type of data in the static geological data and the historical collapse data of a region can be compared and analyzed, and the difference between the geological conditions of the region and the historical collapse data is quantified based on the difference value of each type of static data.

[0073] S103, compare the mining activity data of each region with the mining activity data of the plurality of regions to determine the dynamic fusion weight of each region.

[0074] It should be understood that by comparing the mining activity data of a region with the mining activity data of other regions, the relative intensity of the mining activity of the region in the plurality of regions can be evaluated, and the hotspot region of the current mining activity can be identified. When the mining activity of a region is more active (i.e. faster mining progress, larger mining amount), it means that the geological environment of the region is undergoing more rapid and more intense human changes, and at this time, a higher dynamic fusion weight needs to be given to the region, so that the dynamic real-time data (such as GNSS deformation, remote sensing subsidence data) from the region is preferentially and more fully utilized in the data fusion process, so that the deformation signal caused by high-intensity mining can be captured in time.

[0075] Optionally, each type of data in the mining activity data can be compared respectively, and the data of each type in the plurality of regions is normalized to obtain the normalized value of each type of data of each region, and then the sum of the normalized values of the plurality of types of each region is determined. The sum of the normalized values of the plurality of regions is normalized again, which is mapped to [0, 1] to obtain the dynamic fusion weight of each region.

[0076] Optionally, the normalization processing can be maximum-minimum normalization processing.

[0077] S104, fuse the static geological data of each region and the dynamic real-time data of each region based on the static fusion weight of each region and the dynamic fusion weight of each region to obtain the ground collapse prediction result of each region.

[0078] Optionally, the static geological data of each region can be adjusted based on the static fusion weight of each region, and the dynamic real-time data of each region can be adjusted based on the dynamic fusion weight of each region to obtain the fused data. Then, the ground subsidence prediction result of each region can be obtained based on the fused data.

[0079] It should be understood that static geological data mainly reflects the inherent risk characteristics of a region, while dynamic real-time data mainly reflects the current deformation development trend. By adjusting the weights through static fusion weights and dynamic fusion weights, regions with poor geological conditions or strong disturbances from mining activities can receive higher data attention during the fusion process.

[0080] Optionally, static fusion weights, dynamic fusion weights, static geological data, and dynamic real-time data from different regions can be input into the neural network model to achieve the fusion of static geological data and dynamic real-time data for each region, thereby obtaining the detection results for each ground subsidence.

[0081] For example, the neural network model can be a convolutional neural network (CNN) or a recurrent neural network (RNN).

[0082] It should be noted that the methods for data fusion and prediction based on fused data are existing technologies, and will not be elaborated upon in this solution.

[0083] In this embodiment, the mining area is divided into multiple regions and weights are calculated separately, achieving differentiated processing for different regions and avoiding the limitations of unified analysis of the entire mining area. Then, static fusion weights are determined by comparing static geological data with historical subsidence data, and dynamic fusion weights are determined by the relative differences in mining activity data. This allows the data fusion process to adaptively reflect changes in the actual working conditions of the mining area. Furthermore, the weights are adjusted in real time based on the real-time changes in the data, assigning different weights to static geological data and dynamic real-time data. The resulting fusion results can more accurately predict the ground subsidence prediction results for each region, improving the accuracy of the ground subsidence prediction results and the responsiveness to changes in the mining environment.

[0084] Combination Figure 1 ,like Figure 2 As shown, in one implementation of this application embodiment, the static geological data includes multiple types of static data. Taking the first region as an example, the above S102 can be specifically implemented through S201-S203.

[0085] S201. Based on the differences between static data of each type and historical collapse data in the static geological data of the first region, determine the anomaly index of static data of each type in the first region.

[0086] The first region can be any one of multiple regions.

[0087] It should be understood that since a region contains multiple coordinate points, the mean of a certain type of static data (such as rock layer thickness) in the region (and a certain historical collapse case) can be determined first, and then the differences between the means can be compared to determine the anomaly index.

[0088] Understandably, historical subsidence data includes multiple historical subsidence cases. Therefore, the static geological data of a region can be compared with each historical subsidence case to obtain the anomaly index between each type of static data and each historical subsidence case in the first region.

