A Method and System for Cave 3D Modeling Based on Multi-Sensor Fusion

Abstract

This invention relates to the field of cave 3D modeling technology, specifically a method and system for cave 3D modeling based on multi-sensor fusion. The method includes: identifying a set of simulated feature mesh points; calculating simulation accuracy based on the simulated feature mesh point set and the actual verification mesh point set; plotting points based on the unit test mesh area and simulation accuracy to obtain an area accuracy point set; performing regression analysis on the area accuracy point set to obtain an area accuracy regression line; constructing a triangular plane current mesh based on the current cave region to be measured; performing secondary sensing sampling on the inner wall of the current 3D unit mesh based on the predicted regression accuracy to obtain a secondary 3D sampling point set; using the secondary 3D sampling point set to perform 3D correction on the current 3D unit mesh to obtain an initial cave 3D model; and fusing the cave sensing image and the initial cave 3D model to obtain a target cave 3D model. This invention can improve laser sampling efficiency and cave 3D modeling accuracy.

Description

Technical Field

[0001] This invention relates to the field of cave 3D modeling technology, and in particular to a cave 3D modeling method and system based on multi-sensor fusion. Background Technology

[0002] In fields such as cave exploration, underground space mapping, and geological disaster monitoring, 3D modeling technology is a key means to achieve environmental digitization and intelligent analysis.

[0003] Traditional mesh reconstruction algorithms perform 3D sampling of caves using a fixed laser sampling frequency. This sampling method is difficult to adapt to the multi-scale structure of caves and does not fully consider the influence of cave wall features on point cloud sampling, resulting in excessive redundancy in flat areas and insufficient resolution in steep areas. Therefore, current methods for 3D sampling and modeling caves suffer from low laser sampling efficiency and poor accuracy in 3D cave modeling. Summary of the Invention

[0004] This invention provides a method and system for cave 3D modeling based on multi-sensor fusion, the main purpose of which is to improve laser sampling efficiency and cave 3D modeling accuracy.

[0005] To achieve the above objectives, the present invention provides a method for 3D cave modeling based on multi-sensor fusion, comprising:

[0006] Based on the preset triangular plane test grid, the cave interior wall is sampled using a pre-constructed inertial measurement sensor to obtain the cave's three-dimensional test grid.

[0007] Planar unit test grids are extracted sequentially from the triangular plane test grid, and the corresponding three-dimensional unit test grids are identified in the cave three-dimensional test grid.

[0008] The planar verification point set is uniformly set within the planar unit test grid according to the preset number of accuracy verification points;

[0009] Identify the set of simulated feature mesh points corresponding to the set of planar verification points within the three-dimensional unit test mesh;

[0010] Based on the aforementioned planar verification point set, the cave interior wall is sensor-sampled to obtain the actual verification mesh point set;

[0011] The simulation accuracy is calculated based on the simulated feature mesh point set and the actual verification mesh point set.

[0012] Calculate the area of ​​the three-dimensional unit test grid, and plot points based on the unit test grid area and simulation accuracy to obtain the area accuracy point set;

[0013] Regression analysis was performed on the area accuracy point set to obtain the area accuracy regression line;

[0014] Receive the current cave region to be measured, and construct a triangular plane mesh based on the current cave region to be measured;

[0015] Based on the current mesh of the triangular plane, the cave interior wall is sampled to obtain the current mesh of the three-dimensional unit;

[0016] Identify the current grid area of ​​the current mesh of the three-dimensional unit, and extract the predicted regression accuracy from the area accuracy regression line based on the current grid area of ​​the unit;

[0017] Based on the predicted regression accuracy, perform secondary sensing sampling of the inner wall of the current mesh of the three-dimensional unit to obtain a secondary three-dimensional sampling point set.

[0018] The current mesh of the three-dimensional unit is corrected using the secondary three-dimensional sampling point set to obtain the initial three-dimensional model of the cave.

[0019] Cave sensing images are acquired using a pre-built sensing camera, and the cave sensing images and the initial cave 3D model are fused to obtain the target cave 3D model.

[0020] Optionally, the step of obtaining a three-dimensional test grid of the cave by using a pre-constructed inertial measurement sensor to perform sensing sampling of the cave's inner wall based on a preset triangular plane test grid includes:

[0021] Identify the triangular plane grid points in the triangular plane test grid, identify the surface normal vectors of the triangular plane grid points to the cave wall, and obtain the surface normal vector set;

[0022] Extract the surface normal vectors sequentially from the surface normal vector set;

[0023] Laser sampling of the cave wall is performed on the surface normal vector using a lidar, and the attitude of the lidar is calibrated using the inertial measurement sensor to obtain a set of three-dimensional sampling points for testing.

[0024] Identify the set of planar element mesh points in the planar element test mesh, and identify the set of element normal vectors corresponding to the set of planar element mesh points in the set of surface normal vectors;

[0025] Identify the test unit sampling point group corresponding to the unit normal vector group in the test 3D sampling point set, wherein the number of test unit sampling points in the test unit sampling point group is 3;

[0026] By connecting the sampling points of the test unit in pairs, a three-dimensional test grid for the cave is obtained.

[0027] Optionally, the step of uniformly setting a set of planar verification points within the planar unit test grid according to a preset number of accuracy verification points includes:

[0028] The number of unit grid test split blocks of the planar unit test grid is determined based on the number of precision verification points, wherein the number of precision verification points is: n is a preset positive integer;

[0029] The planar unit test grid is equally divided according to the number of test split blocks of the unit grid to obtain a test unit grid sub-block set, wherein the test unit grid sub-blocks in the test unit grid sub-block set are equilateral triangles;

[0030] Identify the center point of each test sub-block in the test unit grid sub-block set, and use the center point of the test sub-block as a plane verification point to obtain a plane verification point set.

[0031] Optionally, identifying the set of simulated feature mesh points corresponding to the set of planar verification points within the three-dimensional unit test mesh includes:

[0032] Identify the verification normal vector corresponding to each plane verification point in the plane verification point set to obtain a verification normal vector set, wherein the verification normal vector refers to the surface normal vector of the plane verification point to the inner wall of the cave.

[0033] Identify the line-plane intersection points of each verification normal vector in the verification normal vector set with the three-dimensional unit test mesh to obtain the line-plane intersection point set;

[0034] The set of intersection points of the line and the surface is used as the set of points of the simulated feature network.

[0035] Optionally, the step of calculating the simulation accuracy based on the simulated feature mesh point set and the actual verification mesh point set includes:

[0036] Identify the associated simulated mesh points and associated actual mesh points corresponding to the planar verification points in the simulated feature mesh point set and the actual verification mesh point set;

[0037] Identify the heights of the simulated mesh points and the actual mesh points along the verification normal vector direction to obtain the set of simulated mesh point heights and the set of actual mesh point heights.

[0038] Based on the simulated mesh point height set and the actual mesh point height set, the simulation accuracy is calculated using the following formula:

[0039]

[0040] in, This represents the simulation accuracy of the i-th planar element test mesh. Indicates the precision adjustment index. This represents the number of planar verification points in the i-th planar unit test grid. This represents the actual height of the mesh point corresponding to the j-th planar verification point in the i-th planar unit test mesh. This represents the height of the simulated mesh point corresponding to the j-th planar verification point in the i-th planar unit test mesh. Represents the absolute value symbol.

[0041] Optionally, the step of performing regression analysis on the area precision point set to obtain the area precision regression line includes:

[0042] Identify the area precision coordinates of each area precision point in the area precision point set to obtain the area precision coordinate set;

[0043] Based on the area precision coordinate set, the slope of the regression line is calculated using the following formula:

[0044]

[0045] Where k represents the slope of the regression line. This indicates the number of test meshes for planar elements. This represents the x-coordinate of the p-th area precision coordinate in the area precision coordinate set. Represents the ordinate of the p-th area precision coordinate in the area precision coordinate set;

[0046] The intercept of the regression line is calculated using the following formula based on the area accuracy coordinate system:

[0047]

[0048] in, Indicates the intercept of the regression line;

[0049] The area precision regression line is determined based on the slope and intercept of the regression line, wherein the equation of the area precision regression line is as follows:

[0050]

[0051] in, This represents the value of the regression accuracy function. This represents the value of the grid area function.

