A cross-section borehole positioning method based on central projection

Abstract

This invention relates to the field of cross-section borehole positioning technology, specifically disclosing a cross-section borehole positioning method based on central projection. By acquiring the 3D point cloud reconstruction mesh model of the cross-section and its projection pose parameters, the ray intersections and topological constraints are simultaneously calculated within a preset projection domain. An initial intersection set with occlusion determination and distortion ratio form a complete spatial mapping chain. An adaptive correction loop is constructed to relax the mesh protrusion occlusion and borehole spacing distortion in a linked manner, deriving the individualized optimal 3D borehole position distribution in reverse. This method breaks through the rigid limitations of traditional 2D templates, achieving accurate projection based on the true geometric shape of the cross-section, and providing a new high-fidelity pre-distortion guidance path for complex curved surface blasting operations.

Description

Technical Field

[0001] This invention belongs to the field of cross-section borehole positioning technology, and relates to a cross-section borehole positioning method based on center projection. Background Technology

[0002] Tunnel drilling and blasting is a crucial technology in underground engineering excavation and shaping. The accuracy of borehole positioning directly affects the uniform distribution of blasting energy, the smooth formation of the tunnel face, and the safe implementation of subsequent support operations. Since the tunnel face is the rock interface directly exposed after blasting, it bears both the blasting unloading effect and the combined influence of the original geological structure and joints and fissures. Its actual morphology is often not an ideal plane, but rather exhibits complex geometric features such as local depressions caused by under-excavation, local protrusions formed by joint penetrations, and step-like fractures resulting from uneven blasting. When positioning boreholes on such irregular surfaces, the spatial distribution of the borehole positions must not only meet the design hole spacing and resistance line requirements but also take into account the geometrical offsets caused by local changes in the tunnel face's posture. Otherwise, problems such as borehole misalignment and the coexistence of under-excavation and over-excavation can easily occur, directly affecting the quality of blasting and the efficiency of construction.

[0003] Existing center projection methods typically use an ideal plane as the projection receiving surface, directly mapping the two-dimensional hole mesh onto the target cross-section. While this maintains a certain degree of positioning consistency when the cross-section has minimal undulations, significant shifts occur when the tunnel face exhibits obvious unevenness or local large-angle joints. This leads to significant deviations in the landing points of the same projected light rays at different depths, resulting in distortion of hole spacing and deviation of the contour boundaries. Chinese Patent Publication No. CN120444083A discloses a laser galvanometer projection correction method and system suitable for uneven tunnel cross-sections. This method can perform projection correction on blasted cross-sections with known contours. However, it relies on the complete acquisition and subsequent correction of the cross-section morphology. Therefore, it is more suitable for post-blast cross-section verification and correction scenarios and is difficult to directly apply to the real-time hole layout process during the initial drilling positioning stage when the tunnel face has not yet been scanned and the contour information is incomplete. Summary of the Invention

[0004] In view of the problems existing in the prior art, the present invention provides a cross-sectional borehole positioning method based on center projection to solve the above-mentioned technical problems.

[0005] To achieve the above and other objectives, the technical solution adopted by the present invention is as follows: This invention provides a method for locating boreholes in cross-sections based on center projection, the method comprising: Acquire the 3D point cloud data of the cross section and convert it into a 3D triangular mesh model; Obtain the three-dimensional spatial coordinates of the projection center and the two-dimensional standard mesh image to be projected; An initial projection ray is emitted from the projection center to each first borehole pixel in the two-dimensional standard aperture mesh image, and the initial spatial intersection set of the initial projection ray and the three-dimensional triangular mesh model is calculated. Iterate through the initial set of spatial intersection points and enter the adaptive correction loop. The adaptive correction loop includes: Construct a connection vector from the projection center to the initial spatial intersection point of the target, and extract the depth buffer data on the connection vector to determine whether there is mesh protrusion occlusion; If there are mesh protrusions that obstruct the view, search for nearby visible mesh nodes on the surface of the 3D triangular mesh model and update them as the initial spatial intersection points of the target. If there is no mesh protrusion occlusion, calculate the 3D topological distance between the initial spatial intersection of the target and its adjacent spatial intersections; Based on the distortion ratio between the 3D topological distance and the preset standard aperture distance, mesh relaxation iteration is performed to drive the initial spatial intersection of the target to slide on the surface of the 3D triangular mesh model to an equilibrium position where the distortion ratio is lower than the preset distortion threshold, thereby generating an optimized set of 3D target aperture positions. Extract the 3D coordinate data of the optimized 3D target aperture set, and construct the reverse perspective transformation matrix based on the optical parameters of the projection device; A pre-distorted two-dimensional projection image is generated by performing dimension reduction and remapping on the three-dimensional coordinate data using the inverse perspective transformation matrix. Output and control the projection device to project the pre-distorted two-dimensional projection image onto the physical cross-section.

[0006] As described above, the cross-sectional borehole positioning method based on center projection provided by the present invention has at least the following beneficial effects: This invention first reconstructs the three-dimensional point cloud data of the cross-section by meshing, and then combines the precise spatial coordinates of the projection center with the two-dimensional standard borehole mesh image to establish a unified mapping link from the designed borehole position to the actual cross-section. This enables the center projection method, which was originally only applicable to an ideal plane, to have a basis for perceiving the local concave and convex morphology of the working face. On this basis, this invention does not directly project the two-dimensional borehole mesh onto the three-dimensional cross-section all at once. Instead, it first emits initial projection rays for each first borehole pixel and calculates the initial set of spatial intersections. Then, it combines the depth buffer data on the connecting vector to determine whether there is mesh protrusion occlusion. Thus, in the early stage of projection, it identifies borehole positions that are easily disturbed by local joints, stepped fractures, and protruding rock masses, avoiding the problems of distorted borehole positions and the inability to correct occlusion in traditional fixed projection methods. For intersections with occlusion, this invention further searches for and replaces nearby visible nodes on the surface of the 3D triangular mesh model, enabling the boreholes to automatically bypass unworkable areas. For unoccluded areas, the distortion ratio between the 3D topological distance and the standard borehole spacing drives mesh relaxation iteration, allowing adjacent boreholes to gradually slide on the curved surface to a balanced position that satisfies the design borehole mesh constraints. This approach not only elevates borehole positioning from a simple geometric projection to an adaptive process that considers occlusion recognition, neighborhood replacement, and topological constraint optimization, but also significantly reduces the disruption to borehole spacing uniformity on uneven working faces, improves the accuracy of surrounding borehole contour forming, and enhances the consistency between the overall borehole layout and the actual cross-section.

