Coal mine goaf three-dimensional laser scanning deformation monitoring data processing method

Abstract

The present application relates to the technical field of safety monitoring, and particularly relates to a coal mine goaf three-dimensional laser scanning deformation monitoring data processing method, comprising multi-period point cloud data acquisition and preprocessing; a three-dimensional influence cone model is constructed based on rock stratum movement angle, strong deformation areas and weak deformation areas are divided and differential sampling is performed; a point cloud topological complex is constructed, a persistence diagram representing crack evolution characteristics is extracted through a persistent homology algorithm, and topological entropy is calculated; a topological entropy time sequence is established, a curve slope and a variation coefficient are calculated, and intensive monitoring is triggered; critical slowing down indexes such as autocorrelation coefficients and system recovery rates are extracted, a comprehensive instability index is constructed, and an instability grade is determined; early warning information is output according to the early warning grade, and the normal direction of a crack ring is extracted from the persistence diagram to identify a failure mode. The present application combines topological data analysis and critical slowing down theory, realizes early identification of surrounding rock instability precursors and intelligent identification of failure modes, and significantly improves the accuracy and intelligence level of monitoring and early warning.

Description

Technical Field

[0001] This invention relates to the field of safety monitoring technology, specifically to a method for processing data from three-dimensional laser scanning deformation monitoring of coal mine goaf areas. Background Technology

[0002] Monitoring the deformation of surrounding rock in coal mine goaf roadways is a crucial aspect of ensuring safe mine production. With the development of 3D laser scanning technology, non-contact monitoring methods based on point cloud data are gradually becoming mainstream. Among existing technologies, Chinese patent CN112282847B discloses a method for monitoring the deformation of surrounding rock in coal mine roadways. This method involves setting up fixed observation points within the roadway and periodically collecting 3D point cloud data of the roadway surface using a 3D laser scanner. By registering and comparing adjacent point cloud data sets, the displacement change of the surrounding rock surface is calculated, thereby achieving deformation monitoring. Compared to traditional total station single-point measurements, this method offers advantages such as high data acquisition efficiency, large information capacity, and no need for contact with the target body, thus improving the automation level of roadway deformation monitoring to a certain extent.

[0003] However, the technical solution disclosed in CN112282847B has significant shortcomings in practical applications. First, this method only performs a simple geometric comparison between two adjacent point clouds, which is a static, post-hoc measurement. It cannot capture the time-series characteristics of deformation evolution, nor can it identify the precursory information of surrounding rock instability, resulting in a serious lag in early warning. Second, this method lacks the ability to analyze the topological structure of point cloud data, and cannot distinguish between different deformation mechanisms such as crack propagation and overall convergence. It simplifies complex discontinuous deformation into a single displacement, obscuring key structural damage information. Third, this method uses fixed observation base points and fixed monitoring cycles, and does not consider the influence of the overlying rock movement law in the goaf on the spatial distribution of deformation, resulting in unreasonable allocation of monitoring resources, insufficient sampling in strong deformation areas and oversampling in weak deformation areas. Fourth, this method does not establish a correlation between deformation data and mechanical state, and cannot quantify the stability state of the surrounding rock system, making it difficult to meet the needs of intelligent mines for disaster prediction and early warning.

[0004] Given the limitations of existing technologies, which can only perform static geometric measurements and cannot predict dynamic instability, there is an urgent need to develop a new monitoring method to provide a new technical approach for safety monitoring of coal mine goaf roadways. Summary of the Invention

[0005] To address the aforementioned shortcomings of existing technologies, this invention provides a three-dimensional laser scanning deformation monitoring data processing method for coal mine goaf areas, which can effectively solve the problems mentioned in existing technologies.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] This invention provides a method for processing three-dimensional laser scanning deformation monitoring data in coal mine goaf areas, the steps of which include:

[0008] S100 Cloud Data Acquisition and Preprocessing: In the goaf roadway, multi-stage three-dimensional point cloud data of the roadway surrounding rock surface are collected according to the monitoring cycle. Each stage of point cloud data is preprocessed to obtain a point cloud set. The monitoring cycle is dynamically adjusted according to the mining depth. The greater the mining depth, the longer the monitoring cycle.

[0009] S200. Monitoring area division based on rock stratum movement angle: Determine the rock stratum movement angle according to the geological and mining conditions parameters of the goaf. Construct a three-dimensional influence cone model with the goaf mining boundary as the vertex and the rock stratum movement angle as the half-cone angle. Divide the roadway into a strong deformation zone inside the influence cone and a weak deformation zone outside the influence cone. Implement a differentiated sampling strategy for the strong deformation zone and the weak deformation zone. The sampling density of the strong deformation zone is greater than that of the weak deformation zone.

