Power equipment three-dimensional model modeling method based on laser point cloud data
Through the three-dimensional laser scanning system, a high-precision three-dimensional model of power equipment was constructed, which solved the problems of complicated modeling and poor modeling of complex objects in the existing technology, and achieved rapid and accurate three-dimensional construction Modular and efficient simulation performance.
Patent Information
- Application Number
- CN202410899070.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-05
- Publication Date
- 2024-12-03
AI Technical Summary
When modeling power equipment, the geometric model-based method is cumbersome and time-consuming, while the image-based method has poor modeling of complex objects, making it difficult to achieve high-precision three-dimensional model construction.
A three-dimensional laser scanning system is used to collect point cloud data, and a planar triangular network is constructed through data preprocessing, filtering and denoising, multi-site data registration and projection to build a planar triangular network.
It realizes the rapid and accurate acquisition of complex three-dimensional models of power equipment, improves modeling accuracy and simulation performance, and can effectively describe irregular corner points and subtle features.
Smart Images

Figure CN119068140A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of electric power equipment, and in particular relates to a method for modeling a three-dimensional model of electric power equipment based on laser point cloud data. Background Art
[0002] The three-dimensional model can present spatial information more directly. Compared with the two-dimensional digital model, the three-dimensional digital model provides a richer display space for spatial information, and can more realistically, objectively, accurately and realistically reflect the status of on-site power equipment and environment, further improve the interactive visual effect, and has a good sense of reality and presence. It can effectively assist staff to better familiarize themselves with power equipment, make quick and accurate judgments on specific problems, improve work efficiency, and ensure the stability of power grid operation.
[0003] Traditional modeling methods are divided into two types according to technical means and forms of expression, including geometric model-based modeling technology and image-based modeling technology. Among them, geometric model-based modeling technology is simple to operate, but the data collection and modeling process is time-consuming and labor-intensive, while image-based modeling technology has high requirements for data accuracy and is better for modeling regular objects, but is not suitable for complex objects.
[0004] Therefore, in order to solve the above problems, it is necessary to develop a three-dimensional modeling method for power equipment based on laser point cloud data. Summary of the invention
[0005] The purpose of the present invention is to overcome the shortcomings of the prior art and to provide a three-dimensional modeling method for power equipment based on laser point cloud data. The point cloud data is collected by a three-dimensional laser scanning system to achieve real-time, efficient and accurate data acquisition, effectively construct a complex and complete power equipment model, and improve modeling accuracy and simulation performance.
[0006] The object of the present invention is achieved by: a method for modeling a three-dimensional model of an electric power equipment based on laser point cloud data, comprising the following steps:
[0007] (1) Equipment data collection: Use a 3D laser scanning system to perform full coverage scanning of power equipment to obtain surface texture information and 3D spatial point cloud data of the power equipment;
[0008] (2) Data preprocessing: including data stitching and coordinate correction. When collecting data, at least three target points are set in adjacent areas to unify the point cloud data obtained by scanning at different angles at different sites into the same coordinate system. Then, based on the obtained geodetic coordinates, the stitched data is converted into geodetic coordinates.
[0009] (3) Data processing, including filtering and denoising, and multi-site data registration;
[0010] Among them, ① filtering and denoising processing includes filtering the point cloud data to remove noise points introduced by sensor errors and environmental interference factors; ② fusing and aligning the point cloud data obtained from multiple sites and performing data registration processing;
[0011] (4) 3D modeling: Based on the point cloud features, a three-dimensional spatial data model is constructed by projecting a plane triangulated network.
[0012] (5) Texture mapping: mapping texture information onto the model to form the real texture of the model.
[0013] Furthermore, the three-dimensional laser scanning system in step (1) adopts a three-dimensional laser scanner or an airborne laser radar.
[0014] Furthermore, the step (1) of collecting data using a three-dimensional laser scanner includes the following steps:
[0015] ① Set the target balls. The number of target balls should be greater than 3 and the target balls should be in different planes. The target balls should not be moved before the scanning is completed.
[0016] ② Scanning station setting: Set up the scanning station at the actual location of the power equipment to be modeled according to the actual situation, and avoid obstruction during the scanning station setting process;
[0017] ③Data acquisition: obtain point cloud data through laser scanner and collect texture mapping data.
