A method, device, equipment and storage medium for predicting surface movement and deformation in a mining area based on airborne LiDAR

Abstract

The present invention belongs to the field of geological exploration, and specifically relates to a method, device, equipment and storage medium for predicting surface movement and deformation in a mining area based on airborne LiDAR, including collecting the original point cloud data of the surface of the target mining area; extracting the point cloud data of non-growing fixed ground objects based on the echo characteristics of airborne LiDAR; traversing the point cloud data of non-growing fixed ground objects after preprocessing by using a moving window method to determine the surface characteristic parameters, determining several subsidence monitoring points in the target mining area and obtaining their working face parameters, and performing probability integral method parameter inversion in combination with a genetic algorithm to obtain probability integral parameters; constructing a settlement prediction model with the probability integral parameters, surface characteristic parameters, and working face parameters for surface deformation prediction. The present invention realizes the settlement monitoring of non-growing fixed ground objects on the mountain surface and the rapid calculation of characteristic parameters of any point on the mountain surface, and has certain reference value and practical significance for mountain terrain subsidence monitoring and prediction.

Description

Technical Field

[0001] The present invention belongs to the field of geological exploration, and particularly relates to a method, device, equipment and storage medium for predicting surface movement and deformation in a mining area based on airborne LiDAR. Background Art

[0002] With the continuous growth of China's demand for coal resources, coal mining activities have gradually shifted to the western mining areas with more complex geological conditions. However, the terrain in mining areas is undulating and the geological structure is complex. Underground coal mining not only easily causes damage to surface buildings (structures), but also easily triggers secondary disasters such as landslides and collapses, seriously expanding the scope and degree of mining impacts, and such disasters are highly unpredictable. According to statistics, more than one-third of coal mines in China are located in mining areas, which makes the contradiction between coal mining in mining areas and surface safety increasingly sharp.

[0003] Then, the existing surface subsidence monitoring technologies still have the following defects:

[0004] 1. Currently, surface subsidence monitoring is mainly divided into two types: point data monitoring and surface data monitoring. Point data monitoring mainly includes leveling, trigonometric leveling and GPS leveling. However, these methods have significant limitations in mining area applications: Traditional leveling: In mining areas, due to complex terrain, the measurement difficulty is large, the monitored positions are limited, and there is a high risk. GPS leveling: Affected by the terrain and vegetation occlusion in mining areas, the GPS signal is relatively weak, and the monitored area and accuracy are greatly limited. Surface data monitoring methods such as InSAR and lidar, although they can achieve the monitoring of two-dimensional surface data, also face challenges: InSAR technology: The imaging conditions are easily affected by the environment, and the accuracy for large deformation monitoring is low. Lidar technology: Although it has high accuracy, comprehensive data collection and strong flexibility, the monitoring data volume is large and the data quality evaluation is difficult, which restricts its wide application.

[0005] 2. Some achievements have been made in the research on the mechanism and prediction of surface movement and deformation caused by coal mining in mining areas. However, although the existing prediction models have high calculation efficiency and simple prediction functions, they require a large number of accurate surface characteristic parameters, and the acquisition of these parameters is complex and time-consuming, which limits the wide application of the models. Summary of the Invention

[0006] The purpose of the present invention is to provide a method, device, equipment and storage medium for predicting surface movement and deformation in a mining area based on airborne LiDAR to solve the problems raised in the background art.

[0007] The present invention achieves the above purpose through the following technical solutions:

[0008] First aspect, the present invention provides a method for predicting surface movement and deformation in a mining area based on airborne LiDAR, and the method includes:

[0009] Scanning a target mining area using airborne LiDAR technology to collect original point cloud data of the surface of the target mining area;

[0010] Processing the point cloud data based on the echo characteristics of airborne LiDAR to extract point cloud data of non-growing fixed ground objects;

[0011] Using the moving window method to traverse the point cloud data of non-growing fixed ground objects after preprocessing to determine surface characteristic parameters, where the surface characteristic parameters include the tilt direction angle, the surface point dip angle, and the surface feature coefficient;

[0012] Determining several subsidence monitoring points in the target mining area based on the point cloud data of non-growing fixed ground objects and obtaining their working face parameters, and performing parameter inversion of the probability integral method in combination with the genetic algorithm to obtain probability integral parameters;

[0013] Constructing a settlement prediction model with the probability integral parameters, the surface characteristic parameters, and the working face parameters, and using the settlement prediction model to predict surface deformation of the point cloud data of the mining area surface obtained in real time.

[0014] Further, the processing of the point cloud data to extract point cloud data of non-growing fixed ground objects includes:

[0015] Dividing the surface point cloud data into grids and recording the number of point clouds and the number of echo points in each grid;

[0016] Setting threshold conditions, regarding the grids that meet the threshold conditions as non-fixed ground objects and filtering them using the PCL slope method to extract point cloud data of non-growing fixed ground objects, where the growing fixed ground objects include roads, building tops, and bare soil.

[0017] Further, the using the moving window method to traverse the point cloud data of non-growing fixed ground objects after preprocessing to determine surface characteristic parameters includes: using a moving window with a set grid spacing to traverse the point cloud data of non-growing fixed ground objects after preprocessing, and calculating the tilt direction angle, the surface dip angle, and the surface feature coefficient in each window, as shown in the following formula:

[0018] (1) Tilt direction angle Obtaining:

[0019]

[0020] (2) Surface dip angle α x,y Obtaining:

[0021]

[0022] (3) Surface feature coefficient D x,y Obtain:

[0023] D x,y =-200*((d + f + b + h) / 2 - 2*e) / ds 2

[0024] In the above formula, a, b, c, d, e, f, g, h, i are the minimum values of the point cloud elevation within the corresponding grid, ds is the grid spacing, and dx and dy are the differentials along the x and y directions respectively.