[0007] This invention reverse-maps the optimized 3D target borehole location set onto the 2D imaging plane of a projection device. A reverse perspective transformation matrix is ​​constructed using the optical parameters of the projection device. The 3D coordinate data is then dimensionality-reduced and remapped to generate a pre-distorted 2D projection image. This image is then projected onto the physical cross-section. The pre-distortion of the 2D image offsets the spatial distortion of the 3D cross-section, resulting in a visual positioning result on-site that is highly consistent with the designed borehole network. Compared to existing center projection methods that rely solely on the assumption of an ideal plane, this invention adds continuous processing steps before projection, including point cloud reconstruction, occlusion determination, neighborhood correction, topological relaxation, and reverse distortion compensation. This ensures that the borehole location output not only meets the requirements for projection visibility but also takes into account the constraints of blasting construction on borehole spacing, angle, and borehole location continuity, thereby improving the stability and anti-interference capability of borehole positioning. Especially in complex conditions with significant local undulations at the tunnel face, obvious protrusions causing occlusion, or steep joint surfaces, this invention effectively reduces problems such as borehole deviation, missed projections, and ghosting, ensuring the geometric consistency and construction operability of the projected pattern on the actual cross-section. Ultimately, it provides a more reliable digital positioning foundation for high-precision blasting borehole layout. Attached Figure Description

[0008] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0009] Figure 1 This is a schematic diagram showing the connections between the steps of the method of the present invention.

[0010] Figure 2 This is a schematic diagram of the adaptive correction loop logic provided by the present invention. Detailed Implementation

[0011] The following description, in conjunction with the implementation of this invention, is merely an example and illustration of the concept of this invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the inventive concept or exceed the scope defined in these claims, all of which should fall within the protection scope of this invention.

[0012] In traditional center-projection section positioning systems, two-dimensional standard borehole meshes are typically directly mapped onto the section receiving surface. The system defaults to projecting an approximate ideal plane, thus ignoring the complex three-dimensional morphology of the drilling face after blasting, such as depressions, protrusions, joint steps, and local occlusions. When there are significant elevation differences or large local dip angles in the section, the actual landing points of the borehole pixels, which were originally uniformly distributed in the two-dimensional image, will experience radial offset and tangential stretching in three-dimensional space. This leads to problems such as hole spacing distortion, contour boundary drift, and local hole overlap, thereby weakening the blasting contour control effect and the consistency of borehole mesh construction. Especially without sufficient perception and pre-compensation of the section, the projection results often only meet the visual rough alignment requirements, making it difficult to guarantee the drilling positioning accuracy required for actual construction.

[0013] For example, in scenarios where there are local protrusions or stepped fractures at the tunnel face, if the system generates hole positions solely based on a fixed mapping relationship between the projection center and the 2D hole mesh, the pixels of boreholes located behind protrusions will have missing landing points due to the truncation of the light path, while the pixels of boreholes located in recessed areas will experience excessive compression or expansion due to depth changes. In this case, the traditional system still uses a static projection model and cannot identify which holes are obscured or which holes have experienced hole spacing distortion. This results in some critical boreholes being incorrectly retained as valid points, while some landing points that should have been corrected are directly output to the physical cross-section. Once the projection results are used for subsequent drilling operations, it will cause hole positions to deviate from the design axis, irregular hole formation around the perimeter, and failure of blasting boundary control, thereby affecting the cross-section formation quality and construction safety.

[0014] If the aforementioned problems are not addressed, the application of the central projection method on complex cross-sections will continue to be limited by the fixed receiving surface assumption and single-path mapping mechanism, making it impossible to respond in real time to changes in the face morphology. The direct consequence is that the geometric error between the projected pattern and the actual cross-section cannot be actively eliminated. Once the hole position has already shifted during the projection stage, subsequent drilling can only passively accept the accumulation of errors, ultimately resulting in a superimposed amplification effect of projection errors, positioning errors, and construction errors. Simultaneously, traditional methods lack a mechanism for replacing obscured areas with nearby visible alternatives and an iterative correction process for hole spacing distortion. This makes it difficult to capture the impact of areas with significant local curvature changes on the stability of the borehole network, thus preventing the formation of a constructable, verifiable, and reviewable closed-loop control during the hole position generation stage.

[0015] When faced with the aforementioned problems, if a fixed threshold and static mapping strategy are still used to project borehole positions onto the cross-section, the borehole location results will continuously mismatch with the actual morphology of the tunnel face, failing to reflect the real-time impact of local concavity and convexity variations on the borehole distribution. To address this, this invention attempts to establish a dynamic coupling relationship between the projection center, the 3D mesh morphology, and the 2D standard borehole mesh through a series of steps, including 3D point cloud reconstruction of the cross-section, initial intersection point solving, occlusion determination, replacement of nearby visible nodes, topological distance-driven mesh relaxation iteration, and reverse perspective distortion compensation. This transforms the borehole generation process from a one-time mapping to a linked mechanism of adaptive correction and reverse remapping. Further analysis reveals that relying solely on 2D patterns or single intersection point calculations is insufficient to capture the synergistic effects of local occlusion and surface distortion at the tunnel face. Therefore, depth buffer determination and 3D topological distance constraints need to be introduced simultaneously in the adaptive correction loop to generate an optimized borehole set that better matches the actual cross-section conditions. Finally, a 2D projection image with pre-distortion compensation features is output, thereby improving the accuracy and construction adaptability of borehole positioning under complex cross-sections.

[0016] After introducing the basic concept of the present invention, the embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0017] Example 1: Please see Figure 1-2 As shown, a method for locating boreholes in a cross-section based on central projection includes the following steps: Preferably, before acquiring the 3D point cloud data of the cross-section and converting it into a 3D triangular mesh model, the method further includes: The drive line structured light sensor emits a scanning beam array toward the cross-section to be projected and receives the image of the beam deformation after reflection. An initial dense point set is calculated by performing a stereo matching operation based on epipolar geometric constraints on the beam deformation image; A voxel downsampling filtering algorithm is applied to spatially mesh and sparse the density of the initial dense point set to remove redundant and overlapping noise points and generate a simplified point set data. The simplified point set data is input into the pre-established projector base coordinate system to perform a three-dimensional affine transformation alignment operation, unifying the spatial reference frame and outputting three-dimensional point cloud data.

