A real-time monitoring system and method for underground goaf filling volume

Abstract

A system and method for real-time monitoring of the volume of filling bodies in underground goafs, the system includes a hardware installation and calibration module, a data acquisition and preprocessing module, a goaf volume monitoring module, a display and monitoring early warning module and a mobile control terminal. The method is to use the hardware installation and calibration module to calibrate the goaf online, and transmit the calibration information to the data acquisition and preprocessing module; the data acquisition and preprocessing module performs goaf point cloud splicing and cloud filtering on the received calibration information, obtains the goaf point cloud with homogeneous density and transmits it to the goaf volume monitoring module; the goaf volume monitoring module calculates the volume of the filling body in the goaf in real time according to the adaptive slicing method; the display and monitoring early warning module displays the relevant information, draws a measurement error diagram and a curve of the volume of the filling material changing with time, issues an early warning and calculates the volume of the filling material. Accurate calculation of the filling volume of the underground goaf can be achieved, the filling progress can be monitored in real time, and the efficiency of coal seam filling and mining can be ensured.

Description

Technical Field

[0001] The invention relates to a real-time monitoring system and method for the volume of a filling body in an underground goaf area, belonging to the technical field of filling mining. Background Art

[0002] As my country's development and utilization of coal resources become increasingly high, coal mining under buildings, under railways, under water bodies, and above pressurized water (three under and one over) has become a major problem facing coal mining in the central and eastern regions. Filling mining is an ideal way to solve the "three under and one over" coal mining problem, and is an effective way to solve mine gangue pollution and handle solids. In the existing filling process, it is difficult to accurately obtain the filling volume and filling progress of the goaf, and there is a lack of high-precision real-time filling monitoring systems to help coal mine workers make decisions. During the filling monitoring process, if the volume calculation of the filling material is inaccurate, it will cause insufficient filling and waste of materials above the well.

[0003] At present, the field of mining and filling mainly uses the method of monitoring the flow rate of filling materials to estimate the volume of the goaf for underground goaf filling monitoring. The initial goaf volume is roughly estimated by the sum of the coal mining volume and the volume of the top collapse counted in the coal mining production stage, and then the flow rate of the filling material during the filling process is monitored to achieve dynamic monitoring of the goaf filling progress. However, there are two main problems with traditional filling monitoring. First, the volume of the goaf filling body cannot be obtained in real time during the filling process; second, due to the expansion of the filling material after cooling, the filling material flow statistics cannot accurately calculate the actual volume of the filled area, which seriously affects the use and transportation of the filling material, and then affects the efficiency of coal seam filling mining, causing economic losses. Therefore, it is necessary to adopt advanced goaf filling monitoring methods to improve the accuracy and efficiency of the filling process. Summary of the invention

[0004] The present invention provides a real-time monitoring system and method for the filling volume of an underground goaf area. The system and method can accurately calculate the filling volume of the underground goaf area and monitor the filling progress in real time to ensure the efficiency of coal seam filling and mining.

[0005] In order to achieve the above-mentioned purpose, the present invention provides a real-time monitoring system for the volume of filling bodies in underground goaf areas, comprising a hardware installation and calibration module, a data acquisition and preprocessing module, a goaf area volume monitoring module, a display and monitoring early warning module and a mobile control terminal, wherein the mobile control terminal is connected to the input end of the hardware installation and calibration module, the output end of the hardware installation and calibration module is connected to the input end of the data acquisition and preprocessing module, and the output end of the data acquisition and preprocessing module is connected to the input end of the goaf area volume monitoring module; the display and monitoring early warning module is connected to the hardware installation and calibration module, the data acquisition and preprocessing module and the goaf area volume monitoring module;

[0006] The hardware installation and calibration module includes a plurality of laser radar sensors and a mounting base for assembling the laser radar sensors, wherein the mounting base is provided with a motor drive mechanism for driving the laser radar sensors to swing in the horizontal direction; and further includes a calibration unit, which is used for online calibrating the internal parameters of different laser radars and the external parameters between adjacent laser radars;

