A mine area comprehensive safety evaluation method and system based on multi-source remote sensing correlation analysis

Abstract

The application discloses a mine area comprehensive safety evaluation method and system based on multi-source remote sensing correlation analysis, and relates to the technical field of mine area safety monitoring and intelligent risk evaluation. The application is invented to improve the accuracy, reliability and interpretability of mine area safety supervision. The mine area comprehensive safety evaluation method based on multi-source remote sensing correlation analysis comprises the following steps: acquiring multi-source monitoring data of a target mine area; pre-processing the multi-source monitoring data, and extracting risk evidence objects from each multi-source monitoring data; performing correlation analysis on the risk evidence objects to generate a risk evidence correlation network; performing reverse reasoning on underground mining activities based on the risk evidence correlation network to construct risk events; calculating evaluation indexes of each risk event, and constructing a comprehensive safety risk evaluation model according to the evaluation indexes to obtain a risk evaluation result of the target mine area. The application is suitable for the scene of mine area comprehensive safety evaluation.

Description

Technical Field

[0001] This application relates to the field of mine safety monitoring and intelligent risk assessment technology, and in particular to a comprehensive mine safety assessment method and system based on multi-source remote sensing correlation analysis. Background Technology

[0002] Mine safety supervision mainly relies on technologies such as surface deformation monitoring, ecological environment monitoring, gas monitoring, and manual inspections. In recent years, some studies have begun to use multi-source data fusion technology for comprehensive evaluation of mine areas, constructing risk assessment models through methods such as indicator standardization, weight assignment, and spatial overlay.

[0003] However, existing technologies still have the following shortcomings: they typically analyze single monitoring information separately, focusing on surface deformation, ecological changes, or gas anomalies, lacking a collaborative analysis mechanism among multi-source risk information, making it difficult to comprehensively reflect the same underground mining activity from multiple monitoring perspectives; they lack a correlation reasoning mechanism between risk evidence, failing to establish spatial neighborhood relationships, directional consistency relationships, temporal continuity relationships, and anomaly intensity correlations between different anomaly objects, making it impossible to determine whether multiple anomalies originate from the same underground mining activity, and also making it difficult to reconstruct the risk events corresponding to underground mining activities and infer the location, scope, and direction of advance of concealed working faces; the comprehensive risk assessment results lack interpretability, making it difficult to explain the sources of correlation between different risk information, and failing to provide regulatory authorities with clear and credible enforcement basis.

[0004] Therefore, there is an urgent need for a technical solution to the problem of being unable to accurately identify and assess the risks of underground mining activities in mining areas due to the lack of a correlation and reasoning mechanism between multi-source monitoring data. Summary of the Invention

[0005] In view of this, this application provides a comprehensive safety assessment method and system for mining areas based on multi-remote sensing correlation analysis, which can solve the technical problem that it is impossible to accurately identify and assess the risks of underground mining activities in mining areas due to the lack of correlation reasoning mechanisms between multi-source monitoring data, and improve the accuracy of safety supervision in mining areas.

[0006] Firstly, this application provides a comprehensive safety assessment method for mining areas based on multi-source remote sensing correlation analysis, including:

[0007] Acquire multi-source monitoring data of the target mining area, wherein the multi-source monitoring data includes at least: surface deformation monitoring data, optical remote sensing monitoring data, greenhouse gas monitoring data, ground monitoring data, and basic geographic data of the mining area;

[0008] The multi-source monitoring data is preprocessed to extract risk evidence objects from each of the multi-source monitoring data.

[0009] A correlation analysis is performed on the risk evidence objects to generate a risk evidence correlation network;

[0010] Based on the aforementioned risk evidence association network, reverse reasoning is performed on underground mining activities to construct risk events; wherein, the risk events are used to characterize a set of risk evidence objects with direct or indirect relationships formed within the same spatiotemporal range by an underground mining activity.

[0011] Calculate the evaluation index for each risk event, and construct a comprehensive safety risk assessment model based on the evaluation index to obtain the risk assessment results for the target mining area.

[0012] Optionally, the preprocessing of the multi-source monitoring data to extract risk evidence objects from each of the multi-source monitoring data includes:

[0013] The monitoring data from different sources are preprocessed in a unified manner. The preprocessing includes data quality control, outlier removal, temporal unification, spatial unification, coordinate unification, and standardization, so that the monitoring data from different sources can be jointly analyzed under the same spatial reference system and temporal reference system.

[0014] Various abnormal regions are extracted from the monitoring data from different sources, and the abnormal regions are uniformly converted into risk evidence objects; the risk evidence objects include at least spatial location attributes, geometric attributes, time attributes, directional attributes, anomaly intensity, and evidence credibility.

[0015] Optionally, the anomaly intensity is determined based on the degree of deviation of the monitoring value of the area corresponding to the risk evidence object from the background statistical value of the area, and the calculation result is normalized to a preset numerical range; the credibility of the evidence is determined by a weighted fusion of data quality indicators, algorithm accuracy and data integrity indicators.

[0016] Optionally, the step of performing correlation analysis on the risk evidence objects to generate a risk evidence correlation network includes:

[0017] Calculate the spatial neighborhood correlation, directional consistency coefficient, temporal continuity coefficient, and anomaly strength correlation between any two of the aforementioned risk evidence objects;

[0018] The reliability of the association is determined by comprehensively considering the calculation results of the spatial neighborhood correlation degree, the directional consistency coefficient, the temporal continuity coefficient, and the anomaly strength correlation degree.

