A method and system for visualizing mine goaf measurement data

Abstract

The present application relates to the technical field of goaf measurement, in particular to a kind of mine goaf measurement data visualization method and system, the combination weight of direction vector and neighborhood edge tangent number value is constructed in node level, node elevation meets smoothness and position constraint in horizontal and vertical direction simultaneously, the elevation change of high difference mutation area obtains controllable interpolation in the range of multiple nodes, by using the numerical propagation characteristics of Laplacian in weight system, the coordinate adjustment quantity presents the change trend that is restricted by neighborhood direction consistency and local curvature boundary in space, the continuous form of wall node and local mutation position are synchronously expressed in the whole surface, the elevation segment generated by Gaussian process regression is written into node position record, and the reliability of through part identification, wall trend judgment and surrounding rock stability parameter reading is improved by constructing multi-scale node connection path combined with distance weight controlled by attenuation factor.

Description

Technical Field

[0001] This invention relates to the field of goaf measurement technology, and in particular to a method and system for visualizing goaf measurement data in mines. Background Technology

[0002] The field of goaf measurement technology includes multiple aspects such as goaf spatial geometry measurement, goaf boundary profile estimation, goaf volume calculation, goaf connectivity identification, surrounding rock stability parameter calculation, and goaf risk level assessment. The core task of this technology is to establish a spatial point set of the goaf, the topological relationships between point sets, and the mapping relationship between surrounding rock deformation parameters. It also organizes the three-dimensional coordinate set composed of measurement sources, including laser scanning measurements, underground positioning coordinates, borehole trajectory coordinates, and geological measurement points, into a point cloud object that can be used for spatial analysis. In the field of goaf measurement technology, it is usually necessary to construct the three-dimensional contour surface of the goaf, the cavern connection relationship, the cavern top subsidence trend, and the long-term evolution trend of the goaf based on mathematical models. This supports the systematic decision-making of mining engineering in goaf management, backfilling planning, ground pressure monitoring layout, and mining path design.

[0003] A visualization method for mine goaf measurement data is a technical solution that uses a set of measurement points and curves in the goaf as input, and generates a three-dimensional graphical representation through spatial geometry construction algorithms and graphics processing algorithms. The method aims to convert the spatial outline, detailed spatial structure, cavern topology, and surrounding rock deformation distribution of the goaf contained in the measurement data into a visual representation, enabling decision-makers to directly read the goaf's shape parameters, volume parameters, local anomalies, and potential risk areas. The desired effects of the method include generating a three-dimensional goaf model with coordinate accuracy and geometric continuity, making the goaf outline, mining... The curved surfaces of the goaf walls, the volume sections of the goaf, and the interconnected areas of the caverns are displayed in a three-dimensional coordinate space in an operable and visual manner. This enables the model to rotate, scale, section, zoom in locally, annotate wall attributes, and display volume partitions. This allows for the execution of multiple assessment tasks in goaf stability determination, goaf backfilling engineering design, mining area recovery plan formulation, and mine geological analysis. Goaf information that originally relied on a large number of numerical tables and two-dimensional profiles for judgment is transformed into visual results with structural readability, scale recognizability, and spatial logical continuity, improving the efficiency, accuracy, and operability of the goaf assessment process.

[0004] Existing technologies, when processing measurement point sets in goaf areas, largely rely on a holistic fitting method that constructs curved surfaces and spatial envelopes from point clouds. The directional differences and geometric relationships between points and their neighbors lack independent numerical records, causing local morphology to appear naturally due to the smoothness of continuous surfaces. This makes it difficult to clearly define abrupt boundaries, sharp turns, and narrow passages. Furthermore, in the point cloud geometric processing, the geometric intervals between nodes and the ratio of neighboring side lengths do not form curvature quantities that can serve as structural criteria. Consequently, local curved areas, collapsed depressions, and tunnel access areas are weakened into gradually changing regions in the model, resulting in insufficiently concentrated structural information. In terms of elevation representation, current technologies typically employ single-surface interpolation and fitting strategies based on overall control points. Elevation changes are primarily driven by global smoothing trends, lacking multi-point weighted control based on distance quantities. This leads to the averaging of local elevation anomalies, and the true distribution of abrupt elevation changes is easily flattened in the model. Although existing technologies can rotate, scale, and section the model, frequent switching of perspectives is still required for small local structures, necessitating manual comparison of multiple sections, resulting in a high visual load. In complex caverns and through areas, key structural details are easily overlooked, affecting the accuracy of risk identification and the reliability of surrounding rock stability analysis. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of existing technologies and to propose a visualization method and system for measuring data in mining goaf areas.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a method for visualizing measurement data of mining goaf areas, comprising the following steps:

[0007] S1: Based on the original measurement point coordinates of the goaf, establish the connection between the measurement point and the coordinates of the neighboring points and extract the direction vector. Combine the direction vector with the cotangent value of the neighboring edge angle to form the direction weight. Write the structural position within the neighborhood and generate the neighborhood structural weight map.

[0008] S2: Based on the neighborhood structure weight map, obtain the geometric interval between the node and the neighboring node and characterize the local curvature amplitude. Convert the ratio of the interval to the neighborhood side length into a curvature index. Combine the goaf roadway connection point criterion and write it into the weight sequence to obtain the curvature boundary discrimination sequence.

[0009] S3: Based on the neighborhood structure weight map and the curvature boundary discrimination sequence, the two types of weights of the nodes are merged, and the merged weights are applied to the node coordinates to obtain the coordinate adjustment amount. The coordinate adjustment amount is incorporated into the original node coordinates and written into the list according to the number to construct the wall node coordinate list.

[0010] S4: Based on the list of wall node coordinates, extract the node coordinates and establish a distance metric with the reference coordinates of the goaf elevation difference section. Establish a weight function with the distance metric and Gaussian process regression and apply it to the node elevation. Collect the weighted results in the order of node number and output the interpolated node elevation segment.

[0011] S5: Based on the interpolated node elevation segment, establish distance quantities with neighboring nodes through node coordinates, introduce the distance quantities into the attenuation factor to construct weights and write them into the node positions along with the node elevations, and then assemble the node coordinate sequence according to the predetermined number to establish a multi-scale wall space block.

[0012] As a further embodiment of the present invention, the neighborhood structure weight map includes direction vector records, angle cotangent value records, and structure position records; the curvature boundary discrimination sequence includes geometric interval records, curvature index records, and tunnel connection point discrimination records; the wall node coordinate list includes original node coordinates, coordinate adjustment amounts, and updated node coordinates; the interpolated node elevation segment includes node distance measurement, weight function results, and weighted elevation; and the multi-scale wall space block includes node coordinate sequences, weights constructed from distance measurements, and node elevation entries.

