Pipeline automatic reverse modeling method based on point cloud data

Abstract

The invention discloses an automatic pipeline reverse modeling method based on point cloud data. The method comprises the following steps: acquiring pipeline data by using a scanner; filtering the point cloud data, and extracting desktop plane, tool and pipeline point clouds by adopting a clustering segmentation method; performing segmented RANSAC cylinder fitting on the pipeline point cloud, and extracting pipe diameter information; extracting a plane pipeline center line by using least square polynomial fitting and quasi-uniform B spline curve fitting; carrying out secondary development by using CAD software, and carrying out automatic modeling according to pipeline characteristics; and performing deviation analysis on the model and the point cloud data, and if the deviation is out of tolerance, performing feature extraction again. Compared with the prior art, the point cloud feature extraction and automatic modeling method is provided, the full-automatic process from scanning to modeling is achieved, the model can be applied to follow-up design and machining assembly application of mainstream CAD software, the defect that a traditional point cloud reconstruction generation file lacks geometric information is overcome, an error analysis and feedback mechanism is provided, and the method has the advantages of being high in practicability and easy to popularize. And the model meets the precision requirement.

Description

Technical Field

[0001] The present application belongs to the field of reverse engineering and computer-aided design (CAD), and specifically is a method for automatic reverse modeling of pipelines based on point cloud data. Background Art

[0002] Aerospace pipelines are an important part of spacecraft, used to transport liquids, gases and other media. They play a key role in the fuel delivery, cooling system, environmental control and other aspects of spacecraft. Due to the special working environment of spacecraft in space, pipelines must have high reliability, high strength and high temperature resistance. Therefore, the design and manufacturing quality of aerospace pipelines directly affect the performance and safety of spacecraft.

[0003] Reverse modeling is a technology that generates digital models from existing objects. It achieves accurate replication of actual objects by measuring and reconstructing the geometric shape and characteristics of objects. In the field of aerospace, reverse modeling of pipelines is of great significance. Due to springback deformation and welding deformation in the manufacturing process, there are often differences between actual pipelines and theoretical design models. Through reverse modeling, a digital model consistent with the actual situation can be obtained to help with subsequent analysis, optimization and improvement, and improve the accuracy of design and the quality of manufacturing.

[0004] Point cloud model reconstruction usually generates STL files, which only contain triangular mesh data and lack geometric information such as points, lines, and surfaces, which are crucial in the process of mechanical processing and assembly. STL models lacking point, line, and surface information cannot be directly edited and used in CAD software (such as SolidWorks, CATIA, UG, etc.). Parametric reverse modeling extracts geometric feature information from point cloud data to generate a three-dimensional CAD model containing point, line, and surface information, which can be used for secondary design, processing, and assembly in mainstream CAD software to improve the practicality and accuracy of the model.

[0005] Traditional reverse modeling methods usually rely on manual measurement and modeling, which is time-consuming and prone to errors. With the development of computer technology and automation technology, automatic reverse modeling has become possible. Through automatic reverse modeling technology, modeling efficiency and accuracy can be significantly improved, manual intervention can be reduced, and costs can be reduced. For complex medium and large diameter plane pipelines in the aerospace field, automatic reverse modeling methods can quickly generate high-precision pipeline models to ensure the consistency between design and reality and meet the high standards and high precision requirements of spacecraft. Summary of the invention

[0006] The technical problem solved by this application is: to overcome the shortcomings of the existing technology and provide an automatic reverse modeling method for pipelines based on point cloud data. For aerospace pipelines, in the actual production process, due to bending rebound deformation and welding deformation, the actual pipelines usually have certain differences from the theoretical models, and their actual appearance deviates from the designed digital model. The design model cannot guide the actual processing or assembly process. This patent solves the above problems.