[0089] In one alternative implementation, a training set of a random forest model containing multiple historical collapse cases can be established first. For each type of static data in the first region, the difference between it and the corresponding static data of the same type in each historical collapse case is calculated. This difference is then processed using a normalization function to obtain the anomaly index of that type of static data under each historical collapse case.

[0090] Optionally, the anomaly index of a type of static data in a historical collapse case satisfies the following formula:

[0091]

[0092] in, Indicates the region The Static data of various types in historical collapse cases Abnormal index below, Indicates the region The The mean of static data of each type Indicating historical collapse cases The Middle The mean of static data of each type This represents a normalization function, such as max-min normalization, used to normalize... Mapping to the standard range of [0, 1] ensures that the anomaly indices of different types of static data are comparable.

[0093] S202. Determine the collapse probability of the first region based on the anomaly index of multiple types of static data in the first region.

[0094] In an optional implementation, the multi-label comprehensive evaluation can be performed by a random forest model. Specifically, each decision tree in the random forest corresponds to a historical collapse case perspective. For each decision tree, an anomaly index threshold is set. When the anomaly index of the historical collapse case corresponding to the decision tree is greater than or equal to the anomaly index threshold, it is determined that the static data of this type meets the decision rule of the decision tree, and the model result is output as 1, indicating that the static data of this type is determined as abnormal data. Finally, the number of model results output as 1 in all decision trees is counted, and the ratio of the number to the total number of decision trees in the random forest is determined as the collapse probability of the first region.

[0095] Optionally, the anomaly index threshold is determined according to statistical analysis of historical collapse data. For example, the anomaly index threshold can be 0.8.

[0096] In another optional implementation, the number of types of abnormal data in the static geological data can be determined, and then the ratio of the number of types of abnormal data to the total number of types is determined as the collapse probability of the first region.

[0097] The abnormal data is static data with an anomaly index greater than an anomaly index threshold in the static geological data.

[0098] It should be understood that the anomaly index threshold is used to distinguish normal geological conditions from abnormal geological conditions. When the anomaly index of static data of a type is greater than or equal to the anomaly index threshold, the static data of the type can be marked as abnormal data. When the anomaly index of static data of a type is less than the anomaly index threshold, the static data of the type can be marked as normal data.

[0099] Optionally, after the abnormal data is determined, the number of types of static data marked as abnormal data in the first region can be counted, and then the ratio of the number of types of abnormal data to the total number of types can be determined.

[0100] It should be understood that the collapse probability of a region reflects the collapse risk level of the geological conditions of the region. When most types of static data in a region exhibit abnormality, the collapse probability tends to 1, indicating a high risk state. Conversely, when most types of static data are within the normal range, the collapse probability tends to 0, indicating a low risk state.

[0101] It can be understood that the method of evaluating the abnormality of each type of data based on the anomaly index and then evaluating the collapse probability based on the number of abnormal data effectively improves the accuracy and reliability of the collapse risk evaluation.

[0102] In another optional implementation, the average of the anomaly indexes of the static data of the types in the first region can be determined as the collapse probability of the first region.

[0103] S203, determine a static fusion weight of the first region based on the collapse probability of the first region, the area proportion of the first region, and the anomaly indexes of the multiple types of static data of the first region.

[0104] The area proportion of the first region is a proportion of an area of the first region in an area of the mining area.

[0105] Optionally, a ratio of a number of coordinate points included in the first region to a total number of coordinate points included in the mining area can be determined as the area proportion of the first region.

[0106] Optionally, the static fusion weight of a region satisfies the following formula:

[0107]

[0108] wherein, represents a static fusion weight of a region , represents a basic weight of the region ; represents a number of coordinate points included in the region , represents a total number of coordinate points included in the mining area, represents an average of anomaly indexes of multiple types of static data included in the region , represents a normalization function, for example, a max-min normalization, is a scaling coefficient of , used for mapping a value of to [0, 1], represents a collapse probability of the region .