[0018] Acquire the 3D point cloud data of the cross section and convert it into a 3D triangular mesh model; Obtain the three-dimensional spatial coordinates of the projection center and the two-dimensional standard mesh image to be projected; An initial projection ray is emitted from the projection center to each first borehole pixel in the two-dimensional standard borehole image, and the initial spatial intersection set of the initial projection ray and the three-dimensional triangular mesh model is calculated.

[0019] In this embodiment of the invention, firstly, a line structured light sensor integrated at the front end of the projection device is driven to emit an infrared scanning beam array of a preset frequency toward the physical cross-section to be projected, and simultaneously, an infrared camera receives the beam deformation image, which is distorted after reflection from the surface topography of the physical cross-section. After receiving the beam deformation image, a stereo matching operation based on epipolar geometric constraints is performed to calculate an initial dense point set. Specifically, for any pixel feature point in the beam deformation image, the epipolar search range is narrowed using pre-calibrated left and right camera fundamental matrix constraints, and disparity data is extracted using a normalized cross-correlation algorithm. Combined with the binocular vision triangulation principle, the disparity data is converted into an initial dense point set containing X, Y, and Z three-dimensional coordinate components. The point cloud coordinates obtained in this step are all in millimeters. To eliminate data redundancy caused by on-site dust interference and high-frequency sampling of the equipment, a voxel downsampling filtering algorithm is used to perform spatial gridding density sparse processing on the initial dense point set. Specifically, the three-dimensional space is divided into a grid with a side length of... A cubic voxel mesh, in which For the preset spatial resolution threshold, it is recommended to set it to an adjustable range of 2 mm to 10 mm. For the multiple point cloud data contained in each voxel, accumulate the three-dimensional coordinates and obtain the geometric centroid. Use the three-dimensional coordinates of the centroid to replace all the original points contained in the voxel, thereby eliminating redundant and overlapping noise points and generating simplified point set data. Subsequently, the generated simplified point set data is input into the pre-established projector base coordinate system, and a three-dimensional affine transformation alignment operation is performed to unify the spatial reference frame. The transformation calculation formula is as follows:

[0020] in The simplified 3D coordinate vector of the point set in the camera coordinate system is expressed in millimeters. R is a 3×3 orthogonal rotation matrix, dimensionless, calculated by the system's pre-shipment stereo calibration program using a checkerboard calibration board. T is a 3×1 translation vector, also in millimeters, representing the physical offset distance between the optical centers of the structured light camera and the projector in 3D space. The data represents the transformed 3D point cloud coordinates in the projector coordinate system, in millimeters. The topological connectivity of the unified 3D point cloud coordinate data is further reconstructed using the Delaunay triangulation algorithm, outputting a continuous 3D triangular mesh model. Next, the 3D spatial coordinates of the projection center are extracted from the system calibration parameter library of the equipment, and the 2D standard aperture mesh image to be projected is read. For each first borehole pixel in the 2D standard aperture mesh image, an initial projection ray is constructed based on the projection center as the origin and the perspective imaging principle of the projector. An intersection solution formula with grazing angle compensation is then constructed.

[0021] Where t is the scalar distance of the initial projected ray from the projection center to the grid intersection point, in millimeters; u and v are the centroid coordinates of the intersection point relative to the vertex in the penetrated triangular grid surface, both of which are dimensionless values; D is the direction vector of the initial projected ray corresponding to the first borehole pixel point, which is a normalized dimensionless vector. and These are the directed edge vectors of the two shared vertices of the currently traversed triangular facet in the 3D triangular mesh model, in millimeters; S is the translation vector from the 3D coordinates of the projection center to the reference vertex of the triangular facet, in millimeters. The adaptive relaxation coefficient for the normal direction, introduced for the steep joint surface of the tunnel, is calculated by scalar multiplication with the vector on the right side of the equation. The formula is as follows: , where n is the unit normal vector of the triangular facet. The pre-defined grazing angle compensation factor is dimensionless and logically pre-defined. The value typically ranges from 0.01 to 0.05, with 0.02 being the preferred value. This variation is used to introduce... This is because when the ray direction vector D is close to perpendicular to the mesh surface normal vector n (i.e. close to grazing state, the dot product result is close to 0), floating point operation error can easily lead to ray penetration failure or intersection coordinate drift. Introducing this relaxation coefficient based on the cosine of the normal angle can slightly widen the tolerance range of effective penetration, thereby improving the success rate of intersection capture under harsh morphology. The parameters calculated using this formula need to be verified by boundary conditions. , and When it is determined that the ray has effectively penetrated the current triangular mesh surface, the calculated distance scalar t is substituted into the equation of the line. ,in Using the coordinates of the projection center, calculate the absolute three-dimensional coordinates of the intersection point. And continue to add it to the initial set of spatial intersections.

[0022] Preferably, before traversing the initial set of spatial intersections and entering the adaptive correction loop, the method further includes a step of perceptually extracting cross-sectional step features from the 3D triangular mesh model: Calculate the local normal vector parameters of each mesh face in a 3D triangular mesh model; Compare the local normal angle values ​​of two adjacent mesh faces that share the same mesh edge; When the angle between local normal vectors is greater than the preset cross-sectional abrupt change threshold, the two adjacent grid surfaces are marked as step fracture surface features in the dataset and stored in the obstacle avoidance prohibited area spatial library. A boundary penalty function is introduced when performing mesh relaxation iterations.

[0023] Preferably, a boundary penalty function is introduced during mesh relaxation iteration, including: Monitor the sliding trajectory coordinates of the initial spatial intersection point of the target on the surface of the three-dimensional triangular mesh model; Determine whether the coordinates of the sliding trajectory fall within the boundary influence range of the obstacle avoidance prohibited area space library; If it falls within the boundary influence range, the boundary penalty function is triggered to generate a penalty repulsion force vector that is reversed along the feature normal of the step fracture surface; The penalty repulsion force vector is superimposed onto the tangential virtual force vector of the current mesh surface to change the target's sliding direction.