[0019] When the association credibility is greater than or equal to a preset association credibility threshold, an association edge is established between the two risk evidence objects, and all the association edges constitute the risk evidence association network.

[0020] Optionally, the step of calculating the spatial neighborhood correlation, directional consistency coefficient, temporal continuity coefficient, and anomaly strength correlation between any two risk evidence objects includes:

[0021] The spatial neighborhood correlation degree is determined based on the ratio of the distance between the center points of two risk objects to the preset influence radius. When the distance between the center points of two risk evidence objects is greater than the preset influence radius, it is determined that the two risk evidence objects do not have a spatial neighborhood correlation relationship.

[0022] The directional consistency coefficient is determined based on the angle between the respective main spatial orientations of the two risk evidence objects.

[0023] The time continuity coefficient is determined based on the ratio of the time difference between the occurrence times of the two risk evidence objects to a preset time association window;

[0024] The correlation degree of anomaly intensity is determined based on the degree of difference between the anomaly intensity of the two objects of risk evidence.

[0025] Optionally, the step of comprehensively determining the association credibility based on the calculation results of the spatial neighborhood correlation degree, the directional consistency coefficient, the temporal continuity coefficient, and the anomaly strength correlation degree includes:

[0026] The correlation confidence is obtained by weighted fusion of the spatial neighborhood correlation degree, the directional consistency coefficient, the temporal continuity coefficient, and the anomaly strength correlation degree.

[0027] Optionally, the step of constructing risk events by performing reverse reasoning on underground mining activities based on the risk evidence association network includes:

[0028] Using the association edges between the risk evidence objects in the risk evidence association network as the basis for reasoning, the risk evidence objects with direct or indirect association relationships are grouped into the same risk event.

[0029] Optionally, the step of grouping risk evidence objects with direct or indirect relationships into the same risk event based on the association edges between the risk evidence objects in the risk evidence association network as the reasoning basis includes:

[0030] Based on the connectivity of the risk evidence association network, all risk evidence objects belonging to the same connected component are clustered to construct risk events.

[0031] Optionally, the step of calculating the evaluation index for each risk event and constructing a comprehensive safety risk assessment model based on the evaluation index to obtain the risk assessment result for the target mining area includes:

[0032] The evaluation indicators include at least the abnormal intensity of the risk event, the credibility of the risk event, the completeness of the risk event, the duration of the risk, and the spatial impact range.

[0033] Optionally, the step of calculating the evaluation index for each risk event and constructing a comprehensive safety risk assessment model based on the evaluation index to obtain the risk assessment result for the target mining area specifically includes:

[0034] The evaluation indicators were weighted and fused to determine the comprehensive risk value;

[0035] The risk level of the risk event is determined based on the preset comprehensive risk value range into which the calculated comprehensive risk value falls.

[0036] Secondly, this application also provides a comprehensive safety assessment system for mining areas based on multi-source remote sensing correlation analysis, including:

[0037] The data acquisition module is used to acquire multi-source monitoring data of the target mining area, which includes at least surface deformation monitoring data, optical remote sensing monitoring data, and greenhouse gas monitoring data.

[0038] The evidence object construction module is used to preprocess the multi-source monitoring data and extract risk evidence objects from each of the multi-source monitoring data.

[0039] The association analysis module is used to perform association analysis on the risk evidence objects and generate a risk evidence association network;

[0040] The event reconstruction module is used to perform reverse reasoning on underground mining activities based on the risk evidence association network to construct risk events; wherein, the risk event is used to characterize a set of risk evidence objects with direct or indirect correlation formed within the same time and space range by an underground mining activity;

[0041] The comprehensive assessment module is used to calculate the evaluation index for each of the risk events, and to construct a comprehensive safety risk assessment model based on the evaluation index to obtain the risk assessment results for the target mining area.

[0042] This application acquires multi-source monitoring data from the target mining area, then preprocesses the data to extract risk evidence objects. This allows monitoring data, which were originally of different formats and independent, to be jointly analyzed within the same framework. This enables the comprehensive utilization of the complementary characteristics of multi-source monitoring information, reducing omissions and misjudgments caused by single monitoring data, and improving the comprehensiveness and accuracy of mine safety status perception. Furthermore, it performs correlation analysis on the risk evidence objects to generate a risk evidence correlation network, realizing a shift from traditional spatial overlay analysis to risk evidence correlation reasoning, providing a reliable basis for accurately identifying underground mining activities. Based on the risk evidence correlation network, it performs reverse reasoning on underground mining activities to construct risk events, restoring scattered single monitoring anomalies to a complete spatial organizational structure of underground mining activities. This elevates the risk assessment object from isolated monitoring anomalies to underground mining risk events with clear physical meaning. Finally, it calculates evaluation indicators for each risk event and constructs a comprehensive safety risk assessment model based on these indicators, ultimately obtaining the risk assessment results for the target mining area, improving the intelligence and precision of mine safety supervision. Therefore, this application can solve the technical problem that it is impossible to accurately identify and assess the risks of underground mining activities in mining areas due to the lack of a correlation reasoning mechanism between multi-source monitoring data. Attached Figure Description

[0043] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art 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.