[0109] In the formula, characterizes an objective contribution of a region size, the greater the area proportion of a region in the mining area, the more significant the stability of the region to the global influence; characterizes a dynamic adjustment contribution of a geological risk, when the abnormal degree of the geological condition is higher (the collapse probability of the region is higher), the value is larger, which can amplify the risk signal brought by the collapse probability of the region; on the contrary, if the geological condition is relatively normal, even if the collapse probability of the region is high, the influence will be appropriately inhibited.

[0110] ​​The method provided in S201-S203 can accurately identify geological features highly related to the collapse risk by calculating the difference between each type of static data and historical collapse data, and determine the static fusion weight in combination with the area proportion and the anomaly index, so that the weight distribution is more reasonable by considering the regional scale factor and the influence of the geological anomaly degree.

[0111] It can be understood that the dynamic real-time data of the mining area is mainly affected by the mining progress and the mining scale of the mining area. Large-scale rapid mining can lead to the formation of connected goaf in the underground goaf, causing uniform ground subsidence or local collapse pit, and the real-time dynamic data of the terrain surface changes relatively sharply. Small-scale slow mining can cause dispersed settlement, and the real-time dynamic data of the terrain surface changes relatively gently. In combination with the above Figure 1 Figure 3 As shown in the implementation manner of the embodiment of the present application, the mining activity data includes the mining progress, the mining amount, and the mined duration. S103 can be implemented by S301-S303.

[0112] S301, determining a mining progress index of the first region based on the mining progress and the mined duration of the first region.

[0113] It should be understood that the mining progress index of a region is used to represent the mining urgency in the time dimension and the resource consumption degree in the space dimension. When the mining progress index of a region is high, it means that the region is experiencing rapid and intensive resource mining, and the underground structure may be changing dramatically. At this time, a higher attention is given to the region.

[0114] It can be understood that the faster the mining progress of a region, the more sufficient the resource mining of the region, and the shorter the mining duration, the faster the mining speed.

[0115] Optionally, the ratio of the mined area of a region to the initial mining area can be determined as the mining progress of the region, and the time length from the initial mining area to the mined area after the region starts mining can be determined as the mined duration.

[0116] Optionally, the unit of the mined duration can be set as days.

[0117] Optionally, the mining progress index of a region satisfies the following formula:

[0118]

[0119] wherein, represents the mining progress index of the region, ​​Indicates the region The mined area, Indicates the region The initial area to be mined, Indicates the region The value indicates the mining progress; the larger the value, the faster the mining progress. Indicates the region The duration of mining, This represents a normalization function, such as max-min normalization. Used to Mapped to the standard range of [0, 1].

[0120] Based on this formula, it can be seen that when the region The greater the mining progress and the shorter the mining duration, the higher the mining progress index. When, indicate the area The mining has been completed, and at this point, the mining progress index can be set to 1.

[0121] S302. Determine the mining scale index of the first region based on the ratio between the mining scale of the first region and the average mining scale of multiple regions.

[0122] It should be understood that a region's mining scale index is used to characterize the relative level of mining intensity in that region, reflecting its importance and activity level within the overall mining pattern of the mining area. When a region's mining scale index is high, it indicates that the region may be a core production area of ​​the mining area, requiring greater attention.

[0123] Alternatively, the average monthly output of a region can be defined as the mining scale of that region.

[0124] Optionally, the mining scale index of a region satisfies the following formula:

[0125]

[0126] in, Indicates the region Mining scale index Indicates the region Average monthly mining volume; This represents the average monthly mining output across all areas of the current mining district.

[0127] S303. Determine the dynamic fusion weight of the first region based on the mining progress index and the mining scale index of the first region.

[0128] In one alternative implementation, the average of the mining progress index of the region and the mining scale index of the first region can be used as the dynamic fusion weight of the first region.

[0129] The method provided by S301-S303 can accurately reflect the relative mining intensity of each region by introducing the mining progress index and the mining scale index, and comparing the regional data with the average value of the mining area. In combination with the mining duration factor, the progress index can reflect the influence of mining efficiency.

[0130] In an implementation form of the embodiment of the application, S303 can specifically include: determining a mining activity change trend of the first region based on the change of the mining activity data of the first region within a historical time window; determining a dynamic influence of the first region based on the mining progress index, the mining scale index and the mining activity change trend of the first region; and determining a dynamic fusion weight of the first region based on the dynamic influence of the first region.