[0024] In this embodiment of the invention, to ensure that the borehole positioning points do not fall on impassable step fractures or deep groove areas under the drastically uneven face topography, a cross-sectional step feature perception and extraction program is first executed before traversing the initial spatial intersection set and entering the adaptive correction loop. This program first calculates each mesh facet in the three-dimensional triangular mesh model. Local normal vector parameters The calculation formula is: ,in and For dough The two edge vectors sharing a vertex are in millimeters. Next, the topological adjacency list is used to retrieve and compare two adjacent mesh faces that share the same mesh edge. and The local normal vector angle value The unit is degrees, when the angle between the local normal vectors is... When the angle exceeds the preset cross-sectional abrupt change threshold, the shared grid edge is determined to be a brittle fracture step formed after rock blasting. The two adjacent grid patches are marked as step fracture surface features in the dataset and stored in the obstacle avoidance prohibited area spatial library. The preset cross-sectional abrupt change threshold is used to effectively filter out the small undulations caused by natural rock joints and lock in steep steps with drilling risks. 45° is recommended as the preferred value. This threshold is used to distinguish between fracture steps formed by blasting and natural rock joints or continuous curved surfaces. Generally, a patch angle greater than 45° indicates the presence of obvious fracture steps, which constitute drilling construction obstacles. Subsequently, during the mesh relaxation iteration, the system monitors in real time the sliding trajectory coordinates of the initial spatial intersection point of the target on the surface of the 3D triangular mesh model. It then determines whether the trajectory coordinates fall within the boundary influence range of the obstacle avoidance prohibited area space library. If they do, it triggers the boundary penalty function to generate a penalty repulsion force vector that is back-inferred along the feature normal of the step fracture surface. The formula for constructing the variant of the penalty repulsion force vector is as follows:

[0025] in The penalty displacement bias vector, in millimeters, is essentially a repulsive force model based on the barrier gradient descent method, which aims to simulate the repulsive effect of the step edge on the hole through nonlinear decay. The preset penalty intensity coefficient, in millimeters, represents the maximum physical displacement of the penalty effect. The preferred value is 10 millimeters to ensure that it can push open the hole without causing the hole mesh structure to collapse. Current trajectory coordinates Euclidean distance to the nearest edge of the step fracture surface, in millimeters; The preset boundary sensing bandwidth, in millimeters, controls the effective range of the penalty force; a value of 10% to 20% of the standard aperture spacing is recommended. The coordinates of the nearest geometric boundary point within the obstacle avoidance restricted area, in millimeters, are given; finally, the calculated penalty repulsion force vector is... The resultant force is directly superimposed on the tangential virtual force vector generated by the hole spacing constraint on the current mesh surface, and together they drive the target's sliding direction.

[0026] Iterate through the initial set of spatial intersection points and enter the adaptive correction loop. The adaptive correction loop includes: Construct a connection vector from the projection center to the initial spatial intersection point of the target, and extract the depth buffer data on the connection vector to determine whether there is mesh occlusion, including: A sequence of sampling points is obtained by discretizing the three-dimensional coordinates along the connecting vector at a preset step size; Extract the local depth value of each sampling point in the projection direction of the 3D triangular mesh model to construct depth buffer data; Compare the local depth value with the linear distance from the sampling point to the projection center; If the local depth value of any sampling point is less than the straight-line distance and the difference exceeds the preset tolerance, it is determined that there is a mesh protrusion occlusion.

[0027] If there are mesh protrusions that obstruct the view, search for nearby visible mesh nodes on the surface of the 3D triangular mesh model and update them as the initial spatial intersection points of the target. Preferably, searching for nearby visible mesh nodes on the surface of the 3D triangular mesh model and updating them as the target initial spatial intersection points includes: Using the initial spatial intersection point of the target as the geometric center, a breadth-first search is performed outward on the three-dimensional triangular mesh model according to the preset topological hierarchy to extract the set of candidate mesh nodes; Spatial cross-validation is performed between the candidate grid node set and the depth buffer data to remove invalid nodes that are occluded and retain valid visible nodes; Calculate the abrupt change value of the normal vector of the local tangent plane where each valid visible node is located; Select the benchmark visible nodes whose normal vector mutation value is less than the preset angle threshold and whose distance from the geodesic line of the initial spatial intersection point of the target is the shortest. The current node is updated by assigning the 3D coordinates of the reference visible node to the initial spatial intersection point of the target.

[0028] If there is no mesh protrusion occlusion, calculate the 3D topological distance between the initial spatial intersection of the target and its adjacent spatial intersections.

[0029] Preferably, calculating the three-dimensional topological distance between the initial spatial intersection point of the target and its adjacent spatial intersection points includes: Extract all mesh edge connectivity paths connecting the initial spatial intersection point of the target and its adjacent spatial intersection points from the 3D triangular mesh model; Calculate the cumulative edge length along the mesh surface for each connected path along the mesh edge; Dijkstra's algorithm is used to select the shortest cumulative edge length value from all grid edge-connected paths as the geodesic distance data between two points. Set the geodesic distance data as a three-dimensional topological distance.

[0030] In this embodiment of the invention, after generating the initial spatial intersection point, the set of initial spatial intersection points is traversed to enter an adaptive correction loop. First, a connecting vector is constructed from the projection center to the target initial spatial intersection point, and the depth buffer data on the connecting vector is extracted to determine whether there is a mesh protrusion occlusion. In specific execution, the system performs three-dimensional coordinate discretization sampling along the connecting vector according to a preset step size to obtain a sampling point sequence. The preset step size is to balance the detection accuracy and the memory throughput burden of the underlying hardware, and is preferably 2 mm. For each discrete sampling point, it is projected onto the device's imaging plane using a pre-established frustum matrix. The local depth value of the corresponding pixel position in the projection direction of the 3D triangular mesh model is extracted to construct depth buffer data. The pre-established frustum matrix is ​​constructed by multiplying the optical intrinsic parameter matrix pre-calibrated by the projection device with the extrinsic parameter matrix representing the spatial pose of the projection center. Specifically, the depth buffer data is a global 2D depth map generated before the adaptive correction loop begins, based on the frustum matrix, using a GPU hardware-accelerated off-screen rendering pipeline or a CPU depth testing algorithm to perform a rasterization projection rendering of the entire 3D triangular mesh model. When extracting local depth values, the system directly multiplies the 3D coordinates of the sampling point by the frustum matrix to transform them into 2D pixel coordinates (u, v) on the imaging plane. It then uses bilinear interpolation to query the corresponding Z-axis depth value in the pre-generated global 2D depth map. This simulates the perspective observation process of the projection device in a 3D scene and significantly improves the efficiency of occlusion determination during ray stepping. Subsequently, the local depth value is compared with the physical straight-line distance from the current sampling point to the projection center. To avoid misjudgment caused by truncation of floating-point operations, this invention adopts a variant formula for occlusion determination and evaluation based on relative error limits:

[0031] in The occlusion determination coefficient is dimensionless. This represents the three-dimensional Euclidean straight-line distance from the current sampling point to the projection center, in millimeters. This is the local depth value of the grid read from the perspective projection of the sampling point onto the (u,v) pixel coordinates, in millimeters; If the calculated occlusion determination coefficient If the value exceeds the preset tolerance (the preset tolerance is typically between 0.01 and 0.03, preferably 0.02), it is determined that the current sampling point is obscured by a preceding protruding rock surface, thus confirming that the entire ray is obscured by a mesh protrusion. If mesh protrusion obscuration is determined, it indicates that the target's initial spatial intersection point is in the optical blind zone. The system then searches for nearby visible mesh nodes on the surface of the 3D triangular mesh model to update them as the target's initial spatial intersection point. Specifically, with the obscured target's initial spatial intersection point as the geometric center, a breadth-first search is performed on the 3D triangular mesh model according to the preset topological hierarchy using the topological adjacency matrix. The system extracts a candidate grid node set, with a preset topology level recommended to be 3 to 10 layers, preferably 5 layers to cover the area of ​​at least one borehole spacing. Then, the 3D coordinates of the obtained candidate grid node set are re-substituted into the aforementioned occlusion determination logic for spatial cross-validation. Invalid nodes that are still occluded are eliminated, and valid visible nodes are retained. To select the optimal alternative landing point, it is required not only to be close but also to have a smooth rock surface for easy drilling. The system calculates the mutation value of the normal vector of the local tangent plane where each valid visible node is located and constructs a variation of the optimization evaluation function based on multi-objective optimization theory.

[0032] in To achieve a comprehensive and optimal score, the value is dimensionless; the smaller the value, the better. The distance from the current candidate node to the initial spatial intersection point of the original target is the surface geodesic distance, in millimeters. This is a preset standard hole spacing constant, in millimeters. The distance penalty weight is dimensionless and is recommended to be set to 0.6. This is the normal smoothness weight, dimensionless, and a recommended value of 0.4. This is the local unit normal vector of the current candidate node, which is dimensionless. The unit normal vector of the original target node is dimensionless. The degree of normal deflection is quantified by the cosine value of the dot product; the necessity of this variant function lies in transforming the single "proximity principle" into a "dual constraint of distance and flatness," avoiding the allocation of alternative hole positions to the slope edge. Based on this formula, a comprehensive optimization score is selected. The lowest (i.e., the normal vector mutation value is less than the preset angle threshold and the distance from the geodesic line of the target initial spatial intersection point is the shortest) reference visible node is assigned its three-dimensional coordinates to the target initial spatial intersection point to complete the update of the current node; Conversely, if it is initially determined that there is no mesh protrusion occlusion, the original point position is directly retained. Then the system calculates the three-dimensional topological distance between the target's initial spatial intersection point and its adjacent spatial intersection points. The adjacent spatial intersection point refers to the initial spatial intersection point corresponding to the first borehole pixel point in the two-dimensional standard hole mesh image, and the initial spatial intersection point corresponding to the target's initial spatial intersection point in the two-dimensional image space. Traditional calculations based on straight-line distances severely underestimate the actual hole spacing on uneven surfaces, leading to failure in the transmission of blasting stress waves. Therefore, this invention extracts all grid edge connectivity paths connecting the initial spatial intersection point of the target with adjacent spatial intersection points in a three-dimensional triangular mesh model, and calculates the cumulative edge length along the mesh surface for each grid edge connectivity path. To incorporate the influence of rock surface micro-roughness on drill rod placement, this embodiment modifies and superimposes the classical graph theory edge length weight formula to construct a single-edge weight formula with surface roughness penalty:

[0033] in The corrected equivalent length of one side of the mesh, in millimeters. This represents the actual Euclidean physical length of the mesh edge in three-dimensional space, in millimeters. and These are the unit normal vectors of two adjacent triangular faces sharing the same grid edge, and are dimensionless. The roughness sensitivity coefficient is a preset dimensionless value, with a recommended value of 1.5. This variant formula allows for a higher distance penalty when traversing paths with jagged protrusions, thus prompting subsequent pathfinding algorithms to automatically avoid these small obstacles. The Dijkstra algorithm is then applied to accumulate the values ​​across all edge-connected paths in the mesh. The shortest cumulative side length value is selected as the geodesic distance data between two points, and this geodesic distance data is precisely set as the three-dimensional topological distance between the current borehole positions.

[0034] Based on the distortion ratio between the 3D topological distance and the preset standard aperture distance, mesh relaxation iteration is performed to drive the initial spatial intersection of the target to slide on the surface of the 3D triangular mesh model to an equilibrium position where the distortion ratio is lower than the preset distortion threshold, thereby generating an optimized set of 3D target aperture positions.

[0035] Preferably, mesh relaxation iteration is performed based on the distortion ratio between the 3D topological distance and the preset standard aperture distance, driving the initial spatial intersection point of the target to slide on the surface of the 3D triangular mesh model to an equilibrium position where the distortion ratio is lower than a preset distortion threshold, including: Calculate the absolute value of the difference between the three-dimensional topological distance and the preset standard hole spacing; The initial distortion ratio is calculated by dividing the absolute value of the difference by the preset standard hole spacing. If the initial distortion ratio is greater than the preset distortion threshold, a virtual physical system model is established with the initial spatial intersection of the target as an independent mass point and the preset standard hole spacing as the natural length of the spring. The absolute value of the difference is transformed into a virtual tangential force vector on the mesh surface acting at the initial spatial intersection point of the target, according to Hooke's Law. The offset of a single sliding position is calculated by multiplying the tangential virtual force vector on the mesh surface by the preset iteration step size parameter.