[0007] The data acquisition and preprocessing module includes a goaf point cloud stitching unit and a goaf point cloud filtering processing unit; the goaf point cloud stitching unit is used for adjacent laser point cloud registration and stitching; the goaf point cloud filtering processing unit is used for processing noise points in the goaf point cloud after stitching and performing downsampling processing;

[0008] The goaf volume monitoring module includes a goaf point cloud surface reconstruction unit and an adaptive volume calculation unit; the goaf point cloud surface reconstruction unit is used to reconstruct the point cloud surface of the goaf filling body; the adaptive volume calculation unit is used to adaptively calculate the goaf surface model volume according to the filling body shape;

[0009] The display and monitoring early warning module includes a display unit and an early warning unit. The display unit is used to display the calibration results, filtering processing effects, filling progress, filling volume ratio curve and filling volume measurement error diagram in real time; the early warning unit is used to calculate the adaptive threshold according to the shape of the goaf, issue an early warning, and provide recommended filling material volume.

[0010] Furthermore, the hardware installation and calibration module includes four 32-line laser radar sensors with a vertical field of view angle of 30°.

[0011] A method for real-time monitoring of filling volume of underground goaf area comprises the following steps:

[0012] Step 1: Use the hardware installation and calibration module to perform online calibration on the goaf, and transmit the calibration information to the data acquisition and preprocessing module;

[0013] Step 2: The data acquisition and preprocessing module performs goaf point cloud splicing and point cloud filtering on the received calibration information to obtain a goaf point cloud with homogeneous density and transmit it to the goaf volume monitoring module;

[0014] Step 3: The goaf volume monitoring module calculates the volume of the goaf filling body in real time according to the adaptive slicing method;

[0015] Step 4: The display and monitoring warning module displays the calibration effect of the lidar monitoring device, the point cloud filtering processing effect, and the real-time display of the filling progress, draws the filling volume ratio curve and the filling body volume measurement error diagram, and fits and draws the curve of the filling material volume changing with time, issues a warning and calculates the filling material volume.

[0016] Furthermore, the method of step 1 is as follows: after multiple laser radar sensors are installed, the calibration unit extracts line features and surface features in the environment and calibrates the internal parameters of each laser radar and the external parameters between adjacent laser radars online based on feature constraints; the specific steps are as follows:

[0017] (1) Principal component analysis is used to calculate the local neighborhood feature vector of the point cloud for multiple lidar data, and based on Calculate the local curvature; λ0, λ1, λ2 are the eigenvalues ​​of the covariance matrix of the local neighborhood of the point cloud, and satisfy λ0≤λ1≤λ2;

[0018] (2) Extract line features and surface features based on local curvature: For line features, select points with smaller curvature. The selection condition for line feature points is k i ≤Th line , where Th line To set the threshold, Th line The value range is 0.1 to 0.5, which is selected and adjusted according to the actual environment; for surface features, points with larger curvature are selected, and the selection condition of surface feature points is k i >Th plane , where Th plane To set the threshold, Th plane The value range is 0~0.1, which can be selected and adjusted according to the actual environment;

[0019] (3) Use the nearest neighbor search method to match different line features and surface features to generate corresponding feature pairs; then use the hand-eye calibration principle A i X=XB i Achieve calibration, where A i is the transformation matrix of the i-th pair of adjacent lidars, B i is the transformation matrix of the i-th pair of adjacent radars, X is the position change required for calibration, including the translation t and the rotation matrix R, expressed as the homogeneous transformation matrix Finally, the nonlinear optimization method is used, and the formula is Get the global optimization transformation matrix X fin , where g i are the line feature points and surface feature points extracted from the target point cloud, u i These are the line feature points and surface feature points that match them in the source point cloud.

