Method for generating spatial structure pipe truss calculation model based on point cloud

Abstract

The invention relates to the technical field of structural engineering digital modeling, and discloses a method for generating a spatial structure pipe truss calculation model based on point clouds, which comprises the following steps: step 1, obtaining candidate circular pipe point clouds through point cloud processing; step 2, obtaining a closed surface; 3, obtaining any separation line; and 4, generating a space framework wireframe model and importing the space framework wireframe model into the SAP. According to the method, full-automatic conversion of the circular pipe truss point cloud, the closed curved surface, the separation line, the center line and the finite element wireframe model at any space angle is achieved, and SAP2000 is directly imported for structural calculation. The spatial structure pipe truss reading point cloud is converted into a calculation model capable of being used for structural analysis, the method is an important premise for achieving damage assessment, reconstruction and extension design and digital operation and maintenance, and the method has important value and significance.

Description

Technical Field

[0001] This invention relates to the field of digital modeling technology for structural engineering, and in particular to a method for generating a computational model of a spatial structural truss based on point clouds. Background Technology

[0002] In the digitalization of structural engineering, the problem of generating wireframes from point cloud models of beam-column structures in orthogonal space can be solved. However, in actual engineering, the point cloud angles of spatial structures are arbitrary and non-orthogonal. Therefore, it is necessary to propose a technology that can solve the problem of generating computational models of spatial structural trusses.

[0003] The existing technology has the following main shortcomings: It can only handle orthogonal beam-column systems and cannot adapt to circular tube trusses at arbitrary spatial angles.

[0004] Directly applying "centerline fitting" to the point cloud of a circular tube results in large errors and makes it difficult to handle noise in areas where adjacent tubes overlap or weld reinforcement is present.

[0005] It cannot automatically generate closed circular tube surfaces, and therefore cannot use surface normals for precise cutting and segmentation.

[0006] 4. The system lacks automatic deduplication, topology reconstruction, and API interfaces from "segment center point" to "finite element linear element", still requiring a lot of manual intervention. Summary of the Invention

[0007] The purpose of this invention is to provide a method for generating a spatial truss calculation model based on point cloud. Its main feature is to solve the problem of generating truss wireframes from point clouds of truss structures at arbitrary spatial angles. It is applicable to the fully automatic construction of a finite element calculation model from measured point clouds of circular truss structures arranged at arbitrary spatial angles, thereby solving the problems mentioned in the background art.

[0008] To achieve the above objectives, the present invention provides a method for generating a spatial structural truss computational model based on point clouds, comprising: Step 1: Obtain the point cloud of the candidate circular tube through point cloud processing; 1.1 Read point cloud: Read the information of the point cloud of the spatial structure truss, including the X, Y, and Z coordinates of the point cloud; 1.2 Denoising: A two-stage strategy of statistical outlier filtering (SOR) + radius outlier filtering (ROR) is adopted, with adaptive point spacing parameters; 1.3 Structuring: The point cloud is divided into "tubular clusters" using region growing. A KD-Tree is built for each cluster and the local cylindrical fitting residual is calculated. Clusters with residuals < δ are retained as candidate cylindrical point clouds. Step 2: Obtain the closed surface; 2.1 Generating triangular surfaces: Extract the boundary from the point cloud of the circular tube, and then use Poisson to reconstruct a triangular mesh circular tube with closed ends and a manifold shape; 2.2 Obtaining the normal from the surface: Calculate the weighted average normal of the annular domain at the vertices of the triangular mesh, and then perform mean-shift clustering along the axis to obtain the uniform outward-facing normal field of the entire tube; 2.3 Circular tube cutting: Establish a local coordinate system (u,v,w) along the normal field, with the w-axis aligned with the tube axis; generate inner and outer hyperboloids by offsetting in the w-direction, and obtain the precise end cutting line using the "surface-to-surface intersection" algorithm; Step 3: Obtain any separator line; 3.1 Obtaining the dividing line by cutting: Project the precise end cutting line obtained in step 2.3 onto the tube axis to obtain the circumferential dividing line; 3.2 Obtaining the center point from the circumferential dividing line: Calculate the geometric center of each circumferential dividing line and interpolate at equal intervals along the w-axis to generate a series of center points {Pi}; 3.3 Connecting into a straight line: Fitting the line {Pi} to obtain the center line L of a single circular tube, and recording the starting point SO and the ending point EO and their coordinates; Step 4: Generate the spatial frame wireframe model and import it into SAP; 4.1 Deduplication and Topology Reconstruction: A 3D R-tree is established, and all centerline endpoints are clustered according to the distance tolerance ε=0.02 m and merged into a global node set {Nk}; the centerline L of each single circular pipe is replaced with the corresponding node index to obtain a spatial frame wireframe with no duplicate nodes and no duplicate units. 4.2 API Import: Call SAP to obtain the start and end coordinates Pi (start coordinates: Pix, Piy, Piz) and Pj (start coordinates: Pjx, Pjy, Pjz) of any wireframe L, and create the following model in sequence: node: ret = SapModel.PointObj.AddCartesian(Pix, Piy, Piz, Pi, ); ret = SapModel.PointObj.AddCartesian(Pjx, Pjy, Pjz, Pj, ); Frame unit: ret = SapModel.FrameObj.AddByPoint(Pi, Pj, L); Save the model; 4.3 Output: Generates a *.sdb file, completing the creation of a spatial structural truss calculation model based on point cloud generation.