[0131] Specifically, one mining activity data includes multiple types of mining activity data, each type of mining activity data can be taken as an index, the average change rate of each index within the historical time window is determined, and then the average change rate of each index is weighted and averaged based on the average change rate of each index and the weight of each index to determine the mining activity change trend of the first region.

[0132] Optionally, the historical time window can be the last 30 days.

[0133] Optionally, the index for evaluating the mining activity change trend can include the number of mining equipment, the number of mining personnel, the daily mining time, the daily mining amount, etc.

[0134] Optionally, the change rate of each index within the historical time window can be determined, and then the average of the multiple change rates within the historical time window is determined as the average change rate.

[0135] For example, assuming that the two adjacent dates are date 1 and date 2, the number of mining equipment of date 1 can be taken as the mining equipment basis, and the ratio of the difference between the number of mining equipment of date 1 and date 2 to the number of mining equipment of date 1 can be determined as the change rate of the number of mining equipment between date 1 and date 2, which can represent the relative growth (or reduction) rate of the number of mining equipment relative to the mining equipment basis from date 1 to date 2.

[0136] Optionally, the weight of each index can be set according to the actual situation. For example, assuming that the index includes the number of mining equipment, the number of mining personnel, the daily mining time, and the daily mining amount, the weight of the number of mining equipment can be set to 0.4, the weight of the number of mining personnel can be set to 0.3, the weight of the daily mining time can be set to 0.3, and the weight of the daily mining amount can be set to 0.1.

[0137] Optionally, the dynamic influence of a region satisfies the following formula:

[0138]

[0139] wherein, represents the dynamic influence of a region , represents the trend of mining activity change of a region , represents the mining progress index of a region , represents the mining scale index of a region , represents a normalization function, for example, max-min normalization, for mapping to a stable, dimensionless numerical interval, such as [0, 1].

[0140] In an optional implementation, the dynamic influence can be determined as the dynamic fusion weight.

[0141] It can be understood that the lower the static fusion weight of a region, the poorer the foundation of the region, that is, the easier the region collapses, and the greater the terrain change intensity of the region when the disturbance caused by mining is greater. Therefore, the dynamic influence can be adjusted based on the static fusion weight to obtain the dynamic fusion weight of a region.

[0142] In an optional implementation, the terrain change intensity index of each region can be determined based on the static fusion weight of each region and the dynamic influence of each region; and then the dynamic fusion weight of the first region is determined based on the terrain change intensity index of the first region and the adjacent regions of the first region.

[0143] The terrain change intensity index is a quantitative prediction index of static-dynamic coupling risk. It is used to represent the total terrain instability possibility that a region may trigger based on its geological foundation and superimposed current mining disturbance.

[0144] Optionally, the terrain change intensity index of a region satisfies the following formula:

[0145]

[0146] wherein, represents the terrain change intensity index of a region , represents the static fusion weight of a region , represents the terrain change intensity index of a region .

[0147] In the formula, This is a dynamic impact amplification factor, which is increased by 1 to ensure it is always greater than 1, so that dynamic disturbances always amplify the impact on top of static risks. This dynamic impact amplification factor is multiplied by the static fusion weight to achieve the effect of increasing the risk level based on the geological foundation (i.e., the lower the risk level). The larger the size of the object, the more sensitive it is to external disturbances.

[0148] Alternatively, the same method can be used to determine the intensity index of topographic change for each region.

[0149] In one alternative implementation, if the terrain change intensity index of the first region is greater than the terrain change intensity indices of all adjacent regions of the first region, the maximum terrain change intensity of the adjacent regions is determined; the difference between the terrain change intensity index of the first region and the maximum terrain change intensity is determined; and the dynamic fusion weight of the first region is determined based on the difference, the terrain change intensity index of the first region, and the average terrain change intensity index of the adjacent regions.