[0020] Furthermore, the specific steps of the goaf point cloud stitching and point cloud filtering processing in step 2 are as follows:

[0021] (1) The point cloud stitching of the goaf area uses the NDT-based point cloud coarse registration algorithm and the GICP-based point cloud fine registration algorithm. The NDT-based point cloud coarse registration algorithm realizes the coarse alignment of the point cloud. The process is as follows: the acquired point cloud data is divided into voxel grids, and the Gaussian distribution of each voxel, that is, the mean and covariance matrix, is calculated to describe the distribution of points within the voxel. The NDT objective function is defined as: where p i is a point in the target point cloud, q i is the matching point in the source point cloud, w i is the weight factor, P is the target point cloud;

[0022] The GICP-based point cloud registration algorithm achieves high-precision matching of adjacent point clouds. The process is as follows: The GICP algorithm models the surface of the sampling points as a Gaussian distribution Define the transformation error as e i =q i -Tp i , and e i The distribution of Find the transformation T with the largest log-likelihood, which is

[0023] (2) Point cloud filtering processing of goaf area The point cloud filtering processing includes denoising and downsampling processing, wherein denoising uses range filtering to remove the collected outlier point cloud, including setting a reasonable distance, height and angle range, applying range filtering to remove outlier point clouds beyond the predetermined range, and obtaining high-quality point cloud data after preliminary filtering. By dividing the point cloud into a voxel grid of fixed size, calculating the distance weight of each point to the voxel center, and using the weighted average method to calculate the representative point, the voxel filtering downsampling based on the distance weight improvement is realized, and the homogeneously distributed point cloud data is obtained by downsampling. Furthermore, the specific steps of the step three adaptive slicing method to calculate the volume of the goaf filling body in real time are as follows:

[0024] (1) Traverse all scattered points, find the bounding box of the point set, obtain the initial triangle as the convex hull of the point set and put it into the triangle list; insert the scattered points in the point set one by one, find the triangle whose circumscribed circle contains the inserted triangle in the triangle list, delete the common edges of the affected triangles, connect the inserted point with all the vertices of the affected triangles, and thus complete the insertion of a point in the Dionysian triangle list;

[0025] (2) According to the maximum and minimum angle criteria, the locally formed triangles are optimized and placed into the Delaunay triangle linked list. After the optimization is completed, the super triangles are removed and the boundaries are processed, and finally the Delaunay triangulation surface model is constructed;

[0026] (3) The generated goaf surface model is first divided into equal parts along the main direction axis. The initial slice thickness is set to be thin, and the initial slice polygon P is generated, where i = 0, 1, 2, 3, ..., m0, and the area S of each slice is calculated. i , where i = 0, 1, 2, 3, ..., n0;

[0027] (4) Calculate the area ratio r and adjacent area difference d between adjacent slices:

[0028]

[0029] d i =S i -S i-1 ;

[0030] If 0.9 <r i <1.1 with d i <Th s At the same time, Th s To set the area difference threshold, adjacent slices are merged to generate the final slice boundary polygon P i ′, where i = 0, 1, 2, 3, ..., m1, and the area of ​​the merged section is S i ′, where i=0, 1, 2, 3, …, n1;

[0031] (5) The filling area volume calculation unit is used to multiply the slices generated by the adaptive slicing unit by the corresponding thickness H i Calculate the volume v of each slice i And add up to get the filling area volume V:

[0032]

[0033] Where n1 represents the number of slices after merging.

[0034] Furthermore, the point cloud filtering and denoising process divides the outlier point cloud in the goaf into large-scale noise and small-scale noise. The large-scale noise refers to isolated points distributed near the main body of the point cloud and drift points far away from the main body of the point cloud. The small-scale noise refers to abnormal points mixed with the main point cloud, which is the reason why the point cloud surface model is inconsistent with the actual surface model. For large-scale noise, statistical filtering and radius filtering are used for joint denoising, and for small-scale noise, bilateral filtering is used for processing.