[0013] As a further aspect of the present invention, the specific steps for generating the neighborhood structure weight map are as follows:

[0014] Based on the original coordinates of the measurement points in the goaf, the measurement points are obtained sequentially by number within the spatial range of the measurement points, and the corresponding neighboring point positions are extracted. The correspondence between the measurement point positions and the neighboring point positions is established and organized into a structural sequence to form a set of neighboring difference vectors.

[0015] Based on the neighborhood difference vector set, the direction of each vector in the set is determined and the corresponding neighborhood edge angle cotangent value is extracted. The direction determination content and cotangent value are combined according to the node neighborhood structure to form the direction weights used to construct the Laplacian operator weights and then written into the weight sequence to obtain the neighborhood direction weight set.

[0016] Based on the neighborhood directional weight set, directional weights are collected in the neighborhood range of each measurement point according to the measurement point number, and the collected content is written into a preset structural position. The records are stored in a fixed index manner to form a continuous distribution record, thereby generating a neighborhood structural weight map.

[0017] As a further aspect of the present invention, the Laplace operator, based on the neighborhood structure weight map, fills the directional weights from each node to the neighboring nodes into the row and column positions of the coefficient matrix according to the node number, writes the node self-weights associated with the combination of neighboring directional weights in the diagonal position, establishes a matrix corresponding to the triangular grid of the goaf, reads the coordinates of each node in the wall node coordinate list according to the number, writes the node coordinates into the coordinate column to be updated and adjusts the coordinates according to the weight order corresponding to each row of the matrix, generates node coordinate increment records, and merges the coordinate increments into the original node coordinates according to the number of the node coordinate increment records and writes them into a new coordinate list to form an updated wall node coordinate list.

[0018] As a further aspect of the present invention, the specific steps for obtaining the curvature boundary discrimination sequence are as follows:

[0019] Based on the neighborhood structure weight map, the spatial coordinates of the nodes and neighboring nodes are extracted and the directional records are established by the direction of the line connecting the nodes to the neighboring nodes. After being arranged into a continuous structure column according to the node number order, the arrangement rules are uniformly defined to generate an interval sequence.

[0020] Based on the interval sequence, the neighborhood polygon contour is read and the spatial morphological differences between nodes are compared in the order of node relationships. The morphological differences are written into the orientation structure column according to the node number and organized into a continuous structure record in the index order to generate a curvature index sequence.

[0021] Based on the curvature index sequence, the information of the roadway connection points in the goaf is read and the spatial morphological characteristics of the nodes are compared in the order of node numbers. The comparison content is written into the boundary record column and stored in a fixed structural order to form a boundary sequence, thereby generating a curvature boundary discrimination sequence.

[0022] As a further aspect of the present invention, the specific steps for constructing the list of wall node coordinates are as follows:

[0023] Based on the neighborhood structure weight map and the curvature boundary discrimination sequence, two types of weight records are extracted by nodes and a weight corresponding path is established based on the node position. After being written into the calibration column in order through the node index, the calibration column is continuously sorted to form a continuous offset record and generate a coordinate adjustment list.

[0024] Based on the coordinate adjustment list, offset records are extracted by node number and embedded at the original node position as a reference. The coordinates are written into the coordinate column in numerical order, and the coordinate column is uniformly organized to form a continuous storage sequence while maintaining the consistency of the index structure, thus generating a list of wall node coordinates.

[0025] As a further aspect of the present invention, the specific steps for outputting the interpolation node elevation segment are as follows:

[0026] Based on the list of wall node coordinates, the node positions are extracted by number and a corresponding path is formed by the node positions and the reference coordinates of the goaf elevation difference section. After recording the relative displacement information through the direction of the node, the information is written into the position column in numerical order and organized to form a structural record, generating the position connection sequence antecedent.

[0027] Based on the location-related sequence antecedent, node association items are extracted by number and an input sequence is established by recording node positions. The nodes are written into the elevation association column according to the node arrangement order and uniformly organized to form a node input structure column, thereby generating the node input sequence.

[0028] Based on the node input sequence, the node position records are extracted by number and written into the predicted elevation column through Gaussian process regression. They are then merged into a continuous elevation structure column and the sequence index is sorted to maintain the consistency of the structural order, generating interpolated node elevation segments.

[0029] As a further embodiment of the present invention, the Gaussian process regression records the node positions in the model input column according to the numbering order of the node input sequence. After establishing the relationship structure between nodes through the node input column and writing the relationship structure into the internal record of the model, it is introduced into the model calculation path according to the node number. The predicted elevation record is generated in the model calculation path and written into the output column according to the number. Then, the output column is organized into a continuous structure sequence to form the predicted elevation structure column.

[0030] As a further aspect of the present invention, the specific steps for establishing the multi-scale wall space block are as follows:

[0031] Based on the interpolated node elevation segment, the node coordinates are extracted by number and a connection path is established with the neighboring node positions through the node coordinates. The orientation between nodes is recorded to form node correspondence items and written into the record column by number. The record column is sorted in order to form a continuous structure sequence, and a node position sequence is generated.

[0032] Based on the node position sequence, the corresponding items of the nodes are extracted by number and written into the weight column by setting the attenuation coefficient. The weight column and the corresponding items of the nodes are merged to form a structural unit. The node coordinate sequence is assembled according to the predetermined number and the sequence is continuously integrated to form a spatial block, generating a multi-scale wall spatial block.

[0033] A visualization system for mine goaf measurement data, the system being used to execute the aforementioned visualization method for mine goaf measurement data, the system comprising:

[0034] Neighborhood weight construction module: Based on the original coordinates of the measurement points in the goaf, establish the connection between the coordinates of the measurement points and the coordinates of the neighboring points, calculate the direction vector and combine it with the cotangent values ​​of the neighboring edge angles, write the direction weights into the node neighborhood records according to the structural positions, and generate a neighborhood structural weight map.

[0035] Curvature boundary calculation module: Based on the neighborhood structure weight map, obtain the geometric interval between the node and the neighboring node, convert the interval and the ratio of the neighborhood side length into a curvature index, and write it into the weight sequence according to the goaf roadway connection point criterion. Integrate the numbered arrangement results to obtain the curvature boundary discrimination sequence.

[0036] Wall coordinate generation module: Based on the neighborhood structure weight map and the curvature boundary discrimination sequence, extract the two types of node weights and merge them to form coordinate adjustment amount, add the coordinate adjustment amount to the original node coordinates, write it into the record list according to the number, organize it into a continuous coordinate sequence, and construct the wall node coordinate list;

[0037] Elevation interpolation calculation module: Based on the list of wall node coordinates, extract the node coordinates and establish a distance metric with the reference coordinates of the goaf elevation difference section. Write the distance metric into the weight function and apply it to the node elevation. Collect the weighted results in sequence according to the node number and output the interpolated node elevation segment.

