Transformer substation modeling method and system based on point cloud data

Abstract

The invention provides a transformer substation modeling method and system based on point cloud data, and the method comprises the steps: constructing a transformer substation equipment model library according to the point cloud data of all equipment in a transformer substation; converting the point cloud data of each device into a two-dimensional planar bounding box topological structure of each device; geometric shape detection is carried out on the two-dimensional planar bounding box topological structure of each device, and a minimum repetitive structure unit is extracted; performing shape retrieval and template fitting on the minimum repetitive structure unit and a substation equipment model library to obtain a three-dimensional point cloud model of the minimum repetitive structure unit, and completing rapid construction of a substation model; according to the method, geometrical shape detection is performed on the two-dimensional planar bounding box topological structure of each device, and the accuracy of extracting the minimum repetitive structure unit can be improved on the basis of the determined geometrical shape, so that the accurate minimum repetitive structure unit can be rapidly extracted on the basis of the clear geometrical shape, and the accuracy of extracting the minimum repetitive structure unit is improved. And completing rapid construction of a subsequent transformer substation model.

Description

Technical Field

[0001] The patent application of this invention belongs to the technical field of power systems, and specifically relates to a substation modeling method and system based on point cloud data. Background Art

[0002] With the development of power systems and the advancement of smart grids, the digitization and intelligence of substations have become a trend. The operation steps are numerous and complex, which leads to reduced efficiency in the implementation process, and it is easy to have erroneous data, making it difficult to quickly complete the modeling later.

[0003] The currently popular VRML (Virtual Reality Modeling Language) modeling method, the manual substation modeling method based on 3D laser point cloud, the 3D substation modeling method based on surveying and mapping and 2D images, the geometric modeling method and the ground lidar modeling method were analyzed and compared. It was found that the VRML modeling method has the following disadvantages: poor intuitiveness, low efficiency, and high requirements for operators; the visual effect and sense of reality of the model are not strong enough to intuitively reflect the actual situation of the substation; the modeling process is time-consuming and difficult to meet the needs of efficient modeling. The operator needs to have high professional skills, which increases the labor cost; the manual substation modeling method based on 3D laser point cloud has the following disadvantages: primary and secondary equipment are created according to point cloud data, and finally the models are combined. This manual modeling method has a large amount of data, complicated operations and a long cycle; the laser point cloud data is huge , high processing difficulty; manually creating primary and secondary equipment has many steps and is complicated; the manual modeling process is time-consuming and it is difficult to quickly complete the model construction; the substation 3D modeling method based on the combination of surveying and mapping and two-dimensional images has the following disadvantages: an algorithm is used to segment and extract objects from the two-dimensional image of the substation, and then the corresponding three-dimensional shape is found from the shape library, and then the position and posture of the three-dimensional model is determined according to the surveying and mapping data; it is difficult to ensure the precise positional relationship between substation equipment, which affects the accuracy of the model; the precision and accuracy of the model are limited by the richness of the shape library and the matching effect. Although this method can improve the modeling efficiency, the surveying and mapping data is relatively cumbersome and cannot ensure the precise positional relationship between substation equipment; therefore, the substation modeling method in the prior art has many and complicated operation steps, which leads to reduced efficiency in the implementation process, and it is easy to have erroneous data, which makes it difficult to quickly complete the modeling later. Summary of the invention

[0004] In order to overcome the above-mentioned deficiencies of the prior art, the present invention patent application proposes a substation modeling method based on point cloud data, including:

[0005] Based on the point cloud data of each device in the substation, combined with the image data and design standards of each device, a substation equipment model library is constructed;

[0006] Converting the point cloud data of each device into a two-dimensional plane bounding box topological structure of each device;

[0007] Performing geometric shape detection on the topological structure of the two-dimensional plane bounding box of each device to extract the minimum repeating structural unit;

[0008] The minimum repeating structural unit is subjected to shape retrieval and template fitting with the substation equipment model library to obtain a three-dimensional point cloud model of the minimum repeating structural unit, and the substation model is quickly constructed based on the three-dimensional point cloud model of the minimum repeating structural unit.

[0009] Preferably, converting the point cloud data of each device into a two-dimensional plane bounding box topological structure of each device includes:

[0010] Projecting the point cloud data of each device onto a two-dimensional coordinate plane to obtain cluster points corresponding to the point cloud data of each device, and generating bounding boxes corresponding to each device according to the coordinates of the cluster points on the two-dimensional coordinate plane;

[0011] The K-means++ data clustering algorithm is used to cluster the bounding boxes of all devices to obtain multiple bounding box cluster sets;

[0012] The bounding boxes in each bounding box cluster set are taken as nodes, and the edges of the nodes are determined according to the spatial relationship between the bounding boxes in each bounding box cluster set. Based on the relationship between the nodes and the edges, the topological structure of each bounding box cluster set is obtained as the two-dimensional plane bounding box topological structure of the corresponding device.

[0013] Preferably, the performing geometric shape detection on the topological structure of the two-dimensional plane bounding box of each device to extract the minimum repeating structural unit includes:

[0014] Using Hough transform, detect the geometric shapes existing in the topological structure of the two-dimensional plane bounding box of each device;

[0015] matching the geometric shape with a predefined device pattern to determine a repeating structure of the geometric shape;

[0016] Based on the repetitive structure, the two-dimensional plane bounding box topological structure of each device is compared to determine the two-dimensional plane bounding box topological structure of similar devices, and the minimum repetitive structure unit is extracted from the topological structure of the similar bounding box cluster set.

