Multi-modal fusion monitoring method and system for closing coal mine surface fissures and environmental factors

Abstract

The invention provides a multi-modal fusion monitoring method and a multi-modal fusion monitoring system for closing coal mine surface fissures and environmental factors. The method specifically comprises the steps that remote sensing images, gas concentration, fracture parameters, earth surface displacement and vibration information of a closed coal mine area are collected; constructing a multi-objective function, performing multi-modal data feature fusion on the remote sensing images, the gas concentration, the fracture parameters, the surface displacement and the vibration information to generate a unified feature matrix, solving the multi-objective function by using the feature matrix, and calculating risk index values of the remote sensing images, the gas concentration, the fracture parameters, the surface displacement and the vibration information; and predicting the risk according to the risk index values of the remote sensing image, the gas concentration, the fracture parameters, the surface displacement and the vibration information. According to the invention, the timeliness, the accuracy and the coverage range of crack monitoring can be improved, scientific early warning is provided in time, and the safety and the environment stability after the mining area is closed are guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of closed coal mine monitoring, and in particular to a multi-modal fusion monitoring method and system for closed coal mine surface fissures and environmental factors. Background Art

[0002] Closed coal mines refer to coal mines that have been closed and abandoned due to depletion of coal resources, non-compliance with safe mining conditions or policies, etc. The coal pillars left in the closed coal mines, unmined coal seams, and coal blocks in the goaf continue to desorb and gush out. Under the influence of underground air pressure and water level changes, gas and other toxic and harmful gases can easily leak to the ground through channels such as the shaft, mining cracks in the overburden, and drainage outlets. After the coal mine is closed, the formation and expansion of surface cracks may pose a threat to surface buildings, infrastructure, and the environment. Existing monitoring methods include optical remote sensing and geological radar monitoring, but they often have problems such as limited monitoring range, delayed data analysis, and incomplete depth information. Summary of the invention

[0003] In order to overcome the defects existing in the above-mentioned prior art, the purpose of the present invention is to provide a multi-modal fusion monitoring method and system for closing coal mine surface fissures and environmental factors.

[0004] In order to achieve the above-mentioned object of the present invention, the present invention provides a multi-modal fusion monitoring method for closing coal mine surface fissures and environmental factors, comprising the following steps:

[0005] Collect remote sensing images, gas concentration, fracture parameters, surface displacement and vibration information of closed coal mine areas;

[0006] Construct a multi-objective function, which includes one or any combination of the following objective functions:

[0007] Objective function 1 aims to maximize the accuracy of crack propagation prediction while reducing the significant abnormal area of ​​crack propagation speed;

[0008] Objective function 2 is to minimize the number of abnormal points of gas concentration and reduce the amplitude of abnormal fluctuation;

[0009] Objective function three, which aims to minimize the spatial heterogeneity of surface subsidence;

[0010] Objective function 4, which aims to minimize the impact of closing a coal mine on the surrounding environment;

[0011] The multimodal data features of remote sensing images, gas concentration, crack parameters, surface displacement and vibration information are fused to generate a unified feature matrix, the feature matrix is ​​used to solve the multi-objective function, and the risk index values ​​of each remote sensing image, gas concentration, crack parameter, surface displacement and vibration information are calculated; the risk is predicted based on the risk index values ​​of remote sensing images, gas concentration, crack parameters, surface displacement and vibration information.

[0012] In this multimodal fusion monitoring method of surface fissures and environmental factors in closed coal mines, objective function one optimizes the fissure distribution model based on fissure parameters and expansion trends; objective function two optimizes the gas extraction layout by combining concentration data with fluctuation variance; objective function three extracts point deviations from settlement data and remote sensing images to reduce the risk of local collapse; objective function four evaluates the vegetation restoration effect through NDVI time series changes to reduce degradation trends; by integrating multi-source data through multi-objective functions, the timeliness, accuracy and coverage of fissure monitoring are improved, scientific warnings are provided in a timely manner, and safety and environmental stability are guaranteed after the mine is closed.