[0052] Optionally, the step of performing secondary sensing sampling of the inner wall of the current mesh of the three-dimensional unit based on the predicted regression accuracy to obtain a secondary three-dimensional sampling point set includes:

[0053] Identify the planar element current mesh corresponding to the current three-dimensional element current mesh in the current triangular plane mesh;

[0054] Based on the predicted regression accuracy, the accuracy filler points are calculated using the following formula:

[0055]

[0056] in, The precision is indicated by the number of fill points, and e represents the natural constant. Indicates the accuracy of the predicted regression. Indicates the floor function;

[0057] The number of precision compensation points is used as the current number of cell grid split blocks of the current grid of the planar element;

[0058] The planar unit current grid is equally divided according to the current number of split blocks of the unit grid to obtain the current unit grid sub-block set, wherein the current unit grid sub-blocks in the current unit grid sub-block set are equilateral triangles;

[0059] Identify the center point of the current sub-block of each current unit grid sub-block in the current unit grid sub-block set, and use the center point of the current sub-block as the plane completion point to obtain the plane completion point set;

[0060] Based on the aforementioned planar supplementary point set, the cave interior wall is sensor-sampled to obtain a secondary three-dimensional sampling point set.

[0061] Optionally, the step of using the secondary three-dimensional sampling point set to perform three-dimensional correction on the current mesh of the three-dimensional unit to obtain the initial three-dimensional cave model includes:

[0062] In the current unit grid sub-block set, identify the group of unit grid sub-blocks sharing the same edge, and identify the center group of the sub-blocks sharing the same edge in the group of unit grid sub-blocks. The group of unit grid sub-blocks sharing the same edge refers to two current unit grid sub-blocks that share the same edge.

[0063] Identify the associated secondary three-dimensional sampling point group corresponding to the center group of the shared-edge sub-block in the secondary three-dimensional sampling point set;

[0064] Connecting the associated secondary three-dimensional sampling points in the associated secondary three-dimensional sampling point group yields a three-dimensional corrected feature line;

[0065] The three-dimensional correction feature lines are used to perform three-dimensional correction on the current mesh of the three-dimensional unit to obtain the initial three-dimensional cave model.

[0066] Optionally, fusing the cave sensing image and the initial cave 3D model to obtain the target cave 3D model includes:

[0067] The texture and color of each triangular plane in the initial 3D cave model are identified using the cave sensing images.

[0068] The initial cave 3D model is supplemented with texture color based on the texture color of the triangular plane to obtain the target cave 3D model.

[0069] To achieve the above objectives, the present invention also provides a cave 3D modeling system based on multi-sensor fusion, comprising:

[0070] The area accuracy regression line identification module is used to obtain a three-dimensional cave test grid by using a pre-constructed inertial measurement sensor to perform sensing sampling on the cave interior wall based on a preset triangular plane test grid; sequentially extract planar unit test grids from the triangular plane test grid, and identify the corresponding three-dimensional unit test grids in the cave three-dimensional test grid; uniformly set a set of planar verification points within the planar unit test grid according to a preset number of accuracy verification points; identify the set of simulated feature mesh points corresponding to the set of planar verification points within the three-dimensional unit test grid; perform sensing sampling on the cave interior wall based on the set of planar verification points to obtain an actual set of verification mesh points; calculate the simulation accuracy based on the simulated feature mesh points and the actual verification mesh points; calculate the area of ​​the unit test grid of the three-dimensional unit test grid; plot points based on the unit test grid area and the simulation accuracy to obtain an area accuracy point set; and perform regression analysis on the area accuracy point set to obtain an area accuracy regression line.

[0071] The 3D unit current mesh acquisition module is used to receive the current cave area to be measured, construct a triangular plane current mesh based on the current cave area to be measured, and perform cave inner wall sensing sampling based on the triangular plane current mesh to obtain the 3D unit current mesh;

[0072] The three-dimensional unit current mesh correction module is used to identify the current mesh area of ​​the three-dimensional unit current mesh, extract the predicted regression accuracy from the area accuracy regression line based on the current mesh area, perform secondary sensing sampling of the inner wall of the three-dimensional unit current mesh based on the predicted regression accuracy to obtain a secondary three-dimensional sampling point set, and use the secondary three-dimensional sampling point set to perform three-dimensional correction of the three-dimensional unit current mesh to obtain an initial cave three-dimensional model.

[0073] The target cave 3D model construction module is used to acquire cave sensing images using a pre-built sensing camera, and to fuse the cave sensing images and the initial cave 3D model to obtain the target cave 3D model.

[0074] To address the above problems, the present invention also provides an electronic device, the electronic device comprising:

[0075] A memory that stores at least one instruction; and a processor that executes the instructions stored in the memory to implement the multi-sensor fusion-based cave 3D modeling method described above.

[0076] To address the aforementioned problems, the present invention also provides a computer-readable storage medium storing at least one instruction, which is executed by a processor in an electronic device to implement the aforementioned multi-sensor fusion-based cave 3D modeling method.

[0077] Beneficial Effects: To address the problems described in the background art, this invention first analyzes the relationship between the unit test grid area and simulation accuracy, obtaining an area accuracy regression line. Then, it identifies the predicted regression accuracy corresponding to the current grid area of ​​different units using this regression line. Finally, based on the predicted regression accuracy, it performs secondary sensing sampling of the inner wall, thereby achieving the effect of three-dimensional correction of the current grid of the three-dimensional unit using the secondary three-dimensional sampling point set obtained from the secondary sensing sampling. When obtaining the area accuracy regression line, it is necessary to first perform cave inner wall sensing sampling using a pre-constructed inertial measurement sensor based on the triangular plane test grid to obtain the cave's three-dimensional test grid. To achieve targeted analysis of the planar unit test grid, it is necessary to perform triangular plane sensing sampling... Planar unit test grids are extracted sequentially from the test grid. Since there is a correspondence between the 3D unit test grid and the planar unit test grid, it is necessary to identify the corresponding 3D unit test grid in the 3D cave test grid. To calculate the simulation accuracy, a set of planar verification points needs to be uniformly set within the planar unit test grid according to a preset number of accuracy verification points. For comparison, the set of simulated feature mesh points corresponding to the set of planar verification points needs to be identified within the 3D unit test grid. Then, cave wall sensing sampling is performed based on the set of planar verification points to obtain the actual verification mesh point set. At this point, the simulation accuracy can be calculated based on the simulated feature mesh point set and the actual verification mesh point set. To analyze the correspondence between the unit test grid area and the simulation accuracy, it is necessary to calculate the unit test grid area of ​​the three-dimensional unit test grid. Then, points are plotted based on the unit test grid area and the simulation accuracy to obtain the area accuracy point set. Finally, regression analysis is performed on the area accuracy point set to obtain the area accuracy regression line. After obtaining the area accuracy regression line, the current cave area to be tested can be received, and a triangular plane current grid can be constructed based on the current cave area to be tested. Then, the cave inner wall is sampled based on the triangular plane current grid to obtain the three-dimensional unit current grid. At this point, three-dimensional correction can be performed on the three-dimensional unit current grid. First, it is necessary to identify the units of the three-dimensional unit current grid. The current grid area is used as the basis for predictive regression accuracy. Since a higher predictive regression accuracy requires fewer secondary 3D sampling points, secondary sensing sampling of the inner wall of the current 3D unit grid can be performed based on the predicted regression accuracy to obtain a secondary 3D sampling point set. Finally, the secondary 3D sampling point set is used to perform 3D correction on the current 3D unit grid to obtain an initial cave 3D model. Since the initial cave 3D model does not contain texture or color information, a pre-built sensing camera can be used to acquire cave sensing images, and the cave sensing images and the initial cave 3D model can be fused to obtain the target cave 3D model. Therefore, this invention can improve laser sampling efficiency and cave 3D modeling accuracy. Attached Figure Description

[0078] Figure 1 This is a flowchart illustrating a method for three-dimensional cave modeling based on multi-sensor fusion, provided in an embodiment of the present invention.