[0010] S300, Point Cloud Topological Feature Extraction: For each period of point cloud data, a topological complex with neighborhood radius as a parameter is constructed. The one-dimensional homology group of the complex is calculated by the persistent homology algorithm to obtain a persistent graph used to characterize the evolution features of closed cracks in the point cloud data. Multi-scale persistent homology analysis is used to determine the optimal analysis scale.

[0011] S400. Topological Entropy Temporal Evolution Analysis: Calculate the topological entropy value of each period of point cloud data based on the continuous graph. The topological entropy value is calculated based on the duration of the crack loop. Establish a topological entropy temporal sequence, fit its evolution curve and calculate the curve slope. Calculate the coefficient of variation of the topological entropy temporal sequence. When the coefficient of variation exceeds a preset variation threshold, trigger the encrypted monitoring mode to shorten the monitoring cycle.

[0012] S500, Critical Slowing Index Extraction and Instability Determination: Autocorrelation analysis is performed on the topological entropy time series to calculate the autocorrelation coefficient and establish an autoregressive model to extract the system recovery rate index; a comprehensive instability index is constructed based on the curve slope, autocorrelation coefficient, and recovery rate index; multi-level threshold criteria are set, and the instability state level of the surrounding rock is determined based on the threshold range of the comprehensive instability index.

[0013] S600, Early Warning Output and Damage Mode Recognition: Output the corresponding early warning level and instability probability prediction value based on the instability determination result; when the early warning level reaches the preset level, extract the crack ring with a duration exceeding the preset length threshold from the persistence map, calculate its geometric center coordinates and normal direction vector, and determine the potential damage mode based on the angle relationship between the normal direction and the coordinate axis.

[0014] Furthermore, the preprocessing in step S100 includes denoising and downsampling; the denoising uses a statistical filtering algorithm to identify and remove noise points by setting the number of neighborhood points and the standard deviation multiplier; the downsampling uses voxel grid filtering.

[0015] Furthermore, the rock strata movement angle in step S200 is determined according to the classification of overlying lithology, which includes hard rock strata, medium-hard rock strata and soft rock strata. Different lithologies correspond to different ranges of rock strata movement angle values. In the differentiated sampling strategy, the strong deformation zone adopts the first sampling density, and the weak deformation zone adopts the second sampling density. The first sampling density is greater than the second sampling density.

[0016] Furthermore, the topological complex in step S300 is a Vietoris-Rips complex; the multi-scale persistent cohomology analysis calculates the persistent graph at each scale by setting a scale parameter sequence of neighborhood radius, and selects the scale with the largest one-dimensional Betti number change rate as the optimal analysis scale; the persistent cohomology calculation adopts a distributed parallel computing architecture, divides the point cloud data into several sub-blocks, calculates the local persistent graph in parallel, and then merges them into a global persistent graph through a boundary stitching algorithm.

[0017] Furthermore, the topological entropy value in step S400 The calculation is based on the following principle: the birth-scale parameter and death-scale parameter of each crack loop are obtained from the persistence graph, and the persistence length of each crack loop is calculated, wherein the persistence length is the difference between the death-scale parameter and the birth-scale parameter; the Shannon entropy of the probability distribution is calculated using the proportion of the persistence length of each crack loop to the sum of the persistence lengths of all crack loops as the probability distribution, and the topological entropy value is obtained; the coefficient of variation is the ratio of the standard deviation to the mean of the topological entropy time series.

[0018] Furthermore, the autocorrelation coefficient in step S500 is a lag-1 autocorrelation coefficient, used to measure the degree of linear correlation between the topological entropy values ​​of two adjacent periods; the autoregressive model is a first-order autoregressive model AR(1), and the specific calculation formula of the model is as follows:

[0019] ;

[0020] in, These are the autoregressive coefficients. The noise is white noise; the system recovery rate index is calculated based on autoregressive coefficients and is used to characterize the system's ability to recover to equilibrium after being disturbed.

[0021] Furthermore, the comprehensive instability index in step S500 is a predicted instability probability value, which is positively correlated with the curve slope and autocorrelation coefficient, and negatively correlated with the system recovery rate index; the multi-level threshold criterion includes multiple threshold intervals that increase sequentially, each threshold interval corresponding to an instability state level, and the instability state level includes at least a stable level, a creep level, an acceleration level, and a critical level.