[0017] Preferably, the substation equipment model library is constructed based on the point cloud data of each device in the substation, in combination with the image data and design standards of each device, including:

[0018] Using 3D modeling and rendering technology, the basic outlines of each device in the substation are created based on the point cloud data of each device in the substation;

[0019] Based on the basic outline of each device, combined with the image data of each device and the device-related design standards, fine modeling is performed to obtain a three-dimensional point cloud model of each device;

[0020] After mapping the three-dimensional point cloud models of the various devices, the three-dimensional point cloud models of the various devices are organized to complete the construction of the substation equipment model library.

[0021] Preferably, before constructing the substation equipment model library based on the point cloud data of each device in the substation, combined with the image data and design standards of each device, the method further includes:

[0022] Use 3D laser scanning technology to collect the original point cloud data of the substation;

[0023] Adopting an adaptive ground detection algorithm, detecting ground point cloud data from the original point cloud data, and cutting the ground point cloud data to obtain point cloud data after cutting the ground point cloud data;

[0024] The K-means++ data clustering algorithm is used to perform a clustering operation on the point cloud data after cutting the ground point cloud data, so as to obtain the point cloud data of each device in the substation.

[0025] Preferably, the method of using an adaptive ground detection algorithm to detect ground point cloud data from the original point cloud data and cutting the ground point cloud data to obtain point cloud data after cutting the ground point cloud data comprises:

[0026] After performing local high-frequency noise denoising on the original point cloud data, denoising original point cloud data is obtained;

[0027] According to the terrain characteristics and point cloud density information of the substation, adaptive threshold adjustment and optimization iteration are performed on the denoised original point cloud data to identify ground point cloud data;

[0028] The ground point cloud data is cut out from the denoised original point cloud data to obtain point cloud data after cutting the ground point cloud data.

[0029] Preferably, the performing shape retrieval and template fitting on the minimum repeating structural unit and the substation equipment model library to obtain a three-dimensional point cloud model of the minimum repeating structural unit, and completing the rapid construction of the substation model according to the three-dimensional point cloud model of the minimum repeating structural unit, includes:

[0030] A three-dimensional model retrieval method is used to extract a feature description operator of the original point cloud data, and combined with the minimum repeating structural unit, a similarity matching algorithm is used to retrieve a three-dimensional point cloud model corresponding to the minimum repeating structural unit from the substation equipment model library;

[0031] After performing template fitting on the three-dimensional point cloud model corresponding to the retrieved minimum repeating structural unit, the model is reconstructed according to the detected repeating pattern to complete the rapid construction of the substation model;

[0032] The feature description operator is generated according to the difference in equipment views and the unique properties of substation equipment.

[0033] Based on the same inventive concept, the patent application of the present invention also provides a substation modeling system based on point cloud data, including: a substation equipment model library construction module, a point cloud data conversion module, a minimum repetitive structural unit extraction module and a substation model construction module;

[0034] The substation equipment model library construction module is used to construct a substation equipment model library based on the point cloud data of each device in the substation, combined with the image data and design standards of each device;

[0035] The point cloud data conversion module is used to convert the point cloud data of each device into a two-dimensional plane bounding box topological structure of each device;

[0036] The minimum repeating structural unit extraction module is used to perform geometric shape detection on the two-dimensional plane bounding box topological structure of each device to extract the minimum repeating structural unit;

[0037] The substation model construction module is used to perform shape retrieval and template fitting on the minimum repeating structural unit and the substation equipment model library to obtain a three-dimensional point cloud model of the minimum repeating structural unit, and complete the rapid construction of the substation model based on the three-dimensional point cloud model of the minimum repeating structural unit.

[0038] Preferably, the point cloud data conversion module includes:

[0039] A bounding box generation submodule is used to project the point cloud data of each device onto a two-dimensional coordinate plane to obtain cluster points corresponding to the point cloud data of each device, and generate a bounding box corresponding to each device according to the coordinates of the cluster points on the two-dimensional coordinate plane;

[0040] The bounding box clustering submodule is used to cluster the bounding boxes of all devices using the K-means++ data clustering algorithm to obtain multiple bounding box cluster sets;

[0041] The bounding box topology structure generation submodule is used to take the bounding boxes in each bounding box cluster set as nodes, determine the edges of the nodes according to the spatial relationship between the bounding boxes in each bounding box cluster set, and obtain the topological structure of each bounding box cluster set according to the relationship between the nodes and the edges as the two-dimensional plane bounding box topological structure of the corresponding device.

[0042] Preferably, the minimum repeating structural unit extraction module is specifically used for:

[0043] Using Hough transform, detect the geometric shapes existing in the topological structure of the two-dimensional plane bounding box of each device;

[0044] matching the geometric shape with a predefined device pattern to determine a repeating structure of the geometric shape;

[0045] Based on the repetitive structure, the two-dimensional plane bounding box topological structure of each device is compared to determine the two-dimensional plane bounding box topological structure of similar devices, and the minimum repetitive structure unit is extracted from the topological structure of the similar bounding box cluster set.

[0046] Preferably, the substation equipment model library construction module is specifically used for:

[0047] Using 3D modeling and rendering technology, the basic outlines of each device in the substation are created based on the point cloud data of each device in the substation;

[0048] Based on the basic outline of each device, combined with the image data of each device and the device-related design standards, fine modeling is performed to obtain a three-dimensional point cloud model of each device;

[0049] After mapping the three-dimensional point cloud models of the various devices, the three-dimensional point cloud models of the various devices are organized to complete the construction of the substation equipment model library.

