A network deployment line three-dimensional scene modeling device and method comprising a storage medium

Abstract

The application discloses a kind of network configuration line three-dimensional scene modeling equipment and method comprising storage medium.Belong to robot, digital twin, power engineering and fire protection engineering technical field;Including point cloud acquisition module, data receiving and filtering module, point cloud storage module, three-dimensional modeling module, three-dimensional model storage library module and visualization terminal;Its modeling steps are: point cloud acquisition, volume filtering, denoising processing, region segmentation processing, feature attribute extraction, model library feature attribute matching, OpenGL library 3D modeling, model assembly, end.The application is based on laser radar to environment information acquisition, without theodolite, GPS, Beidou positioning and other auxiliary measuring tools, low in cost;Without human participation measurement, reduce human uncertain influencing factor;Since laser radar uses active infrared light to measure, not susceptible to environmental illumination influence.Point cloud feature in the application carries out three-dimensional modeling, method is simple, modeling speed is fast, high in accuracy.

Description

Technical Field

[0001] This invention belongs to the fields of robotics, digital twins, power engineering, and fire protection engineering, and relates to a device and method for modeling three-dimensional scenes of distribution network lines including a storage medium; specifically, it relates to a method, device, and storage medium for modeling three-dimensional scenes of distribution network lines. Background Technology

[0002] The concept of digital twins first appeared in Professor Grieves's product lifecycle management course in 2003. In recent years, digital twin technology has become increasingly widespread. Simultaneously, thanks to the development of next-generation information technologies such as communication, computer, and artificial intelligence, the practical application of digital twins in products has gradually become possible.

[0003] With the development of intelligent power systems, digital twin technology has gradually been integrated into power intelligence research. Digital design of distribution network operations can significantly improve operational efficiency and safety. Traditional environmental modeling methods rely on drawings and on-site measurements, requiring tools such as infrared rangefinders, GPS / BeiDou-assisted measurements, and theodolites, resulting in complex processes and a large workload. Traditional methods are also subject to significant human subjectivity, leading to uncertainties in measurement results. Furthermore, discrepancies between the model and the actual physical structure result in insufficient simulation accuracy for distribution network operations, impacting operational safety.

[0004] Chinese patent application CN201811372986.X discloses a method for modeling 3D scenes of power distribution networks based on mixed reality technology. It utilizes multi-angle aerial images captured by drones and processes these images with GPS information to obtain point cloud data. A triangulation algorithm is then used for surface reconstruction to achieve 3D scene modeling. However, this method requires camera calibration and a series of image processing steps to obtain the point cloud, which is cumbersome and heavily influenced by ambient lighting. Furthermore, the accuracy of GPS positioning affects the 3D point cloud data, making environmental modeling difficult.

[0005] Currently, most 3D modeling systems for power distribution networks require auxiliary measurement tools such as theodolites, GPS, and BeiDou positioning. These tools are costly, and manual measurement is subject to uncertainties, resulting in long modeling times, low efficiency, and large model errors. While some systems use binocular cameras to collect environmental information about power distribution networks, these methods are susceptible to ambient lighting conditions, involve cumbersome procedures, are difficult to implement, and have low modeling accuracy. Summary of the Invention

[0006] Purpose of the Invention: The purpose of this invention is to provide a three-dimensional environmental modeling device and method for power distribution lines. This invention is based on LiDAR for environmental information collection, eliminating the need for auxiliary measurement tools such as theodolites, GPS, and BeiDou positioning, thus reducing costs; it also eliminates the need for manual measurement, reducing human uncertainty; since LiDAR uses active infrared light for measurement, it is less affected by ambient lighting; furthermore, the point cloud features used in this invention for three-dimensional modeling are simple, fast, and highly accurate.

[0007] Technical solution: The modeling method of the three-dimensional scene modeling device for distribution network lines including storage medium described in this invention has the following specific steps:

[0008] Step (1) Point cloud acquisition;

[0009] Specifically:

[0010] The power distribution network operation environment is scanned using a 3D lidar, and point cloud data C1 is obtained after passing through the data receiving and filtering module;

[0011] Wherein, the point cloud data C1 is based on the point cloud system coordinate system. A set of spatial coordinate points;

[0012] Step (2), volume filtering;

[0013] Step (3), noise reduction processing;

[0014] Step (4): Region segmentation processing;

[0015] Step (5), feature attribute extraction;

[0016] Specifically:

[0017] By constructing a kd-tree on the subset point cloud, the start and end points of the subset point cloud are found, and the length is calculated using Euclidean distance. The subset of point clouds, from beginning to end, is defined by a step size. Slicing is performed towards the end, and curvature fitting is performed on the cross-sections of each slice to obtain the curvature set K. Simultaneously, circle center fitting is performed to obtain the circle center set. The fitting radius is obtained by fitting the curvature set K. ; through length and fitted radius Distinguish the subset point cloud types;

[0018] The feature attributes include type, length, fitting radius, and set of circle centers;

[0019] The types mentioned include overhead lines, poles, crossarms, insulators, and other types;

[0020] Step (6): Matching feature attributes in the model library;

[0021] Specifically:

[0022] The feature attributes of the subset point cloud are obtained from the feature attribute extraction module and compared with the feature attributes in the model library;

[0023] If the feature attributes match successfully, the model in the model library is called directly and the model assembly in step (8) is performed; otherwise, the process jumps to step (7).

[0024] Step (7): 3D modeling using the OpenGL library;

[0025] Specifically, it involves using the feature attributes of a subset of point clouds to perform geometric modeling;

[0026] Based on the set of centers of the subset point cloud and fitted radius Create a set of sliced ​​circles, combine every two sliced ​​circles to get segmented cylinders, connect the segmented cylinders together to get a 3D model of the subset point cloud, and save it to the 3D model repository module.

[0027] At the same time, its feature attributes are saved in text form to the corresponding 3D model repository module;

[0028] Step (8), model assembly;

[0029] Step (9) End: Once the model assembly of the current distribution network line environment is completed, the distribution network digital twin environment modeling process ends.

[0030] Furthermore, in step (2), the volume filtering specifically involves:

[0031] Volumetric spatial filtering is applied to the point cloud data C1 of the distribution network operation environment to remove point clouds that are not related to the distribution network, thereby reducing the amount of point cloud data processing computation.

[0032] The filtering rule refers to using multiple cuboids to enclose the regional point cloud of the distribution network line, crossarm, insulator and tower, and extracting only the point cloud data C2 within the cuboid space.

[0033] Furthermore, in step (3), the noise reduction process specifically includes:

[0034] Denoising the point cloud data C2 and removing outliers will yield point cloud data C3.

[0035] The denoising rule is as follows: query each point in the point cloud data C2 one by one, and perform a search within a sphere with the radius r of the point in the point cloud data C2. If the number of points in the sphere is less than the value n, then the point in the point cloud data C2 is determined to be an outlier and removed from the dataset.

[0036] Furthermore, in step (4), the region segmentation process specifically involves:

[0037] The point cloud data C3 is segmented using a region growing method to obtain multiple subset point clouds.

[0038] Furthermore, in step (8), the model assembly specifically involves:

[0039] First, establish a model space reference coordinate system. ; Obtain the set of circle centers in the model's feature attributes starting point With the last site Set the starting position of the model with the center of the circle as the set. starting point The value is fixed in the model space, and the end positions of the model are set with the center of the circle. The last site The value is fixed in the model space;

[0040] The model mentioned above refers to the crossbar, tower, crossarm, insulator and other related models of the distribution network line.

[0041] Furthermore, a three-dimensional scene modeling device for distribution network lines including a storage medium, prepared by the modeling method, includes a point cloud acquisition module, a data receiving and filtering module, a point cloud storage module, a three-dimensional modeling module, a three-dimensional model storage module, and a visualization terminal.

[0042] The output of the point cloud acquisition module is connected to the input of the data receiving and filtering module.

[0043] The output terminal of the data receiving and filtering module is connected to the input terminal of the 3D modeling module.

[0044] The output of the 3D modeling module is connected to the input of the visualization terminal;

[0045] The other end of the data receiving and filtering module is connected to the point cloud storage module.

[0046] The other end of the 3D modeling module is connected to the 3D model storage module;

[0047] The point cloud acquisition module is composed of a 3D lidar sensor.

[0048] The data receiving and filtering module consists of a data filtering module, a wireless transmitting module, and a wireless receiving module;

[0049] Both the point cloud storage module and the 3D model storage module are computer-readable storage media;

[0050] The visualization terminal is a visualization device used to display a three-dimensional environmental model of the power distribution network.