[0079] Figure 2 A functional block diagram of a cave 3D modeling system based on multi-sensor fusion provided in an embodiment of the present invention;

[0080] Figure 3 This is a schematic diagram of the structure of an electronic device that implements the multi-sensor fusion-based cave 3D modeling method according to an embodiment of the present invention.

[0081] Explanation of reference numerals in the attached figures:

[0082] 10. Electronic device; 11. Processor; 12. Memory; 13. Bus.

[0083] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0084] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0085] This application provides a method for 3D cave modeling based on multi-sensor fusion. The executing entity of this method includes, but is not limited to, at least one electronic device that can be configured to execute the method provided in this application, such as a server or a terminal. In other words, the method can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster.

[0086] Reference Figure 1 The diagram shown is a flowchart illustrating a cave 3D modeling method based on multi-sensor fusion according to an embodiment of the present invention. In this embodiment, the cave 3D modeling method based on multi-sensor fusion includes:

[0087] S1. Based on the preset triangular plane test grid, use a pre-constructed inertial measurement sensor to perform sensing sampling on the cave wall to obtain the cave three-dimensional test grid.

[0088] Understandably, the triangular plane test mesh refers to a triangular mesh located on a preset plane, used to test the area accuracy regression line. The triangular plane test mesh consists of multiple equilateral triangular units. The side length of the equilateral triangular unit can be set according to the user's accuracy requirements for 3D modeling, for example, it can be 2cm. The area accuracy regression line refers to the regression line representing the relationship between the unit mesh area and the simulation accuracy. See the following embodiment for details. The plane refers to a reference horizontal plane passing through a preset location of the cave's interior surface.

[0089] Furthermore, the inertial measurement unit (IMU) refers to an integrated sensor that measures the acceleration and angular velocity of a carrier in an inertial reference frame, thereby calculating the carrier's motion state. The carrier can be a lidar (LiDAR). The cave interior wall sensing sampling refers to the sampling process of determining the spatial positions of the cave interior wall using the inertial measurement unit and lidar based on the surface normal vectors of points within the triangular plane test grid, as detailed in the following embodiment. The surface normal vector refers to the normal vector perpendicular to the plane and pointing upwards, with the triangular plane grid points as the vector origin.

[0090] In detail, the cave three-dimensional test grid refers to a triangular grid surface formed by connecting spatial locations on the cave wall obtained from sensing and sampling of the cave wall.

[0091] In this embodiment of the invention, the step of obtaining a three-dimensional test grid of the cave by using a pre-constructed inertial measurement sensor to perform sensing sampling of the cave's inner wall based on a preset triangular plane test grid includes:

[0092] Identify the triangular plane grid points in the triangular plane test grid, identify the surface normal vectors of the triangular plane grid points to the cave wall, and obtain the surface normal vector set;

[0093] Extract the surface normal vectors sequentially from the surface normal vector set;

[0094] Laser sampling of the cave wall is performed on the surface normal vector using a lidar, and the attitude of the lidar is calibrated using the inertial measurement sensor to obtain a set of three-dimensional sampling points for testing.

[0095] Identify the set of planar element mesh points in the planar element test mesh, and identify the set of element normal vectors corresponding to the set of planar element mesh points in the set of surface normal vectors;

[0096] Identify the test unit sampling point group corresponding to the unit normal vector group in the test 3D sampling point set, wherein the number of test unit sampling points in the test unit sampling point group is 3;

[0097] By connecting the sampling points of the test unit in pairs, a three-dimensional test grid for the cave is obtained.

[0098] Explained, the triangular plane grid points refer to the grid points of the planar test grid. The surface normal vector set refers to the set of surface normal vectors corresponding to each triangular plane grid point. The cave interior wall laser sampling refers to the process of using a lidar to determine the spatial position of the intersection point of the surface normal vector and the cave interior wall. Since it is necessary to obtain spatial information such as the attitude, velocity, and position of the lidar, it is necessary to use an inertial measurement sensor to calibrate the attitude of the lidar. The test three-dimensional sampling point set refers to the set of three-dimensional spatial positions of the cave interior wall determined based on the normal vectors of the triangular plane grid points during the regression relationship between the test grid area and the simulation accuracy of the test unit.

[0099] Further, the planar unit test mesh refers to the unit mesh within the triangular planar test mesh. Since the triangular planar test mesh is composed of multiple equilateral triangles, the planar unit test mesh is the smallest equilateral triangle within the triangular planar test mesh. The planar unit mesh point set refers to the set of three mesh points of the planar unit test mesh. The unit normal vector set refers to the three surface normal vectors originating from the planar unit mesh points in the planar unit mesh point set. The test unit sampling point set refers to the three three-dimensional spatial intersections between the unit normal vectors in the unit normal vector set and the cave wall.

[0100] S2. Extract planar unit test grids sequentially from the triangular plane test grid, and identify the corresponding three-dimensional unit test grids from the cave three-dimensional test grid.

[0101] Furthermore, the three-dimensional unit test grid refers to a spatial three-dimensional grid obtained by connecting the test unit sampling point groups corresponding to the planar unit grid point set in the planar unit test grid.

[0102] S3. Set up a set of planar verification points evenly within the planar unit test grid according to the preset number of accuracy verification points.

[0103] Understandably, the number of accuracy verification points refers to the preset number of sites used to verify the simulation accuracy of the three-dimensional element test mesh. The planar verification point set refers to the set of sites used to verify the simulation accuracy of the three-dimensional element test mesh.

[0104] In this embodiment of the invention, the step of uniformly setting a set of planar verification points within the planar unit test grid according to a preset number of precision verification points includes:

[0105] The number of unit grid test split blocks of the planar unit test grid is determined based on the number of precision verification points, wherein the number of precision verification points is: n is a preset positive integer;

[0106] The planar unit test grid is equally divided according to the number of test split blocks of the unit grid to obtain a test unit grid sub-block set, wherein the test unit grid sub-blocks in the test unit grid sub-block set are equilateral triangles;

[0107] Identify the center point of each test sub-block in the test unit grid sub-block set, and use the center point of the test sub-block as a plane verification point to obtain a plane verification point set.

[0108] Understandably, the number of test cell blocks refers to the number of splits in a planar unit test grid. The number of test cell blocks is equal to the number of precision verification points. Since the planar unit test grid, which is equilaterally triangular, is split into multiple smaller, equal equilateral triangles, the number of test cell blocks should be a power function to the base 4. The test cell grid sub-set refers to the set of grid sub-blocks obtained after the planar unit test grid is equally split.

[0109] S4. Identify the set of simulated feature mesh points corresponding to the set of planar verification points within the three-dimensional unit test mesh.

[0110] It should be understood that the simulated feature mesh point set refers to the set of intersection points of the surface normal vector passing through the plane verification point and the three-dimensional unit test mesh.

[0111] In this embodiment of the invention, identifying the set of simulated feature mesh points corresponding to the set of planar verification points within the three-dimensional unit test mesh includes:

[0112] Identify the verification normal vector corresponding to each plane verification point in the plane verification point set to obtain a verification normal vector set, wherein the verification normal vector refers to the surface normal vector of the plane verification point to the inner wall of the cave.

[0113] Identify the line-plane intersection points of each verification normal vector in the verification normal vector set with the three-dimensional unit test mesh to obtain the line-plane intersection point set;

[0114] The set of intersection points of the line and the surface is used as the set of points of the simulated feature network.