[0044] Figure 1 A schematic diagram of the process for a comprehensive safety assessment method for mining areas based on multi-source remote sensing correlation analysis provided in this application embodiment;

[0045] Figure 2 This is a flowchart illustrating step S120 of the comprehensive safety assessment method for mining areas based on multi-source remote sensing correlation analysis in this application.

[0046] Figure 3 This is a flowchart illustrating step S130 of the comprehensive safety assessment method for mining areas based on multi-source remote sensing correlation analysis in this application.

[0047] Figure 4 This is a flowchart illustrating step S150 of the comprehensive safety assessment method for mining areas based on multi-source remote sensing correlation analysis in this application.

[0048] Figure 5This is a schematic diagram of the architecture of a comprehensive safety assessment system for mining areas based on multi-source remote sensing correlation analysis, provided in an embodiment of this application. Detailed Implementation

[0049] The embodiments of this application will now be described in detail with reference to the accompanying drawings.

[0050] It should be understood that the described embodiments are merely some, not all, of the embodiments in this application. All other embodiments obtained by those skilled in the art based on the embodiments in this application without inventive step are within the scope of protection of this application.

[0051] See Figure 1 The present application provides a method for comprehensive safety assessment of mining areas based on multi-source remote sensing correlation analysis, which includes the following steps.

[0052] S110. Obtain multi-source monitoring data of the target mining area, wherein the multi-source monitoring data includes at least: surface deformation monitoring data, optical remote sensing monitoring data, greenhouse gas monitoring data, ground monitoring data, and basic geographic data of the mining area.

[0053] First, multi-source monitoring data of the target mining area is acquired. This multi-source monitoring data includes at least surface deformation monitoring data, optical remote sensing monitoring data, greenhouse gas monitoring data, ground monitoring data, and basic geographic data of the mining area. Surface deformation monitoring data can be obtained through methods such as synthetic aperture radar interferometry, global navigation satellite positioning, and leveling, and is used to identify the location, extent, and trend of surface subsidence funnels in the mining area. Optical remote sensing monitoring data can be acquired through high-resolution optical satellite imagery and is used to monitor changes in vegetation cover, land use, and carbon sequestration. Greenhouse gas monitoring data can be acquired through greenhouse gas remote sensing or ground sensors and is used to monitor changes in the concentration of gases such as methane in the mining area. Ground monitoring data includes ground sensor monitoring data and leveling data. Basic geographic data of the mining area includes basic geographic information such as mining area boundaries, tunnel layout, and mining planning.

[0054] S120. Preprocess the multi-source monitoring data and extract risk evidence objects from each of the multi-source monitoring data.

[0055] See Figure 2 Step S120 further includes:

[0056] S121. Perform unified preprocessing on monitoring data from different sources. The preprocessing includes data quality control, outlier removal, time unification, spatial unification, coordinate unification, and standardization, so that the monitoring data from different sources can be jointly analyzed under the same spatial reference system and time reference system.

[0057] Specifically, monitoring data with different temporal and spatial resolutions are unified into the same event benchmark and spatial coordinate system. Data quality is checked, obvious outliers are removed, and monitoring indicators with different dimensions are standardized to make different types of data comparable in subsequent steps.

[0058] S122. Extract various abnormal regions from the monitoring data from different sources, and convert the abnormal regions into risk evidence objects in a unified manner; the risk evidence objects include at least spatial location attributes, geometric attributes, time attributes, direction attributes, anomaly intensity, and evidence credibility.

[0059] Specifically, risk evidence objects are spatial anomalies that reflect safety risks in mining areas. Each risk evidence object possesses independent attributes such as spatial location, geometric boundaries, occurrence time, anomaly intensity, and credibility. Risk evidence objects include, but are not limited to, subsidence risk evidence objects, carbon sink risk evidence objects, greenhouse gas anomaly risk evidence objects, and other anomalies that reflect the safety status of mining areas. For example, subsidence funnel areas can be extracted from surface deformation monitoring data as subsidence risk evidence objects; areas of vegetation degradation or carbon sink loss can be extracted from optical remote sensing monitoring data as carbon sink risk evidence objects; and areas of abnormal gas concentrations can be extracted from greenhouse gas monitoring data as greenhouse gas anomaly risk evidence objects.

[0060] The risk evidence object includes at least spatial location attributes, geometric attributes, temporal attributes, directional attributes, anomaly intensity, and evidence credibility. The i-th risk evidence object can be defined as:

[0061] REO i =( L i G i ,T i D i ,S i C i );

[0062] Among them, REO i For the i-th risk evidence object, L i For spatial location attributes (including center coordinates and spatial location), G i For geometric attributes (including boundaries, area, and spatial shape), T i For time attributes (including the time of occurrence and duration), D i For direction attribute, S i For abnormal intensity, C i To ensure the credibility of the evidence, all types of risk evidence are expressed using the aforementioned unified structure, and their specific monitoring attributes are standardized and categorized into the aforementioned unified attributes.