[0022] Furthermore, the preset level in step S600 is the critical level; the determination of the failure mode is based on the angle relationship between the crack ring normal direction vector and the three coordinate axis directions, and the roof collapse mode, side spalling mode and shear slip mode are distinguished according to the principal component characteristics of the normal direction; among them, the roof collapse mode corresponds to the vertical component being dominant in the normal direction, the side spalling mode corresponds to the horizontal component being dominant in the normal direction, and the shear slip mode corresponds to none of the three components of the normal direction being dominant.

[0023] A three-dimensional laser scanning deformation monitoring data processing system for coal mine goaf areas includes:

[0024] The point cloud data acquisition module is used to collect three-dimensional point cloud data of the surrounding rock surface of the roadway according to the monitoring cycle;

[0025] The data preprocessing module is used to preprocess point cloud data;

[0026] The region division module is used to construct a three-dimensional influence cone model based on the rock strata movement angle and divide the strong deformation zone and the weak deformation zone.

[0027] The topology analysis module is used to construct topological complexes, calculate persistent graphs, and calculate topological entropy.

[0028] The critical slowdown analysis module is used to extract critical slowdown characteristic indicators and determine the state of the surrounding rock.

[0029] The early warning output module is used to output the early warning level and damage mode based on the judgment result.

[0030] Furthermore, the region partitioning module is data-connected to the topology analysis module. The boundary coordinates of the strong deformation zone output by the region partitioning module are transmitted to the topology analysis module. The topology analysis module performs continuous cohomology calculation on the first neighborhood radius for the strong deformation zone and on the second neighborhood radius for the weak deformation zone. The first neighborhood radius is smaller than the second neighborhood radius.

[0031] The technical solution provided by this invention has the following advantages compared with the known prior art:

[0032] This invention is based on topological data analysis for identifying precursors of surrounding rock instability. By constructing a Vietoris-Rips complex of point cloud data and calculating continuous cohomology, the topological evolution characteristics of fracture rings are extracted, elevating traditional geometric measurements to topological structural analysis. This enables the capture of critical slowing phenomena in the surrounding rock system before macroscopic deformation occurs, achieving a leap from post-detection to early warning.

[0033] This invention integrates quantitative early warning of critical slowing indicators with intelligent determination of failure modes. It constructs an instability probability prediction model by combining the slope of the topological entropy curve, autocorrelation coefficient, and recovery rate indicators. Based on the crack ring normal direction, it intelligently determines failure modes such as roof collapse, spalling, or shear slip, and outputs the spatial location of potential instability areas, providing precise guidance for on-site support decisions. Attached Figure Description

[0034] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0035] Figure 1 This is a schematic diagram of the method flow of the present invention;

[0036] Figure 2 This is a flowchart of the topological feature extraction process of the present invention;

[0037] Figure 3 This is a flowchart of the critical slowdown analysis of the present invention;

[0038] Figure 4 This is a schematic diagram of the system structure of the present invention. Detailed Implementation

[0039] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0040] The present invention will be further described below with reference to embodiments.

[0041] Reference Figure 1 , Figure 1 This is a flowchart illustrating a method for processing data from three-dimensional laser scanning deformation monitoring in coal mine goaf areas, as provided in an embodiment of the present invention. Figure 1 As shown, the method includes the following steps:

[0042] Step S100: Multi-phase point cloud data acquisition and preprocessing.

[0043] In some embodiments, a 3D laser scanner is deployed in the goaf roadway to collect multi-period 3D point cloud data of the roadway surrounding rock surface according to the monitoring cycle. The 3D laser scanner can be a ground-based or track-mounted scanning system. The track-mounted platform can be deployed along the roadway axis to achieve full coverage scanning of the entire roadway cross-section.

[0044] Monitoring cycle depends on mining depth The monitoring cycle is dynamically adjusted; the greater the mining depth, the longer the monitoring period. Specifically, the principle for setting the monitoring cycle T is: when the mining depth... At that time, it was shallow mining, and the surrounding rock deformation rate was relatively fast. The monitoring cycle was set to... Heaven; when At that time, it belonged to medium-deep mining, and the monitoring cycle was set to... Heaven; when At that time, it was deep mining, and the deformation rate of the surrounding rock was relatively slow. The monitoring cycle was set to... It should be noted that the above values ​​are for illustrative purposes only and may be adjusted according to specific geological conditions and monitoring needs in actual applications.

[0045] Preprocessing of each point cloud data session, including denoising, filtering, and downsampling, yields a point cloud set. ,in For the number of monitoring periods, To ensure the reliability of time series analysis.