[0050] Preferably, the system further comprises: a point cloud data processing module, the point cloud data processing module comprising:

[0051] The data acquisition submodule is used to collect the original point cloud data of the substation using 3D laser scanning technology;

[0052] A data cutting submodule is used to detect ground point cloud data from the original point cloud data using an adaptive ground detection algorithm, and cut the ground point cloud data to obtain point cloud data after cutting the ground point cloud data;

[0053] The data clustering submodule is used to use the K-means++ data clustering algorithm to perform a clustering operation on the point cloud data after cutting the ground point cloud data, so as to obtain the point cloud data of each device in the substation.

[0054] Preferably, the data cutting submodule is specifically used for:

[0055] After performing local high-frequency noise denoising on the original point cloud data, denoising original point cloud data is obtained;

[0056] According to the terrain characteristics and point cloud density information of the substation, adaptive threshold adjustment and optimization iteration are performed on the denoised original point cloud data to identify ground point cloud data;

[0057] The ground point cloud data is cut out from the denoised original point cloud data to obtain point cloud data after cutting the ground point cloud data.

[0058] Preferably, the substation model building module is specifically used for:

[0059] A three-dimensional model retrieval method is used to extract a feature description operator of the original point cloud data, and combined with the minimum repeating structural unit, a similarity matching algorithm is used to retrieve a three-dimensional point cloud model corresponding to the minimum repeating structural unit from the substation equipment model library;

[0060] After performing template fitting on the three-dimensional point cloud model corresponding to the retrieved minimum repeating structural unit, the model is reconstructed according to the detected repeating pattern to complete the rapid construction of the substation model;

[0061] The feature description operator is generated according to the difference in equipment views and the unique properties of substation equipment.

[0062] Based on the same inventive concept, the patent application of the present invention also provides an electronic device, comprising: at least one processor and a memory; the memory and the processor are connected via a bus;

[0063] The memory is used to store one or more programs;

[0064] When the one or more programs are executed by the at least one processor, a substation modeling method based on point cloud data as described above is implemented.

[0065] Based on the same inventive concept, the patent application of the present invention also provides a readable storage medium on which a computer program and an execution program are stored. When the execution program is executed, a substation modeling method based on point cloud data as described above is implemented.

[0066] Compared with the closest prior art, the patent application of this invention has the following beneficial effects:

[0067] The patent application of the present invention provides a substation modeling method and system based on point cloud data, including: constructing a substation equipment model library according to the point cloud data of each device in the substation, combining the image data and design standards of each device; converting the point cloud data of each device into a two-dimensional plane bounding box topological structure of each device; performing geometric shape detection on the two-dimensional plane bounding box topological structure of each device to extract the minimum repeating structural unit; performing shape retrieval and template fitting on the minimum repeating structural unit and the substation equipment model library to obtain a three-dimensional point cloud model of the minimum repeating structural unit, and completing the rapid construction of the substation model based on the three-dimensional point cloud model of the minimum repeating structural unit; the present invention converts the point cloud data of each device into a two-dimensional plane bounding box topological structure of each device, and performs geometric shape detection on the two-dimensional plane bounding box topological structure of each device, based on the determined geometric shape, so as to improve the accuracy of extracting the minimum repeating structural unit, and then on the basis of a clear geometric shape, can quickly extract the accurate minimum repeating structural unit, compared with the prior art, reduces the operation process, and completes the rapid construction of the subsequent substation model. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] Figure 1 A schematic flow chart of a substation modeling method based on laser point cloud data provided for the patent application of the present invention;

[0069] Figure 2 A schematic diagram of a substation modeling system based on laser point cloud data provided in the patent application of the present invention;

[0070] Figure 3 A schematic structural diagram of an electronic device provided for the patent application of the present invention. DETAILED DESCRIPTION

[0071] The specific implementation methods of the patent application of the present invention are further described in detail below in conjunction with the accompanying drawings.

[0072] Embodiment 1:

[0073] The present invention patent application provides a substation modeling method based on laser point cloud data. Figure 1 As shown, including:

[0074] Step 1: Build a substation equipment model library based on the point cloud data of each device in the substation, combined with the image data and design standards of each device;

[0075] Step 2: converting the point cloud data of each device into a two-dimensional plane bounding box topological structure of each device;

[0076] Step 3: Perform geometric shape detection on the two-dimensional plane bounding box topological structure of each device to extract the minimum repeating structural unit;

[0077] Step 4: Perform shape retrieval and template fitting on the minimum repeating structural unit and the substation equipment model library to obtain a three-dimensional point cloud model of the minimum repeating structural unit, and complete the rapid construction of the substation model based on the three-dimensional point cloud model of the minimum repeating structural unit.

[0078] Before building a substation equipment model library based on the point cloud data of each device in the substation, combined with the image data and design standards of each device as described in step 1 above, the following steps are also included:

[0079] Use 3D laser scanning technology to collect the original point cloud data of the substation;

[0080] For example, the information included in each raw point cloud data includes: point coordinate data, reflection intensity, color information, timestamp, point classification information, intensity value echo number and normal vector; among them, the point coordinate data is the three-dimensional space coordinate (X, Y, Z) of each raw point cloud data, which is the core part of the raw point cloud data and defines the position of each raw point cloud data in three-dimensional space; the reflection intensity is the reflection signal intensity value returned by each raw point cloud data during the laser ranging process, which is used to indicate the reflection ability of the point, which helps to distinguish different materials or surface features; the color information also includes RGB value or other color information for each raw point cloud data Color data in color space; timestamp, which is the collection time information of each raw point cloud data, which is very important in dynamic point cloud or when it needs to be fused with other time-synchronized data; point classification information: some point cloud data may be accompanied by preliminary classification information of each raw point cloud data (such as ground, buildings, vegetation, etc.), which is usually generated by the lidar system or post-processing software; intensity value: number of echoes, which represents the number of multiple reflections and their intensity when the laser hits the surface of different objects, and is suitable for object detection in complex environments; normal vector, which is the surface normal direction information of each raw point cloud data, is used for surface smoothness analysis or subsequent point cloud processing.