[0115] Understandably, the verification normal vector refers to the surface normal vector passing through the plane verification point. The verification normal vector set refers to the set of verification normal vectors corresponding to each plane verification point. The line-plane intersection point refers to the intersection point of the verification normal vector with the plane containing the 3D element test mesh. The line-plane intersection point set refers to the set of line-plane intersection points of each verification normal vector with the 3D element test mesh.

[0116] S5. Based on the planar verification point set, perform sensing sampling on the cave wall to obtain the actual verification network point set.

[0117] Understandably, the actual set of verification network points refers to the set of intersection points of the surface normal vector passing through the plane verification points and the inner wall of the cave.

[0118] S6. Calculate the simulation accuracy based on the simulated feature network point set and the actual verification network point set.

[0119] In detail, the simulation accuracy refers to the degree of accuracy of the simulation of the three-dimensional unit test mesh.

[0120] In this embodiment of the invention, calculating the simulation accuracy based on the simulated feature mesh point set and the actual verification mesh point set includes:

[0121] Identify the associated simulated mesh points and associated actual mesh points corresponding to the planar verification points in the simulated feature mesh point set and the actual verification mesh point set;

[0122] Identify the heights of the simulated mesh points and the actual mesh points along the verification normal vector direction to obtain the set of simulated mesh point heights and the set of actual mesh point heights.

[0123] Based on the simulated mesh point height set and the actual mesh point height set, the simulation accuracy is calculated using the following formula:

[0124]

[0125] in, This represents the simulation accuracy of the i-th planar element test mesh. Indicates the precision adjustment index. This represents the number of planar verification points in the i-th planar unit test grid. This represents the actual height of the mesh point corresponding to the j-th planar verification point in the i-th planar unit test mesh. This represents the height of the simulated mesh point corresponding to the j-th planar verification point in the i-th planar unit test mesh. Represents the absolute value symbol.

[0126] In detail, the associated simulated mesh point refers to a simulated feature mesh point that lies on the same plane normal vector as the plane verification point, and the associated actual mesh point refers to an actual verification mesh point that lies on the same plane normal vector as the plane verification point. The simulated mesh point height refers to the height distance of the associated simulated mesh point from the plane in the direction of the verification normal vector. The actual mesh point height refers to the height distance of the associated actual mesh point from the plane in the direction of the verification normal vector. The simulated mesh point height set refers to the set of simulated mesh point heights corresponding to each associated simulated mesh point, and the actual mesh point height set refers to the set of actual mesh point heights corresponding to each associated actual mesh point. The precision adjustment index refers to an index that adjusts the value of the simulation precision; the larger the precision adjustment index, the larger the value of the simulation precision, and it can be set according to actual conditions.

[0127] S7. Calculate the area of ​​the three-dimensional unit test grid, and plot the points according to the area of ​​the unit test grid and the simulation accuracy to obtain the area accuracy point set.

[0128] Explained, the unit test grid area refers to the grid area of ​​the three-dimensional unit test grid, and the area accuracy point set refers to the set of coordinate points with the unit test grid area as the independent variable and the simulation accuracy as the dependent variable.

[0129] S8. Perform regression analysis on the area accuracy point set to obtain the area accuracy regression line.

[0130] Furthermore, the area accuracy regression line refers to the regression line that describes the relationship between the area of ​​the unit test grid and the simulation accuracy.

[0131] In this embodiment of the invention, the step of performing regression analysis on the area accuracy point set to obtain the area accuracy regression line includes:

[0132] Identify the area precision coordinates of each area precision point in the area precision point set to obtain the area precision coordinate set;

[0133] Based on the area precision coordinate set, the slope of the regression line is calculated using the following formula:

[0134]

[0135] Where k represents the slope of the regression line. This indicates the number of test meshes for planar elements. This represents the x-coordinate of the p-th area precision coordinate in the area precision coordinate set. Represents the ordinate of the p-th area precision coordinate in the area precision coordinate set;

[0136] The intercept of the regression line is calculated using the following formula based on the area accuracy coordinate system:

[0137]

[0138] in, Indicates the intercept of the regression line;

[0139] The area precision regression line is determined based on the slope and intercept of the regression line, wherein the equation of the area precision regression line is as follows:

[0140]

[0141] in, This represents the value of the regression accuracy function. This represents the value of the grid area function.

[0142] Understandably, the least squares method can be used to perform regression analysis on the area accuracy point set to obtain the area accuracy regression line. The least squares method is an existing technology and will not be described in detail here.

[0143] Specifically, the area accuracy coordinates refer to the coordinates of the area accuracy points determined in a pre-constructed grid area-simulation accuracy coordinate system, where the unit test grid area is the abscissa and the simulation accuracy is the ordinate. The area accuracy coordinate set refers to the collection of area accuracy coordinates for each area accuracy point. The regression line slope refers to the slope of the area accuracy regression line. The regression line intercept refers to the intercept of the area accuracy regression line.

[0144] S9. Receive the current cave region to be tested, and construct the current triangular plane grid based on the current cave region to be tested.

[0145] Explained, the "current cave region to be tested" refers to the cave region that currently needs to be 3D modeled, and the current cave region to be tested is located in a plane. The "current triangular plane mesh" refers to an equilateral triangular mesh located in the plane, used for 3D modeling of the cave interior wall corresponding to the current cave region to be tested, and the triangular plane test mesh is composed of multiple equilateral triangular units.

[0146] S10. Based on the current mesh of the triangular plane, perform sensing sampling of the cave wall to obtain the current mesh of the three-dimensional unit.

[0147] Furthermore, the current three-dimensional unit grid refers to the three-dimensional unit spatial grid obtained by sensing and sampling the cave interior wall based on the grid points of a single equilateral triangle unit in the current grid of the triangular plane.

[0148] S11. Identify the current grid area of ​​the current grid of the three-dimensional unit, and extract the predicted regression accuracy from the area accuracy regression line based on the current grid area of ​​the unit.

[0149] It should be understood that the current grid area of ​​the cell refers to the area of ​​the triangle in the current grid of the three-dimensional cell. The prediction regression accuracy refers to the simulation accuracy of the current grid of the three-dimensional cell predicted based on the area accuracy regression line.

[0150] S12. Based on the predicted regression accuracy, perform secondary sensing sampling of the inner wall of the current mesh of the three-dimensional unit to obtain a secondary three-dimensional sampling point set.

[0151] Explained, the secondary three-dimensional sampling point set refers to the set of sampling points used to perform three-dimensional correction on the current mesh of the three-dimensional unit. The secondary sensing sampling of the inner wall refers to the sampling process of using inertial measurement sensors and lidar to determine the spatial position of the inner wall of the current mesh of the three-dimensional unit.

[0152] In this embodiment of the invention, the step of performing secondary sensing sampling of the inner wall of the current mesh of the three-dimensional unit based on the predicted regression accuracy to obtain a secondary three-dimensional sampling point set includes:

[0153] Identify the planar element current mesh corresponding to the current three-dimensional element current mesh in the current triangular plane mesh;

[0154] Based on the predicted regression accuracy, the accuracy filler points are calculated using the following formula:

[0155]

[0156] in, The precision is indicated by the number of fill points, and e represents the natural constant. Indicates the accuracy of the predicted regression. Indicates the floor function;

[0157] The number of precision compensation points is used as the current number of cell grid split blocks of the current grid of the planar element;

[0158] The planar unit current grid is equally divided according to the current number of split blocks of the unit grid to obtain the current unit grid sub-block set, wherein the current unit grid sub-blocks in the current unit grid sub-block set are equilateral triangles;

[0159] Identify the center point of the current sub-block of each current unit grid sub-block in the current unit grid sub-block set, and use the center point of the current sub-block as the plane completion point to obtain the plane completion point set;

[0160] Based on the aforementioned planar supplementary point set, the cave interior wall is sensor-sampled to obtain a secondary three-dimensional sampling point set.