[0038] Multi-scale block construction module: Based on the interpolated node elevation segment, extract the node coordinates and establish distance with neighboring nodes, write the distance into the attenuation factor to generate weights, and write them into the node position record along with the node elevation. Assemble the node coordinate sequence according to the predetermined number to establish a multi-scale wall space block.

[0039] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0040] 1. In this invention, by constructing a combination of weights of direction vectors and cotangent values ​​of neighborhood edge angles at the node level, the spatial relationship between nodes is transformed into a weight system that can be used to describe directional differences and geometric constraints. Moreover, the node elevation satisfies both smoothness and position constraints in both the horizontal and vertical directions, and the elevation change in the elevation change zone of the elevation change abrupt change is controllable within the range of multiple nodes.

[0041] 2. In this invention, by utilizing the numerical propagation characteristics of the Laplacian operator in the weight system, the directional weights and curvature boundary discrimination sequence are applied together to the node coordinates, so that the coordinate adjustment amount in space presents a change trend that is constrained by both the consistency of the neighborhood direction and the local curvature boundary. The continuous shape of the wall node and the position of the local abrupt change are synchronously expressed in the overall surface.

[0042] 3. In this invention, by writing the elevation fragments generated by Gaussian process regression into the node position record, and constructing multi-scale node connection paths by combining the distance weight controlled by the attenuation factor, the resulting spatial block maintains consistency in elevation continuity, local boundary expression, and overall structural coordination. This enables the goaf wall to have higher geometric continuity, abnormal structure presentation, and elevation control accuracy in three-dimensional expression, thereby improving the reliability of identifying penetration points, judging wall trend, and reading surrounding rock stability parameters. Attached Figure Description

[0043] Figure 1 This is a schematic diagram of the workflow of the present invention;

[0044] Figure 2 This is a system flowchart of the present invention. Detailed Implementation

[0045] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0046] Example 1

[0047] Please see Figure 1 This invention provides a technical solution: a method for visualizing measurement data of mining goaf areas, comprising the following steps:

[0048] S1: Based on the original measurement point coordinates of the goaf, establish the connection between the measurement point and the coordinates of the neighboring points and extract the direction vector. Combine the direction vector with the cotangent value of the neighboring edge angle to form the direction weight. Write the structural position within the neighborhood and generate the neighborhood structural weight map.

[0049] S2: Based on the neighborhood structure weight map, obtain the geometric interval between the node and the neighboring node and characterize the local bending amplitude. Convert the interval and the ratio of the neighborhood side length into a curvature index. Combine the goaf roadway connection point criterion and write it into the weight sequence to obtain the curvature boundary discrimination sequence.

[0050] S3: Based on the neighborhood structure weight map and curvature boundary discrimination sequence, the two types of weights of the nodes are merged. The merged weights are applied to the node coordinates to obtain the coordinate adjustment amount. The coordinate adjustment amount is incorporated into the original node coordinates and written into the list according to the number to construct the wall node coordinate list.

[0051] S4: Based on the list of wall node coordinates, extract the node coordinates and establish a distance metric with the reference coordinates of the goaf elevation difference section. Establish a weight function with the distance metric and Gaussian process regression and apply it to the node elevation. Collect the weighted results in the order of node number and output the interpolated node elevation segment.

[0052] S5: Based on the interpolated node elevation segment, establish distance quantities between the node coordinates and neighboring nodes, introduce the distance quantities into the attenuation factor to construct weights and write them into the node positions along with the node elevations, and then assemble the node coordinate sequence according to the predetermined number to establish a multi-scale wall space block.

[0053] The neighborhood structure weight map includes direction vector records, angle cotangent numerical records, and structural position records. The curvature boundary discrimination sequence includes geometric interval records, curvature index records, and tunnel connection point discrimination records. The wall node coordinate list includes the original node coordinates, coordinate adjustment amounts, and updated node coordinates. The interpolated node elevation fragment includes node distance metrics, weight function results, and weighted elevations. The multi-scale wall space block includes node coordinate sequences, weights constructed from distance metrics, and node elevation entries.

[0054] The specific steps for generating the neighborhood structure weight map are as follows:

[0055] Based on the original coordinates of the measurement points in the goaf, the measurement points are obtained sequentially by number within the spatial range of the measurement points, and the corresponding neighboring point positions are extracted. The correspondence between the measurement point positions and the neighboring point positions is established and organized into a structural sequence to form a set of neighboring difference vectors.

[0056] Based on the neighborhood difference vector set, the direction of each vector in the set is determined and the corresponding neighborhood edge angle cotangent value is extracted. The direction determination content and cotangent value are combined according to the node neighborhood structure to form the direction weights used to construct the Laplacian operator weights and then written into the weight sequence to obtain the neighborhood direction weight set.

[0057] Based on the neighborhood directional weight set, the directional weights are collected in the neighborhood range of each measurement point according to the measurement point number, and the collected content is written into the preset structure position. The records are stored in a fixed index to form a continuous distribution record, generating a neighborhood structure weight map.

[0058] Based on the original coordinates of the measurement points in the goaf, the three-dimensional coordinate values ​​of each measurement point are obtained sequentially by number within the spatial range of the measurement points. For each measurement point, the location of neighboring points is retrieved within a 3-meter radius. During the retrieval process, the three-axis coordinate differences of all neighboring points are recorded. The difference ranges are: x-axis difference between 0.5 meters and 2.4 meters, y-axis difference between 0.3 meters and 1.8 meters, and z-axis difference between 0.6 meters and 3.1 meters. The three-axis differences between the measurement point coordinates and the neighboring point coordinates are combined into a single vector record in vector order. Each item in the vector record consists of three sets of spatial differences and is written into a structure sequence according to the measurement point number. Typically, 4 to 7 vector records are formed for one measurement point. All the formed vector records are arranged in order from 1 to 300 according to the measurement point number, so that the records are arranged in ascending order of number and organized into a structured vector sequence in a continuous storage manner, forming a set of neighborhood difference vectors.

[0059] Based on the neighborhood vector set, the direction category of each vector in the set is determined by the positive and negative combination of the differences of the x, y, and z axes. When the x difference is greater than 1 meter and is positive, it is recorded as the forward direction; when the y difference is less than 0.4 meters and is positive, it is recorded as the lateral rightward direction; and when the z difference is negative and the absolute value is greater than 1.5 meters, it is recorded as the downward direction. The cotangent values ​​of the neighborhood edge angles are extracted through fixed angle samples, using 30 degrees, 45 degrees, and 60 degrees as reference angles, with corresponding cotangent values ​​of 1.732, 1.000, and 0.577. The direction category and cotangent value are combined item by item according to the order of the node's neighborhood, so that each node forms a direction weight sequence consisting of direction determination items and cotangent combination items. Each measurement point is usually written with 5 to 9 direction weights. The direction weights are written into the record column according to the node number order to obtain the neighborhood direction weight set.