[0081] Adopting an adaptive ground detection algorithm, detecting ground point cloud data from the original point cloud data, and cutting the ground point cloud data to obtain point cloud data after cutting the ground point cloud data;

[0082] The K-means++ data clustering algorithm is used to perform clustering operations on the point cloud data after cutting the ground point cloud data, so as to obtain the point cloud data of each device in the substation; the present invention collects the original point cloud data of the substation through three-dimensional laser scanning technology, realizes rapid modeling based on large-scale cloud data, and performs preprocessing such as point cloud segmentation and clustering; the original point cloud data of the substation is collected by three-dimensional laser scanning technology, and in order to perform repeated structure detection, shape retrieval and template fitting operations in the later stage, the point cloud data of each device in the substation is separated from a large number of original electric cloud data through an adaptive ground detection algorithm and a K-means++ data clustering algorithm, and the same monomer or device unit point is separated from the entire environment, and ground segmentation and point cloud clustering are performed. For point cloud clustering, in order to perform block operations on point cloud data in the later stage, that is, to obtain the electric cloud data of each device, it is necessary to perform spatial clustering on the overall point cloud, and finally obtain the point cloud data after cutting the ground point cloud data, so as to improve the efficiency of data processing and the accuracy of subsequent modeling, and provide basic data for later repeated structure detection, shape retrieval and template fitting.

[0083] In one implementation, the adaptive ground detection algorithm is used to detect ground point cloud data from the original point cloud data, and the ground point cloud data is cut to obtain point cloud data after the ground point cloud data is cut, including:

[0084] After performing local high-frequency noise denoising on the original point cloud data, denoising original point cloud data is obtained;

[0085] For example, for the extraction of ground point cloud data, an adaptive ground detection algorithm was developed to pre-process the collected raw point cloud data, perform specific denoising and coordinate transformation according to the specific environment of the substation, and use methods such as Gaussian filtering to reduce local high-frequency noise.

[0086] According to the terrain characteristics and point cloud density information of the substation, adaptive threshold adjustment and optimization iteration are performed on the denoised original point cloud data to identify ground point cloud data;

[0087] For example, according to the terrain characteristics and point cloud density information of the substation, an adaptive threshold adjustment module and an iterative optimization module are added to repeatedly adjust the recognition results of the ground points to approximate the real ground and obtain the final ground point cloud data.

[0088] The ground point cloud data is cut out from the denoised original point cloud data to obtain point cloud data after cutting the ground point cloud data; the present invention can extract ground point cloud data close to the real one through an adaptive ground detection algorithm, thereby obtaining point cloud data after cutting the ground point cloud data with higher accuracy, and can obtain point cloud data of the device with perfect data through a K-means++ data clustering algorithm;

[0089] In the above step 1, based on the point cloud data of each device in the substation, combined with the image data and design standards of each device, a substation equipment model library is constructed, including:

[0090] Using 3D modeling and rendering technology, the basic outlines of each device in the substation are created based on the point cloud data of each device in the substation;

[0091] Based on the basic outline of each device, combined with the image data of each device and the device-related design standards, fine modeling is performed to obtain a three-dimensional point cloud model of each device;

[0092] After mapping the three-dimensional point cloud models of the various devices, the three-dimensional point cloud models of the various devices are organized to complete the construction of the substation equipment model library.

[0093] For example, in the process of building a substation equipment model library, first use 3D modeling and rendering production technology, such as 3ds Max software, to create a basic outline of the substation equipment model library based on the point cloud data of each device, and then refer to the image data and the relevant design standards of the equipment to carry out fine modeling of each device in the substation, generate a three-dimensional point cloud model of each device, map the three-dimensional point cloud model of each device to make the three-dimensional point cloud model more realistic in appearance, and then use the corresponding management system to organize the established three-dimensional point cloud models to build a usable substation equipment model library. The present invention obtains a three-dimensional point cloud model of an independent device through pre-processing steps such as noise reduction and coordinate conversion to ensure high quality and consistency of data;

[0094] The present invention also introduces real-time video streaming technology for collecting substation equipment to achieve continuous monitoring and data updating of substation equipment. Through real-time video streaming, dynamic changes of equipment and environment can be captured to ensure the latest and accurate three-dimensional point cloud models and data. The substation equipment model library can be updated in real time according to changes in equipment in actual situations. Therefore, the present invention combines multiple data sources such as real-time video streaming and laser point cloud data to provide more comprehensive and accurate three-dimensional modeling information. The fusion of multi-source data can improve the accuracy and reliability of the model. The substation equipment model library of the present invention provides sufficient three-dimensional point cloud models of equipment of different models, and performs detailed feature extraction on each three-dimensional point cloud model, thereby improving the efficiency of substation modeling and avoiding the workload of manual modeling.

[0095] The step 2 above converts the point cloud data of each device into a two-dimensional plane bounding box topological structure of each device, including:

[0096] Projecting the point cloud data of each device onto a two-dimensional coordinate plane to obtain cluster points corresponding to the point cloud data of each device, and generating bounding boxes corresponding to each device according to the coordinates of the cluster points on the two-dimensional coordinate plane;

[0097] For example, the bounding boxes of different types of devices are significantly different, while devices with the same structure not only have similar bounding boxes, but also the arrangement of the bounding boxes has a certain regularity. Therefore, the present invention application proposes a repeated structure detection algorithm based on the topological structure of a two-dimensional plane bounding box. In order to reduce the difficulty of detection and improve the detection efficiency, the repeated structure detection of the plane point cloud is converted into the repeated structure detection of the bounding box; then, the clustered point cloud data is orthogonally projected along the Z axis onto the X-Y plane. After the projection, clusters of points on the two-dimensional plane can be obtained, where each cluster of points represents a device, and the bounding box of each cluster of point clouds on the two-dimensional plane can be obtained according to the coordinates.