[0063] Abnormal intensity S i This is used to characterize the degree to which a risk evidence object deviates from the normal background state of the region. Let the monitoring value corresponding to the risk evidence object be X. i The mean of the regional background is μ i The background standard deviation is σ i Then the abnormal standardized value Z i for:

[0064] ;

[0065] Further mapping the normalized anomaly values ​​to a unified dimension yields the anomaly intensity S. i :

[0066] ;

[0067] In the formula, A higher value indicates a stronger degree of abnormality. For different types of monitoring data, corresponding monitoring indicators can be selected as X. i For example, deformation, carbon sink loss, abnormal gas concentration, or other indicators that can reflect the safety status of the mining area.

[0068] Credibility of evidence C i This is used to characterize the reliability of the risk evidence itself. Let Q be the data quality index for the i-th risk evidence object. i The algorithm accuracy index is P. i The data integrity metric is M. i Then the credibility of the evidence is C. i Defined as:

[0069] ;

[0070] Where a is the data quality index Q i The weighting coefficients, b is the algorithm accuracy index P. i The weighting coefficient, c, is the data integrity index M. i The weighting coefficients, and Q i P i M i All values ​​are normalized to the [0,1] interval. Higher evidence credibility indicates more reliable data quality and identification results for the risk evidence object. Data quality indicators reflect the signal-to-noise ratio, positioning accuracy, and other quality characteristics of the original monitoring data; algorithm accuracy indicators reflect the detection rate and false detection rate of the anomaly identification algorithm; data integrity indicators reflect the spatial coverage integrity and temporal sampling density of the monitoring data.

[0071] In addition, to determine the spatial correlation search range of risk evidence objects, the influence radius R of each risk evidence object is adaptively calculated based on the object's spatial scale and anomaly intensity.i :

[0072] ;

[0073] In the formula, R i Let A be the radius of influence of the i-th risk evidence object. i S represents the area of ​​the object. i It is of abnormal strength. The empirical coefficient for the area term. The empirical coefficient for the anomaly intensity term. This is an empirical coefficient for the constant term. The physical meaning of the radius of influence is: the maximum distance within which a risk evidence object can have a related impact on other risk evidence objects in space. If it exceeds this range, it is considered to be too far away in space and not within the possible influence range of the same underground mining activity.

[0074] In this embodiment, by unifying the monitoring anomalies from different sources and of different types in mining areas into risk evidence objects and establishing a unified data expression model, the object-oriented expression of multi-source monitoring data is realized, providing a unified benchmark for the collaborative analysis of different types of monitoring data. This step enables multi-source data that were originally independent and had different formats to be jointly analyzed within the same framework, fundamentally solving the problem of multi-source data heterogeneity.

[0075] S130. Perform correlation analysis on the risk evidence objects to generate a risk evidence correlation network.

[0076] See Figure 3 Step S130 specifically includes:

[0077] S131. Calculate the spatial neighborhood correlation, directional consistency coefficient, temporal continuity coefficient, and anomaly strength correlation between any two of the aforementioned risk evidence objects.

[0078] S132. The correlation confidence is determined by comprehensively considering the calculation results of the spatial neighborhood correlation degree, the directional consistency coefficient, the temporal continuity coefficient, and the anomaly strength correlation degree.

[0079] S133. When the association credibility is greater than or equal to the preset association credibility threshold, an association edge is established between the two risk evidence objects, and all the association edges constitute the risk evidence association network.

[0080] Specifically, let the risk evidence object REO be... i with REO j The relationship between them is denoted as E. ij The degree of correlation is jointly determined by spatial neighborhood relationship, directional consistency relationship, temporal continuity relationship and anomaly intensity relationship.

[0081] The spatial neighborhood correlation degree is determined based on the ratio of the distance between the center points of two risk evidence objects to a preset influence radius. Let the distance between the center points of two risk evidence objects be d. ij Risk evidence object REO i The radius of influence is R i The spatial neighborhood correlation degree is defined as:

[0082] ;

[0083] In the formula, N s (i,j) represents the spatial neighborhood correlation degree. When d ij >R i At this time, no association is established. For complex spatial objects, the shortest distance between the object boundaries can be used instead of the distance between the center points of the risk evidence objects for calculation. Spatial neighborhood correlation characterizes whether different anomalous objects are spatially adjacent, and is used to determine whether different anomalous objects may originate from the same underground mining activity. This is because different monitoring anomalies generated by the same underground mining activity (such as underground goaf areas, surface subsidence, gangue discharge outlets, and gas escape points) should be within a reasonable radius of influence in three-dimensional space, even though their spatial locations are different.

[0084] The directional consistency coefficient is determined based on the angle between the principal directions of the spatial distribution of the two risk evidence objects. Let the principal directions of the two risk evidence objects be... and Then, directional consistency is defined as:

[0085] ;

[0086] In the formula, N d (i,j) represents the directional consistency coefficient. The directional consistency coefficient ranges from [0,1], with a value of 1 when the included angle is 0° and a value of 0 when the included angle is 90°. For risk evidence objects without obvious directional attributes, their main direction can be determined based on their spatial distribution direction, center of gravity migration direction, or historical change trajectory. Directional consistency characterizes the degree of consistency among different anomalies in their spatial distribution direction and is used to evaluate the spatial organizational relationship between different anomalies. For example, the advancing direction of the same underground mining face is usually spatially consistent with the main direction of surface subsidence, the direction of gangue stacking, and the direction of gas emission. Directional consistency analysis can determine whether different anomalies follow the same spatial organizational rules.