[0060] Based on the neighborhood directional weight set, the neighborhood directional weights are aggregated according to the measurement point number. A preset structural position is set for each measurement point. The preset structural position uses a range of starting index value 20 and ending index value 80 to store the aggregated content, so that each measurement point obtains 61 consecutive weight positions. All directional weights from the same measurement point are written into a fixed index range in ascending order of number. During the writing process, the corresponding positions of each weight item are kept consistent and do not overlap or skip indexes. The aggregated records of all measurement points are arranged in numerical order, so that the overall record is arranged from the 1st measurement point to the 300th measurement point, and the index range is kept consistent, generating a neighborhood structural weight map in a continuous distribution form.

[0061] The Laplace operator, based on the neighborhood structure weight map, fills the directional weights from each node to its neighboring nodes into the row and column positions of the coefficient matrix according to the node number. It writes the node self-weights associated with the combination of neighborhood directional weights in the diagonal position, establishes a matrix corresponding to the triangular grid of the goaf, and reads the coordinates of each node in the wall node coordinate list according to the number. It writes the node coordinates into the coordinate column to be updated and adjusts the coordinates according to the weight order of each row of the matrix to generate node coordinate increment records. The coordinate increment records are merged into the original node coordinates according to the number and written into the new coordinate list to form the updated wall node coordinate list.

[0062] The Laplace operator, according to the formula:

[0063]

[0064] in: For the first Incremental vector of node coordinates For the first The three-dimensional coordinate vectors recorded in the list of wall node coordinates in the current iteration of each node. For the first The original 3D coordinate vectors written by each node in the initial stage of constructing the wall node coordinate list. In order to be with the first The node has a triangular mesh connection relationship and is involved in the calculation in the current iteration. Three-dimensional coordinate vectors of neighboring nodes The first value is read from the neighborhood structure weight map according to the node number and neighborhood number, and matched with the direction weight. The node to the first The directional weights of each neighboring node. For the first A set of neighboring node indexes for each node. The first node read from the curvature boundary discrimination sequence by node number Individual node curvature coefficients To pass the first triangle on the corresponding triangular mesh The coordinates of the nth node and its adjacent nodes are calculated and normalized to obtain the nth node. Unit normal vector of local wall surface at each node To read and indicate the first curvature boundary discrimination sequence The boundary marker coefficient indicating whether a node is located at the boundary of a roadway connection is determined when the node is... When a node is determined to be a boundary node Values When the first When a node is determined to be a non-boundary node Values , These are the Laplace smoothing weights, used to control the numerical scale of the Laplace term during coordinate updates. These are curvature correction weighting coefficients used to control the curvature coefficient. With normal vector The strength of the participation of the combined term in coordinate updates Weighting coefficients are used to restore the boundary and control the boundary marking coefficients. Difference term from original coordinates The intensity of participation in coordinate updates;

[0065] Execution process: Update the node numbers in the coordinate list of the traversed wall nodes for each round of coordinate iteration. (This is done while processing the first...) When there are nodes, first, in the neighborhood structure weight map, according to the node number... Retrieve Neighborhood Index Set And read the directional weights from the node to each neighboring node in the set. At the same time, read the first node from the wall node coordinate list. Current coordinates of each node and the current coordinates of each neighboring node Calculate the differences of each vector according to the order of the neighborhood list. The three-dimensional components and their corresponding After performing term-by-term product, summation is performed within the neighborhood to obtain the Laplace term, which is then obtained by numbering nodes from the curvature boundary discrimination sequence. Extracting curvature coefficients With boundary marker coefficient In the triangular mesh structure, the first Calculate the local wall normal vector based on the coordinates of each node and its adjacent nodes, and then normalize the vectors by length to obtain the unit normal vector. During the initialization phase, the original coordinates of the nodes are read from the list of wall node coordinates and the original measurement data. Before this iteration, the directional weights of all nodes are used. Calculate the metric for each node And determine the maximum metric. ,set up for Calculate the average curvature coefficient of all nodes according to the curvature coefficient set. and set Then, the number of boundary nodes is determined based on the curvature boundary discrimination sequence. With the total number of nodes And based on the ratio Location range setting for and and In the computation phase of a single node, the Laplace term is multiplied by curvature coefficient With normal vector Multiply by Original coordinate difference With boundary marker coefficient Multiply by The coordinate increment vector of the node is obtained by performing component-by-component addition of the three-dimensional coordinate components of the vector. and will Write the node coordinates into the node coordinate increment record in order of node number, so that they can be superimposed onto the corresponding node coordinates in the wall node coordinate list in subsequent steps.

[0066] The specific steps to obtain the curvature boundary discrimination sequence are as follows:

[0067] Based on the neighborhood structure weight map, the spatial coordinates of nodes and neighboring nodes are extracted and the directional records are established by the direction of the line connecting the nodes to the neighboring nodes. After being arranged into a continuous structure column according to the node number order, the arrangement rules are uniformly defined to generate an interval sequence.

[0068] Based on the interval sequence, the contour of the neighborhood polygon is read and the spatial morphological differences between nodes are compared in the order of node relationship. The morphological differences are written into the orientation structure column according to the node number and organized into a continuous structure record in the index order to generate the curvature index sequence.

[0069] Based on the curvature index sequence, the information of the roadway connection points in the goaf is read and the spatial morphological characteristics of the nodes are compared in the order of node numbers. The comparison content is written into the boundary record column and stored in a fixed structure order to form a boundary sequence, thereby generating a curvature boundary discrimination sequence.

[0070] Based on the neighborhood structure weight map, the spatial coordinates of each node and its neighboring nodes are extracted in numerical order. The x, y, and z coordinates of each node and the corresponding neighboring nodes are read. The difference between the lines connecting the node and its neighboring nodes is calculated, with the x-direction difference set between 0.1 meters and 5.0 meters, the y-direction difference set between 0.1 meters and 5.0 meters, and the z-direction difference set between 0.1 meters and 5.0 meters. The three-axis differences are written into the azimuth record in a fixed order. The direction of the connection is written into the azimuth number in the order from the node to the neighboring node. All connection records of the same node are numbered sequentially from 1 to the maximum number of node connections of 20. The connection records of each node are spliced ​​into a single structural column in the order of node number. The structural column is rounded down by a step size of 0.01 meters, so that the distance value in each record retains two decimal places and corresponds to the corresponding node number. The distance values ​​of each record in the structural column are written into the numerical sequence in order to generate the interval sequence.