[0161] Explained, the planar unit current mesh refers to the equilateral triangular unit mesh that is perpendicularly projected onto the plane by the 3D unit current mesh. The number of precision compensation points refers to the number of mesh points used for 3D correction of the 3D unit current mesh. The number of current unit mesh split blocks refers to the number of blocks that equally split the planar unit current mesh. The current unit mesh sub-block set refers to the set of sub-blocks into equilateral triangular units of equal area that are equally split from the planar unit current mesh. The current sub-block center point refers to the center point of the current unit mesh sub-block, i.e., the center point of the equilateral triangular unit. The planar compensation point refers to the location within the planar unit current mesh used for secondary sensing sampling of the inner wall. The planar compensation point set refers to the set of planar compensation points corresponding to each current unit mesh sub-block.

[0162] S13. Use the secondary three-dimensional sampling point set to perform three-dimensional correction on the current mesh of the three-dimensional unit to obtain the initial three-dimensional cave model.

[0163] Understandably, the initial cave 3D model refers to the cave 3D model obtained by using secondary sensing sampling of the inner wall to perform 3D correction on the current mesh of the 3D unit.

[0164] In this embodiment of the invention, the step of using the secondary three-dimensional sampling point set to perform three-dimensional correction on the current mesh of the three-dimensional unit to obtain an initial three-dimensional cave model includes:

[0165] In the current unit grid sub-block set, identify the group of unit grid sub-blocks sharing the same edge, and identify the center group of the sub-blocks sharing the same edge in the group of unit grid sub-blocks. The group of unit grid sub-blocks sharing the same edge refers to two current unit grid sub-blocks that share the same edge.

[0166] Identify the associated secondary three-dimensional sampling point group corresponding to the center group of the shared-edge sub-block in the secondary three-dimensional sampling point set;

[0167] Connecting the associated secondary three-dimensional sampling points in the associated secondary three-dimensional sampling point group yields a three-dimensional corrected feature line;

[0168] The three-dimensional correction feature lines are used to perform three-dimensional correction on the current mesh of the three-dimensional unit to obtain the initial three-dimensional cave model.

[0169] Understandably, the shared-edge sub-block center group refers to the center points of two shared-edge unit grid sub-blocks in the shared-edge unit grid sub-block group, and the shared-edge unit grid sub-block is an equilateral triangle. The associated secondary three-dimensional sampling point group refers to the two secondary three-dimensional sampling points corresponding to the shared-edge sub-block center group in the direction of the surface normal vector. The three-dimensional correction feature line refers to the three-dimensional spatial line segment formed by connecting the associated secondary three-dimensional sampling point groups.

[0170] S14. Acquire cave sensing images using a pre-built sensing camera, and fuse the cave sensing images and the initial cave 3D model to obtain the target cave 3D model.

[0171] Understandably, the cave sensing image refers to an image of the cave wall in the current cave area to be tested, captured by a sensing camera. The target cave 3D model refers to the cave 3D model obtained by adding texture and color to the initial cave 3D model using the cave sensing image.

[0172] In this embodiment of the invention, the step of fusing the cave sensing image and the initial cave 3D model to obtain the target cave 3D model includes:

[0173] The texture and color of each triangular plane in the initial 3D cave model are identified using the cave sensing images.

[0174] The initial cave 3D model is supplemented with texture color based on the texture color of the triangular plane to obtain the target cave 3D model.

[0175] Understandably, the triangular plane refers to the smallest spatial triangular unit in the initial cave 3D model.

[0176] To address the problems described in the background art, this invention first analyzes the relationship between the unit test grid area and simulation accuracy, obtaining an area accuracy regression line. Then, it identifies the predicted regression accuracy corresponding to the current grid area of ​​different units using this regression line. Finally, based on the predicted regression accuracy, it performs secondary sensing sampling of the inner wall, thereby achieving the effect of three-dimensional correction of the current grid of the three-dimensional unit using the secondary three-dimensional sampling point set obtained from the secondary sensing sampling. When obtaining the area accuracy regression line, it is necessary to first perform cave inner wall sensing sampling using a pre-constructed inertial measurement sensor based on the triangular plane test grid to obtain the cave's three-dimensional test grid. To achieve targeted analysis of the planar unit test grid, it is necessary to further analyze the triangular plane test grid... The planar unit test mesh is extracted sequentially. Since there is a correspondence between the 3D unit test mesh and the planar unit test mesh, it is necessary to identify the corresponding 3D unit test mesh in the 3D cave test mesh. To calculate the simulation accuracy, a set of planar verification points is uniformly set within the planar unit test mesh according to a preset number of accuracy verification points. For comparison, the set of simulated feature mesh points corresponding to the planar verification point set is identified within the 3D unit test mesh. Then, cave wall sensing sampling is performed based on the planar verification point set to obtain the actual verification mesh point set. At this point, the simulation accuracy can be calculated based on the simulated feature mesh point set and the actual verification mesh point set. To analyze the correspondence between the unit test grid area and the simulation accuracy, it is necessary to calculate the unit test grid area of ​​the three-dimensional unit test grid, then plot points based on the unit test grid area and the simulation accuracy to obtain the area accuracy point set. Finally, regression analysis is performed on the area accuracy point set to obtain the area accuracy regression line. After obtaining the area accuracy regression line, the current cave area to be tested can be received, and a triangular plane current grid can be constructed based on the current cave area to be tested. Then, the cave inner wall is sampled based on the triangular plane current grid to obtain the three-dimensional unit current grid. At this point, three-dimensional correction can be performed on the three-dimensional unit current grid. First, it is necessary to identify the unit current of the three-dimensional unit current grid. The area of ​​the current grid cell is calculated, and then the predicted regression accuracy is extracted from the area accuracy regression line based on the current grid area of ​​the cell. Since the higher the predicted regression accuracy, the fewer secondary 3D sampling points are needed, the inner wall of the current 3D cell can be sampled secondaryly based on the predicted regression accuracy to obtain a secondary 3D sampling point set. Finally, the current 3D cell is 3D corrected using the secondary 3D sampling point set to obtain an initial cave 3D model. Since the initial cave 3D model does not contain texture and color information, a pre-built sensing camera can be used to acquire cave sensing images, and the cave sensing images and the initial cave 3D model can be fused to obtain the target cave 3D model. Therefore, this invention can improve laser sampling efficiency and cave 3D modeling accuracy.

[0177] like Figure 2 The diagram shown is a functional block diagram of a cave 3D modeling system based on multi-sensor fusion provided in an embodiment of the present invention.

[0178] The cave 3D modeling system 100 based on multi-sensor fusion described in this invention can be installed in an electronic device. Depending on the functions implemented, the cave 3D modeling system 100 based on multi-sensor fusion may include an area accuracy regression line recognition module 101, a 3D element current mesh acquisition module 102, a 3D element current mesh correction module 103, and a target cave 3D model construction module 104. The module described in this invention can also be called a unit, which refers to a series of computer program segments that can be executed by the processor of an electronic device and can perform a fixed function, and is stored in the memory of the electronic device.

[0179] The area accuracy regression line identification module 101 is used to obtain a three-dimensional cave test grid by using a pre-constructed inertial measurement sensor to perform cave wall sensing sampling based on a preset triangular plane test grid; sequentially extract planar unit test grids from the triangular plane test grid, and identify the three-dimensional unit test grids corresponding to the planar unit test grids in the three-dimensional cave test grid; uniformly set a set of planar verification points within the planar unit test grid according to a preset number of accuracy verification points; identify the set of simulated feature mesh points corresponding to the set of planar verification points within the three-dimensional unit test grid; perform cave wall sensing sampling based on the set of planar verification points to obtain an actual set of verification mesh points; calculate the simulation accuracy based on the set of simulated feature mesh points and the actual set of verification mesh points; calculate the area of ​​the unit test grid of the three-dimensional unit test grid; plot points based on the area of ​​the unit test grid and the simulation accuracy to obtain an area accuracy point set; and perform regression analysis on the area accuracy point set to obtain an area accuracy regression line.