[0051] Beneficial effects: Compared with the prior art, the present invention has the following characteristics: the present invention has low development cost, simple data collection process, low requirements for the number or skills of operators, and greatly reduces the input of human and material resources; the modeling steps are simple, fast, and the model accuracy is high, and the mapping of the digital twin system model space to the working environment is more realistic. Attached Figure Description

[0052] Figure 1 This is a schematic diagram of the present invention;

[0053] Figure 2 This is a flowchart of the method of the present invention;

[0054] Figure 3 This is a schematic diagram of a subset point cloud slice of a three-dimensional modeling method for distribution network lines in this invention;

[0055] Figure 4 This is a schematic diagram of a three-dimensional scene embodiment of the power distribution network in this invention;

[0056] Among them, 1 is the data acquisition system, 2 is the tower, 3 is the crossarm, 4 is the insulator, and 5 is the overhead line. Detailed Implementation

[0057] To more clearly illustrate the technical solution of the present invention, the technical solution of the present invention will be further described in detail below with reference to the accompanying drawings:

[0058] like Figure 1-3 The present invention provides a three-dimensional scene modeling device for power distribution lines that includes a storage medium, comprising a point cloud acquisition module, a data receiving and filtering module, a point cloud storage module, a three-dimensional modeling module, a three-dimensional model storage module, and a visualization terminal.

[0059] The point cloud acquisition module consists of a 3D lidar sensor and is used to acquire environmental point cloud data C0.

[0060] The data receiving and filtering module consists of a data filtering module, a wireless transmitting module, and a wireless receiving module;

[0061] The data filtering module is responsible for filtering out invalid data and data outside the working space from the point cloud data C0 of the point cloud acquisition module to obtain point cloud data C1.

[0062] The function of the wireless transmission module is to broadcast the point cloud data C1 from the data filtering module via a wireless protocol.

[0063] The function of the wireless receiving module is to receive data transmitted by the wireless transmitting module via a wireless protocol. The received data is still point cloud data C1.

[0064] The point cloud storage module is a computer-readable storage medium whose function is to store point cloud data C1 and to store the raw data.

[0065] The three-dimensional modeling module takes point cloud data C1 and uses the invention's three-dimensional environment modeling method for distribution network lines to establish a three-dimensional model of the distribution network line.

[0066] The 3D model repository module is a computer-readable storage medium.

[0067] Its function is to create a 3D model of the distribution network line using the 3D modeling module, and store it according to categories such as line, tower, crossarm, and insulator;

[0068] Furthermore, each model file is associated with its own attribute information, which can be used to index the corresponding model file.

[0069] The attribute information of the line includes name, length, cross-sectional radius, and key points of the line's centerline; the attribute information of the tower includes name, length, cross-sectional radius, and key points of the tower's centerline; the attribute information of the crossarm includes name, length, cross-sectional type, and cross-sectional dimensions; the attribute information of the insulator includes name, insulator type, radius, and height.

[0070] The 3D model storage rule compares the model attribute information in the 3D modeling module with the model attribute information in the 3D model storage module. If the relevant model attribute information already exists, the model is not stored to avoid duplicate storage and save storage space; otherwise, the model is stored.

[0071] The visualization terminal is a visualization device used to display a three-dimensional environmental model of the power distribution network.

[0072] A modeling method for a 3D scene modeling device for power distribution lines including a storage medium, comprising the following steps:

[0073] S1, Point Cloud Acquisition:

[0074] A 3D LiDAR is used to scan the power distribution network operation environment. After passing through the data receiving and filtering module, point cloud data C1 is obtained (point cloud data is a set of spatial coordinate points based on the point cloud system coordinate system).

[0075] S2, Volume Filtering:

[0076] Volumetric spatial filtering is applied to the point cloud data C1 of the distribution network operation environment to remove point clouds that are not related to the distribution network, thereby reducing the amount of point cloud data processing computation.

[0077] Among them, the filtering rule is: use multiple cuboids to surround the point cloud of the distribution network line, crossarm, insulator, tower and other areas, and only extract the point cloud data C2 within the cuboid space;

[0078] S3, Noise Reduction Processing:

[0079] Denoising the point cloud data C2 and removing outliers will yield point cloud data C3.

[0080] The noise reduction rule is as follows: each point in the point cloud data C2 is queried one by one, and a search is performed within the sphere with the radius r of the point. If the number of points in the sphere is less than the value n, the point is considered an outlier and is removed from the dataset.

[0081] S4. Region segmentation processing:

[0082] The point cloud data C3 is segmented using a region growing method to obtain multiple subset point clouds;

[0083] S5. Feature Attribute Extraction:

[0084] First, a kd-tree is constructed from the subset point cloud to find the start and end points of the subset point cloud, and the length is calculated using Euclidean distance. The subset of point clouds, from beginning to end, is defined by a step size. Slicing is performed towards the end, and curvature fitting is performed on the cross-sections of each slice to obtain the curvature set K. Simultaneously, circle center fitting is performed to obtain the circle center set. (Where, the set of center points is based on the point cloud system coordinate system) The set of spatial coordinate points); the fitting radius is obtained by fitting the curvature set K. ; through length and fitting radius region Differentiate the point cloud types from the subsets;

[0085] The feature attributes include type, length, fitting radius, and set of circle centers;

[0086] The types mentioned include line type, tower type, crossarm type, insulator type, and other types;

[0087] S6. Model library feature attribute matching:

[0088] The feature attributes of the subset point cloud are obtained from the feature attribute extraction module and compared with the feature attributes in the model library;

[0089] If the feature attributes match successfully, the model in the model library is called directly and the process is switched to S8 model assembly; otherwise, the process is switched to S7 and OpenGL library 3D modeling.