[0071] Based on the interval sequence, the interval values ​​associated with each node are read in numerical order. The three adjacent interval values ​​of the same node are combined into a set of side length data. The coordinates of the vertices corresponding to the node in the neighborhood contour are read. The coordinates of the node and its two adjacent vertices are written into the contour three-point sequence. The length of the line connecting the three points in the contour three-point sequence is compared with the side length data in the interval sequence. The angle range between each set of three points is set to 5 degrees to 175 degrees. The angle is divided into several levels in 5-degree increments and each level is assigned an integer code. The side length codes and angle codes in all contour three-point sequences of the same node are written into the orientation structure column in sequence. The node number and local sequence number index are corresponding to each position in the orientation structure column. In the local range of each node, the three-point coordinates and three side lengths are used as input. The fitting neighborhood radius is set to 1.0 meter and the arc length segment length is set to 0.5 meters. The curvature value is obtained by combining the arc length and angle code segment by segment. The curvature value is written into the structure record in the order of node number and local sequence number to generate the curvature index sequence.

[0072] Based on the curvature index sequence, the information of roadway connection points in the goaf is read. The three-dimensional coordinates of each roadway connection point are arranged in numerical order. The curvature index value corresponding to each node is compared with the preset curvature segmentation threshold in a hierarchical manner. The preset curvature segmentation threshold is set with a first-level threshold of 0.10, a second-level threshold of 0.50, and a third-level threshold of 1.00. Curvature less than or equal to 0.10 is recorded as first-level code 1, curvature greater than 0.10 and less than or equal to 0.50 is recorded as second-level code 2, and curvature greater than 0.50 is recorded as third-level code 3. The empty space of each node is then compared. The distance between the coordinates and the coordinates of all roadway connection points is calculated. The matching radius is set to 3.0 meters. The case where the distance is less than or equal to 3.0 meters is recorded as connection mark 1, and the case where the distance is greater than 3.0 meters is recorded as connection mark 0. The curvature code and connection mark are written into the boundary record column in the order of node number. Each node record in the boundary record column is stored in a fixed structure order, where the first position is the curvature code and the second position is the connection mark. The node boundary records are arranged continuously from number 1 to the maximum number to form a boundary sequence, and a curvature boundary discrimination sequence is generated.

[0073] The specific steps for constructing the list of wall node coordinates are as follows:

[0074] Based on the neighborhood structure weight map and curvature boundary discrimination sequence, two types of weight records are extracted by nodes and weight corresponding paths are established based on node positions. After being written sequentially into the calibration column through node index, the calibration column is continuously organized to form continuous offset records and generate a coordinate adjustment list.

[0075] Based on the coordinate adjustment list, the offset records are extracted by node number and embedded in the position based on the original node position. The coordinates are written into the coordinate column in numerical order, and the coordinate column is uniformly organized to form a continuous storage sequence while maintaining the consistency of the index structure, thus generating a list of wall node coordinates.

[0076] Based on the neighborhood structure weight map and curvature boundary discrimination sequence, two types of weight records are extracted for each node. The direction weight value and boundary discrimination value are read for each node. The direction weight is set to a range of 0.01 to 2.00 according to the record order, and the boundary discrimination value is set to integers 0 and 1. For each node, the x, y, and z coordinates are read according to its 3D position, and the position is used as the reference position for the path corresponding to the weight. The two types of weights are written into the calibration column in the order of the node's internal structure. Each record in the calibration column contains the direction weight value and the boundary discrimination value. The numerical values ​​and node numbers are used to sequentially arrange the calibration columns from 1 to the maximum node number while maintaining a continuous arrangement. When processing the calibration columns, the numerical translation step size in the weight path registration algorithm is set to 0.05, and the direction weights and boundary discrimination values ​​are translated item by item. The translated direction weights and boundary discrimination values ​​are written into the offset record one by one, so that each node forms a record segment containing no less than 3 offset records. The node offset records are concatenated in the order of node numbers to form a continuous offset record sequence, generating a coordinate adjustment list.

[0077] Based on the coordinate adjustment list, offset records are extracted by node number, and the offset values ​​in the offset records of each node are read. The offset values ​​are set to range from 0.01 meters to 1.20 meters in the x-direction, from 0.01 meters to 1.00 meters in the y-direction, and from 0.01 meters to 0.80 meters in the z-direction. The offsets are embedded using the original node positions as a reference, with the actual node number as the sequence control condition. The embedded coordinates are then written into a coordinate column, where each item contains... The coordinate column includes node number, offset x-coordinate, offset y-coordinate, and offset z-coordinate. When organizing the coordinate column, the fixed index structure setting method in the sequence coordinate assembly method is adopted. The starting position of the coordinate index of each node is set to the current node number multiplied by 3 plus 1, and the ending position is set to the current node number multiplied by 3 plus 3, ensuring that each node coordinate occupies three consecutive positions. The coordinate columns of all nodes are merged and stored in ascending order of node number from 1 to the maximum number, so that the entire coordinate list maintains a continuous arrangement structure, generating a wall node coordinate list.

[0078] The specific steps for outputting the interpolation node elevation segment are as follows:

[0079] Based on the list of wall node coordinates, the node positions are extracted by number and a corresponding path is formed by the node positions and the reference coordinates of the goaf elevation difference section. After recording the relative displacement information through the direction of the node, the information is written into the position column in numerical order and organized to form a structural record, generating the position relationship sequence antecedent.

[0080] Based on the location-related sequence antecedents, node association items are extracted by number and an input sequence is established by recording node positions. The nodes are written into the elevation association column according to the node arrangement order and uniformly organized to form a node input structure column, thus generating the node input sequence.

[0081] Based on the node input sequence, the node position records are extracted by number and written into the predicted elevation column through Gaussian process regression. Then, they are merged into a continuous elevation structure column and the sequence index is sorted to maintain the consistent structural order, generating interpolated node elevation segments.

[0082] Based on the list of wall node coordinates, the node positions are extracted by number, and the x, y, and z coordinates of each node are read. The node coordinates are then matched with the reference coordinates of the goaf elevation difference section to construct a corresponding path. The number of sampling points for the reference coordinates is set to 50, and each point contains three-dimensional coordinates arranged in numerical order. The coordinate differences between the node and the reference point are recorded item by item in the x-direction (0.1 m to 4.0 m), y-direction (0.1 m to 4.0 m), and z-direction (0.1 m to 6.0 m). The difference records are written into the direction record item according to the direction of the node. The direction record item is divided into forward code 1, lateral code 2, and vertical code 3 according to the preset direction encoding rules in the node path construction algorithm and written into the direction field. The direction record and difference record of each node are combined and written into the position column. The position column is arranged into a continuous structure record according to the node number order, so that each node corresponds to a continuous record segment. All node records are spliced ​​in sequence to form a structure column, generating the position relationship sequence antecedent.