[0025] Furthermore, determining several subsidence monitoring points in the target mining area based on the point cloud data of non-growing fixed ground objects and obtaining their working face parameters, and combining with the genetic algorithm to perform parameter inversion of the probability integral method to obtain probability integral parameters, including:

[0026] Comparing the point cloud data of non-growing fixed ground objects at different time points, analyzing the deformation of the ground surface, identifying the subsidence area and determining several subsidence monitoring points and obtaining their working face parameters;

[0027] Taking the working face parameters of the subsidence monitoring points as input data for parameter inversion of the probability integral method to solve the probability integral parameters of several subsidence monitoring points.

[0028] Furthermore, the working face parameters include corner points, coal seam dip angle, and mining thickness; the probability integral parameters include subsidence coefficient q, horizontal movement coefficient b, main influence angle tangent tanβ, mining propagation angle θ, and left upper - right lower inflection point offset s.

[0029] Furthermore, constructing a settlement prediction model with the probability integral parameters, the surface characteristic parameters, and the working face parameters, and using the settlement prediction model to predict the surface deformation of the real - time obtained mining area surface point cloud data, including:

[0030] Loading a working face prediction file containing probability integral parameters, surface characteristic parameters, and working face parameters in the data processing system, and initializing the settlement prediction model according to the parameters in the prediction file;

[0031] Using an airborne LiDAR monitoring device to regularly obtain the real - time surface original point cloud data of the mining area, inputting the pre - processed real - time point cloud data into the settlement prediction model, simulating the surface deformation situation of the mining activities in real - time in the data processing system, and outputting the surface deformation prediction results including the surface settlement map and the horizontal movement map.

[0032] In the second aspect, the present invention proposes a device for predicting surface movement and deformation in a mining area based on airborne LiDAR, and the device includes:

[0033] A data acquisition module, which is used to scan a target mining area using airborne LiDAR technology and collect the original point cloud data of the surface of the target mining area;

[0034] A data processing module, which is used to process the point cloud data based on the echo characteristics of airborne LiDAR and extract the point cloud data of non-growing fixed ground objects;

[0035] A parameter extraction module, which is used to traverse the point cloud data of non-growing fixed ground objects after preprocessing by using the moving window method to determine the surface characteristic parameters, and the surface characteristic parameters include the tilt direction angle, the surface point dip angle, and the surface feature coefficient;

[0036] A parameter inversion module, which is used to determine several subsidence monitoring points in the target mining area based on the point cloud data of non-growing fixed ground objects and obtain their working face parameters, and combine the genetic algorithm to perform parameter inversion of the probability integral method to obtain the probability integral parameters;

[0037] A real-time prediction module, which is used to construct a settlement prediction model with the probability integral parameters, the surface characteristic parameters, and the working face parameters, and use the settlement prediction model to predict the surface deformation of the real-time acquired mining area surface point cloud data.

[0038] Furthermore, the data acquisition module includes a drone equipped with an airborne LiDAR device, and the drone is used to perform aerial scanning on the target mining area and transmit the acquired original point cloud data of the mining area surface to the data processing module in real time.

[0039] In a third aspect, the present invention proposes an electronic device, including:

[0040] A processor; a memory for storing instructions executable by the processor;

[0041] Wherein, the processor is configured to execute the instructions to implement the prediction method as described in any one of the above.

[0042] In a fourth aspect, the present invention proposes a computer-readable storage medium, when the instructions in the computer-readable storage medium are executed by the processor of an electronic device, the electronic device can execute the prediction method as described in any one of the above.

[0043] The beneficial effects of the present invention are as follows:

[0044] In view of the problems of large amount of data obtained by airborne LiDAR and difficulty in evaluating data quality, the present invention proposes a method for extracting point clouds of non-growing fixed ground objects in mountainous areas based on echo characteristics, thereby realizing the settlement monitoring of non-growing fixed ground objects on the mountainous surface and improving the monitoring accuracy; and in view of the fact that a large number of surface characteristic parameters are required for the surface movement and deformation prediction model in mining areas, a point cloud calculation method based on a moving window is proposed, realizing the rapid calculation of characteristic parameters of any point on the mountainous surface, which has certain reference value and practical significance for mountain terrain subsidence monitoring and prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 It is a schematic flow chart of a method for predicting surface movement and deformation in mining areas based on airborne LiDAR in the present invention.

[0046] Figure 2 It is another schematic flow chart of a method for predicting surface movement and deformation in mining areas based on airborne LiDAR in the present invention.

[0047] Figure 3 It is a schematic diagram of multiple echoes of the airborne LiDAR system in the specific embodiment of the present invention.

[0048] Figure 4 It is the overlaid point cloud and echo ratio threshold diagram of Experiment 1 in the specific embodiment of the present invention; wherein, (a) is the overlaid point cloud of Experiment 1, (b) is the echo ratio threshold of 0%, (c) is the echo ratio threshold of 10%, and (d) is the echo ratio threshold of 20%.

[0049] Figure 5 It is the overlaid point cloud and echo ratio threshold diagram of Experiment 2 in the specific embodiment of the present invention; wherein, (a) is the overlaid point cloud of Experiment 2, (b) is the echo ratio threshold of 0%, (c) is the echo ratio threshold of 10%, and (d) is the echo ratio threshold of 20%.

[0050] Figure 6 It is a schematic diagram of a virtual grid in the specific embodiment of the present invention.

[0051] Figure 7 It is a schematic diagram of a moving window in the virtual grid in the specific embodiment of the present invention.