[0007] The technical solutions provided by this application are as follows:

[0008] A pipeline automatic reverse modeling method based on point cloud data, comprising:

[0009] S1. Fix the pipeline on the scanning workbench through the tooling, scan the pipeline fixed on the scanning workbench, and obtain the point cloud data of the pipeline in the scanning coordinate system;

[0010] S2, preprocessing and feature extraction of point cloud data to obtain desktop plane point cloud, tooling point cloud and pipeline point cloud;

[0011] S3, extracting the diameter of the pipeline according to the pipeline point cloud;

[0012] S4. According to the desktop plane point cloud, the desktop plane is fitted and obtained, and the desktop plane is used to cut the pipeline point cloud above its own axis to obtain the upper pipeline data;

[0013] S5. Obtain the pipeline centerline according to the upper pipeline data fitting;

[0014] S6. Import the pipeline centerline into the CAD software and generate a pipeline model based on the pipe diameter and wall thickness.

[0015] The S4 includes: S41, taking the desktop plane as the XOY plane and the direction perpendicular to the XOY plane as the Z axis, converting the coordinates of the pipeline point cloud in the scanning coordinate system into a coordinate system taking the desktop plane as the XOY plane to obtain pipeline data;

[0016] S42, according to the tooling size and the diameter of the pipeline, determine the Z coordinate of the pipeline axis, and the Z coordinate of the pipeline axis is marked as Z0;

[0017] S43, moving the XOY plane upward along the Z axis to exceed Z0, cutting the pipeline data, using straight-through filtering to obtain the upper pipeline data, and then projecting the upper pipeline data onto the XOY plane to obtain symmetrical planar pipeline data;

[0018] S44. A fitting polynomial curve equation is obtained according to the planar pipeline data, and the fitting polynomial curve equation is smoothed to be used as the pipeline centerline.

[0019] In the above S43, the XOY plane is moved upward along the Z axis to the position of Z0+1 / 6×tube diameter±1 / 20×tube diameter.

[0020] In S44, based on the planar pipeline data, the pipeline centerline is segmentally fitted using the least squares polynomial fitting method to obtain a fitting polynomial curve equation; and the fitting polynomial curve equation is smoothed using a quasi-uniform B-spline curve.

[0021] In S2, preprocessing the point cloud data includes: performing statistical filtering on the point cloud data to remove outliers and using voxel grid filtering to simplify and downsample the point cloud data; extracting features from the point cloud data includes: segmenting the point cloud data through Euclidean clustering to obtain desktop plane point cloud, tooling point cloud and pipeline point cloud.

[0022] The Euclidean clustering segmentation includes constructing a search structure of point cloud data using a KD tree, setting clustering parameters, which include a minimum cluster size, a maximum cluster size, and a cluster tolerance, performing Euclidean distance clustering on the point cloud data, and segmenting the point cloud data into multiple clusters.

[0023] The S6 includes:

[0024] Import the pipeline centerline into the CAD software and create points on the pipeline centerline;

[0025] Select a point and create a reference surface that is perpendicular to the centerline of the pipeline at that location;

[0026] On the reference surface, take the pipe diameter as the cross-section circle, set the wall thickness, and generate the pipeline model with the pipeline centerline as the trajectory.

[0027] After S6, there is step S7, which includes: performing distance deviation analysis on the generated pipeline model and the point cloud data, calculating the distance between the point cloud data and the pipeline model, and if the distance deviation exceeds the preset accuracy requirement, repeating steps S2-S7, adjusting the voxel grid filtering or cluster segmentation parameters when repeating the above steps, and preprocessing and feature extracting the point cloud data again until a pipeline model that meets the accuracy requirements is generated.

[0028] A pipeline automatic reverse modeling method based on point cloud data includes the following steps:

[0029] S1. Obtaining point cloud data: Use a 3D scanner to scan the pipeline fixed on the desktop to obtain the point cloud data of the pipeline.

[0030] S2. Preprocessing of point cloud data: Filter the point cloud data to remove noise, improve the quality of the point cloud data, and perform clustering and segmentation to extract the desktop plane, tooling point cloud and pipeline point cloud respectively.

[0031] S3. Pipeline feature extraction: (1) Extract pipe diameter information based on segmented cylinder fitting; (2) Extract the center line of the planar pipeline based on least squares polynomial fitting and quasi-uniform B-spline curve fitting.