[0035] The present invention sets a hardware installation and calibration module, a data acquisition and preprocessing module, a goaf volume monitoring module, and a display and monitoring early warning module. Through the hardware installation and calibration module, a suitable intrinsically safe laser radar assembly monitoring device can be selected according to different installation positions and filling area shapes. The calibration unit can accurately obtain accurate internal and external parameters of the radar monitoring device and collect goaf point clouds with high geometric accuracy; the data acquisition and preprocessing module performs denoising and downsampling preprocessing on the original point cloud data of the goaf; the goaf volume monitoring module is used to reconstruct the goaf surface model and calculate the goaf volume based on the improved adaptive slicing method of the present invention; the display and monitoring early warning module is connected with the hardware installation and calibration module, the data acquisition and preprocessing module and the goaf point cloud processing module, and is used to display the calibration results, the filtering processing effect and the real-time display of the filling progress, and calculate the recommended filling material volume after the early warning based on the statistical information. The present invention realizes the accurate calculation of the filling volume of the underground goaf, and monitors the filling progress in real time, thereby ensuring the efficiency of coal seam filling and mining. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 It is a schematic diagram of the system structure of the present invention;

[0037] Figure 2 is a schematic diagram of installing the laser radar sensor of the present invention;

[0038] Figure 3 is a flow chart of the method of the present invention;

[0039] Figure 4 (a) is a point cloud diagram of a goaf area according to an embodiment of the present invention, and (b) is a schematic diagram of the installation of a monitoring system according to an embodiment of the present invention;

[0040] Figure 5 Schematic diagram of the integral area of ​​goaf volume calculation in an embodiment of the present invention, (a), (b), and (c) respectively represent V 锥1 、V 锥2 、V 锥3 ;

[0041] Figure 6 4 is a volume measurement error curve diagram of an embodiment of the present invention.

[0042] In the figure: 1. Hardware installation and calibration module, 2. Data acquisition and preprocessing module, 3. Goaf volume monitoring module, 4. Display and monitoring early warning module, 101. LiDAR sensor, 102. Installation base. DETAILED DESCRIPTION

[0043] The present invention will be further described below in conjunction with the accompanying drawings.

[0044] like Figure 1As shown, a real-time monitoring system for the volume of filling body in underground goaf area comprises a hardware installation and calibration module 1, a data acquisition and preprocessing module 2, a goaf area volume monitoring module 3, a display and monitoring early warning module 4 and a mobile control terminal, wherein the mobile control terminal is connected to the input end of the hardware installation and calibration module 1, the output end of the hardware installation and calibration module 1 is connected to the input end of the data acquisition and preprocessing module 2, and the output end of the data acquisition and preprocessing module 2 is connected to the input end of the goaf area volume monitoring module 3; the display and monitoring early warning module 4 is connected to the hardware installation and calibration module 1, the data acquisition and preprocessing module 2 and the goaf area volume monitoring module 3;

[0045] The hardware installation and calibration module 1 includes a plurality of laser radar sensors 101 and a mounting base 102 for assembling the laser radar sensors, wherein the mounting base 102 is provided with a motor drive mechanism for driving the laser radar to swing in the horizontal direction; and further includes a calibration unit, which is used for online calibrating the internal parameters of different laser radars and the external parameters between adjacent laser radars;

[0046] The data acquisition and preprocessing module 2 includes a goaf point cloud stitching unit and a goaf point cloud filtering processing unit; the goaf point cloud stitching unit is used for adjacent laser point cloud registration and stitching; the goaf point cloud filtering processing unit is used for processing noise points in the goaf point cloud after stitching and performing downsampling processing;

[0047] The goaf volume monitoring module 3 includes a goaf point cloud surface reconstruction unit and an adaptive volume calculation unit; the goaf point cloud surface reconstruction unit is used to reconstruct the goaf filling body point cloud surface; the adaptive volume calculation unit is used to adaptively calculate the goaf surface model volume according to the filling body shape;

[0048] The display and monitoring warning module 4 includes a display unit and a warning unit. The display unit is used to display the calibration results, filtering processing effects, filling progress, filling body volume ratio curve and filling body volume measurement error diagram in real time; the warning unit is used to calculate the adaptive threshold according to the shape of the goaf, issue a warning, and provide recommended filling material volume.