[0150] Optionally, if the topographic change intensity index of a region is greater than the topographic change intensity indices of all its neighboring regions, the dynamic fusion weight of that region satisfies the following formula:

[0151]

[0152] in, Indicates the area Dynamic fusion weights, Indicates the area The intensity index of topographic change, Indicates the area The maximum value of the topographic change intensity index in the adjacent region, Indicates the area The number of adjacent regions, Indicates the area The The intensity index of topographic change in adjacent areas This represents a normalization function, such as max-min normalization, used to normalize... It is mapped to a stable, dimensionless numerical range, such as [0,1].

[0153] Based on this formula, it should be understood that, due to It is the maximum value in the neighboring region, therefore Reflects the region The topographic change intensity index represents the "prominence" of a region relative to its adjacent areas. When this difference is small (e.g., less than 1), it indicates that the region... The topographic change intensity index is only slightly higher than that of adjacent areas, and its characteristics as a change center are not obvious; when the difference is large (e.g., greater than 1), it indicates that the area... The topographic change intensity index is significantly higher than that of adjacent areas, making it a clear center of change.

[0154] It should be understood that if a region is only slightly higher than its neighboring regions, it should not be overemphasized as a center of change; however, if it is significantly higher than its neighboring regions, it should be given higher weight to highlight its risk. Therefore, the formula compares this difference with... Multiplication is used to achieve the following: The scaling up or down ensures that the weight adjustment and region are consistent. The relative importance is directly proportional to the relative importance.

[0155] In another alternative implementation, if the terrain change intensity index of any adjacent region of a region is greater than or equal to the terrain change intensity index of that region, the dynamic fusion weight of a region satisfies the following formula:

[0156]

[0157] in, Indicates the area Dynamic fusion weights, Indicates the area The intensity index of topographic change, Indicates the area The number of adjacent regions, Indicates the area The The intensity index of topographic change in adjacent areas This represents a normalization function, such as max-min normalization, used to normalize... It is mapped to a stable, dimensionless numerical range, such as [0,1].

[0158] In this formula, Indicates the area The average risk level of the surrounding environment; the higher the value, the higher the risk in the vicinity of the first region. The greater the likelihood of being affected, the more attention is needed.

[0159] Optionally, the area bordering the boundary of the first area can be defined as the adjacent area of ​​the first area.

[0160] It should be understood that when the topographic change intensity index of the first region is significantly higher than that of all other adjacent regions, it indicates that the first region is located at the center of topographic change, and at this time, the first region needs to be given greater attention.

[0161] In the embodiments of the present application, the coupling relationship between the geological basic conditions and the mining disturbance is accurately reflected by multiplying the static fusion weight and the dynamic influence by constructing the terrain change intensity index. Further, the dynamic fusion weight is determined by combining the terrain change intensity index of the adjacent region, which reflects the spatial propagation characteristics of the collapse risk. Through the comprehensive local features and regional relevance, the dynamic fusion weight accurately reflects the dynamic risk of the region.

[0162] It can be understood that in the process of collecting real-time dynamic data of the mining area, the complex personnel flow and environmental changes of the mining area will have a certain influence on the collected data, and with the improvement of the mining progress and the mining scale, this influence will gradually expand. When the error accumulation of the collected data reaches a certain degree, it will continuously affect the subsequent detection results of the ground collapse of the mining area. In an implementation manner of the embodiments of the present application, a time window can be established for the dynamic real-time data, the influence coefficient of error propagation is determined according to the data change trend of different types of dynamic data, and then the dynamic fusion weight is corrected. Combined with the above-mentioned S103, the dynamic fusion weight is multiplied by the static fusion weight to obtain the final fusion weight of the mining area. Figure 1 As shown in Figure 4 S104 can be implemented by the following S401-S405.

[0163] S401, acquire the dynamic real-time data of the mining area in a historical time period.

[0164] It should be understood that the historical time period includes the current time, and the historical time period is a preset time length before the current time.

[0165] It is exemplary that since the historical time period is used to divide a plurality of time windows, the historical time period should be the length of at least two time windows.

[0166] It is exemplary that the historical time period can be 48 hours, and the time window can be a time window with a scale of 24 hours.

[0167] S402, divide the historical time period into a plurality of time windows.

[0168] Among them, the plurality of time windows includes a current time window.

[0169] S403, determine the error probability of each type of dynamic data in each time window based on the difference of the plurality of types of dynamic data in the same pair of time windows.