[0032] S4. Based on the pipeline feature information extracted in S3, secondary development is carried out in combination with SOLIDWORKS to complete the automatic reverse modeling of the pipeline.

[0033] The distance deviation between the established model and the point cloud data is analyzed. If the deviation exceeds the deviation requirement, return to step S2 and process the point cloud data again.

[0034] In summary, this application at least includes the following beneficial technical effects:

[0035] The present invention provides a method for point cloud feature extraction and automatic parametric CAD modeling, realizing a fully automatic process from scanning to modeling. The model can be applied to subsequent design and processing and assembly applications of mainstream CAD software, overcoming the shortcomings of traditional point cloud reconstruction to generate STL files with lack of geometric information. At the same time, it has error analysis and feedback mechanisms to ensure that the model meets the accuracy requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 It is a schematic diagram of the workflow of the present invention;

[0037] Figure 2 It is a schematic diagram of a scanning workbench of the present invention;

[0038] Figure 3 This is a schematic diagram of pipeline point cloud extraction according to the present invention. DETAILED DESCRIPTION

[0039] In order to make the objectives, technical solutions and advantages of the present invention more clear, the embodiments disclosed in the present invention will be further described in detail below with reference to the accompanying drawings.

[0040] The present application embodiment discloses a pipeline automatic reverse modeling method based on point cloud data, such as Figure 1 As shown, including:

[0041] S1. Fix the pipeline on the scanning workbench through the tooling. The pipeline is a plane pipeline, that is, the axis of the pipeline is on a plane, such as Figure 2 As shown;

[0042] Get point cloud data: Use a 3D scanner to scan the pipeline fixed on the desktop to get the point cloud data of the pipeline. Make sure the pipeline is fixed firmly during the scanning process to avoid movement and errors during data acquisition. Use a high-precision 3D scanner to ensure the accuracy of the point cloud data. The pipeline is fixed, and due to the scanning limit, the scanner can only get the point cloud data of the upper half of the pipeline (the side of the pipeline facing the desktop cannot be scanned). Figure 3 shown.

[0043] S2. Preprocessing of point cloud data:

[0044] (1) Filter the point cloud data and use statistical outlier removal to remove outliers to ensure the purity of the point cloud data and improve the quality of the point cloud data. Use voxel grid filter to simplify and downsample the point cloud data to reduce the amount of data and improve computing efficiency.

[0045] (2) The Euclidean Cluster Extraction algorithm is used to cluster the point cloud data. The KD tree is used to construct the search structure of the point cloud data to improve the search efficiency. The clustering parameters, including the minimum cluster size (MinClusterSize), the maximum cluster size (MaxClusterSize) and the cluster tolerance (ClusterTolerance), are set to perform Euclidean distance clustering on the point cloud data. The point cloud data is divided into multiple clusters, each cluster representing an independent part. The desktop plane point cloud, tooling point cloud and pipeline point cloud are extracted respectively.

[0046] S3. Feature extraction of pipelines, including pipe diameter and centerline:

[0047] (1) Extracting pipe diameter: Perform segmented RANSAC cylindrical fitting on the extracted pipeline point cloud to extract the pipe diameter information.

[0048] (2) Extract center line

[0049] a. Point cloud spatial position transformation

[0050] Perform RANSAC plane fitting on the desktop point cloud to obtain the desktop plane; take the desktop plane as the XOY plane, the fitted plane normal vector P, the rotated plane normal vector Q, and the rotation axis Calculation formula: Rotation angle calculation formula: Calculate the Rodriguez rotation matrix and transform the pipeline point cloud into a spatial position.

[0051]

[0052] The desktop plane point cloud, tooling point cloud and pipeline point cloud obtained by scanning are all located in the scanning coordinate system; according to the above rotation axis and rotation angle to transform the pipeline point cloud into a coordinate system with the desktop plane as the XOY plane.