[0052] Figure 8 It is a relative position diagram of the working face and the monitoring point in the actual processing routine of the specific embodiment of the present invention.

[0053] Figure 9 It is a distribution map of the extraction of point clouds of non-growing fixed ground objects on the working face surface in the actual processing routine of the specific embodiment of the present invention.

[0054] Figure 10This is a comparison chart of the actual measured subsidence and the subsidence extracted from point clouds in the actual processing routine of the specific embodiment of the present invention.

[0055] Figure 11 This is a point distribution map of the point cloud extracted by inverting the parameters of the mined working face in the actual processing routine of the specific embodiment of the present invention.

[0056] Figure 12 This is a distribution map of the leveling monitoring points obtained by inverting the parameters of the mined working face in the actual processing routine of the specific embodiment of the present invention.

[0057] Figure 13 This is the predicted fitting subsidence curve of the points extracted from the point cloud in the parameter solution in the actual processing routine of the specific embodiment of the present invention.

[0058] Figure 14 This is the predicted fitting subsidence curve of the leveling monitoring points in the parameter solution in the actual processing routine of the specific embodiment of the present invention.

[0059] Figure 15 This is the surface dip distribution map in the distribution map of the surface characteristic parameters and the predicted subsidence contour lines in the actual processing routine of the specific embodiment of the present invention.

[0060] Figure 16 This is the distribution map of the inclination direction angle in the distribution map of the surface characteristic parameters and the predicted subsidence contour lines in the actual processing routine of the specific embodiment of the present invention.

[0061] Figure 17 This is the distribution map of the surface characteristic coefficient in the distribution map of the surface characteristic parameters and the predicted subsidence contour lines in the actual processing routine of the specific embodiment of the present invention.

[0062] Figure 18 This is the predicted subsidence contour line distribution map in the distribution map of the surface characteristic parameters and the predicted subsidence contour lines in the actual processing routine of the specific embodiment of the present invention. Specific Embodiment

[0063] The following further describes the present application in detail with reference to the accompanying drawings. It is necessary to point out here that the following specific embodiments are only used to further illustrate the present application and cannot be construed as limiting the protection scope of the present application. Those skilled in the art can make some non-essential improvements and adjustments to the present application based on the above application content.

[0064] Example 1

[0065] As Figure 1-2 shown, this embodiment proposes a method for predicting the surface movement and deformation in a mining area based on airborne LiDAR (Light Detection and Ranging), and the method includes the following steps:

[0066] S1. Use airborne LiDAR technology to scan the target mining area and collect the original point cloud data of the surface of the target mining area;

[0067] S2. Process the point cloud data based on the echo characteristics of airborne LiDAR and extract the point cloud data of non-growing fixed ground objects;

[0068] S3. Use the moving window method to traverse the point cloud data of non-growing fixed ground objects after preprocessing to determine the surface characteristic parameters, where the surface characteristic parameters include the tilt direction angle, the surface point dip angle, and the surface characteristic coefficient;

[0069] S4. Determine several subsidence monitoring points in the target mining area based on the point cloud data of non-growing fixed ground objects and obtain their working face parameters, and combine with the genetic algorithm to perform parameter inversion of the probability integral method to obtain the probability integral parameters;

[0070] S5. Construct a settlement prediction model with the probability integral parameters, the surface characteristic parameters, and the working face parameters, and use the settlement prediction model to predict the surface deformation of the real-time obtained point cloud data of the mining area surface.

[0071] In this embodiment, due to the complex terrain in mountainous areas, traditional filtering methods are difficult to effectively handle this special terrain. At the same time, the study of surface deformation laws requires multi-period data, and the existence of time intervals will lead to tree growth and crop alternation, thus further reducing the reliability of the point cloud processing results. In contrast, non-growing fixed ground objects in the mining area (such as planar data such as roads, building tops, and bare soil) are less affected by time differences, so they are more reliable as basic data. Therefore, this embodiment proposes a method for extracting the point cloud of non-growing fixed ground objects based on the echo characteristics of airborne LiDAR.

[0072] It can be understood that in combination with Figure 3 , the basic principle of airborne LiDAR technology is to calculate the distance of the target object by emitting laser pulses and recording the time difference from the emission to the reception of the laser pulses. Laser has strong penetrability. When the laser irradiates the ground, it will penetrate some ground objects and generate laser pulse signals with multiple reflections. According to factors such as the type, structure, and material of the ground objects, there are significant differences in the reflection frequencies. For example, when the laser signal penetrates vegetation, multiple echoes are usually generated; while when the laser signal irradiates a smooth surface building or road, usually only a single echo is generated. By analyzing the echo information, it can help to judge the type of point cloud.

[0073] Further preferably, the point cloud data is processed to extract the point cloud data of non-growing fixed ground objects, including: dividing the ground surface point cloud data into grids, and recording the number of point clouds and the number of echo points in each grid; setting threshold conditions (echo ratio threshold), and regarding the grids that meet the threshold conditions as non-fixed ground objects and filtering them using the PCL slope method to extract the point cloud data of non-growing fixed ground objects, where the growing fixed ground objects include roads, building tops, and bare soil.

[0074] More specifically, combined with Figure 4 , according to the echo characteristics of the point cloud, first divide the point cloud into grids, record the total number of point clouds M in the grid, and judge the number of echo points N in the grid, and calculate the proportion it occupies. If the proportion reaches a certain threshold (N / M≥10%), then this grid is regarded as a non-fixed ground object and filtered. According to the probability integral method, the trend of ground subsidence is continuous. At the same time, according to the third law of geography (the more similar the geographical environment, the more similar the geographical target characteristics), when expressing the subsidence trend, the subsidence amount of the mining area surface in a certain area can be regarded as the same. And the virtual grid is composed of adjacent small area grids, so the subsidence amount in the grid is generally the same, so the grid can be averaged.