[0170] Among them, a pair of time windows includes two adjacent time windows.

[0171] In one alternative implementation, for each type of dynamic data, a least-squares linear fit is performed on the data within a certain time window (e.g., time window 1) to obtain a fitted straight line. The slope of the fitted straight line is determined as the dynamic data change trend of that type of dynamic data within that time window. Then, the absolute value of the difference between the dynamic data change trends of time window 1 and the previous time window (e.g., time window 2) is determined as the trend change index of time window 1. Based on the mining activity change trend and trend change index of time window 1, a trend anomaly probability index is determined. Based on the difference between the trend anomaly probability index of each type and the trend anomaly probability index of other types within the same time window, the error probability of each type of dynamic data within time window 1 is determined.

[0172] It should be understood that this trend anomaly probability index is used to characterize the degree of data anomaly of a type of dynamic data within a time window.

[0173] Optionally, the method for determining the trend of mining activities over a time window is the same as or similar to the method for determining the trend of mining activities in the first region, and will not be repeated here.

[0174] Optionally, the probability index of trend anomalies within a time window satisfies the following formula:

[0175]

[0176] in, Indicates the first Various types of dynamic data in the time window The probability index of trend anomalies Indicates the first Various types of dynamic data in the time window Trend change index Indicates the first Various types of dynamic data in the time window The changing trend of mining activities express The absolute value, This represents a preset, small positive integer used to ensure that the denominator is not zero, such as 10. -8 , This represents a normalization function, such as max-min normalization, used to normalize... It is mapped to a stable, dimensionless numerical range, such as [0,1].

[0177] In this formula, for This indicates the degree of stability of changes in mining activities, when mining... When the value is relatively small, it indicates that mining activities are relatively stable. If the trend change index is relatively high, and the dynamic data shows significant fluctuations (i.e., the trend change index is high), the fluctuations are more likely to originate from anomalies rather than genuine changes, and the trend anomaly probability index is greater.

[0178] It is understandable that when the probability index of trend anomalies of a certain type of dynamic data is greater than the average probability index of trend anomalies of all types of dynamic data, it can be considered that there may be data errors.

[0179] Optionally, the mean of the trend anomaly probability index for all types of dynamic data can be determined first, and then the error probability for each type of dynamic data can be determined.

[0180] Optionally, the error probability of a type of dynamic data satisfies the following formula:

[0181]

[0182] in, Indicates the first Various types of dynamic data in the time window The error probability, Indicates the first Various types of dynamic data in the time window The probability index of trend anomalies Indicates time window The mean of the trend anomaly probability index for all types of dynamic data. This represents a normalization function, such as max-min normalization, used to normalize... It is mapped to a stable, dimensionless numerical range, such as [0,1].

[0183] In this formula, Characterizing the first The deviation of the trend anomaly probability index of a particular type of dynamic data from the average trend anomaly probability index of all types of dynamic data. The larger the deviation, the stronger the trend anomaly probability index of that particular type of dynamic data. This type of dynamic data is more likely to exhibit anomalies compared to other types of dynamic data.

[0184] S404. Based on the error probability of each type of dynamic data in the current time window, the dynamic fusion weight of the first region is corrected to obtain the dynamic fusion weight of each type of dynamic data in the current dynamic real-time data of the first region.

[0185] In one alternative implementation, the product of the error probability of each type of dynamic data and the dynamic fusion weight of the first region can be determined as the dynamic fusion weight of each type of dynamic data in the first region.

[0186] In another optional implementation, the error propagation influence coefficient of the first type of dynamic data in the current time window can be determined based on the error probability of the first type of dynamic data in the current time window and the data acquisition frequency; and the dynamic fusion weight of the first region is corrected based on the error propagation influence coefficient of the first type of dynamic data in the current time window, to obtain the dynamic fusion weight of each type of dynamic data in the current dynamic real-time data of the first region.

[0187] wherein the first type is any type included in the dynamic real-time data.

[0188] It should be understood that the data acquisition frequency reflects the time density of data update, and the faster the data acquisition frequency, the faster the error propagation.