[0036] Preferably, the optimized three-dimensional target pore location set is generated, including: The sliding intermediate node is obtained by superimposing the single sliding position offset onto the current coordinate parameters of the target initial spatial intersection point; A projection operation perpendicular to the local mesh surface is performed on the sliding intermediate node to make it re-fit onto the surface of the 3D triangular mesh model, thus completing a single iteration of sliding; Repeatedly calculate the tangential virtual force vector of the virtual physical system model and perform a single iterative sliding until the spatial displacement between two consecutive sliding intermediate nodes is lower than the preset stopping threshold, at which point the mesh relaxation iteration is exited. After exiting the mesh relaxation iteration, the sliding intermediate state nodes are extracted as optimized 3D target holes and merged to generate an optimized 3D target hole set.

[0037] In this embodiment of the invention, after obtaining the geodesic distance, a mesh relaxation iteration is performed based on the distortion ratio between the three-dimensional topological distance and the preset standard aperture distance. This drives the initial spatial intersection point of the target to slide on the surface of the three-dimensional triangular mesh model to an equilibrium position where the distortion ratio is lower than the preset distortion threshold, generating an optimized set of three-dimensional target aperture locations. The specific process is as follows: First, the previously obtained three-dimensional topological distance is calculated. With respect to the pre-set standard hole spacing of the blasting design absolute value of the difference ; Next, divide the absolute value of the difference by the preset standard hole spacing. Calculate the initial distortion ratio The initial distortion ratio characterizes the relative stretching or compression of the aperture spacing due to surface undulations. The system determines whether the initial distortion ratio is greater than a preset distortion threshold, which is recommended to be set to an adjustable range of 0.03 to 0.08. If the initial distortion ratio is greater than the preset distortion threshold, the system establishes a system with the target initial spatial intersection point as an independent mass point and a preset standard aperture distance. The system is modeled as a virtual physical system based on the natural length of a spring. Subsequently, a manifold space variation is derived based on the classical Hooke's law, transforming the absolute value of the difference into a virtual tangential force vector acting on the mesh surface at the initial spatial intersection point of the target. The specific formula is as follows:

[0038] in The equivalent tangential virtual force vector is represented by spatial displacement, with units of millimeters; K is the preset system spring stiffness coefficient, a dimensionless coefficient used to simulate the basic strength of virtual attraction or repulsion, and is recommended to be between 0.3 and 0.7, preferably 0.5; is the symbol extraction function, dimensionless, used to determine whether the actual distance is greater than or less than the standard hole spacing, thereby determining the direction of the force, whether it is pulling inward or repelling outward; I is a 3×3 identity matrix, dimensionless; The unit normal vector of the triangular facet containing the initial spatial intersection point of the target, matrix combination This constitutes an orthogonal tangential projection operator, which is necessary because it can forcibly filter out the components of attraction or repulsion in the direction of normal to the grid surface, ensuring that the sliding tendency is always tangent to the rock surface. is a dimensionless tangential unit direction vector pointing along the shortest geodesic path to the intersection of adjacent spatial points; Using the above formula, the three-dimensional driving vector that is completely attached to the current rock surface undulations can be directly calculated, and then the tangential virtual force vector is... With preset iteration step size parameters Multiply to calculate the position offset of a single slide. ; Based on this, the system superimposes the single sliding position offset onto the current absolute coordinate parameters of the target's initial spatial intersection point. Above, we obtain the sliding intermediate state node after the initial translation. Considering that the sliding operation only occurs within the tangent plane of the local discrete patch, large-step sliding can easily lead to... Detached from the piecewise linear curved shell, the system then performs a projection operation on the sliding intermediate node perpendicular to the local mesh surface, i.e., along the negative normal direction. A ray is fired downwards to search for the latest physical intersection with the 3D triangular mesh model, so that it is absolutely reattached to the surface of the 3D triangular mesh model, thus completing a rigorous single-iteration sliding operation; The system iterates through the updated 3D topological distance of nodes and repeatedly calculates the tangential virtual force vector of the virtual physical system model to perform a single iterative sliding operation until the absolute spatial displacement between intermediate nodes in two consecutive sliding operations is reached. When the value falls below the preset stopping threshold, the mesh relaxation iteration is exited. Finally, the system extracts the final sliding intermediate state node after exiting the mesh relaxation iteration as the optimized 3D target hole location and merges all solved hole locations to generate an optimized 3D target hole location set.

[0039] Extract the 3D coordinate data of the optimized 3D target aperture set, and construct the reverse perspective transformation matrix based on the optical parameters of the projection device; A pre-distorted 2D projection image is generated by performing dimensionality reduction and remapping on 3D coordinate data using a reverse perspective transformation matrix, including: Extract the optical intrinsic parameter matrix of the projection device and the extrinsic parameter matrix based on the spatial coordinates of the projection center; Construct the reverse perspective transformation matrix by multiplying the intrinsic optical parameter matrix and the extrinsic optical parameter matrix; The inverse perspective transformation matrix is ​​used to perform matrix multiplication dimensionality reduction calculation on each three-dimensional coordinate data in the optimized three-dimensional target hole position set to map out the corresponding distorted two-dimensional pixel position coordinates. Extract the aperture shape and size parameters of the two-dimensional standard mesh image, re-render and draw them onto the coordinates of each distorted two-dimensional pixel point to generate a pre-distorted two-dimensional projection image.

[0040] Output and control the projection device to project the pre-distorted two-dimensional projection image onto the physical cross-section.