[0075] The determination of the echo ratio threshold is based on the following Experiment 1 and Experiment 2:

[0076] The key to extracting the point cloud data of non-growing ground objects lies in determining the echo ratio threshold in the grid. To determine the echo ratio threshold, an area near a teaching building of a certain university in the plain area was selected as the research object for the experiment (Experiment 1), and the LiDAR point cloud was orthogonally acquired using the Pegasus D2000. The total number of points is 8,569,590, the average point cloud density is 95 points / m2, and the proportion of the first echo is 94.51%. As Figure 4 (a) shows, the experimental area contains non-growing fixed ground objects such as roads, building tops, and bare soil. Echo ratio thresholds of 0%, 10%, and 20% are set for the experimental area to extract the point clouds of non-growing fixed ground objects, and the results are shown in Figure 4 (b), (c), (d).

[0077] The results of Experiment 1 show that: echo ratio thresholds of 0%, 10%, and 20% can all achieve the extraction of fixed ground objects, but some bare soil data will be lost under the condition of 0%, and the data extracted from 10% to 20% tends to be the same and the accuracy is comparable. Therefore, considering the time cost and space complexity of the calculation, it is considered that setting the echo ratio threshold at 10% is more reasonable in the plain area.

[0078] Considering that different topographic conditions may affect the setting of the echo ratio threshold, in the experiment, LiDAR point clouds obtained above a working face in a mountainous area (Experiment 2) were further selected. The total number of points is 4,527,915, the average point cloud density is 15 points / m2, and the proportion of the first echo is 95.13%. As Figure 5 shown in (a), the experimental area includes non-growing fixed ground features such as snow-covered roads and bare soil. The echo ratio thresholds of 0%, 10%, and 20% are also set for extracting the point clouds of non-growing fixed ground features, and the results are shown in Figure 5 (b), (c), and (d).

[0079] The results of Experiment 2 show that: when the echo ratio threshold is 0%, the extraction of the snow-covered road is incomplete, and some bare soil data are missing; when the echo ratio threshold is 10%, the extraction of the snow-covered road is basically complete, and there is also some loss of bare soil data; when the condition is 20%, the extraction of the snow-covered road is more complete, and the loss of some bare soil data is less, but the amount of extracted data does not change significantly compared with the case of 10%. Therefore, setting the echo ratio threshold to 10% under the mountainous topographic conditions can still meet the basic extraction of the point clouds of non-growing fixed ground features.

[0080] In summary, when the echo ratio threshold is set at 10%, for both plain areas and mountainous areas, ground features such as shrubs and crops in the point clouds can be removed, so as to obtain data of non-growing fixed ground features such as roads, building tops, and bare soil.

[0081] In this embodiment, the point cloud data of the mining area monitoring range is obtained to determine the range of the subsidence basin. Specifically, the initial (original) point cloud data is processed with a small-scale grid spacing (ds = 1m), and combined with the echo characteristics of airborne LiDAR and the PCL slope method, the point cloud data of non-growing fixed ground features such as roads, building tops, and bare soil is obtained, and the difference between the two-phase fixed ground feature data is calculated to obtain the subsidence basin.

[0082] It can be understood that in the case of mining horizontal and gently inclined coal seams with an average slope angle of the hillside less than 30°, the subsidence value W s (x,y) of any point in the mountainous area can be predicted by referring to the following formula:

[0083]

[0084] In the formula, W(x,y) refers to the subsidence displacement of any point (x,y) on the flat ground under the same geological and mining technical conditions. D x,y is the surface characteristic coefficient of the point (x,y). P[x] and P[y] are the slip influence functions of the strike and dip main sections. φ is the inclination direction angle (°) of the point (x,y). α x,y is the inclination angle (°) of the surface point (x,y).

[0085] Among them, the slip influence function of the point (x, y) along the x-direction:

[0086]

[0087] The slip influence function of the point (x, y) along the y-direction:

[0088]

[0089] In the above formula, r is the main influence radius, W max is the maximum subsidence value under flat ground conditions, and A, t, P are slip influence parameters.

[0090] It can be seen from the above that the prerequisite for accurately predicting the surface movement and deformation in mountainous areas is not only to obtain the predicted parameters of the prediction model, but also the surface characteristic parameters (surface characteristic coefficient D x,y , dip direction angle surface point dip angle α x,y ) need to be accurately calculated.

[0091] Combined with Figure 7 , in order to simply and quickly obtain the surface characteristic parameters required for the above mountainous area prediction, according to the concept of virtual grid in 1.1, a 3*3 moving window is used to traverse the LiDAR point cloud to calculate the mountainous area prediction parameters, where a, b, c, d, e, f, g, h, i are the minimum values of the point cloud elevation within the corresponding grid, and ds is the grid spacing. Since the measurement area is much larger than the working face range, the problem of undetected monitoring of the boundary of the surface movement basin caused by underground coal mining can be avoided.

[0092] Further preferably, the point cloud data of non-growing fixed ground objects after preprocessing is traversed by the moving window method to determine the surface characteristic parameters, including: using a moving window with a set grid spacing to traverse the point cloud data of non-growing fixed ground objects after preprocessing, and calculating the dip direction angle, surface dip angle and surface characteristic coefficient within each window, as shown in the following formula:

[0093] (1) Dip direction angle Obtain:

[0094]

[0095] (2) Surface dip angle α xy Obtain:

[0096]

[0097] (3) Surface characteristic coefficient D x,y Obtain:

[0098] D x,y=-200 * ((d + f + b + h) / 2 - 2 * e) / ds 2

[0099] In the above formula, a, b, c, d, e, f, g, h, and i are the minimum values of the elevation of the point cloud within the corresponding grid, ds is the grid spacing, and dx and dy are the differentials along the x and y directions respectively.