[0087] The time continuity coefficient is determined based on the ratio of the time difference between the occurrence times of two risk evidence objects to a preset time association window. Let the occurrence times of the two risk evidence objects be t. i and t jIf the preset time-related window is T, then time continuity is defined as follows:

[0088] ;

[0089] In the formula, N t (i,j) represents the temporal continuity coefficient. For continuous monitoring data, the overlap rate of abnormal durations can be used to correct for temporal continuity. Temporal continuity characterizes the correspondence between the occurrence time and duration of different abnormal objects, and is used to determine whether multiple anomalies occur sequentially according to the mining pattern in the mining area. For example, after underground mining subsidence occurs, surface subsidence usually gradually appears over several weeks to several months, and greenhouse gas emissions may occur synchronously with mining activities or with a slight lag. This temporal sequence and regularity are important bases for determining whether different anomalies belong to the same risk event.

[0090] The correlation degree of anomaly strength is determined based on the degree of difference between the anomaly strengths of the two risk evidence objects. Let the anomaly strengths of the two risk evidence objects be S... i and S j The correlation strength of anomalies is defined as follows:

[0091] ;

[0092] In the formula, N i (i,j) represents the correlation value of anomaly intensity. The value is 1 when the intensities of two anomalies are exactly equal, and the value decreases as the difference increases. The physical meaning of the anomaly intensity correlation is that the intensity of the same underground mining activity is usually reflected in multiple monitoring indicators. If the intensity difference between two anomalies is too large, the probability that they originate from the same activity source is low.

[0093] The reliability of the association is determined by comprehensively considering the calculation results of spatial neighborhood correlation, directional consistency coefficient, temporal continuity coefficient, and anomaly strength correlation. A reliability model for association is established by integrating spatial neighborhood relationships, directional consistency relationships, temporal continuity relationships, and anomaly strength correlation relationships.

[0094] ;

[0095] In the formula, AC ij REO, the object of risk evidence i with REO j The credibility of the association between them For correlation fusion function, For spatial neighborhood correlation, The directional consistency coefficient. For time continuity coefficient, This represents the correlation strength of anomalies.

[0096] To facilitate understanding, two preferred methods for calculating association credibility are given below.

[0097] In a preferred embodiment, a weighted summation model is used to calculate the association confidence level:

[0098] ;

[0099] in, The weighting coefficient for spatial neighborhood correlation. The weighting coefficient for the time continuity coefficient. This is the weighting coefficient for the directional consistency coefficient. is the weighting coefficient for the abnormal correlation strength, and The weighted summation model is easy to calculate, and the physical meaning of the weights in each dimension is clear, making it easy to adjust according to actual application scenarios.

[0100] In another preferred embodiment, a weighted geometric mean model is used to calculate the association confidence level:

[0101] ;

[0102] When the association credibility satisfies AC ij When ≥AC0, in the risk evidence object REO i with REO j Establish connecting edges between them:

[0103]

[0104] In the formula, AC0 represents the association credibility threshold. The association credibility threshold can be set according to the actual application scenario and regulatory requirements. A higher threshold means stricter conditions for establishing associations, resulting in higher accuracy in risk event identification but potentially lower recall. Conversely, a lower threshold leads to higher recall but may introduce more false associations. All association edges that meet the conditions together constitute the risk evidence association network. In this network, nodes represent risk evidence objects, edges represent the association relationships between risk evidence objects, and the weight of each edge can be set as the association credibility value.

[0105] In this embodiment, a risk evidence association network is established by comprehensively considering the spatial neighborhood relationship, directional consistency relationship, temporal continuity relationship, and anomaly strength correlation relationship among risk evidence objects. An association credibility model is then used to quantitatively evaluate the degree of association between different risk evidence objects, thus realizing a shift from traditional spatial overlay analysis to risk evidence association reasoning. This method can effectively identify whether there is an inherent correlation between different monitored anomalies, solving the problems of missed and false positives caused by traditional methods relying solely on spatial overlap.

[0106] S140. Based on the risk evidence association network, reverse reasoning is performed on underground mining activities to construct risk events; wherein, the risk events are used to characterize a set of risk evidence objects with direct or indirect relationships formed within the same time and space range by an underground mining activity.

[0107] Specifically, using the association edges between risk evidence objects in the aforementioned risk evidence association network as the basis for reasoning, risk evidence objects with direct or indirect relationships are grouped into the same risk event. The risk event is used to characterize the set of risk evidence objects with direct or indirect relationships formed within the same spatiotemporal range during an underground mining activity. A risk event typically corresponds to a complete mining activity and can simultaneously manifest as multiple monitoring characteristics such as underground mining voids, surface subsidence, ecological disturbance, and gas anomalies. For example, an underground mining activity may simultaneously generate a mining void, an overlying surface subsidence funnel, surrounding vegetation degradation, and gas emissions. Although these monitoring anomalies do not overlap spatially, they are all responses of the same mining activity in different monitoring dimensions and should be grouped into the same risk event.

[0108] The specific implementation method for risk event reconstruction is as follows. Let the risk evidence association network be:

[0109] ;

[0110] in, A collection of risk evidence objects. This represents the set of edges connecting risk evidence objects. Based on the connectivity of this network, the risk evidence objects are clustered to obtain the k-th risk event.