[0180] The three-dimensional unit current mesh acquisition module 102 is used to receive the current cave area to be measured, construct a triangular plane current mesh based on the current cave area to be measured, and perform cave inner wall sensing sampling based on the triangular plane current mesh to obtain the three-dimensional unit current mesh;

[0181] The three-dimensional unit current mesh correction module 103 is used to identify the current mesh area of ​​the three-dimensional unit current mesh, extract the predicted regression accuracy from the area accuracy regression line based on the current mesh area, perform secondary sensing sampling of the inner wall of the three-dimensional unit current mesh based on the predicted regression accuracy to obtain a secondary three-dimensional sampling point set, and perform three-dimensional correction of the three-dimensional unit current mesh using the secondary three-dimensional sampling point set to obtain an initial cave three-dimensional model.

[0182] The target cave 3D model construction module 104 is used to acquire cave sensing images using a pre-built sensing camera, and fuse the cave sensing images and the initial cave 3D model to obtain the target cave 3D model.

[0183] In detail, the modules in the multi-sensor fusion-based cave 3D modeling system 100 described in this embodiment of the invention employ the same methods as described above. Figure 1 The method uses the same techniques as the multi-sensor fusion-based cave 3D modeling method described in the article and can produce the same technical effects, so it will not be repeated here.

[0184] like Figure 3 The diagram shown is a schematic representation of an electronic device for implementing a multi-sensor fusion-based cave 3D modeling method according to an embodiment of the present invention.

[0185] The electronic device 1 may include a processor 10, a memory 11 and a bus 12, and may also include a computer program stored in the memory 11 and capable of running on the processor 10, such as a cave 3D modeling method program based on multi-sensor fusion.

[0186] The memory 11 includes at least one type of readable storage medium, such as flash memory, portable hard drive, multimedia card, card-type memory (e.g., SD or DX memory), magnetic memory, disk, optical disk, etc. In some embodiments, the memory 11 can be an internal storage unit of the electronic device 1, such as a portable hard drive. In other embodiments, the memory 11 can be an external storage device of the electronic device 1, such as a plug-in portable hard drive, smart media card (SMC), secure digital card (SD), flash card, etc., equipped on the electronic device 1. Furthermore, the memory 11 includes both internal storage units and external storage devices of the electronic device 1. The memory 11 can be used not only to store application software and various types of data installed on the electronic device 1, such as the code of a cave 3D modeling method program based on multi-sensor fusion, but also to temporarily store data that has been output or will be output.

[0187] In some embodiments, the processor 10 may be composed of integrated circuits, such as a single packaged integrated circuit or multiple integrated circuits with the same or different functions, including combinations of one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and various control chips. The processor 10 is the control unit of the electronic device, connecting various components of the entire electronic device through various interfaces and lines. It executes programs or modules stored in the memory 11 (e.g., a cave 3D modeling method program based on multi-sensor fusion) and calls data stored in the memory 11 to perform various functions of the electronic device 1 and process data.

[0188] The bus 12 can be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus 12 can be divided into an address bus, a data bus, a control bus, etc. The bus 12 is configured to realize the connection and communication between the memory 11 and at least one processor 10, etc.

[0189] Figure 3 Only electronic devices with components are shown; it will be understood by those skilled in the art that... Figure 3 The structure shown does not constitute a limitation on the electronic device 1, and may include fewer or more components than shown, or combine certain components, or have different component arrangements.

[0190] For example, although not shown, the electronic device 1 may also include a power supply (such as a battery) to power the various components. Preferably, the power supply can be logically connected to the at least one processor 10 through a power management device, thereby enabling functions such as charging management, discharging management, and power consumption management. The power supply may also include one or more DC or AC power supplies, recharging devices, power fault detection circuits, power converters or inverters, power status indicators, and other arbitrary components. The electronic device 1 may also include various sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be described in detail here.

[0191] Furthermore, the electronic device 1 may also include a network interface. Optionally, the network interface may include a wired interface and / or a wireless interface (such as a Wi-Fi interface, a Bluetooth interface, etc.), which is typically used to establish communication connections between the electronic device 1 and other electronic devices.

[0192] Optionally, the electronic device 1 may further include a user interface, which may be a display, an input unit (such as a keyboard), and optionally, a standard wired interface or a wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen, etc. The display may also be appropriately referred to as a screen or display unit, used to display information processed in the electronic device 1 and to display a visual user interface.

[0193] The cave 3D modeling method program based on multi-sensor fusion, stored in the memory 11 of the electronic device 1, is a combination of multiple instructions. When run in the processor 10, it can achieve the following:

[0194] Based on the preset triangular plane test grid, the cave interior wall is sampled using a pre-constructed inertial measurement sensor to obtain the cave's three-dimensional test grid.

[0195] Planar unit test grids are extracted sequentially from the triangular plane test grid, and the corresponding three-dimensional unit test grids are identified in the cave three-dimensional test grid.

[0196] The planar verification point set is uniformly set within the planar unit test grid according to the preset number of accuracy verification points;

[0197] Identify the set of simulated feature mesh points corresponding to the set of planar verification points within the three-dimensional unit test mesh;

[0198] Based on the aforementioned planar verification point set, the cave interior wall is sensor-sampled to obtain the actual verification mesh point set;

[0199] The simulation accuracy is calculated based on the simulated feature mesh point set and the actual verification mesh point set.

[0200] Calculate the area of ​​the three-dimensional unit test grid, and plot points based on the unit test grid area and simulation accuracy to obtain the area accuracy point set;

[0201] Regression analysis was performed on the area accuracy point set to obtain the area accuracy regression line;

[0202] Receive the current cave region to be measured, and construct a triangular plane mesh based on the current cave region to be measured;

[0203] Based on the current mesh of the triangular plane, the cave interior wall is sampled to obtain the current mesh of the three-dimensional unit;

[0204] Identify the current grid area of ​​the current mesh of the three-dimensional unit, and extract the predicted regression accuracy from the area accuracy regression line based on the current grid area of ​​the unit;

[0205] Based on the predicted regression accuracy, perform secondary sensing sampling of the inner wall of the current mesh of the three-dimensional unit to obtain a secondary three-dimensional sampling point set.

[0206] The current mesh of the three-dimensional unit is corrected using the secondary three-dimensional sampling point set to obtain the initial three-dimensional model of the cave.

[0207] Cave sensing images are acquired using a pre-built sensing camera, and the cave sensing images and the initial cave 3D model are fused to obtain the target cave 3D model.

[0208] Specifically, the processor 10's implementation method for the above instructions can be found in [reference needed]. Figures 1 to 3 The descriptions of the relevant steps in the corresponding embodiments are not repeated here.

[0209] Furthermore, if the modules / units integrated in the electronic device 1 are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. The computer-readable storage medium can be volatile or non-volatile. For example, the computer-readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, or a read-only memory (ROM).

[0210] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor of an electronic device, can perform the following:

[0211] Based on the preset triangular plane test grid, the cave interior wall is sampled using a pre-constructed inertial measurement sensor to obtain the cave's three-dimensional test grid.

[0212] Planar unit test grids are extracted sequentially from the triangular plane test grid, and the corresponding three-dimensional unit test grids are identified in the cave three-dimensional test grid.

[0213] The planar verification point set is uniformly set within the planar unit test grid according to the preset number of accuracy verification points;

[0214] Identify the set of simulated feature mesh points corresponding to the set of planar verification points within the three-dimensional unit test mesh;

[0215] Based on the aforementioned planar verification point set, the cave interior wall is sensor-sampled to obtain the actual verification mesh point set;

[0216] The simulation accuracy is calculated based on the simulated feature mesh point set and the actual verification mesh point set.