[0009] Preferably, the equal spacing in step 3.2 is a spacing of 0.08m to 1.5m.

[0010] Preferably, the equal spacing is 0.1m.

[0011] Preferably, the linear fitting of {Pi} in step 3.3 is performed using RANSAC fitting.

[0012] Preferably, the SAP in step 4 is SAP2000.

[0013] Preferably, the API mentioned in step 4.2 is SAP2000 OAPI.

[0014] Preferably, the offset mentioned in step 2.3 is the design wall thickness + 0.05 m.

[0015] Preferably, the boundary extraction of the circular tube point cloud in step 2.1 is performed using the "rolling ball method" α-shape extraction.

[0016] Compared with the prior art, the beneficial effects of the present invention are: For the first time, the entire process of "point cloud of circular tube truss at any spatial angle → SAP2000 calculation model" has been fully automated with zero human intervention. The strategy of "closed surface-normal cutting-center point segmentation" improves the centerline fitting accuracy by ≥3 times compared with the traditional method. By using R-tree deduplication and topology reconstruction, it can process ≥100,000 circular pipes and millions of nodes at once, with a processing time of <5 minutes; The system directly generates a *.sdb format structural calculation model for SAP2000 based on the API interface of the finite element software SAP2000, which can be seamlessly integrated with subsequent structural analysis and optimization design, significantly improving the efficiency of inspection, reinforcement, renovation and expansion. Attached Figure Description

[0017] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Please see Figure 1 The method of generating a spatial structure truss calculation model based on point cloud in this invention includes: Step 1: Obtain the point cloud of the candidate circular tube through point cloud processing.

[0020] 1.1 Reading Point Clouds: Read the information of the point cloud, including the X, Y, and Z coordinates of the point cloud; three-dimensional laser scanning and photogrammetry technology can be used to obtain the measured point cloud of the spatial structure pipe truss.

[0021] 1.2 Denoising: A two-stage strategy of statistical outlier filtering (SOR) + radius outlier filtering (ROR) is adopted, with adaptive point spacing parameters; 1.3 Structuring: The point cloud is divided into "tubular clusters" using region growing. A KD-Tree is built for each cluster and the local cylindrical fitting residual is calculated. Clusters with residuals < δ are retained as candidate cylindrical point clouds.

[0022] Step 2: Obtain the closed surface of the point cloud and finally obtain the cutting line of the circular tube.

[0023] 2.1 Generating triangular surfaces: The boundary of the circular tube point cloud is extracted using the "rolling ball method" α-shape, and then Poisson is used to reconstruct the triangular mesh circular tube with closed ends and a manifold. 2.2 Obtaining the normal from the surface: Calculate the weighted average normal of the annular domain at the vertices of the triangular mesh, and then perform mean-shift clustering along the axis to obtain the uniform outward-facing normal field of the entire tube; 2.3 Circular tube cutting: Establish a local coordinate system (u,v,w) along the normal field, with the w axis aligned with the tube axis; offset in the w direction by "design wall thickness + 0.05 m" to generate inner and outer hyperboloids, and use the "surface-surface intersection" algorithm to obtain the accurate cutting line at the end of the circular tube.

[0024] Step 3: Obtain the circumferential dividing line, and finally obtain the center line L of any circular tube; 3.1 Obtaining the dividing line by cutting: Project the precise end cutting line obtained in step 2.3 onto the tube axis to obtain the circumferential dividing line; 3.2 Obtaining the center point from the circumferential dividing line: Calculate the geometric center of each circumferential dividing line and interpolate at equal intervals along the w-axis to generate a series of center points {Pi}; the equal interval can be a suitable interval from 0.08m to 1.5m, and the preferred embodiment of the present invention is 0.1m.

[0025] 3.3 Connecting into a straight line: Apply RANSAC straight line fitting to {Pi} to obtain the center line L of a single circular pipe, and record the starting point SO and the ending point EO and their coordinates.

[0026] Step 4: Generate the spatial truss wireframe and import it into SAP2000.

[0027] 4.1 Deduplication and Topology Reconstruction: A 3D R-tree is established, and all centerline starting points are clustered according to the distance tolerance ε=0.02m and merged into a global node set {Nk}; the centerline L of each single circular tube is replaced with the corresponding node index to obtain the spatial tube truss wireframe LL (containing nodes and wireframe) with no duplicate nodes and no duplicate elements. 4.2 API Import: Using SAP2000 OAPI, for any spatial tubular truss wireframe LL, obtain the start and end coordinates Pi (start coordinates: Pix, Piy, Piz) and Pj (start coordinates: Pjx, Pjy, Pjz) of the wireframe, and then create the following spatial structure tubular truss wireframe model in SAP2000 software: a) Node: ret = SapModel.PointObj.AddCartesian(Pix, Piy, Piz, Pi, ); ret = SapModel.PointObj.AddCartesian(Pjx, Pjy, Pjz, Pj, ); b) Frame unit: ret = SapModel.FrameObj.AddByPoint(Pi, Pj, L); c) Save the wireframe model of the spatial structure tubular truss; 4.3 Output: Generates a *.sdb file, completing the creation of the computational model.