[0111] ;

[0112] In the formula, RE k Let m be the number of risk evidence objects contained in the k-th risk event. A risk event consists of two or more related risk evidence objects; a single, isolated risk evidence object does not constitute a risk event. A connected component is the largest subgraph in a risk evidence association network where any two nodes are connected by a path. Risk evidence objects belonging to the same connected component are considered to originate from the same underground mining activity. After reconstructing the risk events, output the risk event set. As input for subsequent comprehensive security risk assessment.

[0113] Further determine the credibility of each risk event. To evaluate the stability of the internal correlations among risk events, calculate the average correlation credibility of the risk events:

[0114] ;

[0115] in,

[0116] ;

[0117] This represents the total number of internal related edges of a risk event.

[0118] Let the average confidence level of the data for the k-th risk event be:

[0119] ;

[0120] Let m be the credibility of the i-th risk evidence, m be the total number of risk evidence objects included in the k-th risk event, and i be the index of the risk evidence object. [1,m].

[0121] Risk event completeness is as follows:

[0122] ;

[0123] In the formula, n k The completeness represents the number of risk evidence object types actually included in the k-th risk event, where N is the total number of risk evidence object types predefined by the system (e.g., it may include three types: sedimentation risk evidence objects, carbon sink risk evidence objects, and greenhouse gas anomaly risk evidence objects). Higher completeness indicates that the risk event is more comprehensively covered by monitoring methods, and its credibility is higher.

[0124] Further calculation of the credibility of risk events:

[0125] ;

[0126] In the formula, The weighting coefficient for the average association confidence. The weighting coefficient for integrity. is the weighting coefficient for the credibility of the evidence, and RC k To determine the credibility of the k-th risk event, Let be the average confidence level of the data for the k-th risk event.

[0127] In this embodiment, based on a risk evidence association network, multiple risk evidence objects are clustered and correlated to automatically reconstruct risk events corresponding to the same underground mining activity. The location, spatial impact range, direction, duration, completeness, and credibility of the risk events are calculated, achieving a technological innovation from single anomaly identification to underground mining risk event identification. Through risk event reconstruction, the complete spatial organization structure of underground mining activities can be restored, providing strong technical support for inferring hidden working faces, cross-border mining, and illegal mining.

[0128] S150. Calculate the evaluation index for each of the risk events, and construct a comprehensive safety risk assessment model based on the evaluation index to obtain the risk assessment results for the target mining area.

[0129] See Figure 4 Step S150 specifically includes:

[0130] S151. The comprehensive risk value is determined by weighted fusion of the evaluation indicators;

[0131] S152. Determine the risk level of the risk event based on the preset comprehensive risk value range into which the calculated comprehensive risk value falls.

[0132] Specifically, after the risk event reconstruction is completed, the risk events are used as the objects of comprehensive safety risk assessment. Evaluation indicators for each risk event are calculated, and a comprehensive safety risk assessment model is constructed based on these indicators to obtain the risk assessment results for the target mining area.

[0133] The evaluation indicators include at least the abnormal intensity of the risk event, the credibility of the risk event, the completeness of the risk event, the duration of the risk, and the spatial impact range.

[0134] The anomaly intensity of a risk event reflects the comprehensive impact of underground mining activities on the surface, ecology, and environment. Assuming a risk event contains m risk evidence objects, the anomaly intensity is defined as:

[0135] ;

[0136] In the formula, I k S represents the abnormal intensity of the k-th risk event. i For the i-th risk evidence object, the anomaly strength is... The weights are assigned to the corresponding risk evidence objects. The weights for different monitoring data types can be set according to their indicative importance to mine safety. For example, abnormal surface deformation may have a higher weight, while abnormal ecological changes may have a moderate weight.

[0137] Underground mining activities have a certain duration; therefore, the duration of risk is an important component of overall risk. Let the start time of the risk event be t. s (The earliest occurrence time among all evidence objects), with an end time of t. e (The latest end time among all evidence objects), then the duration. Defined as:

[0138] ;

[0139] To facilitate comparisons between different risk events, normalization can be further applied:

[0140] ;

[0141] Among them, T max This represents the longest duration during historical monitoring. The longer the duration, the more stable and continuous the underground mining activities are, and the higher the risk level is generally.

[0142] Let the spatial range corresponding to the i-th risk evidence object be... The spatial impact range of the k-th risk event is defined as follows:

[0143] ;

[0144] The impact area of ​​the risk event is:

[0145] ;

[0146] In the formula, For the spatial impact range of risk events, A k This represents the area affected by the risk event. For risk evidence objects with spatial overlap, the actual coverage area is calculated only once. The spatial impact range reflects the spatial breadth of the impact of the mining activity on the surface and ecological environment of the mining area.

[0147] The overall risk value is determined by a weighted fusion of the aforementioned evaluation indicators. Let the overall risk value of the k-th risk event be R. k ,but:

[0148] ;

[0149] In the formula, RC k and E k Calculated by step S140, This is a comprehensive risk fusion function.

[0150] In a preferred embodiment, the overall risk value is calculated using a linear weighted model:

[0151] ;

[0152] in, The weighting coefficient for abnormal intensity. This is a weighting coefficient for the credibility of risk events. The weighting coefficient for integrity. The weighting coefficients for normalized duration, Let be the weighting coefficient for the spatial influence area, and .