[0013] Optionally, it is characterized by extracting surface subsidence and NDVI changes from remote sensing images; deploying microseismic sensors in closed coal mine areas to collect surface vibration information; deploying inclination sensors in closed coal mine areas to collect surface displacement; using geological radar modules to collect fracture parameters; and using gas sensors to collect gas concentrations in closed coal mine areas.

[0014] Optionally, objective function 1 is: Among them, n1 is the total number of crack points, L i is the actual length of the crack at the i1th crack point; is the predicted crack length of the i1th crack point; Δv is the crack width change rate, Δt is the time interval for crack extension monitoring; w1 is the penalty weight for abnormal crack change;

[0015] The second objective function is: Where n2 is the number of gas sensor points, c safe is the safety concentration threshold; w2 is the penalty weight for abnormal gas fluctuations, c i is the gas concentration at the i2th gas sensor point, Var(C) is the variance of the gas concentration;

[0016] The third objective function is: Among them, n3 is the total number of settlement points, Δh i is the settlement value of the i3th settlement point; is the average settlement value;

[0017] The objective function four is: Among them, n4 is the total number of vegetation monitoring points, T is the length of the time series, and NDVIi,t is the normalized vegetation index of the i4th vegetation monitoring point at time t.

[0018] Optionally, the solution method for the multi-objective function is:

[0019] Dynamically adjust the multi-objective weights of the multi-objective function:

[0020] Each objective function corresponds to a specific weight, and the weight of each objective function is calculated through data fluctuations;

[0021] Based on the weights, each objective function is linearly weighted to merge the multiple objective functions into a multi-objective fitness function;

[0022] The multi-objective fitness function is solved to obtain the risk index values ​​of remote sensing images, gas concentration, crack parameters, surface displacement and vibration information.

[0023] Optionally, K-means is used to cluster the risk indicators of remote sensing images, gas concentration, fracture parameters, surface displacement and vibration information, and in each iteration, a local search operation is introduced.

[0024] Optionally, the risk includes crack expansion trend risk, which is predicted as follows:

[0025] If at least one of the displacement and vibration frequency exceeds the respective preset thresholds, the crack expansion trend enters the first-level warning state;

[0026] If both the displacement and vibration frequency exceed their respective preset thresholds, the crack expansion trend enters the second-level warning.

[0027] Optionally, calculate the comprehensive risk value R:

[0028] R=a1×D+a2×F f

[0029] Among them, D is the displacement monitoring value, F f is the vibration frequency, a1 and a2 are weight parameters, satisfying a1+a2=1;

[0030] When R exceeds the preset risk threshold, the risk of crack expansion trend enters the second-level warning.

[0031] The present application also proposes a closed coal mine area monitoring system, including a remote sensing imaging module for acquiring images of the closed coal mine area, a geological radar module for detecting the depth and location of cracks, and a microseismic sensor and an inclination sensor for acquiring micro displacement and vibration information of the surface;

[0032] The remote sensing image module, geological radar module, microseismic sensor and inclination sensor are all electrically connected to the processing module, and respectively send to the processing module the closed coal mine area images, crack depth and position, surface micro-displacement and vibration information collected by themselves; the processing module monitors the surface cracks of the closed coal mine according to the above-mentioned multimodal fusion monitoring method of closed coal mine surface cracks and environmental factors.

[0033] Optionally, the remote sensing image module includes a drone and / or a satellite;

[0034] The UAV regularly flies to collect images of closed coal mine areas; and regularly obtains satellite images of closed coal mine areas from satellites.

[0035] Optionally, the drone flies according to a planned flight route, and the planned flight route covers key fissure areas and possible hidden danger points in the closed coal mine area;

[0036] The drone flies according to a planned flight route, and the planned flight route covers key fissure areas and possible hidden danger points in the closed coal mine area;

[0037] The drone also dynamically adjusts the flight plan route based on real-time monitoring data, giving priority to high-risk areas, which are areas where surface vibration parameters, surface displacement, crack parameters, and gas concentration exceed corresponding thresholds.