[0098] The bounding box is usually the smallest enclosing rectangle of the point cloud data, which can effectively represent the footprint and shape characteristics of each device. In this step, the type and shape of the component will be reflected in the size and position of the bounding box; the data of the bounding box includes the position (XY coordinates), size (width, height) and direction of the bounding box.

[0099] The K-means++ data clustering algorithm is used to cluster the bounding boxes of all devices to obtain multiple bounding box cluster sets;

[0100] For example, initialization is first performed: an initial cluster center is selected, with priority given to the point farthest from the current cluster center to ensure that the initial cluster center is reasonably distributed; distance calculation: the distance between each bounding box and the cluster center is calculated based on the geometric features of the bounding box (position, size, etc.); iterative adjustment: each bounding box is assigned to the cluster center closest to it, and the cluster center position is continuously and iteratively adjusted until the bounding boxes within the cluster reach the optimal allocation, and the bounding box clustering set is determined.

[0101] The bounding boxes in each bounding box cluster set are taken as nodes, and the edges of the nodes are determined according to the spatial relationship between the bounding boxes in each bounding box cluster set. Based on the relationship between the nodes and the edges, the topological structure of each bounding box cluster set is obtained as the two-dimensional plane bounding box topological structure of the corresponding device.

[0102] For example, each bounding box corresponds to a node, and the node contains information such as coordinates and size of the bounding box; according to the spatial relationship between components, such as adjacency or repetition, the edges between nodes (indicating the spatial association between bounding boxes) are determined.

[0103] The step 3 described above performs geometric shape detection on the topological structure of the two-dimensional plane bounding box of each device to extract the minimum repeating structural unit, including:

[0104] Using Hough transform, detect the geometric shapes existing in the topological structure of the two-dimensional plane bounding box of each device;

[0105] For example, the Hough transform is used to detect geometric shapes existing in the bounding box topology structure, and the bounding box topology structure in which the geometric shapes exist is determined.

[0106] matching the geometric shape with a predefined device pattern to determine a repeating structure of the geometric shape;

[0107] For example, the detected geometric shapes are matched with predefined equipment patterns to identify whether there are repeated structures in the substation equipment layout.

[0108] Based on the repetitive structure, the two-dimensional plane bounding box topological structure of each device is compared to determine the two-dimensional plane bounding box topological structure of similar devices, and the minimum repetitive structure unit is extracted from the topological structure of the similar bounding box cluster set.

[0109] For example, through observation, it is found that the layout of substation equipment in adjacent areas has a high degree of repetition. In order to improve the efficiency of substation reconstruction, it is necessary to extract the smallest repetitive structural unit of the equipment in the substation. The repetitive structure in the substation scene is detected using a two-dimensional plane bounding box and a topological structure algorithm, which can accurately identify the repetitive components in the scene; therefore, the geometric shapes identified by the Hough transform provide a basis for subsequent topological structure analysis. These geometric shapes help identify and confirm potential repetitive units in the component layout; based on the repetitive structure, by comparing the topological structures of all bounding boxes, the same or similar connection patterns, node and edge configurations in the structure are identified, and the topological structures of the bounding boxes with similar topological structures are extracted. Based on the extracted similar topological structures, the geometric shapes are used to identify the potential repetitive units in the component layout. The topological structure of the bounding box of the puff structure generates the smallest repeating structural unit, which may be a modular combination of equipment, reflecting the regular distribution of equipment in the substation; the present invention adopts Hough transform to determine the geometric shape in the topological structure, and the geometric shape can intuitively show the shape characteristics of each bounding box cluster set. Therefore, based on the geometric shape, by comparing the topological structures of all bounding box cluster sets, the smallest repeating structural unit of similar bounding box cluster sets can be quickly and accurately extracted; the point cloud data is analyzed by the above technical means to detect the repeating structure and equipment mode in the substation. The smallest repeating structural unit utilizes the repeatability of substation equipment and extracts common equipment and structures, thereby significantly improving the modeling efficiency.

[0110] The above step 4 performs shape retrieval and template fitting on the minimum repeating structural unit and the substation equipment model library to obtain a three-dimensional point cloud model of the minimum repeating structural unit, and completes the rapid construction of the substation model based on the three-dimensional point cloud model of the minimum repeating structural unit, including:

[0111] A three-dimensional model retrieval method is used to extract a feature description operator of the original point cloud data, and combined with the minimum repeating structural unit, a similarity matching algorithm is used to retrieve a three-dimensional point cloud model corresponding to the minimum repeating structural unit from the substation equipment model library;

[0112] For example, in order to distinctively describe the point cloud data of the scanned objects in the scene, it is first necessary to define a set of features, namely feature descriptors, also known as shape descriptors; previously, some classic feature descriptors, such as rotated images, curvature, and scale-invariant feature transformations, can better solve the feature definition problem; however, these algorithms are not suitable for noisy substation cloud data; the present invention utilizes the differences in device views and the unique properties of substation equipment to define feature description operators, which are insensitive to noise, have good recognition of substation equipment, and can handle the common problems of occlusion and uneven point cloud distribution in the scanning process.

[0113] The feature description operator of the equipment is extracted from the original point cloud data, and combined with the minimum repeating structural unit, the similarity matching algorithm is used to retrieve the most suitable three-dimensional point cloud model in the substation equipment model library. The retrieval and fitting method can quickly and accurately find the three-dimensional electric cloud model that best matches the point cloud data.