[0153] ;

[0154] Each weight can be determined based on historical sample statistics, expert experience, or model training results. For example, for key mining areas, the weights of the abnormal intensity and credibility of risk events can be increased; for areas with limited regulatory capacity, the weight of the spatial impact range can be appropriately increased to prioritize the handling of large-scale risk events.

[0155] Based on the comprehensive risk value, mining area risk events are classified into different risk levels. Let the risk level thresholds be R1, R2, and R3, then:

[0156] ;

[0157] The risk levels are categorized as follows: Level I indicates low risk, Level II indicates moderate risk, Level III indicates relatively high risk, and Level IV indicates high risk. Each level corresponds to a pre-defined range of comprehensive risk values. Risk level thresholds can be determined based on different mining area types, historical regulatory experience, or statistical analysis results.

[0158] Based on the comprehensive risk value and spatial impact range of each risk event, a comprehensive risk field for the mining area is constructed:

[0159] ;

[0160] In the formula, Let K be the comprehensive risk value for any spatial location, and K be the total number of risk events. Let be the spatial weighting function for the k-th risk event. The final output is a comprehensive safety risk distribution map of the mining area and a list of risk events, providing risk screening results and spatial location data for mining area safety supervision.

[0161] In this embodiment, risk events are used as the object of comprehensive safety risk assessment. A comprehensive risk assessment model is constructed by comprehensively considering factors such as the intensity of risk event anomalies, the credibility of risk events, the completeness of risk events, the duration of risk, and the spatial impact range. This enables quantitative assessment and classification of comprehensive safety risks in the mining area. Because the assessment object is a reconstructed risk event rather than an isolated monitoring anomaly, the assessment results have clear spatial orientation and physical meaning, providing regulatory authorities with more accurate, reliable, and interpretable risk assessment results and enforcement basis.

[0162] See Figure 5 This application provides a comprehensive safety assessment system for mining areas based on multi-source remote sensing correlation analysis. It includes: Data acquisition module 51 is used to acquire multi-source monitoring data of the target mining area, wherein the multi-source monitoring data includes at least surface deformation monitoring data, optical remote sensing monitoring data and greenhouse gas monitoring data; The evidence object construction module 52 is used to preprocess the multi-source monitoring data and extract risk evidence objects from each of the multi-source monitoring data. The association analysis module 53 is used to perform association analysis on the risk evidence objects and generate a risk evidence association network; The event reconstruction module 54 is used to perform reverse reasoning on underground mining activities based on the risk evidence association network to construct risk events; wherein, the risk event is used to characterize a set of risk evidence objects with direct or indirect correlation formed within the same time and space range by an underground mining activity; The comprehensive assessment module 55 is used to calculate the evaluation index for each of the risk events, and to construct a comprehensive safety risk assessment model based on the evaluation index to obtain the risk assessment results for the target mining area.

[0163] In summary, this application utilizes multi-source monitoring data for collaborative analysis, which can more comprehensively reflect the safety status of the mining area, reduce omissions and misjudgments caused by single monitoring information, and improve the accuracy of comprehensive risk assessment results. Secondly, by leveraging the complementary characteristics between different monitoring data, it achieves joint analysis of multi-source monitoring information, improving the efficiency of multi-source data utilization and enhancing the ability to identify safety risks in complex mining areas. Furthermore, it can automatically complete comprehensive safety risk analysis of the mining area, outputting risk levels and spatial distribution of risks, providing risk screening results for regulatory authorities, reducing the workload of large-scale manual inspections, and improving the safety supervision of the mining area. This approach improves efficiency by comprehensively considering multiple risk factors to conduct risk assessments, generating risk event and risk level results. This makes the comprehensive risk assessment process more objective and the sources of risk clearer, providing a reliable basis for subsequent on-site verification. Furthermore, it is applicable to comprehensive safety risk assessments of different types of mining areas and can access different monitoring data according to actual application needs, demonstrating good versatility, scalability, and engineering application value. Finally, this application can generate comprehensive risk distribution results for mining areas, providing basic data support for mining rights management, mining planning verification, safety supervision, and on-site law enforcement, thus helping to improve the intelligence and precision of mining area safety supervision.

[0164] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0165] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A comprehensive safety assessment method for mining areas based on multi-source remote sensing correlation analysis, characterized in that, include: Acquire multi-source monitoring data of the target mining area, wherein the multi-source monitoring data includes at least: surface deformation monitoring data, optical remote sensing monitoring data, greenhouse gas monitoring data, ground monitoring data, and basic geographic data of the mining area; The multi-source monitoring data is preprocessed to extract risk evidence objects from each of the multi-source monitoring data. A correlation analysis is performed on the risk evidence objects to generate a risk evidence correlation network; Based on the aforementioned risk evidence association network, reverse reasoning is performed on underground mining activities to construct risk events; wherein, the risk events are used to characterize a set of risk evidence objects with direct or indirect relationships formed within the same spatiotemporal range by an underground mining activity. Calculate the evaluation index for each risk event, and construct a comprehensive safety risk assessment model based on the evaluation index to obtain the risk assessment results for the target mining area.