[0053] b. Pipeline point cloud projection:

[0054] According to the tooling size and the pipe diameter, determine the Z coordinate of the pipeline axis, which is marked as Z0;

[0055] The pipeline point cloud is incomplete data and only contains 1 / 2 pipeline data. First, the XOY plane is moved to the Z0+1 / 6 pipe diameter position, and the pipeline data is cut. Through filtering is used to obtain the upper 1 / 3 data of the pipeline. Then, the 1 / 3 data on the pipeline is projected onto the XOY plane to obtain symmetrical planar pipeline data. Voxel filtering is performed on the projected point cloud to achieve uniform distribution of the point cloud, reduce the amount of data, and improve calculation efficiency and fitting accuracy.

[0056] c. Least squares method to fit the center line: This paper adopts the least squares method to perform polynomial segmented fitting on the center line of the pipeline.

[0057] The pipeline data projected onto the XOY plane is P = [(x1, y1), (x2, y2), (x3, y3), ..., (x m ,y m )]

[0058] According to the pipeline data projected onto the XOY plane, the fitting polynomial curve equation is obtained as follows:

[0059] d. Quasi-uniform B-spline smoothing: The least square method of the quasi-uniform B-spline curve is used to smooth the fitting polynomial curve equation, which is used as the center line of the pipeline.

[0060] Quasi-uniform B-spline curve equation:

[0061] Cubic B-spline basis functions

[0062]

[0063] Obtain centerline point set information based on the pipeline centerline.

[0064] S4. Conduct secondary development of SolidWorks:

[0065] (1) Importing curve point set: Import the centerline point set information into the CAD software, InsertCurvefile (centerline file), and obtain the centerline curve;

[0066] (2) Create a point on the curve: Select the curve Extension.SelectByID2 (curve), create a point InsertReferencePoint (curve position setting)

[0067] (3) Create a reference surface: Select the point SelectByName (point), select the curve Extension.SelectByID2(), and create a reference surface InsertRefPlane()

[0068] (4) Create a sketch: Select the reference plane SelectByNmae(), insert the sketch InsertSketch(), create a circle CreateCircleByRadius, and exit the sketch InsertSketch().

[0069] (5) Generate pipeline model: Select the section sketch extension.SelectById2(), select the curve Extension.SelectByID2(), and set the wall thickness to generate the pipeline model InsertProtrusionSwept3().

[0070] (6) Automatically save the model and exit the program

[0071] The model is a parametric 3D CAD model with geometric information such as points, lines, and surfaces, which is convenient for subsequent design, processing, and assembly applications.

[0072] Analyze the distance deviation between the established model and the point cloud data

[0073] (1) Use point cloud data to calculate the distance between the generated CAD model and evaluate the accuracy and consistency of the model.

[0074] (2) If it is found that the distance deviation between the model and the point cloud data exceeds the preset accuracy requirement, return to step S2, adjust the voxel grid filtering or cluster segmentation parameters, and preprocess and extract features of the point cloud data again until a model that meets the accuracy requirements is generated.

[0075] The present invention provides a method for point cloud feature extraction and automatic parametric CAD modeling, realizing a fully automatic process from scanning to modeling. The model can be applied to subsequent design and processing and assembly applications of mainstream CAD software, overcoming the shortcomings of traditional point cloud reconstruction to generate STL files with lack of geometric information. At the same time, it has error analysis and feedback mechanisms to ensure that the model meets the accuracy requirements.

[0076] The contents not described in detail in this application specification belong to the common knowledge of those skilled in the art.

[0077] The present application is described in detail above in conjunction with specific implementation methods and exemplary examples, but these descriptions cannot be understood as limiting the present application. Those skilled in the art understand that, without departing from the spirit and scope of the present application, a variety of equivalent replacements, modifications or improvements can be made to the technical solution of the present application and its implementation methods, all of which fall within the scope of the present application. The scope of protection of the present application shall be subject to the attached claims.