[0100] The surface characteristic coefficient of the mountainous area is divided into the following four categories according to its topsoil and ground vegetation characteristics, as shown in Table 1. It can be seen that the topsoil and ground vegetation characteristics have concave and convex landform classifications. Combining with the definition of curvature in ArcGIS, that is, positive values represent convex landforms, negative values represent concave landforms, and 0 represents flat land.

[0101] Table 1 Surface Characteristic Coefficient D of the Mountainous Area x,v Classification Table

[0102]

[0103] Combined with the point cloud echo characteristics in this embodiment, the ratio of multiple echoes (N / M) is used as a parameter to judge the ground vegetation characteristics, corresponding to types I, II, III, and IV in the surface types at 10%, 20%, 30%, and 40% respectively.

[0104] Further preferably, based on the point cloud data of non-growing fixed ground objects, several subsidence monitoring points in the target mining area are determined and their working face parameters are obtained. Combining with the genetic algorithm, parameter inversion of the probability integral method is carried out to obtain the probability integral parameters, including:

[0105] Compare the point cloud data of non-growing fixed ground objects at different time points, analyze the deformation of the surface, identify the subsidence area, determine several subsidence monitoring points and obtain their working face parameters; use the working face parameters of the subsidence monitoring points as input data for parameter inversion of the probability integral method to solve the probability integral parameters of several subsidence monitoring points.

[0106] Further preferably, the working face parameters include corner points, coal seam dip angle, and mining thickness; the probability integral parameters include subsidence coefficient q, horizontal movement coefficient b, main influence angle tangent tanβ, mining propagation angle θ, and left upper - right lower inflection point offset s.

[0107] Further preferably, a settlement prediction model is constructed with the probability integral parameters, surface characteristic parameters, and working face parameters, and the settlement prediction model is used to predict the surface deformation of the real - time obtained mining area surface point cloud data, including:

[0108] Load the working face prediction file containing the probability integral parameters, surface characteristic parameters, and working face parameters in a data processing system (such as a host computer system), and initialize the settlement prediction model according to the parameters in the prediction file;

[0109] Regularly obtain the real-time surface original point cloud data of the mining area using an airborne LiDAR monitoring device, input the preprocessed real-time point cloud data into the settlement prediction model, simulate the surface deformation situation of the mining activities in real time in the data processing system, and output the surface deformation prediction results including the surface settlement map and the horizontal movement map.

[0110] Combined with Figure 2 , according to the above embodiments of the present invention, by processing the point cloud data obtained by airborne LiDAR, combined with the rapid acquisition method of the mountain surface movement deformation model and its characteristic parameters, an integrated scheme for mountain subsidence monitoring and prediction is constructed, and the detailed steps are as follows:

[0111] (1) Obtain the point cloud data of the monitoring range of the mining area and determine the range of the subsidence basin. Process the initial (original) point cloud data with a small-scale grid spacing (ds = 1m), and combine the echo characteristics of airborne LiDAR and the PCL slope method to obtain the point cloud data of non-growing fixed ground objects such as roads, building tops, and bare soil. Take the difference between the two-phase fixed ground object data to obtain the subsidence basin.

[0112] (2) Calculate the surface characteristic parameters. First, set the grid spacing, and traverse the grid data through a moving window to calculate the surface characteristic parameters, namely the surface dip angle, dip direction angle, and surface characteristic coefficient.

[0113] (3) Obtain the working face information. Consult the data to obtain the working face parameters, such as corner points, coal seam dip angle, mining thickness, etc.

[0114] (4) Extract the settlement data of non-growing ground objects and parameter inversion. Select the reliable bare point cloud data from the subsidence basin and use the genetic algorithm for probability integral method parameter inversion.

[0115] (5) Predict the surface movement and deformation of coal mining under the mountain. Combine the obtained probability integral method parameters, surface characteristic parameters, and working face parameters to form a working face prediction file, and predict the subsidence of the working face.

[0116] The above prediction method will be further elaborated in detail below with a practical processing routine.

[0117] 1. Overview of the study area and data collection:

[0118] The study area is located in a certain mining area in Jincheng, Shanxi. The average mining depth is about 675m. The coal seam mined is No. 3 coal seam, with an average dip angle of 5°, a strike length of 1113m, a dip length of 165m, and an average mining height of 2.82m. The fully caving method is used for mining in the working face. The terrain is medium and low mountains, mainly mountainous in the north, there is an east-west mountain ridge in the middle, and mainly gentle mountains and hilly areas in the south. The maximum height difference within the working face is about 200m, and the vegetation is dense.

[0119] Due to the dense vegetation and complex terrain above the working face, it is difficult to lay the monitoring line on the main section. Therefore, it is dispersed along the mountain road around the working face. The layout of the monitoring line is as Figure 8 shown: Line al is laid along the dirt road in the strike direction, deviating about 200 m from the main section; Line c is laid along the dirt road in the strike direction and crosses the working face obliquely, with points located in the middle and north of the working face; Line z is laid along the bare rock in the strike direction, with points located in the middle of the working face; Line a is laid along the strike, mainly concentrated in the middle and south; Line bl is laid along the hardened road in the dip direction, with points located in the southwest of the working face; Lines bb and bc are both laid along the dip and are located in the east of the working face. A total of 168 monitoring points are laid, and the spacing between the measuring points is 50 m. The conventional observation period is one month. Among them, the elevation is based on the third-class leveling, and the XY coordinates are based on GNSS RTK measurement.