[0046] The denoising process employs a statistical filtering algorithm, identifying and removing noisy points by setting the number of neighboring points and the standard deviation multiplier. Specifically, for each point in the point cloud, the average distance to its k nearest neighbors is calculated. It is assumed that the average distances of all points follow a Gaussian distribution with a mean of [value missing]. The standard deviation is Then the average distance is retained in Points within the specified range are selected, while other points are discarded as noise. In this embodiment, the number of neighboring points is set. Standard deviation multiplier .

[0047] Downsampling processing employs voxel grid filtering, which performs spatially uniform sampling of the point cloud by setting the voxel size. The 3D space is divided into a voxel grid of a specified size, with each voxel retaining a centroid or center point, thereby compressing the data volume. In this embodiment, the voxel size is set to... .

[0048] After preprocessing, standardized point cloud data is obtained, laying the foundation for subsequent analysis.

[0049] Step S200: Delineation of monitoring areas based on rock strata movement angle.

[0050] In some embodiments, the rock strata movement angle is determined based on geological and mining condition parameters of the goaf. The rock strata movement angle is an important parameter characterizing the extent of mining impact; its magnitude depends on the overlying lithology. Geological and mining condition parameters include mining depth. Thick Overburden density internal friction angle Information such as mining area geological exploration reports and mining design data can be obtained.

[0051] The angle of movement of rock strata is determined according to the classification of overlying lithology: for hard rock strata (such as sandstone and limestone). For medium-hard rock formations (such as sandy shale and argillaceous sandstone). For weak rock formations (such as shale and mudstone). In practical applications, values ​​can be selected within the above range based on the physical and mechanical properties of the specific rock strata, or determined through engineering analogy.

[0052] A three-dimensional influence cone model is constructed using the goaf boundary as the vertex and the rock strata movement angle δ as the semi-cone angle. This model uses the goaf boundary as the starting line and forms a cone-shaped influence area along the roadway extension direction, with the cone's axis perpendicular to the goaf boundary. The roadway is divided into a high-deformation zone located within the influence cone. and the weak deformation zone located outside the influence cone The boundaries between the strong deformation zone and the weak deformation zone are determined by the intersection of the three-dimensional influence cone model and the centerline of the roadway.

[0053] A differentiated sampling strategy is applied to regions of strong deformation and weak deformation, with a higher sampling density in the strong deformation region than in the weak deformation region. Specifically, a first sampling density is used in the strong deformation region, and a second sampling density is used in the weak deformation region. In this embodiment, the point cloud density in the strong deformation region is set to... The point cloud density in the weakly deformable region is set to By using differentiated sampling, the total amount of data can be effectively controlled while ensuring the monitoring accuracy of key areas, thus improving processing efficiency.

[0054] Step S300: Point cloud topological feature extraction.

[0055] In some embodiments, for each period of point cloud data A topological complex with a neighborhood radius ε as a parameter is constructed, specifically using the Vietoris-Rips complex construction method. The Vietoris-Rips complex is a simple complex construction method based on point cloud distance relationships: for any two points in the point cloud, if their Euclidean distance is less than or equal to... Then connect them with an edge; for three points, if the distance between any two points is less than or equal to... If the first triangle is formed, then a triangle is formed; and so on, a higher-dimensional simplex is constructed.

[0056] The one-dimensional homology group of the complex is calculated using the persistent homology algorithm. This yields a persistent graph used to characterize the evolution features of closed crack loops in point cloud data. The core idea of ​​continuous cohomology is to track topological features (such as connected components, cycles, holes, etc.) as the scale parameter changes. The process of change's "birth" and "death." Specifically, the Betty number of a one-dimensional homology group. Characterizes the number of closed crack loops in point cloud data; each crack loop corresponds to a generator ( , ),in The scale parameter indicates the first appearance of the crack ring. The scale parameter indicates the time it takes for the crack ring to disappear.

[0057] Multi-scale persistence cohomology analysis was employed, and persistence maps at each scale were calculated by setting a series of scale parameters for the neighborhood radius. A one-dimensional Betty number was selected. The scale with the largest rate of change is used as the optimal analytical scale. In this embodiment, the neighborhood radius The range of values ​​is The scale parameter sequence is set to Calculate the continuity plots at each scale and analyze them. Follow Based on the changing patterns, the scale with the largest rate of change is selected as the optimal analysis scale, at which the topological features have the highest discernibility.

[0058] To improve computational efficiency, continuous cohomology computation employs a distributed parallel computing architecture. The point cloud data is divided into... voxel blocks, Determined based on the point cloud size. Where N is the total number of points in the point cloud. Each sub-block independently calculates its local persistent graph, which is then merged into a global persistent graph using a boundary stitching algorithm. The boundary stitching algorithm handles the simplex at the boundaries of the sub-blocks to ensure the continuity of topological features.