[0049] As a preferred embodiment, Figure 2 As shown, considering that the angle between the vertical and horizontal directions of the goaf is close to 90°, and considering the splicing problem of adjacent laser point clouds, four 32-line laser radar sensors 101 with a vertical field of view angle of 30° are assembled and spliced, and are equipped with a motor drive device so that the radar monitoring device can swing in the horizontal direction.

[0050] Figure 2 This is a schematic diagram of the laser radar monitoring device and its sensor placement.

[0051] like Figure 3As shown, a method for real-time monitoring of the filling volume of a goaf area in a mine comprises the following steps:

[0052] Step 1: Use the hardware installation and calibration module to perform online calibration on the goaf, and transmit the calibration information to the data acquisition and preprocessing module;

[0053] Step 2: The data acquisition and preprocessing module performs goaf point cloud splicing and point cloud filtering on the received calibration information to obtain a goaf point cloud with homogeneous density and transmit it to the goaf volume monitoring module;

[0054] Step 3: The goaf volume monitoring module calculates the volume of the goaf filling body in real time according to the adaptive slicing method;

[0055] Step 4: The display and monitoring warning module displays the calibration effect of the lidar monitoring device, the point cloud filtering processing effect, and the real-time display of the filling progress, draws the filling volume ratio curve and the filling body volume measurement error diagram, and fits and draws the curve of the filling material volume changing with time, issues a warning and calculates the filling material volume.

[0056] Example: Figure 4 As shown in (a), the filling area is mostly an irregular geometric body similar to a rectangular parallelepiped with a length of 150 meters, a width of 50 meters, and a height of 3 meters. For the convenience of description, the filling area is simplified to a rectangular parallelepiped with a length of 150 meters, a width of 50 meters, and a height of 3 meters, and the radar monitoring device is placed as follows Figure 4 (b) The center position is shown. According to the symmetry, the volume of the filled area can be calculated by calculating the volume V of the three tetrahedrons. 锥1 、V 锥2 、V 锥3 , three tetrahedrons such as Figure 5 As shown, (a), (b), and (c) represent V 锥1 、V 锥2 、V 锥3 ; The total volume of the filled area is:

[0057] V 总 =2((V 锥1 +V 锥2 +V 锥3 )

[0058] For ease of description, let the length of the filling area be a, the width be b, and the height be c. Further assume that the ranging error of the laser radar monitoring device is proportional to the measured distance, and the ranging of the radar monitoring device is expressed as:

[0059]

[0060] Among them, x, y, z represent the corresponding measurement values ​​of the laser point on the x-axis, y-axis, and z-axis relative to the origin of the radar monitoring device coordinate system, in meters, and k is the ranging error factor, which is determined according to the laser radar ranging error.

[0061] Four-sided pyramid V 锥1 The volume can be calculated as:

[0062]

[0063] Considering the tetrahedron V 锥1 The height from the bottom to the radar monitoring device is a constant value c, that is, z = c.

[0064]

[0065] make V' can be calculated 锥1 for:

[0066]

[0067] Four-sided pyramid V 锥2 The volume can be calculated as:

[0068]

[0069] Considering the tetrahedron V 锥2 The height from the bottom to the radar monitoring device is a fixed value Right now have to

[0070]

[0071] make V' can be calculated 锥2 for:

[0072]

[0073] Four-sided pyramid V 锥3 The volume is calculated as:

[0074]

[0075] Considering the tetrahedron V 锥3 The height from the bottom to the radar monitoring device is a constant, that is, y = a.