[0038] The beneficial effects of the present invention are:

[0039] 1. Improve monitoring accuracy and timeliness

[0040] By integrating multiple data sources such as remote sensing images, geological radar and sensor networks, it is possible to achieve all-round and real-time monitoring of surface cracks, improve the accuracy and timeliness of monitoring data, and overcome the limitations that may exist in a single monitoring method.

[0041] 2. Comprehensive early warning mechanism and risk assessment

[0042] By setting multiple thresholds and trigger mechanisms, combining real-time data and machine learning algorithms, the risk of crack development can be comprehensively assessed, and when key parameters are monitored to exceed the standard, an early warning can be automatically triggered.

[0043] 3. Promote the improvement of mining area environment and safety management

[0044] By achieving efficient crack monitoring and timely early warning, the system helps to improve the level of environmental protection and safety management in mining areas, reduce the risk of accidents caused by the expansion of surface cracks, contribute to the sustainable closure and abandonment management of mining areas, and ensure the environmental stability of mining areas.

[0045] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the description of the embodiments in conjunction with the following drawings, in which:

[0047] Figure 1 It is a schematic diagram of the process of the present invention. DETAILED DESCRIPTION

[0048] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be understood as limiting the present invention.

[0049] In the description of the present invention, unless otherwise specified and limited, it should be noted that the terms "installed", "connected" and "connected" should be understood in a broad sense. For example, it can be a mechanical connection or an electrical connection, or it can be the internal connection between two components. It can be a direct connection or an indirect connection through an intermediate medium. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to the specific circumstances.

[0050] like Figure 1 As shown, the present invention provides a multi-modal fusion monitoring method for closing coal mine surface fissures and environmental factors, which specifically includes the following steps:

[0051] Collect remote sensing images, gas concentration, fracture parameters, surface displacement and vibration information of closed coal mine areas.

[0052] In this embodiment, a UAV flight route is planned in the closed coal mine area, and the UAV flight is started regularly to collect images to obtain remote sensing images of the closed coal mine area; satellite data can also be obtained regularly to obtain remote sensing images of the closed coal mine area; a gas sensor is used to collect the gas concentration in the closed coal mine area; a geological radar module is used to detect and identify cracks and crack parameters, including but not limited to the location, depth, length, width, etc. of the cracks; a sensor network is installed in the closed coal mine area, including but not limited to microseismic sensors and inclination sensors, to obtain surface vibration information and surface displacement in real time, the surface vibration information includes but is not limited to microseismic frequency (whether there is energy release), amplitude (intensity of vibration), duration, and location, and the surface displacement includes but is not limited to surface horizontal displacement, surface vertical displacement, displacement location, displacement time, and surface inclination.

[0053] Construct a multi-objective function, which includes one or any combination of the following objective functions:

[0054] Objective function 1 aims to maximize the accuracy of crack propagation prediction while reducing the significant abnormal area of ​​crack propagation speed.

[0055] The objective function 1 is: Among them, n1 is the total number of crack points, L i is the actual length of the crack at the i1th crack point; is the predicted crack length of the i1th crack point; Δv is the crack width change rate, Δt is the time interval for crack extension monitoring; w1 is the penalty weight for abnormal crack change.

[0056] Objective function 2 aims to minimize the number of abnormal points in gas concentration and reduce the amplitude of abnormal fluctuations.

[0057] The second objective function is: Where n2 is the number of gas sensor points, c safe is the safety concentration threshold; w2 is the penalty weight for abnormal gas fluctuations, c i is the gas concentration at the i2th gas sensor point, Var(C) is the variance of the gas concentration, Refers to the fluctuation range of the gas concentration at the i2th gas sensor point.

[0058] Objective function III with the goal of minimizing the spatial heterogeneity of surface subsidence.

[0059] The third objective function is: Among them, n3 is the total number of settlement points, Δh i is the settlement value of the i3th settlement point; is the average sedimentation value.

[0060] Objective function 4 aims to minimize the impact of closing a coal mine on the surrounding environment.