[0059] Step S400: Topological entropy time-series evolution analysis.

[0060] In some embodiments, according to the continuous graph Calculate the topological entropy value of each period's point cloud data. Topological entropy values, calculated based on the persistence length of crack loops, are used to quantify the complexity of crack loop development. Specifically, the birth-scale parameters of each crack loop are obtained from the persistence graph. and mortality scale parameters The duration of each crack ring is calculated using the following formula:

[0061] ;

[0062] in, The duration length of each crack ring represents the survival time of the crack ring in scale space, reflecting the stability of the crack structure.

[0063] Using the proportion of each crack ring's duration to the total duration of all crack rings as a probability distribution, calculate the Shannon entropy of this probability distribution to obtain the topological entropy value. The calculation formula is:

[0064] ;

[0065] Where m is the total number of crack rings detected in the i-th period. This represents the sum of the duration lengths of all crack rings. A larger topological entropy value indicates a more uniform distribution of crack ring duration lengths and a more complex crack development; a smaller topological entropy value indicates a more concentrated distribution of crack ring duration lengths and a simpler crack development pattern.

[0066] Establish topological entropy time series Fit its evolution curve And calculate the slope of the curve. Curve fitting can be performed using methods such as least squares fitting, polynomial fitting, or spline interpolation. The slope k of the curve reflects the rate of change of topological entropy over time. This indicates that the complexity of the cracks has increased. This indicates that the complexity of the cracks has decreased.

[0067] Calculate the coefficient of variation of the topological entropy time series. ,in The standard deviation of the topological entropy time series is... The mean is used. The coefficient of variation reflects the relative amplitude of topological entropy fluctuations. When the coefficient of variation exceeds a preset variation threshold, it indicates that the surrounding rock deformation has entered an unstable stage, triggering an encrypted monitoring mode that shortens the monitoring cycle. In this embodiment, the preset variation threshold is set to 0.3, and the monitoring cycle is shortened to half of the original cycle under encrypted monitoring mode.

[0068] Step S500: Extraction of critical slowing index and determination of instability.

[0069] In some embodiments, autocorrelation analysis is performed on the topological entropy time series to calculate the lag-1 autocorrelation coefficient ACF (1). The lag-1 autocorrelation coefficient is used to measure the degree of linear correlation between the topological entropy values ​​of two adjacent periods, and the calculation formula is as follows:

[0070] ;

[0071] in, This represents the mean of the topological entropy time series. An autocorrelation coefficient close to 1 indicates that the series has strong memory, meaning the current state is strongly dependent on the state at the previous time step, which is a typical characteristic of critical slowing down.

[0072] A first-order autoregressive model AR(1) is established to extract the system recovery rate index. The model form is as follows:

[0073] ;

[0074] in, These are the autoregressive coefficients. White noise. Autoregressive coefficients. It can be estimated using the least squares method or the Yule-Walker equation. System recovery rate index , The range of values ​​is , The smaller the value, the weaker the system's ability to recover to equilibrium after being disturbed, and the higher the risk of instability.

[0075] A comprehensive instability index was constructed. The comprehensive instability index is positively correlated with the curve slope k and the autocorrelation coefficient ACF (1), and is related to the system recovery rate index. They show a negative correlation. In this embodiment, the comprehensive instability index uses the predicted instability probability value. The calculation formula is:

[0076] ;

[0077] in, This is a historical statistical coefficient for the mining area, determined based on statistical analysis of historical instability events in the area. The range of values ​​is It can be calibrated according to the specific conditions of the mining area.

[0078] A multi-level threshold criterion is set up to determine the level of surrounding rock instability based on the threshold range of the comprehensive instability index. In this embodiment, a four-level warning level is set, and the determination rule is as follows:

[0079] when When the time is right, it is classified as stable (green).

[0080] when and When the time is right, it is classified as creep stage (blue).

[0081] when or When the time is right, it is classified as acceleration level (yellow);

[0082] when and and When this occurs, it is classified as critical (red).

[0083] The above thresholds can be adaptively adjusted according to the specific conditions of different mining areas. When an accelerated increase in the curve slope k is detected ( ), autocorrelation coefficient enhancement ( And the recovery rate is slower. When the surrounding rock is in a critical instability state, early warning measures must be taken immediately.

[0084] Step S600: Early warning output and damage mode identification.

[0085] In some embodiments, the corresponding warning level and instability probability prediction value are output based on the instability determination result. The warning levels include stable (green), creep (blue), accelerated (yellow), and critical (red), and warning information is sent to on-site managers and workers through the monitoring system interface, SMS, and audible and visual alarms.