[0041] In this embodiment of the invention, the optical intrinsic parameter matrix of the projection device is first extracted from the underlying hardware memory. and the extrinsic parameter matrix based on the spatial coordinates of the projection center Since the classic perspective projection model will cause the light spot to be abruptly magnified and torn due to the drastic change in depth when facing a large angle face, a reverse perspective mapping system with dynamic scale scaling constraints was constructed. First, the system will use the optical intrinsic parameter matrix. With extrinsic matrix Multiply to construct a reverse perspective transformation matrix, and use this reverse perspective transformation matrix to apply the absolute 3D coordinate data of each target hole in the optimized 3D target hole set. (Units are all millimeters) For homogeneous matrix multiplication dimensionality reduction calculation, the variant mapping formula is:

[0042] Where the extrinsic parameter matrix It is a 3×4 transformation matrix, responsible for transforming points in the world coordinate system to the optical coordinate system of the projector; The intrinsic parameter matrix is ​​3×3, and its internal diagonal elements are... and These represent the equivalent pixel focal lengths of the projector's X and Y axes, respectively, with pixels as the unit of measurement. is the depth scale factor in the homogeneous coordinate system, which is equal to the physical depth of the 3D point along the Z-axis in the projector coordinate system, with the dimension of millimeters; After obtaining the left-hand vector through matrix multiplication, divide the first and last terms by the depth scale factor. This allows us to escape the homogeneous space and calculate the base coordinates of the corresponding distorted two-dimensional pixel points. Its dimension is pixels; Furthermore, to ensure that the light spot size of each hole projected onto the uneven rock surface is uniform and to avoid difficulties in visual identification by workers, the system extracts the aperture shape and size parameters of the two-dimensional standard hole mesh image and introduces an adaptive scale compensation formula for the light spot based on the surface normal and depth:

[0043] in The compression semi-axis length of the pre-distorted elliptical spot drawn onto the projected image for final rendering, in pixels, along the tilt gradient direction; This is the preset standard physical hole diameter size, in millimeters. It is recommended to set it to the range of 30 mm to 50 mm, preferably 40 mm to match the standard drill bit size. for and The arithmetic mean of the units, with pixels; It is the cosine of the angle between the direction of the projected optical axis ray and the local normal vector of the three-dimensional mesh surface where the target hole is located.

[0044] To address the objective phenomenon in actual optical projection where oblique projection causes the physical light spot to be unidirectionally elongated (diffuse) into an ellipse, this system generates a pre-distorted elliptical light spot for reverse compensation during 2D image rendering: within the imaging plane, the uncompressible semi-axis length perpendicular to the rock surface's slope gradient remains constant. The length of the compression semi-axis along the inclined gradient direction; Then multiply by the above formula Pre-compression is performed. When this pre-compressed two-dimensional ellipse is projected onto the inclined rock surface, it is stretched and magnified in one direction by physical oblique projection, which just cancels out the pre-compression amount, restoring it to an approximately uniform standard circular light spot on the cross-section.

[0045] In addition, to prevent extreme large tilt angles (grazing angles close to 90 degrees) from causing The light spot's shape tends towards zero, causing it to collapse or disappear. A preset lower cutoff threshold is set, preferably between 0.15 and 0.2.

[0046] Based on the calculated coordinates of each distorted two-dimensional pixel With the corresponding spot pixel radius The system re-renders and draws the pre-distorted 2D projection image in the memory buffer. In order to eliminate the diffraction interference of the stepped jagged edges generated by pixel drawing on the optical projection, a Gaussian kernel function is used to perform anti-aliasing smoothing filtering on the edge of the light spot with a preset intensity. The preset parameter of the Gaussian kernel size is recommended to be set to an odd matrix range of 3×3 to 7×7, preferably a 5×5 matrix to balance the rendering smoothness and the computational load. Finally, the system outputs the pre-distorted 2D projection image through the HDMI / DP video bus interface and drives the control projection device to project it onto the actual physical cross-section rock layer. Utilizing the optical cancellation characteristics of three-dimensional spatial distortion and two-dimensional image pre-distortion, the blast hole mesh arrangement is naturally presented on the concave and convex working face.

[0047] It should be noted that the interval and threshold sizes are set for ease of comparison. The size of the threshold depends on the amount of sample data and the base number set by those skilled in the art for each set of sample data, as long as it does not affect the proportional relationship between the parameter and the quantized value. Furthermore, the above formulas are all dimensionless calculations, and the formulas are derived from software simulations using a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0048] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0049] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

[0050] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for locating boreholes in cross-sections based on central projection, characterized in that, include: Acquire the 3D point cloud data of the cross section and convert it into a 3D triangular mesh model; Obtain the three-dimensional spatial coordinates of the projection center and the two-dimensional standard mesh image to be projected; An initial projection ray is emitted from the projection center to each first borehole pixel in the two-dimensional standard aperture mesh image, and the initial spatial intersection set of the initial projection ray and the three-dimensional triangular mesh model is calculated. Iterate through the initial set of spatial intersection points and enter the adaptive correction loop. The adaptive correction loop includes: Construct a connection vector from the projection center to the initial spatial intersection point of the target, and extract the depth buffer data on the connection vector to determine whether there is mesh protrusion occlusion; If there are mesh protrusions that obstruct the view, search for nearby visible mesh nodes on the surface of the 3D triangular mesh model and update them as the initial spatial intersection points of the target. If there is no mesh protrusion occlusion, calculate the 3D topological distance between the initial spatial intersection of the target and its adjacent spatial intersections; Based on the distortion ratio between the 3D topological distance and the preset standard aperture distance, mesh relaxation iteration is performed to drive the initial spatial intersection of the target to slide on the surface of the 3D triangular mesh model to an equilibrium position where the distortion ratio is lower than the preset distortion threshold, thereby generating an optimized set of 3D target aperture positions. Extract the 3D coordinate data of the optimized 3D target aperture set, and construct the reverse perspective transformation matrix based on the optical parameters of the projection device; A pre-distorted two-dimensional projection image is generated by performing dimension reduction and remapping on the three-dimensional coordinate data using the inverse perspective transformation matrix. Output and control the projection device to project the pre-distorted two-dimensional projection image onto the physical cross-section.

2. The cross-sectional borehole positioning method based on central projection according to claim 1, characterized in that, Construct a connection vector from the projection center to the initial spatial intersection point of the target, and extract the depth buffer data on the connection vector to determine whether there is mesh occlusion, including: A sequence of sampling points is obtained by discretizing the three-dimensional coordinates along the connecting vector at a preset step size; Extract the local depth value of each sampling point in the projection direction of the 3D triangular mesh model to construct depth buffer data; Compare the local depth value with the linear distance from the sampling point to the projection center; If the local depth value of any sampling point is less than the straight-line distance and the difference exceeds the preset tolerance, it is determined that there is a mesh protrusion occlusion.