[0090] S7, OpenGL library 3D modeling:

[0091] Geometric modeling is performed using the feature attributes of a subset of point clouds;

[0092] Based on the set of centers of the subset point cloud and fitted radius Create a set of sliced ​​circles, combine every two sliced ​​circles to get segmented cylinders, connect the segmented cylinders together to get a 3D model of the subset point cloud, and save it to the 3D model repository module.

[0093] At the same time, its feature attributes are saved in text form to the corresponding 3D model repository module;

[0094] S8. Model Assembly:

[0095] First, establish a model space reference coordinate system. ; Obtain the set of circle centers in the model's feature attributes starting point With the last site Set the starting position of the model with the center of the circle as the set. starting point The value is fixed in the model space, and the end positions of the model are set with the center of the circle. The last site The value is fixed in the model space;

[0096] The model mentioned refers to the crossbar, tower, crossarm, insulator, and other related models in the distribution network line;

[0097] S9, End:

[0098] Once the model assembly for the current distribution network line environment is completed, the distribution network digital twin environment modeling process is finished. Example

[0099] like Figure 4 It includes the acquisition system 1, the tower 2, the crossarm 3, the insulator 4, and the running line 5.

[0100] First, move the acquisition system 1 below the distribution network line to be acquired. The acquisition system will collect environmental point cloud data from the distribution network line to obtain environmental point cloud data C0.

[0101] The data filtering module is used to filter out invalid data and data outside the working space in the point cloud data C0 from the point cloud acquisition module, so as to obtain point cloud data C1.

[0102] The point cloud data C1 from the data filtering module is broadcast and transmitted via a wireless protocol through the wireless transmission module.

[0103] The wireless receiving module receives data transmitted by the wireless transmitting module via a wireless protocol. The received data is still point cloud data C1.

[0104] The point cloud data C1 is saved using the point cloud storage module; at the same time, the point cloud data C1 is processed into a 3D model.

[0105] Secondly, volume filtering:

[0106] Volumetric spatial filtering is applied to the point cloud data C1 of the distribution network operation environment to remove point clouds that are not related to the distribution network, thereby reducing the amount of point cloud data processing computation.

[0107] Third, noise reduction processing:

[0108] Denoising the point cloud data C2 and removing outliers will yield point cloud data C3.

[0109] Then, region segmentation processing:

[0110] The point cloud data C3 was segmented using the region growing method, resulting in 8 subset point clouds: 3 rows of 5-point clouds, 3 insulators of 4-point clouds, 1 tower of 2-point clouds, and 1 crossarm of 3-point clouds.

[0111] Next, feature attribute extraction:

[0112] The type, length, fitting radius, and center set of the 8 subset point clouds were extracted respectively; thus, the features of the line, tower, crossarm, and insulator were obtained.

[0113] Next, feature attribute matching of the model library:

[0114] The features of line 5, tower 2, crossarm 3, and insulator 4 obtained in the previous step are matched with the feature attributes of the model library to check whether the corresponding model has been saved in the model library.

[0115] If so, do not perform modeling; otherwise, proceed to the next step to perform modeling.

[0116] Secondly, 3D modeling using the OpenGL library:

[0117] Based on the set of centers of the subset point cloud and fitted radius Create a set of sliced ​​circles, perform modeling, and obtain models of line 5, tower 2, crossarm 3, and insulator 4; and save them to the expert database;

[0118] Finally, model assembly:

[0119] Based on the start and end points of the center set in the model's feature attributes, the model's orientation in space can be obtained;

[0120] The models of line 5, tower 2, crossarm 3, and insulator 4 are fixed in the model space according to their respective spatial orientations, thereby completing the three-dimensional scene modeling of the distribution network line.