[0086] When the warning level reaches the critical level, the damage mode identification function is activated. Crack rings with a duration exceeding a preset length threshold are extracted from the current persistence map. In this embodiment, the preset length threshold is set to 0.15m. These crack rings represent potential fracture surfaces with significant developmental characteristics.

[0087] Calculate the geometric center coordinates of each crack ring. and normal direction vector The coordinates of the geometric center are obtained by averaging the coordinates of all points that make up the crack ring; the normal direction vector is obtained by least-squares fitting of the plane containing the crack ring.

[0088] Potential failure modes are determined based on the angles between the normal direction and the coordinate axes. The determination of failure modes is based on the angles between the crack ring's normal vector and the three coordinate axes, and different failure modes are distinguished according to the principal component characteristics of the normal direction.

[0089] when When the condition is determined to be a top plate collapse mode, it indicates that the crack ring mainly develops along the horizontal direction, and the normal direction is dominated by the vertical component, which may lead to top plate delamination or collapse.

[0090] when or When the crack ring is identified as a sidewall fracture pattern, it indicates that the crack ring mainly develops along the vertical direction, and the normal direction is dominated by the horizontal component, which may lead to sidewall fractures on both sides of the roadway.

[0091] Other cases are classified as shear slip mode, indicating that the crack ring normal direction does not significantly dominate in any of the three components, and shear slip failure along the structural plane may occur.

[0092] It outputs the spatial location and failure mode prediction of potential instability areas, providing precise guidance for on-site support and emergency response.

[0093] It should be noted that the various thresholds, weighting coefficients, and parameters in the calculation formulas involved in this invention can be adaptively adjusted according to the specific engineering geological conditions and monitoring data characteristics of the mining area, so as to be optimally applicable to actual engineering scenarios.

[0094] Figure 2 This is a flowchart illustrating another method for processing three-dimensional laser scanning deformation monitoring data in coal mine goaf areas, provided by an embodiment of the present invention. It further details the specific implementation process of topological feature extraction in step S300. For example... Figure 2 As shown, the method includes:

[0095] S310. Construct the Vietoris-Rips complex, inputting the preprocessed point cloud data. Set neighborhood radius Calculate the Euclidean distance between all pairs of points in the point cloud. We construct edges from the points and then construct the simplex.

[0096] S320 uses a distributed parallel computing architecture to divide the point cloud into voxel sub-blocks, with each sub-block independently computing the local persistent graph.

[0097] S330, Boundary Stitching: The boundary stitching algorithm is used to process the simplex at the boundaries of sub-blocks to ensure the continuity of topological features and merge them to obtain a global persistent graph.

[0098] S340. Optimal scale selection: Traversing the scale parameter sequence Calculate the continuity plot at various scales and analyze the one-dimensional Betty number. Follow Based on the changing patterns, the scale with the largest rate of change was selected as the optimal analytical scale. .

[0099] S350, Continuity plot output: Outputs the continuity plot at the optimal analysis scale. , including each crack ring ( , )information.

[0100] Figure 3 This is a flowchart illustrating another method for processing three-dimensional laser scanning deformation monitoring data in coal mine goaf areas, provided by an embodiment of the present invention. It further details the specific implementation process of the critical slowdown analysis in step S500. For example... Figure 3 As shown, the method includes:

[0101] S510. Autocorrelation analysis, input topological entropy time series. Calculate the lag 1 autocorrelation coefficient ACF (1).

[0102] S520 and AR(1) models were fitted to establish a first-order autoregressive model. Estimate the autoregressive coefficients .

[0103] S530, Calculation of recovery rate index, calculation of recovery rate index .

[0104] S540. Calculate the curve slope by fitting the topological entropy evolution curve S(t). .

[0105] S550, instability probability calculation, based on the formula:

[0106] ;

[0107] Calculate the predicted probability of instability.

[0108] S560, Warning Level Determination: Determines the warning level based on multi-level threshold criteria and outputs the result.

[0109] Figure 4 This is a schematic diagram of a three-dimensional laser scanning deformation monitoring system for coal mine goaf areas, provided as an embodiment of the present invention. Figure 4 As shown, the system includes:

[0110] The point cloud data acquisition module is used to collect three-dimensional point cloud data of the surrounding rock surface of the tunnel according to the monitoring cycle. The point cloud data acquisition module includes a three-dimensional laser scanner and a track-based moving platform, which can realize automated scanning of the entire cross-section of the tunnel.

[0111] The data preprocessing module is used to perform noise reduction, filtering, and downsampling preprocessing on point cloud data.