[0061] The objective function four is: Among them, n4 is the total number of vegetation monitoring points, T is the length of the time series, and NDVI i,t is the normalized vegetation index of the i4th vegetation monitoring point at time t.

[0062] The multi-modal data features of remote sensing images, gas concentration, crack parameters, surface displacement and vibration information are fused to generate a unified feature matrix X. In this embodiment, feature alignment and dimensionality reduction algorithms are used to fuse the multi-modal data to generate a unified feature matrix X.

[0063] In this embodiment, the characteristic matrix X=[crack target parameters, surface settlement data, NDVI changes, gas leakage risk parameters].

[0064] Among them, NDVI changes can be extracted from remote sensing images. Fracture target parameters include the collected fracture parameters and the parameters obtained after calculation and processing through the fracture parameters, including fracture expansion rate, etc. The difference is that the fracture target parameters can quantify the fracture expansion trend and abnormal characteristics, while the collected fracture parameters only reflect the fracture geometry characteristics at a single moment. Surface subsidence data can be extracted from remote sensing images, and can also be obtained after calculation and processing of the collected surface displacement. The surface subsidence data extracted from remote sensing images is used as the overall surface displacement trend information of the closed coal mine area, and the surface displacement collected by the inclination sensor is used to monitor local changes or short-term dynamic characteristics near the fracture.

[0065] Gas leakage risk parameters include the collected gas concentration and the fluctuation range and gas concentration variance of the gas concentration obtained after gas concentration calculation.

[0066] The multi-objective function is solved using the characteristic matrix X to calculate the risk index values ​​of each remote sensing image, crack parameter, surface displacement and vibration information. The specific solution method is as follows:

[0067] (1) Dynamically adjust the weights of multi-objective functions

[0068] Each objective function corresponds to a specific weight, and the weight of each objective function is calculated by data fluctuation to reflect the priority of different objective functions at a specific time t. In this embodiment, it is preferred but not limited to use the entropy weight method to dynamically adjust the weight, and the formula for dynamically adjusting the weight is:

[0069]

[0070] Where n is the total number of objective functions, w i is the objective function f i The weight, H i is the objective function f i The information entropy of m is the target function f i The total number of data points collected at a specific time t used to calculate the weight of objective function 2. For example, when calculating the weight of objective function 2, m = 100 gas concentration data are collected at a specific time t for calculation; k is the kth data point currently being calculated; x i,k is the objective function f i The value of the kth data point, j is used to sum the total value of all data points during normalization, x i,j It refers to the objective function f i The value of the jth data point.

[0071] When a parameter fluctuates violently (such as abnormal fluctuations in gas concentration), a higher weight is assigned to the corresponding objective function; when the parameter is stable (such as uniformity of settlement), the weight of the corresponding objective function is lower. For example: when the crack expansion is abnormal, the weight of objective function one increases, and high-risk crack areas are optimized first; when the gas concentration fluctuates more, the weight of objective function two increases, and priority is given to monitoring areas with excessive concentrations; when the environment recovers and stabilizes, the weight of objective function four decreases, reducing the allocation of computing resources for the ecological restoration goal.

[0072] (2) Constructing a multi-objective fitness function

[0073] Based on the weights, each objective function is linearly weighted using a linear weighting method, and each objective function is fused into a multi-objective fitness function F:

[0074]

[0075] Wherein, n is the number of objective functions, and in this embodiment, n=4.

[0076] The multi-objective fitness function F is solved to obtain the risk index value based on remote sensing images, crack parameters, surface displacement and vibration information. In this embodiment, the Pareto frontier solution set X is solved based on the improved NSGA-II or MOEA / D algorithm. * :

[0077] X * =Optimize(F(X))

[0078] Where F(X) is the multi-objective fitness function.

[0079] Clustering and local search operations are used to improve the diversity of solutions. Specifically,

[0080] (1) Use K-means to cluster the solutions and evaluate the target coverage of the current group.

[0081] (2) In each iteration, the distribution breadth of the solution is improved by introducing a local search operation (L-BFGS).