[0018] Furthermore, the data content obtained by using the three-dimensional laser scanner in step (1) is as follows:
[0019] The vertical and horizontal angles of the laser beam emitted by the 3D laser scanner are obtained through the angle values of two continuously rotating mirrors with reflected pulsed lasers. The two angles of the laser beam and the detected distance are used to obtain the relative coordinates of each scanning point and the corresponding space of the 3D laser scanner, as well as the radiation intensity of the scanning point of the 3D laser scanner. The scanning point obtained by the 3D laser scanner is represented by (x, y, z, h). The formula for obtaining the coordinates (x, y, z) of the 3D laser scanning point by the 3D laser scanner is as follows:
[0020]
[0021] Where α and β represent the vertical and horizontal angles of the laser beam, respectively; L and h represent the slant distance from the 3D laser scanner to the scanning point and the reflection intensity, respectively.
[0022] Furthermore, the filtering and denoising process in step (3) specifically adopts median filtering, slides a fixed value window on the distance image, arranges the points in the window in order according to the Y value, and takes the midpoint value as the center point of the window.
[0023] Furthermore, the data registration process in step (3) aligns and splices the data of different sites, specifically by arranging control points with the same name on the surface of the power equipment to achieve the alignment and splicing of the point cloud data;
[0024] Assume that there are overlapping areas where the corresponding data points at different viewing angles are represented by A and B, and the subset {m i |m i ∈A, i = 1, 2, ..., N} is extracted from the data point set A, and the subset {m ′ i |m ′ i ∈B, i = 1, 2, ..., N} is extracted from the data point set B, and the two points correspond to each other. The point cloud registration is realized by obtaining the translation and rotation coordinate transformation relationship of the data points collected from each perspective; C and D represent the translation matrix and rotation matrix respectively. The essence of point cloud registration is to minimize the objective function, and the objective function that satisfies the translation and rotation transformation matrix is established as shown in the following formula:
[0025] f(D,C)=min∑[D·a i +Cb i ],
[0026] In the formula, a i , b i These are the collection points that need to be registered.
[0027] Furthermore, in the step (4), three-dimensional modeling is performed by constructing a plane triangulated network after projection, specifically by using Delaunay triangulation modeling, projecting the point cloud data onto plane coordinates, constructing a two-dimensional triangulated network using X and Y coordinates, and then assigning a Z value to each point to obtain a three-dimensional model.
[0028] Furthermore, the step (4) also includes model simplification, specifically, by keeping the topological relationship and geometric model of the model unchanged after Delaunay triangulation modeling through data extraction (vtkDecimatePro), and reducing the number of triangles in the triangulation network, which mainly includes the following steps:
[0029] ①Simply classify the points according to their local geometric and topological features, and determine whether the points can be deleted;
[0030] ② Determine the vertex to be deleted from the candidate deletion point set, and delete the point and all triangles passing through the point;
[0031] ③ After deleting the vertex and the triangle connected to the vertex, a blank area will be left. First, use the initialized splitting plane to split the blank area into two sub-areas. If the point set in the sub-area is distributed in the cutting plane, it is necessary to recursively split it again to fill the blank area.
[0032] Due to the adoption of the above technical scheme, the beneficial effects of the present invention are: data of power equipment is collected through a three-dimensional laser scanning system to ensure the acquisition of various data of the power equipment including irregular corner points, to achieve the construction of a complex and complete power equipment model, and to provide a detailed description of subtle features, thereby improving modeling accuracy and simulation performance. BRIEF DESCRIPTION OF THE DRAWINGS
[0033] Figure 1 It is a principle block diagram of the present invention. DETAILED DESCRIPTION
[0034] The technical solution of the present invention is further specifically described below through embodiments and in conjunction with the accompanying drawings.
[0035] like Figure 1 As shown, a method for modeling a three-dimensional model of power equipment based on laser point cloud data includes the following steps:
[0036] (1) Equipment data collection: Use a 3D laser scanning system to perform full coverage scanning of the power equipment to obtain surface texture information and 3D spatial point cloud data of the power equipment.
[0037] The three-dimensional laser scanning system in step (1) adopts a three-dimensional laser scanner or an airborne laser radar. Specifically, the three-dimensional laser scanner is preferably used to collect data in step (1), which includes the following steps:
[0038] ① Set target balls. The number of target balls should be greater than 3 and the target balls should be in different planes. The target balls cannot be moved before the scanning is completed.
[0039] ② Scanning station setting: Set up the scanning station at the actual location of the power equipment to be modeled based on the actual situation, and avoid obstruction during the scanning station setting process.
[0040] ③Data acquisition: obtain point cloud data through laser scanner and collect texture mapping data.