2. The comprehensive safety assessment method for mining areas based on multi-source remote sensing correlation analysis according to claim 1, characterized in that, The preprocessing of the multi-source monitoring data, and the extraction of risk evidence objects from each of the multi-source monitoring data, includes: The monitoring data from different sources are preprocessed in a unified manner. The preprocessing includes data quality control, outlier removal, temporal unification, spatial unification, coordinate unification, and standardization, so that the monitoring data from different sources can be jointly analyzed under the same spatial reference system and temporal reference system. Various abnormal regions are extracted from the monitoring data from different sources, and the abnormal regions are uniformly converted into risk evidence objects; the risk evidence objects include at least spatial location attributes, geometric attributes, time attributes, directional attributes, anomaly intensity, and evidence credibility.

3. The comprehensive safety assessment method for mining areas based on multi-source remote sensing correlation analysis according to claim 2, characterized in that, The anomaly intensity is determined based on the degree of deviation of the monitoring value of the area corresponding to the risk evidence object from the background statistical value of the area, and the calculation result is normalized to a preset numerical range; the credibility of the evidence is determined by a weighted fusion of data quality indicators, algorithm accuracy and data integrity indicators.

4. The comprehensive safety assessment method for mining areas based on multi-source remote sensing correlation analysis according to claim 1, characterized in that, The step of performing correlation analysis on the risk evidence objects to generate a risk evidence correlation network includes: Calculate the spatial neighborhood correlation, directional consistency coefficient, temporal continuity coefficient, and anomaly strength correlation between any two of the aforementioned risk evidence objects; The reliability of the association is determined by comprehensively considering the calculation results of the spatial neighborhood correlation degree, the directional consistency coefficient, the temporal continuity coefficient, and the anomaly strength correlation degree. When the association credibility is greater than or equal to a preset association credibility threshold, an association edge is established between the two risk evidence objects, and all the association edges constitute the risk evidence association network.

5. The comprehensive safety assessment method for mining areas based on multi-source remote sensing correlation analysis according to claim 4, characterized in that, The calculation of spatial neighborhood correlation, directional consistency coefficient, temporal continuity coefficient, and anomaly strength correlation between any two risk evidence objects includes: The spatial neighborhood correlation degree is determined based on the ratio of the distance between the center points of two risk evidence objects to the preset influence radius. When the distance between the center points of two risk evidence objects is greater than the preset influence radius, it is determined that the two risk evidence objects do not have a spatial neighborhood correlation relationship. The directional consistency coefficient is determined based on the angle between the respective main spatial orientations of the two risk evidence objects. The time continuity coefficient is determined based on the ratio of the time difference between the occurrence times of the two risk evidence objects to a preset time association window; The correlation degree of anomaly intensity is determined based on the degree of difference between the anomaly intensity of the two objects of risk evidence.

6. The comprehensive safety assessment method for mining areas based on multi-source remote sensing correlation analysis according to claim 4, characterized in that, The step of comprehensively determining the association credibility based on the calculation results of the spatial neighborhood correlation degree, the directional consistency coefficient, the temporal continuity coefficient, and the anomaly strength correlation degree includes: The correlation confidence is obtained by weighted fusion of the spatial neighborhood correlation degree, the directional consistency coefficient, the temporal continuity coefficient, and the anomaly strength correlation degree.

7. The comprehensive safety assessment method for mining areas based on multi-source remote sensing correlation analysis according to claim 1, characterized in that, The process of constructing risk events by reverse reasoning about underground mining activities based on the aforementioned risk evidence association network includes: Based on the connectivity of the risk evidence association network, all risk evidence objects belonging to the same connected component are clustered to construct risk events.

8. The comprehensive safety assessment method for mining areas based on multi-source remote sensing correlation analysis according to claim 1, characterized in that, The calculation of evaluation indicators for each risk event, and the construction of a comprehensive safety risk assessment model based on the evaluation indicators to obtain the risk assessment results for the target mining area, includes: The evaluation indicators include at least the abnormal intensity of the risk event, the credibility of the risk event, the completeness of the risk event, the duration of the risk, and the spatial impact range.

9. The method for comprehensive safety assessment of mining areas based on multi-source remote sensing correlation analysis according to claim 1, characterized in that, The calculation of evaluation indicators for each risk event, and the construction of a comprehensive safety risk assessment model based on the evaluation indicators to obtain the risk assessment results for the target mining area, specifically includes: The evaluation indicators are weighted and fused to determine the comprehensive risk value; The risk level of the risk event is determined based on the preset comprehensive risk value range into which the calculated comprehensive risk value falls.

10. A comprehensive safety assessment system for mining areas based on multi-source remote sensing correlation analysis, characterized in that, include: The data acquisition module is used to acquire multi-source monitoring data of the target mining area. The multi-source monitoring data includes at least surface deformation monitoring data, optical remote sensing monitoring data, greenhouse gas monitoring data, ground monitoring data, and basic geographic data of the mining area. The evidence object construction module is used to preprocess the multi-source monitoring data and extract risk evidence objects from each of the multi-source monitoring data. The association analysis module is used to perform association analysis on the risk evidence objects and generate a risk evidence association network; The event reconstruction module is used to perform reverse reasoning on underground mining activities based on the risk evidence association network to construct risk events; wherein, the risk event is used to characterize a set of risk evidence objects with direct or indirect correlation formed within the same time and space range by an underground mining activity; The comprehensive assessment module is used to calculate the evaluation index for each of the risk events, and to construct a comprehensive safety risk assessment model based on the evaluation index to obtain the risk assessment results for the target mining area.

Images