[0076]

[0077] make V' can be calculated 锥3 for:

[0078]

[0079] like Figure 6 As shown in the figure, when the shape of the goaf is an approximate rectangular area with a length of 150 meters, a width of 50 meters and a height of 3 meters, the measurement error of the lidar monitoring system is related to the volume of the filling body. As the volume of the filling body gradually increases, the measurement error of the lidar monitoring system gradually decreases. During the filling process, the maximum proportion of the filling body volume measurement error is 2%, which meets the requirements of the filling task, and as the volume of the filling body increases, the proportion of the filling body volume measurement error eventually approaches 0.1%, providing high-precision goaf volume information to the staff in the later stage of the filling work.

Claims

1. A method for real-time monitoring of the volume of filling body in underground goaf, wherein the real-time monitoring system for the volume of filling body in underground goaf based on the method comprises a hardware installation and calibration module (1), a data acquisition and preprocessing module (2), a goaf volume monitoring module (3), a display and monitoring early warning module (4) and a mobile control terminal; The hardware installation and calibration module (1) comprises a plurality of laser radar sensors (101) and a calibration unit, wherein the calibration unit is used for online calibrating the internal parameters of different laser radars and the external parameters between adjacent laser radars; The data acquisition and preprocessing module (2) comprises a goaf point cloud splicing unit and a goaf point cloud filtering processing unit; The goaf volume monitoring module (3) comprises a goaf point cloud surface reconstruction unit and an adaptive volume calculation unit; The display and monitoring warning module (4) comprises a display unit and a warning unit; It is characterized in that The steps include: Step 1: Use the hardware installation and calibration module (1) to perform online calibration on the goaf, and transmit the calibration information to the data acquisition and preprocessing module (2); the method is: after the installation of multiple laser radar sensors (101), the calibration unit extracts line features and surface features in the environment and calibrates the internal parameters of each laser radar and the external parameters between adjacent laser radars online based on feature constraints, and the specific steps are as follows: (1.1) The principal component analysis is used to calculate the local neighborhood feature vector of the point cloud for multiple lidar data, and the Calculate the local curvature; λ0, λ1, λ2 are the eigenvalues ​​of the covariance matrix of the local neighborhood of the point cloud, and satisfy λ0≤λ1≤λ2; (1.2) Extract line features and surface features based on local curvature: For line features, select points with smaller curvature. The selection condition for line feature points is k i ≤Th line , where Th line is to set the threshold; for the surface feature, the point with larger curvature is selected, and the selection condition of the surface feature point is k i >Th plane , where Th plane To set the threshold; (1.3) Use the nearest neighbor search method to match different line features and surface features to generate corresponding feature pairs; then use the hand-eye calibration principle A i X=XB i Achieve calibration, where A i is the transformation matrix of the i-th pair of adjacent lidars, B i is the transformation matrix of the i-th pair of adjacent radars, X is the position change required for calibration, including the translation t and the rotation matrix R, expressed as the homogeneous transformation matrix Finally, the nonlinear optimization method is used, and the formula is Get the global optimization transformation matrix X fin , where g i are the line feature points and surface feature points extracted from the target point cloud, u i are the line feature points and surface feature points that match it in the source point cloud; Step 2, the data acquisition and preprocessing module (2) performs goaf point cloud splicing and point cloud filtering on the received calibration information to obtain a goaf point cloud with homogeneous density and transmit it to the goaf volume monitoring module (3); Step 3, the goaf volume monitoring module (3) calculates the volume of the goaf filling body in real time according to the adaptive slicing method; Step 4, the display and monitoring warning module (4) displays the calibration effect of the laser radar monitoring device, the point cloud filtering processing effect, and the real-time display of the filling progress, draws a filling volume ratio curve and a filling body volume measurement error diagram, and fits and draws a curve of the change of the filling material volume over time, issues a warning and calculates the filling material volume.