[0028] As can be seen above, the method of generating spatial structural truss calculation models based on point clouds in this invention realizes the fully automatic conversion of point clouds of circular trusses at arbitrary spatial angles into closed curved surfaces, dividing lines, center lines, and finite element wireframe models, and directly imports them into SAP2000 for structural calculations.

[0029] Transforming point clouds into computational models that can be used for structural analysis is an important prerequisite for damage assessment, renovation and expansion design, and digital operation and maintenance. This invention has significant value and importance.

[0030] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A method for generating a spatial structure tube truss computational model based on a point cloud, characterized in that, Comprise: Step 1: Obtain candidate circular tube point cloud by point cloud processing; 1.1 Read point cloud: read the information of the spatial structure tube truss point cloud, including the X, Y, Z coordinates of the point cloud; 1.2 De-noising: adopt two-stage strategy of statistical outlier filtering (SOR) + radius outlier filtering (ROR), and the parameter is adaptive point spacing; 1.3 Structuring: divide the point cloud into "tubular clusters" by region growing, establish KD-Tree for each cluster and calculate local cylindrical fitting residual, and keep the cluster with residual <δ as the candidate circular tube point cloud; Step 2: Obtain the closed surface of the point cloud, and finally get the cutting line of the circular tube; 2.1 Generate triangular surface: extract the boundary of the circular tube point cloud, and then generate a triangular mesh circular tube with closed ends and manifold by Poisson reconstruction; 2.2 Obtain normal from surface: calculate the ring area weighted average normal of the triangular mesh vertex, and then do mean shift clustering along the axial direction to get a consistent outward tube normal field; 2.3 Circular tube cutting: establish a local coordinate system (u, v, w) along the normal field, with w axis aligned with the tube axis; generate inner and outer surfaces in w direction according to offset, and obtain the accurate circular tube end cutting line by "surface-surface intersection line" algorithm; Step 3: Obtain any separation line and finally get the circular tube center line L; 3.1 Cutting to get the separation line: project the accurate circular tube end cutting line obtained in step 2.3 to the tube axis to get the ring separation line; 3.2 Get center point from ring separation line: calculate the geometric center of each ring separation line, and interpolate along the w axis to generate a series of center points {Pi}; 3.3 Connect into a straight line: fit a straight line to {Pi} to get a single circular tube center line L, and record the start point SO and end point EO and their coordinates; Step 4: Generate a space framework wireframe model and import SAP; 4.1 De-duplication and topology reconstruction: establish a three-dimensional R-tree, cluster all center line endpoints according to distance tolerance ε=0.02 m, and merge them into a global node set {Nk}; replace each single circular tube center line L with the corresponding node index to obtain a space framework wireframe without duplicate nodes and units; 4.2 API import: call SAP, get the start and end coordinates of the wireframe Pi (start coordinates: Pix, Piy, Piz) and Pj (start coordinates: Pjx, Pjy, Pjz) for any wireframe L, and create the following models in turn: Node: ret = SapModel.PointObj.AddCartesian(Pix, Piy, Piz, Pi, ); ret = SapModel.PointObj.AddCartesian(Pjx, Pjy, Pjz, Pj, ); Frame unit: ret = SapModel.FrameObj.AddByPoint(Pi, Pj, L); Save the model; 4.3 Output: generate *.sdb file, complete the creation of spatial structure tube truss calculation model based on point cloud.

2. The method for generating a spatial structure tube truss computational model based on a point cloud according to claim 1, characterized in that, The equidistant spacing described in step 3.2 is a spacing of 0.08m to 1.5m.

3. The method for generating a spatial structural tube-truss computational model based on a point cloud according to claim 2, characterized in that, The equidistant spacing described is a spacing of 0.1m.

4. The method for generating a spatial structural tube-truss computational model based on point clouds according to claim 1, wherein, The linear fitting to {Pi} described in step 3.3 is a RANSAC fitting.

5. The method for generating a spatial structural tube-truss computational model based on point clouds according to claim 1, wherein, The SAP in the step 4 is SAP2000.

6. The method for generating a spatial structural tube-truss computational model based on point clouds according to claim 1, wherein, The API described in step 4.2 is SAP2000 OAPI.

7. The method for generating a spatial structural tube-truss computational model based on point clouds according to claim 1, wherein, The offset described in step 2.3 is a design wall thickness + 0.05m.

8. The method for generating a spatial structural tube-truss computational model based on point clouds according to claim 1, wherein, The boundary extraction for the circular tube point cloud described in step 2.1 is a "rolling ball method" alpha-shape extraction.

Images