[0217] Calculate the area of ​​the three-dimensional unit test grid, and plot points based on the unit test grid area and simulation accuracy to obtain the area accuracy point set;

[0218] Regression analysis was performed on the area accuracy point set to obtain the area accuracy regression line;

[0219] Receive the current cave region to be measured, and construct a triangular plane mesh based on the current cave region to be measured;

[0220] Based on the current mesh of the triangular plane, the cave interior wall is sampled to obtain the current mesh of the three-dimensional unit;

[0221] Identify the current grid area of ​​the current mesh of the three-dimensional unit, and extract the predicted regression accuracy from the area accuracy regression line based on the current grid area of ​​the unit;

[0222] Based on the predicted regression accuracy, perform secondary sensing sampling of the inner wall of the current mesh of the three-dimensional unit to obtain a secondary three-dimensional sampling point set.

[0223] The current mesh of the three-dimensional unit is corrected using the secondary three-dimensional sampling point set to obtain the initial three-dimensional model of the cave.

[0224] Cave sensing images are acquired using a pre-built sensing camera, and the cave sensing images and the initial cave 3D model are fused to obtain the target cave 3D model.

[0225] In the embodiments provided by this invention, it should be understood that the disclosed devices, systems, and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative, and actual implementations may have other classification methods.

[0226] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0227] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0228] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0229] Finally, it should be noted that 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 preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for 3D cave modeling based on multi-sensor fusion, characterized in that, The method includes: Based on the preset triangular plane test grid, the cave interior wall is sampled using a pre-constructed inertial measurement sensor to obtain a three-dimensional test grid for the cave. The triangular plane test grid refers to the triangular grid located on the preset plane used to test the area accuracy regression line, and the three-dimensional test grid for the cave refers to the triangular grid surface formed by connecting the spatial points of the interior wall obtained from the cave interior wall sensor sampling. Planar unit test grids are extracted sequentially from the triangular plane test grid, and the corresponding three-dimensional unit test grids are identified in the cave three-dimensional test grid; the planar unit test grid refers to the unit grid in the triangular plane test grid; the three-dimensional unit test grid refers to the spatial three-dimensional grid obtained by connecting the test unit sampling point groups corresponding to the planar unit grid point sets in the planar unit test grid. A set of planar verification points is uniformly set within the planar unit test grid according to a preset number of accuracy verification points; the set of planar verification points refers to the set of points used to verify the simulation accuracy of the three-dimensional unit test grid. Identify the set of simulated feature mesh points corresponding to the set of planar verification points within the three-dimensional unit test mesh; the set of simulated feature mesh points refers to the set of intersection points between the surface normal vectors passing through the planar verification points and the three-dimensional unit test mesh. Based on the set of planar verification points, the cave wall is sensor-sampled to obtain the actual verification network point set; the actual verification network point set refers to the set of intersection points of the surface normal vector passing through the planar verification points and the cave wall. The simulation accuracy is calculated based on the simulated feature mesh point set and the actual verification mesh point set; the simulation accuracy refers to the accuracy of the simulation of the three-dimensional unit test mesh. Calculate the area of ​​the three-dimensional unit test grid, and plot points based on the unit test grid area and simulation accuracy to obtain the area accuracy point set; the area accuracy point set refers to the set of coordinate points with the unit test grid area as the independent variable and the simulation accuracy as the dependent variable. Regression analysis was performed on the area accuracy point set to obtain the area accuracy regression line; Receive the current cave region to be measured, and construct a triangular plane mesh based on the current cave region to be measured; Based on the current mesh of the triangular plane, the cave interior wall is sampled to obtain the current mesh of the three-dimensional unit; Identify the current grid area of ​​the current mesh of the three-dimensional unit, and extract the predicted regression accuracy from the area accuracy regression line based on the current grid area of ​​the unit; Based on the predicted regression accuracy, perform secondary sensing sampling of the inner wall of the current mesh of the three-dimensional unit to obtain a secondary three-dimensional sampling point set. The current mesh of the three-dimensional unit is corrected using the secondary three-dimensional sampling point set to obtain the initial three-dimensional model of the cave. Cave sensing images are acquired using a pre-built sensing camera, and the cave sensing images and the initial cave 3D model are fused to obtain the target cave 3D model.

2. The cave 3D modeling method based on multi-sensor fusion as described in claim 1, characterized in that, The process of obtaining a three-dimensional test grid for the cave by using a pre-constructed inertial measurement sensor to perform sensing sampling on the cave wall based on a preset triangular plane test grid includes: Identify the triangular plane grid points in the triangular plane test grid, identify the surface normal vectors of the triangular plane grid points to the cave wall, and obtain the surface normal vector set; Extract the surface normal vectors sequentially from the surface normal vector set; Laser sampling of the cave wall is performed on the surface normal vector using a lidar, and the attitude of the lidar is calibrated using the inertial measurement sensor to obtain a test three-dimensional sampling point set; the test three-dimensional sampling point set refers to the set of three-dimensional spatial positions of the cave wall determined based on the normal vector of the triangular plane grid points during the regression relationship between the test grid area and the simulation accuracy of the test unit. Identify the set of planar element mesh points in the planar element test mesh, and identify the set of element normal vectors corresponding to the set of planar element mesh points in the set of surface normal vectors; the set of planar element mesh points refers to the three sets of mesh points of the planar element test mesh; the set of element normal vectors refers to the three surface normal vectors with the planar element mesh points in the set of planar element mesh points as the vector starting points; Identify the test unit sampling point group corresponding to the unit normal vector group in the test three-dimensional sampling point set, wherein the number of test unit sampling points in the test unit sampling point group is 3; the test unit sampling point group refers to the three three-dimensional spatial intersection points of the unit normal vector in the unit normal vector group and the inner wall of the cave. By connecting the sampling points of the test unit in pairs, a three-dimensional test grid for the cave is obtained.

3. The cave 3D modeling method based on multi-sensor fusion as described in claim 2, characterized in that, The step of uniformly setting a set of planar verification points within the planar unit test grid according to a preset number of precision verification points includes: The number of unit grid test split blocks of the planar unit test grid is determined based on the number of precision verification points, wherein the number of precision verification points is: n is a preset positive integer; the number of precision verification points refers to the preset number of points used to verify the simulation precision of the three-dimensional unit test mesh; The planar unit test grid is equally divided according to the number of unit grid test split blocks to obtain a test unit grid sub-block set, wherein the test unit grid sub-blocks in the test unit grid sub-block set are equilateral triangles; the number of unit grid test split blocks is equal to the number of precision verification points; Identify the center point of each test sub-block in the test unit grid sub-block set, and use the center point of the test sub-block as a plane verification point to obtain a plane verification point set.

4. The cave 3D modeling method based on multi-sensor fusion as described in claim 3, characterized in that, The step of identifying the simulated feature mesh point set corresponding to the planar verification point set within the three-dimensional unit test mesh includes: Identify the verification normal vector corresponding to each plane verification point in the plane verification point set to obtain a verification normal vector set, wherein the verification normal vector refers to the surface normal vector of the plane verification point to the inner wall of the cave. Identify the line-plane intersection points of each verification normal vector in the verification normal vector set with the three-dimensional unit test mesh to obtain the line-plane intersection point set; The set of intersection points of the line and the surface is used as the set of points of the simulated feature network.