2. The monitoring method of the underground goaf filling volume real-time monitoring system according to claim 1 is characterized in that: The specific steps of point cloud stitching and point cloud filtering processing of the goaf area in step 2 are as follows: (2.1) The point cloud stitching of the goaf area uses the NDT-based point cloud coarse registration algorithm and the GICP-based point cloud fine registration algorithm; the NDT-based point cloud coarse registration algorithm realizes the coarse alignment of the point cloud. The process is as follows: the acquired point cloud data is divided into voxel grids, and the Gaussian distribution of each voxel, that is, the mean and covariance matrix, is calculated to describe the distribution of points within the voxel. The NDT objective function is defined as: where p i is a point in the target point cloud, q i is the matching point in the source point cloud, w i is the weight factor, P is the target point cloud; The GICP-based point cloud registration algorithm achieves high-precision matching of adjacent point clouds. The process is as follows: The GICP algorithm models the surface of the sampling points as a Gaussian distribution Define the transformation error as e i =q i -Tp i , and e i The distribution of Find the transformation T with the largest log-likelihood, which is (2.2) Point cloud filtering processing of goaf area The point cloud filtering processing includes denoising and downsampling processing, wherein denoising uses range filtering to remove the collected outlier point cloud, including setting reasonable distance, height and angle ranges, applying range filtering to remove outlier point clouds that exceed the predetermined range, and obtaining high-quality point cloud data after preliminary filtering. The point cloud is divided into voxel grids of fixed size, the distance weight from each point to the voxel center is calculated, and the representative point is calculated using the weighted average method, thereby realizing voxel filtering downsampling based on distance weight improvement, and obtaining homogeneously distributed point cloud data through downsampling.

3. The monitoring method of the underground goaf filling volume real-time monitoring system according to claim 2 is characterized in that: The specific steps of the step 3 of calculating the volume of the goaf filling body in real time using the adaptive slicing method are as follows: (3.1) Traverse all scattered points, find the bounding box of the point set, obtain the initial triangle as the convex hull of the point set and put it into the triangle list; insert the scattered points in the point set one by one, find the triangle whose circumscribed circle contains the inserted triangle in the triangle list, delete the common edges of the affected triangle, connect the inserted point with all the vertices of the affected triangle, and complete the insertion of a point in the Dironian triangle list; (3.2) According to the maximum and minimum angle criteria, the locally formed triangles are optimized and placed into the Dironian triangle linked list. After the optimization is completed, the super triangles are removed and the boundaries are processed, and finally the Dironian triangulation surface model is constructed; (3.3) The generated goaf surface model is first divided into equal parts along the main direction axis, and the initial slice thickness is set to be thin, generating the initial slice polygon P i , where i = 0, 1, 2, 3, ..., m0, and calculate the area S of each section i , where i = 0, 1, 2, 3, ..., n0; (3.4) Calculate the area ratio r and adjacent area difference d between adjacent slices: d i =S i -S i-1 ; If 0.9<r i <1.1 and d i <Th s At the same time, Th s To set the area difference threshold, adjacent slices are merged to generate the final slice boundary polygon P i ′, where i = 0, 1, 2, 3, ..., m1, and the area of ​​the merged section is S i ′, where i=0, 1, 2, 3, …, n1; (3.5) The filling area volume calculation unit is used to multiply the slices generated by the adaptive slicing unit by the corresponding thickness H i Calculate the volume v of each slice i And add up to get the filling area volume V: Where n1 represents the number of slices after merging.

4. The monitoring method of the underground goaf filling volume real-time monitoring system according to claim 2 is characterized in that: The point cloud filtering and denoising process divides the outlier point cloud in the goaf into large-scale noise and small-scale noise. The large-scale noise refers to isolated points distributed near the main body of the point cloud and drift points far away from the main body of the point cloud. The small-scale noise refers to abnormal points mixed with the main point cloud, which is the reason why the point cloud surface model is inconsistent with the actual surface model. For large-scale noise, statistical filtering and radius filtering are used for joint denoising, and for small-scale noise, bilateral filtering is used for processing.

5. The method for real-time monitoring of filling volume of underground goaf area according to claim 1 is characterized in that: The hardware installation and calibration module (1) comprises four 32-line laser radar sensors (101) with a vertical field of view angle of 30°.

Images