[0120] The airborne LiDAR data is orthogonally acquired by the Pegasus D2000 quadcopter drone, and the coordinates of the image control points are obtained through static measurement. The specific information collected is shown in Table 2. The mining time of the working face is August 2023. As of January 20, 2024, the working face has advanced 510 m cumulatively.

[0121] Table 1 Information Table for LiDAR Data Collection of the Working Face

[0122]

[0123]

[0124] 2. Verification of Point Cloud Extraction of Non-growing Fixed Ground Objects Based on the Echo Characteristics of Airborne LiDAR:

[0125] From Figure 8 the image map, it can be seen that there is a large amount of vegetation above the working face, and there are also some hardened roads, bare rocks without vegetation coverage, and buildings far from the working face. Based on these discrete fixed ground objects, the two-phase LiDAR point cloud data was processed according to the above point cloud processing method, and the results are as Figure 9 shown.

[0126] From Figure 9 it can be seen that the southwest of the working face is a residential area, and it can be clearly seen that the roads and roofs have been separated. There is a highway in the southeast of the working face, and it can be seen that part of the highway has been completely extracted. It can be concluded that under the complex mountain terrain conditions, the point cloud extraction of fixed ground objects can be realized through the echo characteristics of airborne LiDAR.

[0127] To further verify the reliability of the data, the subsidence values near the monitoring points were selected and compared with the subsidence values obtained from the actual leveling measurement. Considering that there is no point cloud data for some of the monitoring points, data verification was carried out on 71 monitoring points with point cloud data, and the final results are as follows Figure 8 .

[0128] As Figure 10 shown, (1) Generally speaking, the cumulative mean square error of the extraction accuracy is 35 mm, and the maximum subsidence of the leveling monitoring is about 800 mm (the actual maximum subsidence value of the surface is larger due to the layout of the monitoring points), and the relative error is only 4.4%, meeting the engineering requirements. (2) In terms of accuracy, for the same layout along the dirt road, the cumulative mean square error of line al is 34 mm, and the cumulative mean square error of line c is 51 mm. The accuracy of line c is significantly lower, while the accuracy of line al is higher. Combining with the surface image map ( Figure 8 ), it can be found that the vegetation coverage rate of line c is higher than that of line al. Therefore, the increase in vegetation will interfere with the extraction accuracy of fixed ground objects. (3) The cumulative mean square error of line bl is 20 mm, and the cumulative mean square error of line z is 22 mm; By comparing each observation line, it can be seen that for lines z and bl laid along bare rocks and hardened roads, their accuracy is significantly better than that of lines al and c laid along the dirt road. The reason is that the vegetation coverage rate of bare rocks and hardened roads is low, and on the other hand, the echo reflectivity of bare rocks and hardened roads is better than that of the dirt road.

[0129] In summary, the point cloud extraction method of non-growing fixed ground objects based on the echo characteristics of airborne LiDAR can realize the extraction of non-growing fixed ground objects on the surface, and the accuracy meets the engineering requirements. When the vegetation coverage rate near non-growing fixed ground objects is relatively high, it will lead to a decrease in the extraction accuracy. And the point cloud extraction accuracy along bare rocks and hardened roads is better than that along the dirt road. Therefore, it is recommended to select non-growing fixed ground objects with low vegetation coverage rate and high reliability such as bare rocks and hardened roads for data research in point cloud extraction.

[0130] 3. Parameter inversion

[0131] To further study the influence of the settlement of the monitoring points on the deformation of the mined working face, comparative experiments were carried out on the subsidence values obtained from the leveling monitoring and the subsidence values extracted from the point cloud respectively. As Figure 11 shown, a total of 31 points were selected for the point cloud extraction points, 22 points (1 - 22) were selected for the strike, and 9 points (23 - 31) were selected for the dip; As Figure 10 shown, a total of 33 points were selected for the leveling monitoring points, 21 points (z1 - z9, al1 - al10, al8 - 2, al8 - 3) were selected for the strike, and 12 points (bl17 - bl21, bl23 - bl29) were selected for the dip.

[0132] As Figure 11 and 12As shown in the figure, in terms of inversion data selection, the points extracted from the point cloud can be selected according to the main section, while due to topographical reasons, only the points near the working face can be selected for the leveling points. Therefore, detecting ground deformation through airborne LiDAR has more advantages than traditional monitoring methods in the detection of mining subsidence centers.

[0133] The genetic algorithm is used to solve the parameters of the prediction model. To avoid the contingency of the parameter solving results, the algorithm runs 5 times, and the average value of the 5 running results is used as the final result. The results are shown in Table 3. The comparison diagram of the predicted fitting subsidence curves in the parameter solving is shown in Figure 13 and 14 , and the absolute values of the errors in the figure are the differences between the two groups of data.

[0134] Table 2 Comparison diagram of the inversion results of the parameters of the subsidence prediction model in mountainous areas

[0135]

[0136] Combined with Table 3, Figure 13 and 14 it can be seen that: (1) In terms of the fitting accuracy of subsidence, the mean error of the fitting of the leveling monitoring data is 77 mm, while the mean error of the fitting of the point cloud monitoring data is 100 mm. Relative to the maximum subsidence values of their respective monitoring points, which are 800 mm and 1300 mm, the relative errors are 9.6% and 7.6% respectively, and the fitting accuracies are quite comparable; (2) Looking at the inversion results of the parameters of the two different monitoring data, it can be found that the subsidence coefficient inverted from the point cloud monitoring data is much larger than that of the leveling monitoring data. This not only indicates that the lack of the maximum surface subsidence value will lead to a smaller inversion result of the subsidence coefficient, but also further shows that the parameter values inverted using the airborne LiDAR monitoring data are more in line with the actual situation.