[0083] Based on the position-related sequence antecedent, the node association items are extracted by number and the node number, position coordinate difference and direction code contained in the node position record are read in sequence. The node position records are arranged according to the sequence rearrangement method set by the node sequence recombination algorithm. The step size in the rearrangement method is set to 1 and the node numbers are arranged in a continuously increasing manner. The association items of each node are written into the elevation association column. Each record in the elevation association column is stored in the order of node number. The elevation association columns of each node are combined into node input items. All node input items are written into the node input structure column in the order of number. The index in the node input structure column is set to the initial position starting from 1 and increasing in sequence according to the number of nodes. The overall structure is kept continuous and without skips, so that the input records of the nodes are stored in the preset order, and the node input sequence is generated.

[0084] Based on the node input sequence, a Gaussian process regression algorithm is used to extract node position records by number and add node coordinates as input to the regression process. The kernel function of the Gaussian process regression algorithm is set to radial basis kernel type, the length scale parameter is set to 1.5, the amplitude parameter is set to 2.0, and the variance of the noise term is set to 0.05. The coordinate values ​​of each node in the node input sequence are input into the regression process in the order of node number. The elevation prediction values ​​output from the regression results are written into the predicted elevation column. Each record in the predicted elevation column is stored in the order of node number to form a continuous elevation record segment. The continuous elevation record segments are arranged from node number 1 to the maximum number. The elevation record segments are merged and the sequence index is sorted to maintain the structural order consistency, generating interpolated node elevation segments.

[0085] Gaussian process regression records the node positions in the model input column according to the node input sequence number. The relationship structure between nodes is established through the node input column and written into the internal record of the model. Then, the node is introduced into the model calculation path according to the node number. The predicted elevation record is generated in the model calculation path and written into the output column according to the number. Finally, the output column is organized into a continuous structure sequence to form the predicted elevation structure column.

[0086] Gaussian process regression, according to the formula:

[0087]

[0088] in: Specifically, the first The predicted output value of the interpolation node elevation. Specifically, it refers to the elevation offset parameter. Specifically, it refers to the elevation weighting coefficient. Specifically, the first The interpolation node is associated with the reference elevation value of the elevation datum section. Specifically, it is the strength parameter of the covariance function. Specifically, the first The predicted node's three-dimensional spatial position vector. Specifically, the first The three-dimensional spatial position vector of each training node. Specifically, the quantification standard is a covariance function length scale parameter of 1.5. Specifically, it refers to the total number of training nodes. Specifically, 1 to The range of values ​​is the training node index number. Specifically, the covariance matrix. Specifically, the quantification standard is 0.05 for the noise term variance. Specifically, the identity matrix. Specifically, it is an anisotropic weight diagonal matrix. Specifically, it is a symmetric matrix. Specifically, it is the column vector of observed elevations of the training nodes. Specifically, the first element in the matrix operation result vector... Item element extraction operation, Specifically, the natural constant It is a base exponential function. Specifically, it involves second-order norm operations;

[0089] Execution process: First, extract the coordinates of multiple wall surface nodes and the numerical reference coordinates of the elevation difference section in the goaf, construct a set of multiple training nodes and associate them with the observation elevation column vector. The second step is to apply the radial basis kernel function to calculate the three-dimensional spatial position vectors between multiple training nodes. Euclidean distance squared value To construct the covariance matrix The third step is to specify the quantification standard as 0.05 for the noise term variance. Associated identity matrix Generate noise correction terms The fourth step is to construct a symmetric matrix based on the squared vertical distance between training nodes. And associated anisotropic weight diagonal matrix Generate anisotropic terms The fifth step is to execute the combination matrix. Inverse matrix operation and AND with vector Multiplication yields the regression coefficient sequence; the sixth step is to extract the three-dimensional spatial position vectors of the interpolation nodes. And calculate with Distance function term The seventh step is to input the intensity parameter of the covariance function. Associated with the distance function term and accumulated with the regression coefficient sequence, superimposed with the elevation offset parameter Weighting coefficient of elevation Associated reference elevation This constitutes a linear offset term, outputting the predicted elevation values ​​of the interpolation nodes. .

[0090] The specific steps for constructing a multi-scale wall space block are as follows:

[0091] Based on the interpolated node elevation segment, the node coordinates are extracted by number and a connection path is established with the location of neighboring nodes through the node coordinates. The orientation between nodes is recorded to form corresponding node items and written into the record column by number. The record column is sorted in order to form a continuous structure sequence, and a node position sequence is generated.

[0092] Based on the node position sequence, the corresponding items of the nodes are extracted by number and written into the weight column by setting the attenuation coefficient. The weight column and the corresponding items of the nodes are merged to form structural units. The node coordinate sequence is assembled according to the predetermined number and the sequence is continuously integrated to form a spatial block, generating a multi-scale wall spatial block.

[0093] Based on the interpolated node elevation segments, node coordinates are extracted by number, and the corresponding x, y, and z coordinates of each node are read. The three-dimensional positions of each node are compared with those of all neighboring nodes one by one. The differences between the node coordinates and the neighboring node coordinates are recorded in the ranges of 0.1 meters to 6.0 meters in the x direction, 0.1 meters to 6.0 meters in the y direction, and 0.1 meters to 8.0 meters in the z direction. The three-axis differences are arranged into connecting paths in a fixed order. Each connecting path is written into the path record item according to the node number order. The path record item is encoded using the preset direction encoding method in the spatial connection path construction algorithm. The direction codes are divided into six direction categories from number 1 to number 6 and correspond to different combinations of three-axis differences. The direction codes and path record items are combined into node corresponding items. The node corresponding items are written into the record column according to the node number. The record column is sorted sequentially from number 1 to all node numbers. The sorted record column is combined to form a continuous structure sequence, generating a node position sequence.

[0094] Based on the node position sequence, the corresponding items of the nodes are extracted by number, and the three-dimensional connection difference and direction code in each corresponding item of the nodes are read one by one. The attenuation coefficient is set to three fixed levels of 0.20, 0.35 and 0.50, and matching is performed according to the distance between the node and the neighboring nodes within the range of 0.5 meters to 10.0 meters. The matched attenuation coefficient is written into the weight column. The weight column and the corresponding items of the nodes are merged one by one to form structural units. Each structural unit contains a node number, direction code, three-dimensional difference and corresponding attenuation coefficient. The structural units are written into the node coordinate sequence in a predetermined numbering order. The three-dimensional structural unit assembly method is used to integrate the node coordinate sequence in a continuous manner based on the number. The index range of the coordinate storage position of each node is preset during the integration process. The starting index is set to the current node number multiplied by 4 plus 1, and the ending index is set to the current node number multiplied by 4 plus 4. The structural units of all nodes are written into the continuous storage sequence in the index order, and finally a multi-scale wall space block is generated.