[0041] Specifically, the data content obtained by using the three-dimensional laser scanner in step (1) is as follows:
[0042] The vertical and horizontal angles of the laser beam emitted by the 3D laser scanner are obtained through the angle values of two continuously rotating mirrors with reflected pulsed lasers. The two angles of the laser beam and the detected distance are used to obtain the relative coordinates of each scanning point and the corresponding space of the 3D laser scanner, as well as the radiation intensity of the scanning point of the 3D laser scanner. The scanning point obtained by the 3D laser scanner is represented by (x, y, z, h). The formula for obtaining the coordinates (x, y, z) of the 3D laser scanning point by the 3D laser scanner is as follows:
[0043]
[0044] Where α and β represent the vertical and horizontal angles of the laser beam, respectively; L and h represent the slant distance from the 3D laser scanner to the scanning point and the reflection intensity, respectively.
[0045] (2) Data preprocessing: This includes data stitching and coordinate correction. When collecting data, at least three target points are set in adjacent areas to unify the point cloud data obtained by scanning at different angles at different sites into the same coordinate system. The stitched data are then converted into geodetic coordinates based on the obtained geodetic coordinates.
[0046] (3) Data processing, including filtering and denoising, and multi-site data alignment.
[0047] Among them, ① filtering and denoising processing includes filtering the point cloud data to remove noise points introduced by sensor errors and environmental interference factors; specifically, the filtering and denoising processing adopts median filtering, slides a fixed value window on the distance image, arranges the points in the window in order according to the size of the Y value, and takes the midpoint value as the center point of the window.
[0048] ② The point cloud data acquired from multiple sites are fused and aligned, and data registration processing is performed; the data registration processing aligns and splices the data from different sites, specifically by arranging control points with the same name on the surface of the power equipment to achieve the alignment and splicing of the point cloud data.
[0049] Specifically, suppose that the corresponding data point sets of the measured data points in the overlapping area at different viewing angles are represented by A and B, and the subset {m i |m i ∈A, i = 1, 2, ..., N} is extracted from the data point set A, and the subset {m ′ i |m ′ i∈B, i = 1, 2, ..., N} is extracted from the data point set B, and the two points correspond to each other. The point cloud registration is realized by obtaining the translation and rotation coordinate transformation relationship of the data points collected from each perspective; C and D represent the translation matrix and rotation matrix respectively. The essence of point cloud registration is to minimize the objective function, and the objective function that satisfies the translation and rotation transformation matrix is established as shown in the following formula:
[0050] f(D,C)=min∑[D·a i +Cb i ],
[0051] In the formula, a i , b i These are the collection points that need to be registered.
[0052] (4) Three-dimensional modeling: Based on the features of the point cloud, a three-dimensional spatial data model is constructed by projecting a plane triangulated network.
[0053] Among them, in the step (4), the three-dimensional modeling is performed by constructing a plane triangulated network after projection, and specifically, Delaunay triangulation modeling is used to project the point cloud data onto plane coordinates, and a two-dimensional triangulated network is constructed using X and Y coordinates. Then, a Z value is assigned to each point to obtain a three-dimensional model.
[0054] The step (4) also includes model simplification, which specifically keeps the topological relationship and geometric model of the model unchanged after Delaunay triangulation modeling through data extraction (vtkDecimatePro), and reduces the number of triangles in the triangulation network, which mainly includes the following steps:
[0055] ① Simply classify the points according to their local geometric and topological feature relationships, and determine whether the points can be deleted.
[0056] ② Determine the vertex to be deleted from the set of candidate deletion points, and delete the point and all triangles passing through the point; specifically, determine the point to be deleted by the distance value from the vertex to the local plane or the distance value from the vertex to the local edge. If the distance value is less than the set threshold, delete the point and the triangles passing through it, otherwise retain the point.
[0057] ③ After deleting the vertex and the triangle connected to the vertex, a blank area will be left. First, use the initialized splitting plane to split the blank area into two sub-areas. If the point set in the sub-area is distributed in the cutting plane, it is necessary to recursively split it again to fill the blank area.
[0058] (5) Texture mapping: mapping texture information onto the model to form the real texture of the model.
[0059] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that the specific implementation modes of the present invention can still be modified or replaced by equivalents, and any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be included in the scope of the claims of the present invention.
Claims
1. A method for modeling a three-dimensional model of power equipment based on laser point cloud data, characterized in that: The following steps are involved: (1) Equipment data collection; Use a 3D laser scanning system to perform full coverage scanning of power equipment to obtain surface texture information and 3D space point cloud data of power equipment; (2) Data preprocessing: including data stitching and coordinate correction. When collecting data, at least three target points are set in adjacent areas to unify the point cloud data obtained by scanning at different angles at different sites into the same coordinate system. Then, based on the obtained geodetic coordinates, the stitched data is converted into geodetic coordinates. (3) Data processing, including filtering and denoising, and multi-site data registration; Among them, ① filtering and denoising processing includes filtering the point cloud data to remove noise points introduced by sensor errors and environmental interference factors; ② fusing and aligning the point cloud data obtained from multiple sites to perform data registration processing; (4) 3D modeling: Based on the point cloud features, a three-dimensional spatial data model is constructed by projecting a plane triangulated network. (5) Texture mapping: mapping texture information onto the model to form the real texture of the model.