[0137] Therefore, compared with traditional monitoring methods, the airborne LiDAR monitoring technology is not only not restricted in monitoring, but its monitoring data has better applications for later regular research and deformation prediction, and can provide more reliable data support for later parameter inversion and deformation prediction.

[0138] 4. Calculation of surface characteristic parameters and deformation prediction

[0139] The LiDAR point cloud data collected in the first phase is selected to calculate the surface characteristic parameters of the mountainous area. The collection area is 3 km long and 2 km wide. The grid spacing of the surface prediction points is set to 10 m (ds = 10 m), and the ground prediction range is a rectangular area. According to the above content, the surface characteristic parameters obtained during subsidence prediction are as shown in Figure 15 , 16 and 17.

[0140] Table 3 Comparison diagram of subsidence boundaries

[0141]

[0142]

[0143] The parameters of the probability integral method obtained by the above parameter inversion, as well as the parameters of the mountainous surface characteristics obtained from the LiDAR point cloud, are used to predict the deformation of the working face. The subsidence contour lines are as Figure 18 shown. It can be seen that the subsidence center is located at the center of the mined area, and the maximum influence boundary of the working face is near points bc9, bb10, a10, and bl15. As of January 20, 2024, the comparison of the above subsidence boundary according to the leveling monitoring results is shown in Table 4. As shown in Table 4, for edge detection, the cumulative mean square error is only 3.2 mm.

[0144] In summary, the characteristic parameters of any point on the mountainous surface can be quickly calculated through the moving window, and at the same time, the subsidence of the working face is predicted based on the mountainous surface movement model. The results show that it is consistent with the measured leveling monitoring data in terms of edge detection, proving that the prediction method can achieve relatively reliable subsidence prediction.

[0145] Embodiment 2

[0146] Based on the same inventive concept, this embodiment also proposes a device for predicting the surface movement and deformation of a mining area based on airborne LiDAR. This prediction device is used to implement the method for predicting the surface movement and deformation of a mining area proposed in the above Embodiment 1. The device includes:

[0147] A data acquisition module, which is used to scan the target mining area using airborne LiDAR technology and collect the original point cloud data of the surface of the target mining area;

[0148] A data processing module, which is used to process the point cloud data based on the echo characteristics of airborne LiDAR and extract the point cloud data of non-growing fixed ground objects;

[0149] A parameter extraction module, which is used to traverse the point cloud data of non-growing fixed ground objects after preprocessing by the moving window method to determine the surface characteristic parameters. The surface characteristic parameters include the tilt direction angle, the surface point dip angle, and the surface characteristic coefficient;

[0150] A parameter inversion module, which is used to determine several subsidence monitoring points in the target mining area based on the point cloud data of non-growing fixed ground objects and obtain their working face parameters, and perform probability integral method parameter inversion in combination with the genetic algorithm to obtain probability integral parameters;

[0151] A real-time prediction module, which is used to construct a settlement prediction model with the probability integral parameters, the surface characteristic parameters, and the working face parameters, and use the settlement prediction model to predict the surface deformation of the real-time acquired mining area surface point cloud data.

[0152] It should be noted here that each module in the above prediction system corresponds to steps S1 to S5 in the implementation of the above prediction method. The instances and application scenarios realized by multiple modules and the corresponding steps are the same, but are not limited to the content disclosed in the above Embodiment 1. Moreover, this prediction system has all the beneficial effects of the above prediction method, which will not be elaborated here.

[0153] Further preferably, the data acquisition module includes a drone equipped with an airborne LiDAR device, which is used to perform aerial scanning on the target mining area and transmit the obtained original point cloud data of the mining area surface to the data processing module in real time.

[0154] Embodiment 3

[0155] This embodiment also proposes an electronic device, including:

[0156] A processor; a memory for storing instructions executable by the processor;

[0157] Wherein, the processor is configured to execute instructions to implement the prediction method as described above.

[0158] Embodiment 4

[0159] This embodiment also proposes a computer-readable storage medium. When the instructions in the computer-readable storage medium are executed by the processor of the electronic device, the electronic device can execute the prediction method as described above.

[0160] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed in this article can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0161] In addition, in each embodiment of this application, the various functional modules can be integrated in one processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit.

[0162] When the above-mentioned functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in the embodiments of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs.

[0163] The above embodiments are only used to illustrate the technical solutions of this application, rather than to limit them; although this application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for predicting surface movement and deformation in a mining area based on airborne LiDAR, characterized in that, The method includes: Scanning the target mining area using airborne LiDAR technology to collect the original point cloud data of the surface of the target mining area; Processing the point cloud data based on the echo characteristics of airborne LiDAR to extract the point cloud data of non-growing fixed ground objects; Adopting a moving window method to traverse the point cloud data of non-growing fixed ground objects after preprocessing to determine the surface characteristic parameters, where the surface characteristic parameters include the tilt direction angle, the surface point dip angle, and the surface feature coefficient; Determining a number of subsidence monitoring points in the target mining area based on the point cloud data of non-growing fixed ground objects and obtaining their working face parameters, and performing probability integral method parameter inversion in combination with the genetic algorithm to obtain probability integral parameters; Constructing a settlement prediction model with the probability integral parameters, the surface characteristic parameters, and the working face parameters, and using the settlement prediction model to predict the surface deformation of the real-time obtained mining area surface point cloud data; The adopting a moving window method to traverse the point cloud data of non-growing fixed ground objects after preprocessing to determine the surface characteristic parameters includes: using a moving window with a set grid spacing to traverse the point cloud data of non-growing fixed ground objects after preprocessing, and calculating the tilt direction angle, the surface dip angle, and the surface feature coefficient in each window, as shown in the following formula: (1) Tilt direction angle Obtain: (2) Surface dip angle α x,y Obtain: (3) Surface feature coefficient D x,y Obtain: D x,y =-200*((d + f + b + h) / 2 - 2*e) / ds 2 In the above formula, a, b, c, d, e, f, g, h, i are the minimum values of the point cloud elevation in the corresponding grid, ds is the grid spacing, and dx and dy are the differentials along the x and y directions respectively.