[0082] The risk is predicted based on the risk index values ​​of remote sensing images, gas concentration, crack parameters, surface displacement and vibration information to determine whether to trigger an early warning. When the risk index values ​​of remote sensing images, gas concentration, crack parameters, surface displacement and vibration information exceed their corresponding risk thresholds, the corresponding risk exists and the corresponding early warning is issued.

[0083] Taking the crack parameters as an example, this embodiment mainly focuses on the surface displacement and vibration frequency, and the corresponding change thresholds can be preset:

[0084] 1. The surface displacement threshold is set by, but not limited to, the following methods: Based on historical monitoring data and engineering experience, the typical fracture displacement threshold can be set to a cumulative displacement of no more than 5 mm per hour or day. For areas with higher risks, this threshold can be reduced to 2 mm / day.

[0085] 2. The vibration frequency threshold is set by, but not limited to, the following methods: If the frequency of microseismic activity detected by the sensor network exceeds 50 Hz continuously and is accompanied by a microseismic signal with an amplitude higher than 0.2g, an early warning should be triggered.

[0086] If at least one of the surface displacement risk index value and the vibration frequency risk index value obtained by solving the multi-objective fitness function F(X) exceeds their respective preset change thresholds, the crack expansion trend enters the first-level warning state.

[0087] If the surface displacement risk index value and the vibration frequency risk index value both exceed their respective change thresholds, the crack expansion trend enters the second-level warning.

[0088] In this embodiment, if at least one of the collected surface displacement and vibration frequency exceeds the respective preset change thresholds, the crack expansion trend may also enter a first-level warning state.

[0089] In an optional solution of this embodiment, the prediction accuracy can be further improved by calculating the comprehensive risk value R.

[0090] R=a1×D+a2×F f

[0091] Among them, D is the surface displacement monitoring value, F f is the vibration frequency, a1 and a2 are weight parameters, satisfying a1+a2=1.

[0092] When R exceeds the preset risk threshold, the crack expansion trend enters the second-level warning.

[0093] After the warning is triggered, a report can be automatically generated, including timestamp, exceeded parameters, crack expansion chart and predicted impact range. The report is sent to engineering and technical personnel and mining area management departments through the network, and emergency measures are recommended, such as reinforcing the crack area, evacuating personnel, etc., and launching drones to conduct dynamic inspections of the area corresponding to the warning.

[0094] During the entire monitoring process, through multi-objective optimization algorithms, such as dynamic weight adjustment, linear weighted optimization monitoring resource allocation, etc., deep learning models (such as RNN, CNN) are used to identify and predict the expansion of cracks, predict the future trend of target weights, achieve more efficient dynamic fusion of objective functions, and provide intelligent crack development trend analysis, which can more accurately assess the risk of crack expansion, provide early warning, and prevent the occurrence of sudden safety problems. During the optimization process, the model parameters are optimized according to the collected data and feedback, and the monitoring strategy is adjusted in real time.

[0095] Embodiment 2

[0096] The present application also proposes an embodiment of a system for shutting down a coal mine area monitoring system.

[0097] In this embodiment, shutting down the coal mine area monitoring system includes:

[0098] Remote sensing imaging module for obtaining images of closed coal mine areas, geological radar module for detecting crack parameters, and sensor module for obtaining surface displacement and vibration information;

[0099] The remote sensing image module, the geological radar module, and the sensor module are all electrically connected to the processing module, and respectively send their collected closed coal mine area images, crack parameters, surface displacement and vibration information to the processing module; the processing module monitors the surface cracks of the closed coal mine according to the multimodal fusion monitoring method of closed coal mine surface cracks and environmental factors described in Example 1.

[0100] In this embodiment, the remote sensing image module includes a drone and / or a satellite; the drone regularly flies to collect images of the closed coal mine area, and the drone flies according to a planned flight route.