[0112] The region division module is used to construct a three-dimensional influence cone model based on the rock strata movement angle, and to divide the strong deformation zone and the weak deformation zone.

[0113] The topology analysis module is used to construct topological complexes, calculate persistent graphs, and topological entropy.

[0114] The critical slowing analysis module is used to extract critical slowing characteristic indicators and determine the state of the surrounding rock.

[0115] The specific functions and implementation methods of the above modules have been described in detail in the method embodiments, and will not be repeated here.

[0116] The early warning output module is used to output the early warning level and damage mode based on the judgment result.

[0117] Furthermore, the region partitioning module and the topology analysis module are connected. The boundary coordinates of the strong deformation zone output by the region partitioning module are transmitted to the topology analysis module. The topology analysis module performs continuous cohomology calculation on the first neighborhood radius for the strong deformation zone and on the second neighborhood radius for the weak deformation zone. The first neighborhood radius is smaller than the second neighborhood radius.

[0118] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions will not cause the essence of the corresponding technical solutions to deviate from the protection scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for processing data from three-dimensional laser scanning deformation monitoring in coal mine goaf areas, characterized by the following steps: include: S100 Cloud Data Acquisition and Preprocessing: In the goaf roadway, multi-stage three-dimensional point cloud data of the roadway surrounding rock surface are collected according to the monitoring cycle. Each stage of point cloud data is preprocessed to obtain a point cloud set. The monitoring cycle is dynamically adjusted according to the mining depth. The greater the mining depth, the longer the monitoring cycle. S200. Monitoring area division based on rock stratum movement angle: Determine the rock stratum movement angle according to the geological and mining conditions parameters of the goaf. Construct a three-dimensional influence cone model with the goaf mining boundary as the vertex and the rock stratum movement angle as the half-cone angle. Divide the roadway into a strong deformation zone inside the influence cone and a weak deformation zone outside the influence cone. Implement a differentiated sampling strategy for the strong deformation zone and the weak deformation zone. The sampling density of the strong deformation zone is greater than that of the weak deformation zone. S300, Point Cloud Topological Feature Extraction: For each period of point cloud data, a topological complex with neighborhood radius as a parameter is constructed. The one-dimensional homology group of the complex is calculated by the persistent homology algorithm to obtain a persistent graph used to characterize the evolution features of closed cracks in the point cloud data. Multi-scale persistent homology analysis is used to determine the optimal analysis scale. S400. Topological Entropy Temporal Evolution Analysis: Calculate the topological entropy value of each period of point cloud data based on the continuous graph. The topological entropy value is calculated based on the duration of the crack loop. Establish a topological entropy temporal sequence, fit its evolution curve and calculate the curve slope. Calculate the coefficient of variation of the topological entropy temporal sequence. When the coefficient of variation exceeds a preset variation threshold, trigger the encrypted monitoring mode to shorten the monitoring cycle. S500, Critical Slowing Index Extraction and Instability Determination: Autocorrelation analysis is performed on the topological entropy time series to calculate the autocorrelation coefficient and establish an autoregressive model to extract the system recovery rate index; a comprehensive instability index is constructed based on the curve slope, autocorrelation coefficient, and recovery rate index; multi-level threshold criteria are set, and the instability state level of the surrounding rock is determined based on the threshold range of the comprehensive instability index. The autocorrelation coefficient in step S500 is the lag-1 autocorrelation coefficient, used to measure the linear correlation between the topological entropy values ​​of two adjacent periods; the autoregressive model is a first-order autoregressive model AR(1), and the specific calculation formula of the model is as follows: ； in, Let be the topological entropy value at period t. For the first The topological entropy value of the period, These are the autoregressive coefficients. The noise is white noise; the system recovery rate index is calculated based on autoregressive coefficients and is used to characterize the system's ability to recover to equilibrium after being disturbed. The comprehensive instability index uses the predicted instability probability value. The calculation formula is: ； in, This is a historical statistical coefficient for the mining area, determined based on statistical analysis of historical instability events in the area. The range of values ​​is It can be calibrated according to the specific conditions of the mining area, k is the curvature slope, and ACF(1) is the autocorrelation coefficient with a lag of 1 period. As an indicator of recovery rate; S600, Early Warning Output and Damage Mode Recognition: Output the corresponding early warning level and instability probability prediction value based on the instability determination result; when the early warning level reaches the preset level, extract the crack ring with a duration exceeding the preset length threshold from the persistence map, calculate its geometric center coordinates and normal direction vector, and determine the potential damage mode based on the angle relationship between the normal direction and the coordinate axis.