[0189] Optionally, the error propagation influence coefficient of one type of dynamic data in one time window satisfies the following formula:

[0190]

[0191] wherein, denotes the error propagation influence coefficient of the i-th type of dynamic data in the time window T, denotes the error probability of the i-th type of dynamic data in the time window T, denotes the data acquisition frequency of the i-th type of dynamic data, denotes a normalization function, for example, a max-min normalization, for mapping to a dimensionless numerical interval, such as [0, 1]. Optionally, the product of the error propagation influence coefficient of the first type of dynamic data in the current time window and the dynamic fusion weight of the first region can be determined as the dynamic fusion weight of each data type of the first region in the current time window. It can be understood that the error propagation influence coefficient is determined in combination with the error probability and the data acquisition frequency, which takes into account both the accuracy of the data and the timeliness of the data update. The dynamic fusion weight is corrected by the error propagation influence coefficient, so that the influence of high-frequency low-quality data is effectively suppressed, and the accuracy of the fusion result and the anti-interference ability of the system are improved.

[0192]

[0193] It can be understood that the error propagation influence coefficient is determined in combination with the error probability and the data acquisition frequency, which takes into account both the accuracy of the data and the timeliness of the data update. The dynamic fusion weight is corrected by the error propagation influence coefficient, so that the influence of high-frequency low-quality data is effectively suppressed, and the accuracy of the fusion result and the anti-interference ability of the system are improved.

[0194] ​​​​S405, respectively based on the static fusion weight of the first region and the dynamic fusion weight of each type of dynamic data in the current dynamic real-time data of the first region, fusing the static geological data of the first region and each type of dynamic data in the current dynamic real-time data of the first region to obtain the ground collapse prediction result of each region.

[0195] Specifically, the dynamic fusion weight of each type of dynamic data is different, and the fusion is respectively based on the dynamic fusion weight of each type of dynamic data, which can ensure that the influence of data with high error probability is effectively inhibited in the fusion process, while the important contribution of high-quality data is retained.

[0196] The method provided by the above S401-S405 can recognize data that may have errors by analyzing the trend differences of multiple types of dynamic data, correct the dynamic fusion weight based on the error probability, reduce the influence of unreliable data in the fusion, improve the accuracy of the fused data, and further improve the accuracy of ground collapse detection.

[0197] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description, and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are also possible or can be advantageous.

[0198] Each of the embodiments in the specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other. Each embodiment focuses on the difference from other embodiments.

Claims

1. A multi-source information data fusion method for mine area ground collapse detection, characterized in that, The method comprises: Obtaining current static geological data, dynamic real-time data, historical collapse data and mining activity data of multiple regions in a mining area, wherein the static geological data comprises multiple types of static data, and the mining activity data comprises mining progress, mining scale and mining duration; Comparing the static geological data of each region with the historical collapse data to determine the static fusion weight of each region; Comparing the mining activity data of each region with the mining activity data of the multiple regions to determine the dynamic fusion weight of each region; Fusing the static geological data of each region and the dynamic real-time data of each region based on the static fusion weight of each region and the dynamic fusion weight of each region respectively to obtain the ground collapse prediction result of each region; The comparison of the static geological data of each region with the historical collapse data to determine the static fusion weight of each region comprises: Determine the anomaly index of each type of static data in the first region based on the difference between each type of static data in the static geological data of the first region and the historical collapse data, wherein the first region is any one of the multiple regions; Determine the collapse probability of the first region based on the anomaly indexes of the multiple types of static data in the first region; Determine the static fusion weight of the first region based on the collapse probability of the first region, the area proportion of the first region, and the anomaly indexes of the multiple types of static data in the first region, wherein the area proportion of the first region is the proportion of the area of the first region in the area of the mining area; The comparison of the mining activity data of each region with the mining activity data of the multiple regions to determine the dynamic fusion weight of each region comprises: Determine the mining progress index of the first region based on the mining progress and the mining duration of the first region; Determine the mining scale index of the first region based on the ratio between the mining scale of the first region and the average mining scale of the multiple regions; Determine the dynamic fusion weight of the first region based on the mining progress index of the first region and the mining scale index of the first region.