[0101] The UAV flight route planning strategy is as follows:

[0102] Global coverage inspection

[0103] The planned flight route covers key fissure areas and possible hidden danger points in the closed coal mine area. In this embodiment, the regional block strategy is used to divide the mining area into multiple rectangular grids (each with a side length of 50m), and the flight altitude is set to 100m (wide-area coverage monitoring), the camera angle is 90° wide angle, and the flight speed is 5m / s. A serpentine path is used to scan block by block to collect visible light images and multispectral images to monitor initial cracks on the surface and evaluate areas of abnormal soil moisture.

[0104] Inspection of key areas

[0105] The drone regularly inspects and identifies active fissure areas, and dynamically adjusts the flight plan route based on real-time monitoring data, giving priority to high-risk areas, which are areas where surface vibration parameters, surface displacement, fissure parameters, and gas concentration exceed the corresponding thresholds. In this embodiment, refined monitoring is performed on identified fissures and high-risk areas. The inspection parameters are set by analyzing previous inspection data and marking active fissure areas. Set the flight altitude to 50m (refined monitoring), the camera angle to 30°~45°, and the flight speed to 2m / s. The flight adopts a spiral flight strategy from the center of the fissure to the periphery of the fissure, collecting thermal imaging data and fissure morphology data, which are used to detect the temperature difference of the fissure to determine whether the fissure is active and analyze the changes in the depth and width of the fissure.

[0106] Dynamic response inspection

[0107] The drone can also perform multi-target monitoring of crack morphology, surrounding settlement, and environmental factors (such as humidity and gas). In this embodiment, when the sensor module detects local microseismicity, the crack expansion rate exceeds the preset threshold, and the sensor equipment detects abnormal gas concentration, the drone takes off from the standby area, and the inspection parameters are dynamically adjusted (flight altitude 30-70m, flight speed 3-7m / s) to the target area along the shortest path (based on the Dijkstra algorithm), and uses an "X"-shaped cross flight strategy to scan the target area, collect real-time images and gas concentration, temperature and humidity data, provide direct data to the monitoring center, and assist in determining the cause of the event.

[0108] Random inspection supplement

[0109] In order to make up for the global coverage inspection and avoid missing the scanning area, the drone can also perform random inspections as a supplement. In this embodiment, the flight altitude is set to 80m and the flight speed is set to 4m / s, and a random inspection path is used to cover non-key areas.

[0110] Imagery of closed mine areas is also regularly acquired from satellites.

[0111] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0112] Although the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the claims and their equivalents.

Claims

1. A multi-modal fusion monitoring method for closing coal mine surface fissures and environmental factors, characterized in that: The following steps are involved: Collect remote sensing images, gas concentration, fracture parameters, surface displacement and vibration information of closed coal mine areas; Construct a multi-objective function, which includes one or any combination of the following objective functions: Objective function 1 aims to maximize the accuracy of crack propagation prediction while reducing the significant abnormal area of ​​crack propagation speed; Objective function 2 is to minimize the number of abnormal points of gas concentration and reduce the amplitude of abnormal fluctuation; Objective function three, which aims to minimize the spatial heterogeneity of surface subsidence; Objective function 4, which aims to minimize the impact of closing a coal mine on the surrounding environment; The multimodal data features of remote sensing images, gas concentration, crack parameters, surface displacement and vibration information are fused to generate a unified feature matrix, the feature matrix is ​​used to solve the multi-objective function, and the risk index values ​​of each remote sensing image, gas concentration, crack parameter, surface displacement and vibration information are calculated; the risk is predicted based on the risk index values ​​of remote sensing images, gas concentration, crack parameters, surface displacement and vibration information.

2. The multi-modal fusion monitoring method for closing coal mine surface fissures and environmental factors according to claim 1 is characterized in that: Extract surface subsidence and NDVI changes from remote sensing images; deploy microseismic sensors in closed coal mine areas to collect surface vibration information; deploy inclination sensors in closed coal mine areas to collect surface displacement; use geological radar modules to collect fracture parameters; and use gas sensors to collect gas concentrations in closed coal mine areas.