2. The method for processing three-dimensional laser scanning deformation monitoring data in coal mine goaf areas according to claim 1, characterized in that, The preprocessing in step S100 includes denoising and downsampling. The denoising process uses a statistical filtering algorithm to identify and remove noise points by setting the number of neighboring points and the standard deviation multiplier. The downsampling process uses a voxel grid filter to perform spatial uniform sampling of the point cloud by setting the voxel size.

3. The method for processing three-dimensional laser scanning deformation monitoring data in coal mine goaf areas according to claim 1, characterized in that, The rock strata movement angle in step S200 is determined according to the classification of overlying lithology, which includes hard rock strata, medium-hard rock strata and soft rock strata. Different lithologies correspond to different ranges of rock strata movement angle values. In the differentiated sampling strategy, the strong deformation zone adopts the first sampling density and the weak deformation zone adopts the second sampling density. The first sampling density is greater than the second sampling density.

4. The method for processing data of three-dimensional laser scanning deformation monitoring in coal mine goaf according to claim 1, characterized in that, The topological complex in step S300 is a Vietoris-Rips complex; the multi-scale persistent cohomology analysis calculates the persistent graph at each scale by setting a scale parameter sequence of neighborhood radius, and selects the scale with the largest one-dimensional Betti number change rate as the optimal analysis scale; the persistent cohomology calculation adopts a distributed parallel computing architecture, divides the point cloud data into several sub-blocks, calculates the local persistent graph in parallel, and then merges them into a global persistent graph through a boundary stitching algorithm.

5. The method for processing data of three-dimensional laser scanning deformation monitoring in coal mine goaf areas according to claim 1, characterized in that, The topological entropy value in step S400 The calculation is based on the following principle: the birth-scale parameter and death-scale parameter of each crack ring are obtained from the persistence graph, and the persistence length of each crack ring is calculated, wherein the persistence length is the difference between the death-scale parameter and the birth-scale parameter. The Shannon entropy of the probability distribution is calculated using the proportion of the duration of each crack loop to the total duration of all crack loops, thus obtaining the topological entropy value; the coefficient of variation is the ratio of the standard deviation to the mean of the topological entropy time series.

6. The method for processing three-dimensional laser scanning deformation monitoring data in coal mine goaf areas according to claim 1, characterized in that, The comprehensive instability index in step S500 is the predicted instability probability value. The predicted instability probability value is positively correlated with the curve slope and autocorrelation coefficient, and negatively correlated with the system recovery rate index. The multi-level threshold criterion includes multiple threshold intervals that increase sequentially. Each threshold interval corresponds to an instability state level. The instability state level includes at least a stable level, a creep level, an acceleration level, and a critical level.

7. The method for processing data of three-dimensional laser scanning deformation monitoring in coal mine goaf areas according to claim 1, characterized in that, The preset level in step S600 is the critical level; the determination of the failure mode is based on the angle relationship between the crack ring normal direction vector and the three coordinate axis directions, and the roof collapse mode, the spalling mode and the shear slip mode are distinguished according to the principal component characteristics of the normal direction; among them, the roof collapse mode corresponds to the vertical component as the main component of the normal direction, the spalling mode corresponds to the horizontal component as the main component of the normal direction, and the shear slip mode corresponds to none of the three components of the normal direction being dominant.

8. A three-dimensional laser scanning deformation monitoring data processing system for coal mine goaf areas, characterized in that, The method for processing data of three-dimensional laser scanning deformation monitoring of coal mine goaf as described in claim 1 includes: The point cloud data acquisition module is used to collect three-dimensional point cloud data of the surrounding rock surface of the roadway according to the monitoring cycle; The data preprocessing module is used to preprocess point cloud data; The region division module is used to construct a three-dimensional influence cone model based on the rock strata movement angle and divide the strong deformation zone and the weak deformation zone. The topology analysis module is used to construct topological complexes, calculate persistent graphs, and calculate topological entropy. The critical slowdown analysis module is used to extract critical slowdown characteristic indicators and determine the state of the surrounding rock. The early warning output module is used to output the early warning level and damage mode based on the judgment result.

9. The three-dimensional laser scanning deformation monitoring data processing system for coal mine goaf areas according to claim 8, characterized in that, The region partitioning module is connected to the topology analysis module. The boundary coordinates of the strong deformation zone output by the region partitioning module are transmitted to the topology analysis module. The topology analysis module performs continuous cohomology calculation on the first neighborhood radius of the strong deformation zone and on the second neighborhood radius of the weak deformation zone. The first neighborhood radius is smaller than the second neighborhood radius.

Images