[0114] Define the features of the 3D point cloud model in the substation equipment model library, and retrieve the corresponding 3D point cloud model based on the features of the original point cloud data corresponding to the selected minimum repeating structural unit; perform shape fitting between the retrieved 3D point cloud model and the point cloud; use the 3D model retrieval method to extract the feature information of the original point cloud data as a feature description operator, and combine it with the minimum repeating structural unit to perform similarity comparison with the model in the substation equipment model library, and finally convert the shape retrieval problem into the nearest neighbor search problem in the feature space;

[0115] Through feature extraction, the physical information can be converted into a set of vectors in the feature space. Similar three-dimensional point cloud models correspond to similar distances in the feature space. The process of the similarity matching algorithm is as follows: the feature description operator of the original point cloud data corresponding to the input minimum repeating structural unit is converted into a vector in the feature space through feature definition, and the spatial distance between the vectors of all three-dimensional point cloud models in the model library is calculated, or the similarity between the three-dimensional point cloud models is obtained through the constructed classifier, and finally the three-dimensional point cloud model with the greatest similarity to the feature description operator of the original point cloud data corresponding to the input minimum repeating structural unit is returned, that is, shape-based retrieval is realized, and the minimum repeating unit is modeled through the above-mentioned technical means, and the three-dimensional point cloud model of the minimum repeating unit is accurately constructed.

[0116] After performing template fitting on the three-dimensional point cloud model corresponding to the retrieved minimum repeating structural unit, the model is reconstructed according to the detected repeating pattern to complete the rapid construction of the substation model;

[0117] For example, in the fitting process, the center of the 3D point cloud model corresponding to the smallest repeating structural unit is firstly overlapped with the center of the point cloud cluster to obtain the result of the preliminary registration, and then the fine registration is performed on the basis of the preliminary registration, as follows:

[0118] Template fitting includes two steps: initial registration and fine registration. First, during the initial registration, the system aligns the geometric center of the 3D point cloud model corresponding to the minimum repeating structural unit and the corresponding point cloud data, and clusters the point cloud data using a clustering algorithm to find the point cloud cluster of the 3D point cloud model corresponding to each minimum repeating structural unit, and aligns the geometric center of the 3D point cloud model corresponding to the minimum repeating structural unit with the geometric center of the point cloud cluster; the fine registration process is to further adjust the 3D point cloud model corresponding to the minimum repeating structural unit on the basis of the initial registration to make it accurately match the point cloud data; specifically, iterative adjustment is used to adjust the position and direction of the model in each iteration to reduce the error with the point cloud until the set accuracy requirements are met. Through multiple iterations, the 3D point cloud model corresponding to the minimum repeating structural unit can eventually be made highly consistent with the point cloud data in terms of position, direction and shape; the retrieved 3D point cloud model is fit to the point cloud data in shape, and the 3D point cloud model and the point cloud data are aligned through a fitting algorithm to ensure high-precision matching between the 3D point cloud model and the actual equipment. This step ensures the high precision and consistency of the 3D point cloud model by verifying the accuracy and matching degree of the 3D point cloud model; based on the detected repeating structure, the 3D point cloud model corresponding to the minimum repeating structural unit is reconstructed according to the detected repeating pattern, and the repeating pattern of the substation is extracted through the detection of the repeating structure, which greatly improves the efficiency of later modeling and achieves fast and accurate overall modeling.

[0119] Repeating patterns refer to structures, equipment, or components with similar or identical characteristics that frequently appear in the structure or layout of a substation. These repeating patterns are usually caused by the design standardization of substations or the repetitive requirements of certain functions. Common repeating patterns include: repetition of equipment layout, repetition of topological structure, and repetition of geometric shape.

[0120] Repetition of equipment arrangement: In a substation, certain types of equipment (such as circuit breakers, disconnectors, transformers) may appear multiple times in the same way, and their relative positions and connection methods have a certain regularity or symmetry.

[0121] Repetition of topology: Certain electrical connections or busbar layouts may be repeated within a substation, forming a standardized power topology.

[0122] Repetition of geometric shapes: In point cloud data or 3D electric cloud models, there may be a large number of similar geometric shapes (such as poles, brackets, control cabinets, etc.) in the substation.

[0123] Among them, the feature description operator is generated according to the difference in device views and the unique properties of substation equipment; the feature description operator of the present invention is generated according to the difference in device views and the unique properties of substation equipment, so it is insensitive to noise, has good recognition of substation equipment, and can handle the common problems of occlusion and uneven point cloud distribution in the scanning process. Therefore, in the process of shape retrieval and template fitting based on the extraction of feature description operators, the data processing steps can be reduced to complete the rapid construction of the substation model; the present invention converts the point cloud data of each device into a two-dimensional plane bounding box topological structure of each device, and performs geometric shape detection on the two-dimensional plane bounding box topological structure of each device. Based on the geometric shape, it can improve the extraction of the minimum repeating structural unit. The accuracy rate can be increased, and then on the basis of clear geometric shapes, the accurate minimum repeating structural unit can be quickly extracted. Compared with the existing technology, the operation process is reduced, and the subsequent substation model is quickly constructed. The three-dimensional point cloud modeling of the substation using point cloud data can not only realize the visual management of the substation, but also provide important technical support for equipment detection and maintenance. The existing modeling methods have the problems of low efficiency and insufficient accuracy, and it is difficult to meet the needs of practical applications. The method of the present invention significantly improves the modeling efficiency and accuracy, reduces the data processing volume and operation complexity, and is suitable for rapid three-dimensional modeling of complex substation scenes; the present invention automates the steps of data preprocessing, substation equipment model library retrieval, feature matching and shape fitting, reduces manual intervention, and improves modeling efficiency and accuracy.