3. The multi-modal fusion monitoring method for closing coal mine surface fissures and environmental factors according to claim 1 is characterized in that: The objective function 1 is: Where n1 is the total number of crack points, Li is the actual length of the crack at the i1th crack point; is the predicted crack length of the i1th crack point; Δv is the crack width change rate, Δt The time interval for monitoring crack extension; w1 is the penalty weight for abnormal changes in cracks; The second objective function is: Among them, n2 is the number of gas sensor points, csafe is the safety concentration threshold; w2 is the penalty weight for abnormal gas fluctuation, ci is the gas concentration at the i2th gas sensor point, Var(C) is the variance of the gas concentration; The third objective function is: Among them, n3 is the total number of settlement points, Δhi is the settlement value of the i3th settlement point; is the average settlement value; The objective function four is: Among them, n4 is the total number of vegetation monitoring points, T is the length of the time series, and NDVI i,t is the normalized vegetation index of the i4th vegetation monitoring point at time t.

4. The multi-modal fusion monitoring method for closing coal mine surface fissures and environmental factors according to claim 1 is characterized in that: The solution method for the multi-objective function is: Dynamically adjust the multi-objective weights of the multi-objective function: Each objective function corresponds to a specific weight, and the weight of each objective function is calculated through data fluctuations; Based on the weights, each objective function is linearly weighted to merge the multiple objective functions into a multi-objective fitness function; The multi-objective fitness function is solved to obtain the risk index values ​​of remote sensing images, gas concentration, crack parameters, surface displacement and vibration information.

5. The multi-modal fusion monitoring method for closing coal mine surface fissures and environmental factors according to claim 4 is characterized in that: K-means is used to cluster the risk indicators of remote sensing images, gas concentration, crack parameters, surface displacement and vibration information, and a local search operation is introduced in each iteration.

6. The multi-modal fusion monitoring method for closing coal mine surface fissures and environmental factors according to claim 1 is characterized in that: The risks include the risk of crack expansion tendency, which is predicted as follows: If at least one of the displacement and vibration frequency exceeds the respective preset thresholds, the crack expansion trend enters the first-level warning state; If both the displacement and vibration frequency exceed their respective preset thresholds, the crack expansion trend enters the second-level warning.

7. The multi-modal fusion monitoring method for closing coal mine surface fissures and environmental factors according to claim 6 is characterized in that: Calculate the comprehensive risk value R: <h2 style=";text-align:left;direction:ltr">R = a1×D+a2F<h2 style=";text-align:left;direction:ltr"> f Among them, D is the displacement monitoring value, F f is the vibration frequency, a1 and a2 are weight parameters, satisfying a1+a2=1; When R exceeds the preset risk threshold, the risk of crack expansion trend enters the second-level warning.

8. A closed coal mine area monitoring system, characterized in that: It includes remote sensing imaging modules for obtaining images of closed coal mine areas, geological radar modules for detecting the depth and location of cracks, and microseismic sensors and inclination sensors for obtaining surface displacement and vibration information; The remote sensing image module, geological radar module, microseismic sensor and inclination sensor are all electrically connected to the processing module, and respectively send to the processing module the closed coal mine area images, crack depth and position, surface micro-displacement and vibration information collected by themselves; the processing module monitors the surface cracks of the closed coal mine according to the multimodal fusion monitoring method of closed coal mine surface cracks and environmental factors as described in any one of claims 1-7.

9. The closed coal mine area monitoring system according to claim 8, characterized in that: The remote sensing image module includes a drone and / or a satellite; The UAV regularly flies to collect images of closed coal mine areas; and regularly obtains satellite images of closed coal mine areas from satellites.

10. The closed coal mine area monitoring system according to claim 9, characterized in that: The drone flies according to a planned flight route, and the planned flight route covers key fissure areas and possible hidden danger points in the closed coal mine area; The drone also dynamically adjusts the flight plan route based on real-time monitoring data, giving priority to high-risk areas, which are areas where surface vibration parameters, surface displacement, crack parameters, and gas concentration exceed corresponding thresholds.

Images