2. A method for modeling a three-dimensional model of electric power equipment based on laser point cloud data according to claim 1, characterized in that: In the step (1), the three-dimensional laser scanning system adopts a three-dimensional laser scanner or an airborne laser radar.
3. The method for modeling a three-dimensional model of electric power equipment based on laser point cloud data according to claim 2, characterized in that: The step (1) of collecting data using a three-dimensional laser scanner comprises the following steps: ① Set the target balls. The number of target balls should be greater than 3 and the target balls should be in different planes. The target balls should not be moved before the scanning is completed. ② Scanning station setting: Set up the scanning station at the actual location of the power equipment to be modeled according to the actual situation, and avoid obstruction during the scanning station setting process; ③Data acquisition: obtain point cloud data through laser scanner and collect texture mapping data.
4. The method for modeling a three-dimensional model of electric power equipment based on laser point cloud data according to claim 2, characterized in that: The data content obtained by using the three-dimensional laser scanner in step (1) is as follows: The vertical and horizontal angles of the laser beam emitted by the 3D laser scanner are obtained through the angle values of two continuously rotating mirrors with reflected pulsed lasers. The two angles of the laser beam and the detected distance are used to obtain the relative coordinates of each scanning point and the corresponding space of the 3D laser scanner, as well as the radiation intensity of the scanning point of the 3D laser scanner. The scanning point obtained by the 3D laser scanner is represented by (x, y, z, h). The formula for obtaining the coordinates (x, y, z) of the 3D laser scanning point by the 3D laser scanner is as follows: Where α and β represent the vertical and horizontal angles of the laser beam, respectively; L and h represent the slant distance from the 3D laser scanner to the scanning point and the reflection intensity, respectively.
5. The method for modeling a three-dimensional model of electric power equipment based on laser point cloud data according to claim 1, characterized in that: The filtering and denoising process in step (3) specifically adopts median filtering, slides a fixed value window on the distance image, arranges the points in the window in order of Y value, and takes the midpoint value as the center point of the window.
6. The method for modeling a three-dimensional model of electric power equipment based on laser point cloud data according to claim 1, characterized in that: In the step (3), the data registration process aligns and splices the data of different sites, specifically by arranging control points with the same name on the surface of the power equipment to achieve the alignment and splicing of the point cloud data; Assume that there are overlapping areas where the corresponding data points at different viewing angles are represented by A and B, and the subset {m i |m i ∈A, i = 1, 2, ..., N} is extracted from the data point set A, and the subset {m ′ i |m ′ i ∈B, i = 1, 2, ..., N} is extracted from the data point set B, and the two points correspond to each other. The point cloud registration is realized by obtaining the translation and rotation coordinate transformation relationship of the data points collected from each perspective; C and D represent the translation matrix and rotation matrix respectively. The essence of point cloud registration is to minimize the objective function, and the objective function that satisfies the translation and rotation transformation matrix is established as shown in the following formula: f(D,C)=min∑[D·a i +C-b i ], In the formula, a i 、b i These are the collection points that need to be registered.
7. The method for modeling a three-dimensional model of electric power equipment based on laser point cloud data according to claim 1, characterized in that: In the step (4), three-dimensional modeling is performed by constructing a plane triangulated network after projection, and specifically, Delaunay triangulation modeling is used to project the point cloud data onto plane coordinates, and a two-dimensional triangulated network is constructed using X and Y coordinates. Then, a Z value is assigned to each point to obtain a three-dimensional model.
8. The method for modeling a three-dimensional model of electric power equipment based on laser point cloud data according to claim 1, characterized in that: The step (4) also includes model simplification, which specifically keeps the topological relationship and geometric model of the model unchanged after Delaunay triangulation modeling through data extraction (vtkDecimatePro), and reduces the number of triangles in the triangulation network, which mainly includes the following steps: ①Simply classify the points according to their local geometric and topological features, and determine whether the points can be deleted; ② Determine the vertex to be deleted from the candidate deletion point set, and delete the point and all triangles passing through the point; ③ After deleting the vertex and the triangle connected to the vertex, a blank area will be left. First, use the initialized splitting plane to split the blank area into two sub-areas. If the point set in the sub-area is distributed in the cutting plane, it is necessary to recursively split it again to fill the blank area.
Citation Information
Cited By
Data preset learning-based dynamic material allocation system for multiple stations
CN120317643A