[0095] Please see Figure 2 A visualization system for measurement data of mine goaf areas, the system comprising:

[0096] Neighborhood weight construction module: Based on the original coordinates of the measurement points in the goaf, establish the connection between the coordinates of the measurement points and the coordinates of the neighboring points, calculate the direction vector and combine it with the cotangent values ​​of the neighboring edge angles, write the direction weights into the node neighborhood records according to the structural positions, and generate a neighborhood structural weight map.

[0097] Curvature boundary calculation module: Based on the neighborhood structure weight map, the geometric interval between the node and the neighboring node is obtained, the ratio of the interval to the neighborhood side length is converted into a curvature index, and written into the weight sequence according to the goaf roadway connection point criterion. The numbered arrangement results are integrated to obtain the curvature boundary discrimination sequence.

[0098] Wall coordinate generation module: Based on the neighborhood structure weight map and curvature boundary discrimination sequence, extract the two types of node weights and merge them to form coordinate adjustment amount. Add the coordinate adjustment amount to the original node coordinates, write it into the record list according to the number, organize it into a continuous coordinate sequence, and construct the wall node coordinate list.

[0099] Elevation interpolation calculation module: Based on the list of wall node coordinates, extract the node coordinates and establish a distance metric with the reference coordinates of the goaf elevation difference section. Write the distance metric into the weight function and apply it to the node elevation. Collect the weighted results in sequence according to the node number and output the interpolated node elevation segment.

[0100] Multi-scale block construction module: Based on the interpolated node elevation fragment, extract the node coordinates and establish distance with neighboring nodes. Write the distance into the attenuation factor to generate weights, and write them into the node position record along with the node elevation. Assemble the node coordinate sequence according to the predetermined number to build a multi-scale wall space block.

[0101] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A method for visualizing measurement data of mined-out areas, characterized in that, Includes the following steps: S1: Based on the original measurement point coordinates of the goaf, establish the connection between the measurement point and the coordinates of the neighboring points and extract the direction vector. Combine the direction vector with the cotangent value of the neighboring edge angle to form the direction weight. Write the structural position within the neighborhood and generate the neighborhood structural weight map. S2: Based on the neighborhood structure weight map, obtain the geometric interval between the wall node and the neighboring wall node and characterize the local bending amplitude. Convert the interval and the ratio of the neighborhood side length into a curvature index. Combine the goaf roadway connection point criterion and write it into the weight sequence to obtain the curvature boundary discrimination sequence. S3: Based on the neighborhood structure weight map and the curvature boundary discrimination sequence, the two types of weights of the wall nodes are merged, and the merged weights are applied to the coordinates of the wall nodes to obtain the coordinate adjustment amount. The coordinate adjustment amount is incorporated into the original coordinates of the wall nodes and written into the list according to the number to construct the updated list of wall node coordinates. The specific steps for generating the neighborhood structure weight map are as follows: Based on the original coordinates of the measurement points in the goaf, the measurement points are obtained sequentially by number within the spatial range of the measurement points, and the corresponding neighboring point positions are extracted. The correspondence between the measurement point positions and the neighboring point positions is established and organized into a structural sequence to form a set of neighboring difference vectors. Based on the neighborhood difference vector set, the direction of each vector in the set is determined and the corresponding neighborhood edge angle cotangent value is extracted. The direction determination content and cotangent value are combined according to the wall node neighborhood structure to form the direction weights used to construct the Laplacian operator weights and then written into the weight sequence to obtain the neighborhood direction weight set. Based on the neighborhood directional weight set, directional weights are collected in the neighborhood range of each measurement point according to the measurement point number, and the collected content is written into a preset structural position. The records are stored in a fixed index manner to form a continuous distribution record, thereby generating a neighborhood structural weight map. The Laplace operator, based on the neighborhood structure weight map, fills the direction weights from each wall node to its neighboring wall nodes into the row and column positions of the coefficient matrix according to the wall node number. The wall node self-weights associated with the neighborhood direction weights are written in the diagonal position. A matrix corresponding to the triangular grid of the goaf is established. The matrix reads the coordinates of each wall node in the wall node coordinate list according to the number, writes the wall node coordinates into the coordinate column to be updated, and adjusts the coordinates according to the weight order of each row of the matrix to generate wall node coordinate increment records. The coordinate increments are merged into the original wall node coordinates according to the number of the wall node coordinate increment records and written into the new coordinate list to form the updated wall node coordinate list. The Laplace operator, according to the formula: ， in: For the first Incremental vectors of wall node coordinates For the first The three-dimensional coordinate vectors recorded in the coordinate list of wall nodes in the current iteration. For the first The original 3D coordinate vectors written by each wall node in the initial stage of constructing the wall node coordinate list. In order to be with the first The wall nodes have triangular mesh connections and are involved in the calculation in the current iteration. Three-dimensional coordinate vectors of neighborhood wall nodes The first node is read from the neighborhood structure weight map according to the wall node number and the neighborhood number, and matched with the direction weight. The wall node to the first The directional weights of each neighboring wall node. For the first A set of wall node indices in the neighborhood of each wall node. The first node read from the curvature boundary discrimination sequence according to the wall node number. Curvature coefficients of each wall node To pass the first triangle on the corresponding triangular mesh The coordinates of the first wall node and its adjacent wall nodes were calculated and normalized to obtain the first... Local wall unit normal vector of each wall node To read and indicate the first curvature boundary discrimination sequence The boundary marker coefficient indicating whether a wall node is located at the roadway connection boundary is determined when the first wall node is at the boundary boundary. When a wall node is determined to be a boundary wall node Values When the first When a wall node is determined to be a non-boundary wall node Values , These are the Laplace smoothing weights, used to control the numerical scale of the Laplace term during coordinate updates. These are curvature correction weighting coefficients used to control the curvature coefficient. With normal vector The strength of the participation of the combined term in coordinate updates Weighting coefficients are used to restore the boundary and control the boundary marking coefficients. Difference term from original coordinates The intensity of participation in coordinate updates; S4: Based on the updated list of wall node coordinates, extract the wall node coordinates and establish a distance metric with the reference coordinates of the goaf elevation difference section. Establish a weight function with the distance metric and Gaussian process regression and apply it to the wall node elevation. Collect the weighted results in the order of wall node number and output the interpolated wall node elevation segment. S5: Based on the interpolated wall node elevation segment, establish a distance quantity between the wall node coordinates and neighboring wall nodes, introduce the distance quantity into the attenuation factor to construct weights and write it into the wall node position along with the wall node elevation, and then assemble the wall node coordinate sequence according to the predetermined number to establish a multi-scale wall space block.