5. The cave 3D modeling method based on multi-sensor fusion as described in claim 4, characterized in that, The step of calculating the simulation accuracy based on the simulated feature mesh point set and the actual verification mesh point set includes: Identify the associated simulated mesh points and associated actual mesh points corresponding to the planar verification point in the set of simulated feature mesh points and the set of actual verification mesh points; the associated simulated mesh points refer to the simulated feature mesh points that are on the same surface normal vector as the planar verification point, and the associated actual mesh points refer to the actual verification mesh points that are on the same surface normal vector as the planar verification point. Identify the heights of the associated simulated mesh points and the associated actual mesh points along the verification normal vector direction to obtain the simulated mesh point height set and the actual mesh point height set; the simulated mesh point height refers to the height distance of the associated simulated mesh point from the plane along the verification normal vector direction; the actual mesh point height refers to the height distance of the associated actual mesh point from the plane along the verification normal vector direction. Based on the simulated mesh point height set and the actual mesh point height set, the simulation accuracy is calculated using the following formula: ； in, This represents the simulation accuracy of the i-th planar element test mesh. Indicates the precision adjustment index. This represents the number of planar verification points in the i-th planar unit test grid. This represents the actual height of the mesh point corresponding to the j-th planar verification point in the i-th planar unit test mesh. This represents the height of the simulated mesh point corresponding to the j-th planar verification point in the i-th planar unit test mesh. Represents the absolute value symbol.

6. The cave 3D modeling method based on multi-sensor fusion as described in claim 5, characterized in that, The regression analysis performed on the area accuracy point set to obtain the area accuracy regression line includes: Identify the area precision coordinates of each area precision point in the area precision point set to obtain the area precision coordinate set; Based on the area precision coordinate set, the slope of the regression line is calculated using the following formula: ； Where k represents the slope of the regression line. This indicates the number of planar unit test meshes. This represents the x-coordinate of the p-th area precision coordinate in the area precision coordinate set. Represents the ordinate of the p-th area precision coordinate in the area precision coordinate set; The intercept of the regression line is calculated using the following formula based on the area accuracy coordinate system: ； in, Indicates the intercept of the regression line; The area precision regression line is determined based on the slope and intercept of the regression line, wherein the equation of the area precision regression line is as follows: ； in, This represents the value of the regression accuracy function. This represents the value of the grid area function.

7. The cave 3D modeling method based on multi-sensor fusion as described in claim 6, characterized in that, The step of performing secondary sensing sampling of the inner wall of the current mesh of the three-dimensional unit based on the predicted regression accuracy to obtain a secondary three-dimensional sampling point set includes: Identify the planar cell current mesh corresponding to the current 3D cell mesh in the current triangular plane mesh; the planar cell current mesh refers to the equilateral triangular cell mesh that the current 3D cell mesh is projected perpendicularly onto the plane. Based on the predicted regression accuracy, the accuracy filler points are calculated using the following formula: ； in, The precision is indicated by the number of fill points, and e represents the natural constant. Indicates the accuracy of the predicted regression. The symbol indicates rounding down; the number of precision fill points refers to the number of mesh points used for 3D correction of the current mesh of the 3D element; The number of precision compensation points is used as the current number of cell grid split blocks of the current grid of the planar element; The planar unit current grid is equally divided according to the current number of split blocks of the unit grid to obtain the current unit grid sub-block set, wherein the current unit grid sub-blocks in the current unit grid sub-block set are equilateral triangles; Identify the center point of the current sub-block of each current unit grid sub-block in the current unit grid sub-block set, and use the center point of the current sub-block as the plane completion point to obtain the plane completion point set; Based on the aforementioned planar supplementary point set, the cave interior wall is sensor-sampled to obtain a secondary three-dimensional sampling point set.

8. The cave 3D modeling method based on multi-sensor fusion as described in claim 7, characterized in that, The step of using the secondary three-dimensional sampling point set to perform three-dimensional correction on the current mesh of the three-dimensional unit to obtain the initial three-dimensional cave model includes: In the current unit grid sub-block set, identify the group of unit grid sub-blocks sharing the same edge, and identify the center group of the group of unit grid sub-blocks sharing the same edge. The group of unit grid sub-blocks sharing the same edge refers to two current unit grid sub-blocks sharing the same edge. The center group of the ... Identify the associated secondary three-dimensional sampling point group corresponding to the center group of the shared-edge sub-block in the secondary three-dimensional sampling point set; the associated secondary three-dimensional sampling point group refers to the two secondary three-dimensional sampling points corresponding to the center group of the shared-edge sub-block in the secondary three-dimensional sampling point set in the direction of the surface normal vector. Connecting the associated secondary three-dimensional sampling points in the associated secondary three-dimensional sampling point group yields a three-dimensional corrected feature line; the three-dimensional corrected feature line refers to a three-dimensional spatial line segment formed by connecting the associated secondary three-dimensional sampling point group. The three-dimensional correction feature lines are used to perform three-dimensional correction on the current mesh of the three-dimensional unit to obtain the initial three-dimensional cave model.

9. The cave 3D modeling method based on multi-sensor fusion as described in claim 8, characterized in that, The process of fusing the cave sensing image and the initial cave 3D model to obtain the target cave 3D model includes: The texture and color of each triangular plane in the initial 3D cave model are identified using the cave sensing images. The initial cave 3D model is supplemented with texture color based on the texture color of the triangular plane to obtain the target cave 3D model.

10. A cave 3D modeling system based on multi-sensor fusion, characterized in that, The system includes: The area accuracy regression line recognition module is used to obtain a three-dimensional cave test grid by using a pre-constructed inertial measurement sensor to perform sensing sampling on the cave's inner wall based on a preset triangular plane test grid. The triangular plane test grid refers to a triangular grid located on a preset plane used to test the area accuracy regression line. The three-dimensional cave test grid refers to a triangular grid surface formed by connecting spatial points on the inner wall obtained from the cave's inner wall sensing sampling. Planar unit test grids are sequentially extracted from the triangular plane test grid, and the corresponding three-dimensional unit test grids are identified in the three-dimensional cave test grid. The planar unit test grid refers to the unit grid within the triangular plane test grid. The three-dimensional unit test grid refers to a spatial three-dimensional grid formed by connecting test unit sampling point groups corresponding to the planar unit grid point sets in the planar unit test grid. A planar verification point set is uniformly set within the planar unit test grid according to a preset number of accuracy verification points. The planar verification point set refers to the set of points used to verify the three-dimensional accuracy of the individual units within the test grid. The simulation accuracy of the three-dimensional unit test grid is verified by a set of points; the simulated feature mesh point set corresponding to the planar verification point set is identified within the three-dimensional unit test grid; the simulated feature mesh point set refers to the set of intersection points between the surface normal vector passing through the planar verification point and the three-dimensional unit test grid; the cave wall is sampled using the planar verification point set to obtain the actual verification mesh point set; the actual verification mesh point set refers to the set of intersection points between the surface normal vector passing through the planar verification point and the cave wall; the simulation accuracy is calculated based on the simulated feature mesh point set and the actual verification mesh point set; the simulation accuracy refers to the simulation precision of the three-dimensional unit test grid; the unit test grid area of ​​the three-dimensional unit test grid is calculated, and points are plotted based on the unit test grid area and the simulation accuracy to obtain the area accuracy point set; the area accuracy point set refers to the set of coordinate points with the unit test grid area as the independent variable and the simulation accuracy as the dependent variable; regression analysis is performed on the area accuracy point set to obtain the area accuracy regression line. The 3D unit current mesh acquisition module is used to receive the current cave area to be measured, construct a triangular plane current mesh based on the current cave area to be measured, and perform cave inner wall sensing sampling based on the triangular plane current mesh to obtain the 3D unit current mesh; The three-dimensional unit current mesh correction module is used to identify the current mesh area of ​​the three-dimensional unit current mesh, extract the predicted regression accuracy from the area accuracy regression line based on the current mesh area, perform secondary sensing sampling of the inner wall of the three-dimensional unit current mesh based on the predicted regression accuracy to obtain a secondary three-dimensional sampling point set, and use the secondary three-dimensional sampling point set to perform three-dimensional correction of the three-dimensional unit current mesh to obtain an initial cave three-dimensional model. The target cave 3D model construction module is used to acquire cave sensing images using a pre-built sensing camera, and to fuse the cave sensing images and the initial cave 3D model to obtain the target cave 3D model.

Images