Claims

1. A pipeline automatic reverse modeling method based on point cloud data, characterized in that: include: S1. Fix the pipeline on the scanning workbench through the tooling, scan the pipeline fixed on the scanning workbench, and obtain the point cloud data of the pipeline in the scanning coordinate system; S2, preprocessing and feature extraction of point cloud data to obtain desktop plane point cloud, tooling point cloud and pipeline point cloud; S3, extracting the diameter of the pipeline according to the pipeline point cloud; S4. According to the desktop plane point cloud, the desktop plane is fitted and obtained, and the desktop plane is used to cut the pipeline point cloud above its own axis to obtain the upper pipeline data; S5. Obtain the pipeline centerline according to the upper pipeline data fitting; S6. Import the pipeline centerline into the CAD software and generate a pipeline model based on the pipe diameter and wall thickness.

2. The method for automatic reverse modeling of pipelines based on point cloud data according to claim 1, characterized in that: The S4 includes: S41, taking the desktop plane as the XOY plane and the direction perpendicular to the XOY plane as the Z axis, transforming the coordinates of the pipeline point cloud in the scanning coordinate system into a coordinate system taking the desktop plane as the XOY plane, to obtain pipeline data; S42, according to the tooling size and the diameter of the pipeline, determine the Z coordinate of the pipeline axis, and the Z coordinate of the pipeline axis is marked as Z0; S43, moving the XOY plane upward along the Z axis to exceed Z0, cutting the pipeline data, using straight-through filtering to obtain the upper pipeline data, and then projecting the upper pipeline data onto the XOY plane to obtain symmetrical planar pipeline data; S44. A fitting polynomial curve equation is obtained according to the planar pipeline data, and the fitting polynomial curve equation is smoothed to be used as the pipeline centerline.

3. The method for automatic reverse modeling of pipelines based on point cloud data according to claim 2, characterized in that: In the above S43, the XOY plane is moved upward along the Z axis to the position of Z0+1 / 6×tube diameter±1 / 20×tube diameter.

4. The method for automatic reverse modeling of pipelines based on point cloud data according to claim 2, characterized in that: In S44, based on the planar pipeline data, the pipeline centerline is segmentally fitted using the least squares polynomial fitting method to obtain a fitting polynomial curve equation; and the fitting polynomial curve equation is smoothed using a quasi-uniform B-spline curve.

5. The method for automatic reverse modeling of pipelines based on point cloud data according to claim 1, characterized in that: In S2, preprocessing the point cloud data includes: performing statistical filtering on the point cloud data to remove outliers and using voxel grid filtering to simplify and downsample the point cloud data; extracting features from the point cloud data includes: segmenting the point cloud data through Euclidean clustering to obtain desktop plane point cloud, tooling point cloud and pipeline point cloud.

6. The method for automatic reverse modeling of pipelines based on point cloud data according to claim 5, characterized in that: The Euclidean clustering segmentation includes constructing a search structure of point cloud data using a KD tree, setting clustering parameters, which include a minimum cluster size, a maximum cluster size, and a cluster tolerance, performing Euclidean distance clustering on the point cloud data, and segmenting the point cloud data into multiple clusters.

7. The method for automatic reverse modeling of pipelines based on point cloud data according to claim 1, characterized in that: The S6 includes: Import the pipeline centerline into the CAD software and create points on the pipeline centerline; Select a point and create a reference surface that is perpendicular to the centerline of the pipeline at that location; On the reference surface, take the pipe diameter as the cross-section circle, set the wall thickness, and generate the pipeline model with the pipeline centerline as the trajectory.

8. The method for automatic reverse modeling of pipelines based on point cloud data according to claim 1, characterized in that: After S6, there is step S7, which includes: performing distance deviation analysis on the generated pipeline model and the point cloud data, calculating the distance between the point cloud data and the pipeline model, and if the distance deviation exceeds the preset accuracy requirement, repeating steps S2-S7, adjusting the voxel grid filtering or cluster segmentation parameters when repeating the above steps, and preprocessing and feature extracting the point cloud data again until a pipeline model that meets the accuracy requirements is generated.

Images