[0124] Embodiment 2:

[0125] Based on the same inventive concept, the present invention patent application also provides a substation modeling system based on point cloud data such as Figure 2 As shown, it includes: a substation equipment model library construction module, a point cloud data conversion module, a minimum repetitive structural unit extraction module and a substation model construction module;

[0126] The substation equipment model library construction module is used to construct a substation equipment model library based on the point cloud data of each device in the substation, combined with the image data and design standards of each device;

[0127] The point cloud data conversion module is used to convert the point cloud data of each device into a two-dimensional plane bounding box topological structure of each device;

[0128] The minimum repeating structural unit extraction module is used to perform geometric shape detection on the two-dimensional plane bounding box topological structure of each device to extract the minimum repeating structural unit;

[0129] The substation model construction module is used to perform shape retrieval and template fitting on the minimum repeating structural unit and the substation equipment model library to obtain a three-dimensional point cloud model of the minimum repeating structural unit, and complete the rapid construction of the substation model based on the three-dimensional point cloud model of the minimum repeating structural unit.

[0130] Preferably, the point cloud data conversion module includes:

[0131] A bounding box generation submodule is used to project the point cloud data of each device onto a two-dimensional coordinate plane to obtain cluster points corresponding to the point cloud data of each device, and generate a bounding box corresponding to each device according to the coordinates of the cluster points on the two-dimensional coordinate plane;

[0132] The bounding box clustering submodule is used to cluster the bounding boxes of all devices using the K-means++ data clustering algorithm to obtain multiple bounding box cluster sets;

[0133] The bounding box topology structure generation submodule is used to take the bounding boxes in each bounding box cluster set as nodes, determine the edges of the nodes according to the spatial relationship between the bounding boxes in each bounding box cluster set, and obtain the topological structure of each bounding box cluster set according to the relationship between the nodes and the edges as the two-dimensional plane bounding box topological structure of the corresponding device.

[0134] Preferably, the minimum repeating structural unit extraction module is specifically used for:

[0135] Using Hough transform, detect the geometric shapes existing in the topological structure of the two-dimensional plane bounding box of each device;

[0136] matching the geometric shape with a predefined device pattern to determine a repeating structure of the geometric shape;

[0137] Based on the repetitive structure, the two-dimensional plane bounding box topological structure of each device is compared to determine the two-dimensional plane bounding box topological structure of similar devices, and the minimum repetitive structure unit is extracted from the topological structure of the similar bounding box cluster set.

[0138] Preferably, the substation equipment model library construction module is specifically used for:

[0139] Using 3D modeling and rendering technology, the basic outlines of each device in the substation are created based on the point cloud data of each device in the substation;

[0140] Based on the basic outline of each device, combined with the image data of each device and the device-related design standards, fine modeling is performed to obtain a three-dimensional point cloud model of each device;

[0141] After mapping the three-dimensional point cloud models of the various devices, the three-dimensional point cloud models of the various devices are organized to complete the construction of the substation equipment model library.

[0142] Preferably, the system further comprises: a point cloud data processing module, the point cloud data processing module comprising:

[0143] The data acquisition submodule is used to collect the original point cloud data of the substation using 3D laser scanning technology;

[0144] A data cutting submodule is used to detect ground point cloud data from the original point cloud data using an adaptive ground detection algorithm, and cut the ground point cloud data to obtain point cloud data after cutting the ground point cloud data;

[0145] The data clustering submodule is used to use the K-means++ data clustering algorithm to perform a clustering operation on the point cloud data after cutting the ground point cloud data, so as to obtain the point cloud data of each device in the substation.

[0146] Preferably, the data cutting submodule is specifically used for:

[0147] After performing local high-frequency noise denoising on the original point cloud data, denoising original point cloud data is obtained;

[0148] According to the terrain characteristics and point cloud density information of the substation, adaptive threshold adjustment and optimization iteration are performed on the denoised original point cloud data to identify ground point cloud data;

[0149] The ground point cloud data is cut out from the denoised original point cloud data to obtain point cloud data after cutting the ground point cloud data.

[0150] Preferably, the substation model building module is specifically used for:

[0151] A three-dimensional model retrieval method is used to extract a feature description operator of the original point cloud data, and combined with the minimum repeating structural unit, a similarity matching algorithm is used to retrieve a three-dimensional point cloud model corresponding to the minimum repeating structural unit from the substation equipment model library;

[0152] After performing template fitting on the three-dimensional point cloud model corresponding to the retrieved minimum repeating structural unit, the model is reconstructed according to the detected repeating pattern to complete the rapid construction of the substation model;

[0153] The feature description operator is generated according to the difference in equipment views and the unique properties of substation equipment.

[0154] Example 3

[0155] like Figure 3 As shown, the present invention also provides an electronic device, which may be a computer device, a single-chip device, an intelligent mobile device, etc. The electronic device in this embodiment may include a processor, a memory, a transceiver component, etc. The memory, the processor, and the transceiver component are connected via a bus; the memory may be used to store an execution program, and an exemplary execution program may include instructions; the processor is used to execute the instructions stored in the memory. The memory may also be used to store data, which may be called and / or modified when the instructions are executed.

[0156] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, which is suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions in a storage medium to implement the corresponding method flow or corresponding functions, so as to implement the steps of a substation modeling method based on point cloud data in the above embodiment.