2. The visualization method for measurement data of mining goaf areas according to claim 1, characterized in that, The neighborhood structure weight map includes direction vector records, angle cotangent value records, and structure position records. The curvature boundary discrimination sequence includes geometric interval records, curvature index records, and tunnel connection point discrimination records. The wall node coordinate list includes the original coordinates of the wall nodes, coordinate adjustment amounts, and updated coordinates of the wall nodes. The interpolated wall node elevation fragment includes wall node distance metrics, weight function results, and weighted elevations. The multi-scale wall space block includes wall node coordinate sequences, weights constructed from distance metrics, and wall node elevation entries.

3. The visualization method for measurement data of mining goaf areas according to claim 1, characterized in that, The specific steps to obtain the curvature boundary discrimination sequence are as follows: Based on the neighborhood structure weight map, the spatial coordinates of the wall nodes and the neighboring wall nodes are extracted, and the orientation records are established by the direction of the line connecting the wall nodes to the neighboring wall nodes. After being arranged into a continuous structure column according to the wall node numbering order, the arrangement rules are uniformly defined to generate an interval sequence. Based on the interval sequence, the neighborhood polygon contour is read and the spatial morphological differences between wall nodes are compared with the wall node relationship order. The morphological differences are written into the orientation structure column according to the wall node number and organized into a continuous structure record in index order to generate a curvature index sequence. Based on the curvature index sequence, the information of the roadway connection points in the goaf is read and the spatial morphological characteristics of the wall nodes are compared in the order of the wall node numbers. The comparison content is written into the boundary record column and stored in a fixed structural order to form a boundary sequence, thereby generating a curvature boundary discrimination sequence.

4. The visualization method for measurement data of mining goaf areas according to claim 1, characterized in that, The specific steps for constructing the list of wall node coordinates are as follows: Based on the neighborhood structure weight map and the curvature boundary discrimination sequence, two types of weight records are extracted according to the wall nodes, and a weight corresponding path is established based on the wall node position. After being written into the calibration column in sequence through the wall node index, the calibration column is continuously sorted to form a continuous offset record, and a coordinate adjustment list is generated. Based on the coordinate adjustment list, offset records are extracted according to the wall node number and embedded in the position based on the original wall node position. The coordinates are written into the coordinate column in numerical order, and the coordinate column is uniformly organized to form a continuous storage sequence while maintaining the consistency of the index structure, thus generating a list of wall node coordinates.

5. The visualization method for measurement data of mining goaf areas according to claim 1, characterized in that, The specific steps for outputting the interpolated wall node elevation fragment are as follows: Based on the wall node coordinate list, the wall node positions are extracted by number and a corresponding path is formed by the wall node positions and the reference coordinates of the goaf elevation difference section. After recording the relative displacement information through the direction of the wall node, the information is written into the position column in numerical order and organized to form a structural record, generating the position connection sequence antecedent. Based on the location connection sequence antecedent, wall node association items are extracted by number and an input sequence is established using wall node position records. The wall nodes are written into the elevation association column according to their arrangement order and uniformly organized to form a wall node input structure column, thereby generating the wall node input sequence. Based on the wall node input sequence, the wall node position records are extracted by number and written into the predicted elevation column through Gaussian process regression. They are then merged into a continuous elevation structure column and the sequence index is sorted to maintain the consistency of the structural order, generating interpolated wall node elevation segments.

6. The visualization method for measurement data of mining goaf areas according to claim 1, characterized in that, The Gaussian process regression records the wall node positions in the model input column according to the numbering order of the wall node input sequence. It establishes the relationship structure between wall nodes through the wall node input column and writes the relationship structure into the model internal record. Then, it introduces the model calculation path according to the wall node number. It generates the predicted elevation record in the model calculation path and writes the predicted elevation record into the output column according to the number. Finally, it organizes the output column into a continuous structure sequence to form the predicted elevation structure column.

7. The visualization method for measurement data of mining goaf areas according to claim 1, characterized in that, The specific steps for establishing the multi-scale wall space block are as follows: Based on the interpolated wall node elevation fragment, the coordinates of the wall nodes are extracted by number and a connection path is established between the coordinates of the wall nodes and the positions of the neighboring wall nodes. The orientation between the wall nodes is recorded to form corresponding wall node items and written into the record column by number. The record column is sorted in order to form a continuous structure sequence, and a wall node position sequence is generated. Based on the wall node position sequence, the corresponding items of the wall nodes are extracted by number and written into the weight column by setting an attenuation coefficient. The weight column and the corresponding items of the wall nodes are merged to form a structural unit. The wall node coordinate sequence is assembled according to the predetermined number and the sequence is continuously integrated to form a spatial block, generating a multi-scale wall spatial block.

8. A visualization system for measurement data of mined-out areas, characterized in that, The method for visualizing measurement data of mine goaf areas according to any one of claims 1-7, wherein the system comprises: Neighborhood weight construction module: Based on the original coordinates of the measurement points in the goaf, establish the connection between the coordinates of the measurement points and the coordinates of the neighboring points, calculate the direction vector and combine it with the cotangent values ​​of the neighboring edge angles, write the direction weights into the neighboring records of the wall nodes according to the structural positions, and generate a neighborhood structural weight map. Curvature boundary calculation module: Based on the neighborhood structure weight map, obtain the geometric interval between the wall node and the neighboring wall node, convert the interval and the ratio of the neighborhood side length into a curvature index, and write it into the weight sequence according to the goaf roadway connection point criterion. Integrate the numbered arrangement results to obtain the curvature boundary discrimination sequence. Wall coordinate generation module: Based on the neighborhood structure weight map and the curvature boundary discrimination sequence, extract the two types of weights of the wall nodes and merge them to form a coordinate adjustment amount. Add the coordinate adjustment amount to the original wall node coordinates, write it into the record list according to the number, organize it into a continuous coordinate sequence, and construct the updated wall node coordinate list. Elevation interpolation calculation module: Based on the updated list of wall node coordinates, extract the wall node coordinates and establish a distance metric with the reference coordinates of the goaf elevation difference section. Write the distance metric into the weight function and apply it to the wall node elevation. Collect the weighted results in sequence according to the wall node number and output the interpolated wall node elevation segment. Multi-scale block construction module: Based on the interpolated wall node elevation fragment, extract the wall node coordinates and establish distance with neighboring wall nodes. Write the distance into the attenuation factor to generate weights, and write them into the wall node position record along with the wall node elevation. Assemble the wall node coordinate sequence according to the predetermined number to establish a multi-scale wall space block.

Images