2. The multi-source information data fusion method for mine area ground collapse detection according to claim 1, characterized in that, The method further comprises: Obtaining static geological data of each coordinate point in the mining area; Determine the ground collapse risk index of each coordinate point based on the static geological data of each coordinate point; Cluster the coordinate points based on the ground collapse risk index to obtain multiple clustering clusters; Divide the mining area into multiple regions based on the multiple clustering clusters.

3. The multi-source information data fusion method for mine area ground collapse detection according to claim 1, characterized in that, The determination of the collapse probability of the first region based on the anomaly indexes of the multiple types of static data in the first region comprises: Determine the type number of abnormal data in the static geological data, wherein the abnormal data is the static data with an anomaly index greater than an anomaly index threshold in the static geological data; Determine the collapse probability of the first region based on the ratio between the type number of abnormal data and the total number of the multiple types.

4. The multi-source information data fusion method for mine area ground collapse detection according to claim 1, characterized in that, The mining activity data further comprises mining planning, and the determination of the dynamic fusion weight of the first region based on the mining progress index of the first region and the mining scale index of the first region comprises: determine a mining activity change trend of the first region based on changes in mining activity data of the first region within a historical time window; determine a dynamic influence of the first region based on the mining progress index, the mining scale index, and the mining activity change trend of the first region; determine a dynamic fusion weight of the first region based on the dynamic influence of the first region.

5. The multi-source information data fusion method for mine area ground collapse detection according to claim 4, characterized in that, The determination of the dynamic fusion weight of the first region based on the dynamic influence of the first region comprises: determine a topographic change intensity index of each region based on the static fusion weight of each region and the dynamic influence of each region; determine the dynamic fusion weight of the first region based on the topographic change intensity index of the first region and the topographic change intensity index of the adjacent regions of the first region.

6. The multi-source information data fusion method for mine area ground collapse detection according to claim 5, characterized in that, The determination of the dynamic fusion weight of the first region based on the topographic change intensity index of the first region and the topographic change intensity index of the adjacent regions of the first region comprises: determine a topographic change intensity maximum value of the adjacent regions in a case where the topographic change intensity index of the first region is greater than the topographic change intensity index of all the adjacent regions of the first region; determine a difference value between the topographic change intensity index of the first region and the topographic change intensity maximum value; determine the dynamic fusion weight of the first region based on the difference value, the topographic change intensity index of the first region, and a mean value of the topographic change intensity index of the adjacent regions.

7. The multi-source information data fusion method for mine area ground collapse detection according to claim 1, characterized in that, The dynamic real-time data comprises multiple types of dynamic data, and the fusion processing of the static geological data of each region and the dynamic real-time data of each region based on the static fusion weight of each region and the dynamic fusion weight of each region respectively obtains a ground collapse prediction result of each region, which comprises: obtain dynamic real-time data of a mining area within a historical time period; divide the historical time period into multiple time windows, the multiple time windows comprising a current time window; determine an error probability of each type of dynamic data in each time window based on differences in the multiple types of dynamic data within a same pair of time windows; correct the dynamic fusion weight of the first region based on the error probability of each type of dynamic data in the current time window to obtain a dynamic fusion weight of each type of dynamic data in the current dynamic real-time data of the first region; fuse the static geological data of the first region and each type of dynamic data in the current dynamic real-time data of the first region based on the static fusion weight of the first region and the dynamic fusion weight of each type of dynamic data in the current dynamic real-time data of the first region respectively to obtain a ground collapse prediction result of each region.

8. The multi-source information data fusion method for mine area ground collapse detection according to claim 7, characterized in that, The correction of the dynamic fusion weight of the first region based on the error probability of each type of dynamic data in the current time window to obtain a dynamic fusion weight of each type of dynamic data in the current dynamic real-time data of the first region comprises: determine an error propagation influence coefficient of the first type of dynamic data in the current time window based on the error probability of the first type of dynamic data in the current time window and a data acquisition frequency, the first type being any type comprised in the dynamic real-time data; The dynamic fusion weight of the first region is corrected based on the error propagation influence coefficient of the first type of dynamic data in the current time window, to obtain the dynamic fusion weight of each type of dynamic data in the current dynamic real-time data of the first region.

Images