Claims

1. A modeling method for a three-dimensional scene modeling device for distribution network lines including a storage medium, characterized in that, The specific steps are as follows: Step (1), point cloud acquisition; specifically: The power distribution network operation environment is scanned using a 3D lidar, and point cloud data C1 is obtained after passing through the data receiving and filtering module; Wherein, the point cloud data C1 is based on the point cloud system coordinate system. A set of spatial coordinate points; Step (2), volume filtering; Step (3), noise reduction processing; Step (4): Region segmentation processing; Step (5), feature attribute extraction; Specifically: By constructing a kd-tree on the subset point cloud, the start and end points of the subset point cloud are found, and the length is calculated using Euclidean distance. The subset of point clouds, from beginning to end, is defined by a step size. Slicing is performed towards the end, and curvature fitting is performed on the cross-sections of each slice to obtain a curvature set. Simultaneously, a set of circle centers is obtained by fitting the circle centers. By analyzing the curvature set Fitting calculation, obtain the fitting radius. ; through length and fitted radius Distinguish the subset point cloud types; The feature attributes include type, length, fitting radius, and set of circle centers; The types mentioned include overhead lines, poles, crossarms, insulators, and other types; Step (6): Matching feature attributes in the model library; Specifically: The feature attributes of the subset point cloud are obtained from the feature attribute extraction module and compared with the feature attributes in the model library; If the feature attributes match successfully, the model in the model library is called directly and the model assembly in step (8) is performed; otherwise, the process jumps to step (7). Step (7): 3D modeling using the OpenGL library; Specifically, it involves using the feature attributes of a subset of point clouds to perform geometric modeling; Based on the set of centers of the subset point cloud and fitted radius Create a set of sliced ​​circles, combine every two sliced ​​circles to get segmented cylinders, connect the segmented cylinders together to get a 3D model of the subset point cloud, and save it to the 3D model repository module. At the same time, its feature attributes are saved in text form to the corresponding 3D model repository module; Step (8), model assembly; Step (9) End: Once the model assembly of the current distribution network line environment is completed, the distribution network digital twin environment modeling process ends.

2. The modeling method of a three-dimensional scene modeling device for a distribution network line including a storage medium according to claim 1, characterized in that, In step (2), the volume filtering specifically involves: Volumetric spatial filtering is applied to the point cloud data C1 of the distribution network operation environment to remove point clouds that are not related to the distribution network, thereby reducing the amount of point cloud data processing computation. The filtering rule refers to using multiple cuboids to enclose the regional point cloud of the distribution network line, crossarm, insulator and tower, and extracting only the point cloud data C2 within the cuboid space.

3. The modeling method for a three-dimensional scene modeling device for a distribution network line including a storage medium according to claim 1, characterized in that: In step (3), the noise reduction process specifically includes: Denoising the point cloud data C2 and removing outliers will yield point cloud data C3. The denoising rule is as follows: query each point in the point cloud data C2 one by one, and perform a search within a sphere using the radius r of the points in the point cloud data C2. If the number of points in the sphere is less than the value n, then the points in the point cloud data C2 are identified as outliers and removed from the dataset.

4. The modeling method of a three-dimensional scene modeling device for a distribution network line including a storage medium according to claim 1, characterized in that: In step (4), the region segmentation process specifically involves: The point cloud data C3 is segmented using a region growing method to obtain multiple subset point clouds.

5. The modeling method of a three-dimensional scene modeling device for a distribution network line including a storage medium according to claim 1, characterized in that: In step (8), the model assembly specifically involves: First, establish a model space reference coordinate system. ; Obtain the set of circle centers in the model's feature attributes starting point With the last site Set the starting position of the model with the center of the circle as the set. starting point The value is fixed in the model space, and the end positions of the model are set with the center of the circle. The last site The value is fixed in the model space; The model mentioned above refers to the crossbar, tower, crossarm, insulator and other related models of the distribution network line.

6. A three-dimensional scene modeling device for power distribution lines including a storage medium, characterized in that, Prepared by the modeling method according to any one of claims 1-5, it includes a point cloud acquisition module, a data receiving and filtering module, a point cloud storage module, a 3D modeling module, a 3D model storage module, and a visualization terminal; The output of the point cloud acquisition module is connected to the input of the data receiving and filtering module. The output terminal of the data receiving and filtering module is connected to the input terminal of the 3D modeling module. The output of the 3D modeling module is connected to the input of the visualization terminal; The other end of the data receiving and filtering module is connected to the point cloud storage module. The other end of the 3D modeling module is connected to the 3D model storage module; The point cloud acquisition module is composed of a 3D lidar sensor. The data receiving and filtering module consists of a data filtering module, a wireless transmitting module, and a wireless receiving module; Both the point cloud storage module and the 3D model storage module are computer-readable storage media; The visualization terminal is a visualization device used to display a three-dimensional environmental model of the power distribution network.

Images