3. The cross-sectional borehole positioning method based on central projection according to claim 1, characterized in that, If mesh protrusions cause occlusion, search for nearby visible mesh nodes on the surface of the 3D triangular mesh model and update them as the initial spatial intersection points of the target, including: Using the initial spatial intersection point of the target as the geometric center, a breadth-first search is performed outward on the three-dimensional triangular mesh model according to the preset topological hierarchy to extract the set of candidate mesh nodes; Spatial cross-validation is performed between the candidate grid node set and the depth buffer data to remove invalid nodes that are occluded and retain valid visible nodes; Calculate the abrupt change value of the normal vector of the local tangent plane where each valid visible node is located; Select the benchmark visible nodes whose normal vector mutation value is less than the preset angle threshold and whose distance from the geodesic line of the initial spatial intersection point of the target is the shortest. The current node is updated by assigning the 3D coordinates of the reference visible node to the initial spatial intersection point of the target.

4. The cross-sectional borehole positioning method based on central projection according to claim 1, characterized in that, Calculate the three-dimensional topological distance between the initial spatial intersection point of the target and its adjacent spatial intersection points, including: Extract all mesh edge connectivity paths connecting the initial spatial intersection point of the target and its adjacent spatial intersection points from the 3D triangular mesh model; Calculate the cumulative edge length along the mesh surface for each connected path along the mesh edge; Dijkstra's algorithm is used to select the shortest cumulative edge length value from all grid edge-connected paths as the geodesic distance data between two points. Set the geodesic distance data as a three-dimensional topological distance.

5. The cross-sectional borehole positioning method based on central projection according to claim 1, characterized in that, Mesh relaxation iterations are performed based on the distortion ratio between the 3D topological distance and the preset standard aperture distance, driving the initial spatial intersection point of the target to slide on the surface of the 3D triangular mesh model to an equilibrium position where the distortion ratio is lower than the preset distortion threshold, including: Calculate the absolute value of the difference between the three-dimensional topological distance and the preset standard hole spacing; The initial distortion ratio is calculated by dividing the absolute value of the difference by the preset standard hole spacing. If the initial distortion ratio is greater than the preset distortion threshold, a virtual physical system model is established with the initial spatial intersection of the target as an independent mass point and the preset standard hole spacing as the natural length of the spring. The absolute value of the difference is transformed into a virtual tangential force vector on the mesh surface acting on the initial spatial intersection point of the target, according to Hooke's Law. The offset of a single sliding position is calculated by multiplying the tangential virtual force vector on the mesh surface by the preset iteration step size parameter.

6. The cross-sectional borehole positioning method based on central projection according to claim 5, characterized in that, Generate an optimized set of three-dimensional target pore locations, including: The sliding intermediate node is obtained by superimposing the single sliding position offset onto the current coordinate parameters of the target initial spatial intersection point; A projection operation perpendicular to the local mesh surface is performed on the sliding intermediate node to make it re-fit onto the surface of the 3D triangular mesh model, thus completing a single iteration of sliding; Repeatedly calculate the tangential virtual force vector of the virtual physical system model and perform a single iterative sliding until the spatial displacement between two consecutive sliding intermediate nodes is lower than the preset stopping threshold, at which point the mesh relaxation iteration is exited. After exiting the mesh relaxation iteration, the sliding intermediate state nodes are extracted as optimized 3D target holes and merged to generate an optimized 3D target hole set.

7. The cross-sectional borehole positioning method based on central projection according to claim 1, characterized in that, A pre-distorted 2D projection image is generated by performing dimensionality reduction and remapping on 3D coordinate data using a reverse perspective transformation matrix, including: Extract the optical intrinsic parameter matrix of the projection device and the extrinsic parameter matrix based on the spatial coordinates of the projection center; Construct the reverse perspective transformation matrix by multiplying the intrinsic optical parameter matrix and the extrinsic optical parameter matrix; The inverse perspective transformation matrix is ​​used to perform matrix multiplication dimensionality reduction calculation on each three-dimensional coordinate data in the optimized three-dimensional target hole position set to map out the corresponding distorted two-dimensional pixel position coordinates. Extract the aperture shape and size parameters of the two-dimensional standard mesh image, re-render and draw them onto the coordinates of each distorted two-dimensional pixel point to generate a pre-distorted two-dimensional projection image.

8. The method for locating boreholes in cross-sections based on central projection according to claim 1, characterized in that, Before traversing the initial set of spatial intersection points and entering the adaptive correction loop, the method also includes a step of perceptually extracting cross-sectional step features from the 3D triangular mesh model: Calculate the local normal vector parameters of each mesh face in a 3D triangular mesh model; Compare the local normal angle values ​​of two adjacent mesh faces that share the same mesh edge; When the angle between local normal vectors is greater than the preset cross-sectional abrupt change threshold, the two adjacent grid surfaces are marked as step fracture surface features in the dataset and stored in the obstacle avoidance prohibited area spatial library. A boundary penalty function is introduced when performing mesh relaxation iterations.

9. The cross-sectional borehole positioning method based on central projection according to claim 8, characterized in that, A boundary penalty function is introduced during mesh relaxation iteration, including: Monitor the sliding trajectory coordinates of the initial spatial intersection point of the target on the surface of the three-dimensional triangular mesh model; Determine whether the coordinates of the sliding trajectory fall within the boundary influence range of the obstacle avoidance prohibited area space library; If it falls within the boundary influence range, the boundary penalty function is triggered to generate a penalty repulsion force vector that is reversed along the feature normal of the step fracture surface; The penalty repulsion force vector is superimposed onto the tangential virtual force vector of the current mesh surface to change the target's sliding direction.

10. The method for locating boreholes in cross-sections based on central projection according to claim 1, characterized in that, Before acquiring the 3D point cloud data of the cross-section and converting it into a 3D triangular mesh model, the method also includes: The drive line structured light sensor emits a scanning beam array toward the cross-section to be projected and receives the image of the beam deformation after reflection. An initial dense point set is calculated by performing a stereo matching operation based on epipolar geometric constraints on the beam deformation image; A voxel downsampling filtering algorithm is applied to spatially mesh and sparse the density of the initial dense point set to remove redundant and overlapping noise points and generate a simplified point set data. The simplified point set data is input into the pre-established projector base coordinate system to perform a three-dimensional affine transformation alignment operation, unifying the spatial reference frame and outputting three-dimensional point cloud data.

Images