2. The method for predicting surface movement and deformation in a mining area based on airborne LiDAR according to claim 1, characterized in that: The processing the point cloud data to extract the point cloud data of non-growing fixed ground objects includes: Dividing the surface point cloud data into grids and recording the number of point clouds and the number of echo points in each grid; Setting threshold conditions, regarding the grids that meet the threshold conditions as non-fixed ground objects and filtering them using the PCL slope method to extract the point cloud data of non-growing fixed ground objects, where the growing fixed ground objects include roads, building tops, and bare soil.

3. The surface movement and deformation prediction method for mining areas based on airborne LiDAR according to claim 1, characterized in that: The determining a number of subsidence monitoring points in the target mining area based on the point cloud data of non-growing fixed ground objects and obtaining their working face parameters, and performing probability integral method parameter inversion in combination with the genetic algorithm to obtain probability integral parameters includes: Comparing the point cloud data of non-growing fixed ground objects at different time points, analyzing the deformation situation of the surface, identifying the subsidence area and determining a number of subsidence monitoring points and obtaining their working face parameters; Using the working face parameters of the subsidence monitoring points as input data for probability integral method parameter inversion to solve the probability integral parameters of a number of subsidence monitoring points.

4. A method for predicting surface movement and deformation in a mining area based on airborne LiDAR according to claim 3, characterized in that: The working face parameters include corner points, coal seam dip angle, and mining thickness; the probability integral parameters include subsidence coefficient q, horizontal movement coefficient b, main influence angle tangent tanβ, mining propagation angle θ, and left upper right lower inflection point offset s.

5. The method for predicting surface movement and deformation in a mining area based on airborne LiDAR according to claim 4, wherein: The constructing a settlement prediction model with the probability integral parameters, the surface characteristic parameters, and the working face parameters, and using the settlement prediction model to predict the surface deformation of the real-time obtained mining area surface point cloud data includes: Loading a working face prediction file containing probability integral parameters, surface characteristic parameters, and working face parameters in a data processing system, and initializing the settlement prediction model according to the parameters in the prediction file; Regularly obtain real-time raw ground point cloud data of the mining area using an airborne LiDAR monitoring device, input the preprocessed real-time point cloud data into the settlement prediction model, simulate the ground deformation situation during mining activities in real time in the data processing system, and output the ground deformation prediction results including the ground settlement map and the horizontal movement map.

6. An apparatus for predicting surface movement and deformation in a mining area based on airborne LiDAR, characterized in that, The device includes: A data acquisition module for scanning the target mining area using airborne LiDAR technology and collecting the raw ground point cloud data of the target mining area; A data processing module for processing the point cloud data based on the echo characteristics of airborne LiDAR and extracting the point cloud data of non-growing fixed ground objects; A parameter extraction module for traversing the point cloud data of non-growing fixed ground objects after preprocessing using a moving window method to determine the ground characteristic parameters, where the ground characteristic parameters include the tilt direction angle, the ground point dip angle, and the ground feature coefficient; A parameter inversion module for determining several subsidence monitoring points in the target mining area based on the point cloud data of non-growing fixed ground objects and obtaining their working face parameters, and performing probability integral method parameter inversion in combination with the genetic algorithm to obtain probability integral parameters; A real-time prediction module for constructing a settlement prediction model using the probability integral parameters, the ground characteristic parameters, and the working face parameters, and using the settlement prediction model to perform ground deformation prediction on the real-time obtained ground point cloud data of the mining area; The traversing the point cloud data of non-growing fixed ground objects after preprocessing using a moving window method to determine the ground characteristic parameters includes: traversing the point cloud data of non-growing fixed ground objects after preprocessing using a moving window with a set grid spacing, and calculating the tilt direction angle, the ground dip angle, and the ground feature coefficient within each window, as shown in the following formula: (1) Tilt direction angle Obtain: (2) Surface dip angle α x,y Obtain: (3) Surface feature coefficient D x,y Obtain: D x,y =-200*((d + f + b + h) / 2 - 2*e) / ds 2 In the above formula, a, b, c, d, e, f, g, h, i are the minimum values of the point cloud elevations within the corresponding grid, ds is the grid spacing, and dx and dy are the differentials along the x and y directions respectively.

7. The prediction device for surface movement and deformation in a mining area based on airborne LiDAR according to claim 6, characterized in that, The data acquisition module includes a drone equipped with an airborne LiDAR device, which is used to perform aerial scanning on the target mining area and transmit the obtained raw ground point cloud data of the mining area to the data processing module in real time.

8. An electronic device, characterized in that, Includes: A processor; A memory for storing the executable instructions of the processor; Wherein, the processor is configured to execute the instructions to implement the prediction method according to any one of claims 1 to 5.

9. A computer-readable storage medium, characterized in that, When the instructions in the computer-readable storage medium are executed by the processor of the electronic device, the electronic device is enabled to execute the prediction method according to any one of claims 1 to 5.

Images