[0157] Example 4

[0158] Based on the same inventive concept, the present invention also provides a readable storage medium, specifically an electronic device readable storage medium (Memory), which is a memory device in an electronic device for storing programs and data. It can be understood that the storage medium here can include both built-in storage media in electronic devices and, of course, extended storage media supported by electronic devices. The storage medium provides a storage space, which stores the operating system of the terminal. In addition, one or more instructions suitable for being loaded and executed by a processor are also stored in the storage space, and these instructions can be one or more execution programs (including program codes). It should be noted that the storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk storage. The processor loads and executes one or more instructions stored in the storage medium, which can implement the steps of a substation modeling method based on point cloud data in the above embodiment.

[0159] It should be understood by those skilled in the art that the embodiments of the present invention patent application can be provided as methods, systems, or computer program products. Therefore, the present invention patent application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention patent application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0160] The present invention patent application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present invention patent application. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0161] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0162] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0163] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the patent application of the present invention rather than to limit its protection scope. Although the patent application of the present invention is described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that after reading the patent application of the present invention, those skilled in the art can still make various changes, modifications or equivalent substitutions to the specific implementation methods of the application, but these changes, modifications or equivalent substitutions are all within the protection scope of the claims to be approved.

Claims

1. A substation modeling method based on point cloud data, characterized in that: include: Based on the point cloud data of each device in the substation, combined with the image data and design standards of each device, a substation equipment model library is constructed; Converting the point cloud data of each device into a two-dimensional plane bounding box topological structure of each device; Performing geometric shape detection on the topological structure of the two-dimensional plane bounding box of each device to extract the minimum repeating structural unit; The minimum repeating structural unit is subjected to shape retrieval and template fitting with the substation equipment model library to obtain a three-dimensional point cloud model of the minimum repeating structural unit, and the substation model is quickly constructed based on the three-dimensional point cloud model of the minimum repeating structural unit.

2. The method according to claim 1, characterized in that The step of converting the point cloud data of each device into a two-dimensional plane bounding box topological structure of each device includes: Projecting the point cloud data of each device onto a two-dimensional coordinate plane to obtain cluster points corresponding to the point cloud data of each device, and generating bounding boxes corresponding to each device according to the coordinates of the cluster points on the two-dimensional coordinate plane; The K-means++ data clustering algorithm is used to cluster the bounding boxes of all devices to obtain multiple bounding box cluster sets; The bounding boxes in each bounding box cluster set are taken as nodes, and the edges of the nodes are determined according to the spatial relationship between the bounding boxes in each bounding box cluster set. Based on the relationship between the nodes and the edges, the topological structure of each bounding box cluster set is obtained as the two-dimensional plane bounding box topological structure of the corresponding device.

3. The method according to claim 2, characterized in that The performing geometric shape detection on the topological structure of the two-dimensional plane bounding box of each device to extract the minimum repeating structural unit includes: Using Hough transform, detect the geometric shapes existing in the topological structure of the two-dimensional plane bounding box of each device; matching the geometric shape with a predefined device pattern to determine a repeating structure of the geometric shape; Based on the repetitive structure, the two-dimensional plane bounding box topological structure of each device is compared to determine the two-dimensional plane bounding box topological structure of similar devices, and the minimum repetitive structure unit is extracted from the topological structure of the similar bounding box cluster set.

4. The method according to claim 1, characterized in that The substation equipment model library is constructed based on the point cloud data of each device in the substation, combined with the image data and design standards of each device, including: Using 3D modeling and rendering technology, the basic outlines of each device in the substation are created based on the point cloud data of each device in the substation; Based on the basic outline of each device, combined with the image data of each device and the device-related design standards, fine modeling is performed to obtain a three-dimensional point cloud model of each device; After mapping the three-dimensional point cloud models of the various devices, the three-dimensional point cloud models of the various devices are organized to complete the construction of the substation equipment model library.

5. The method according to claim 4, characterized in that Before constructing the substation equipment model library based on the point cloud data of each device in the substation, combined with the image data and design standards of each device, the method further includes: Use 3D laser scanning technology to collect the original point cloud data of the substation; Adopting an adaptive ground detection algorithm, detecting ground point cloud data from the original point cloud data, and cutting the ground point cloud data to obtain point cloud data after cutting the ground point cloud data; The K-means++ data clustering algorithm is used to perform a clustering operation on the point cloud data after cutting the ground point cloud data, so as to obtain the point cloud data of each device in the substation.

6. The method according to claim 5, characterized in that The adaptive ground detection algorithm is used to detect ground point cloud data from the original point cloud data, and cut the ground point cloud data to obtain point cloud data after cutting the ground point cloud data, including: After performing local high-frequency noise denoising on the original point cloud data, denoising original point cloud data is obtained; According to the terrain characteristics and point cloud density information of the substation, adaptive threshold adjustment and optimization iteration are performed on the denoised original point cloud data to identify ground point cloud data; The ground point cloud data is cut out from the denoised original point cloud data to obtain point cloud data after cutting the ground point cloud data.

7. The method according to claim 1, characterized in that The step of performing shape retrieval and template fitting on the minimum repeating structural unit and the substation equipment model library to obtain a three-dimensional point cloud model of the minimum repeating structural unit, and completing rapid construction of a substation model based on the three-dimensional point cloud model of the minimum repeating structural unit includes: A three-dimensional model retrieval method is used to extract a feature description operator of the original point cloud data, and combined with the minimum repeating structural unit, a similarity matching algorithm is used to retrieve a three-dimensional point cloud model corresponding to the minimum repeating structural unit from the substation equipment model library; After performing template fitting on the three-dimensional point cloud model corresponding to the retrieved minimum repeating structural unit, the model is reconstructed according to the detected repeating pattern to complete the rapid construction of the substation model; The feature description operator is generated according to the difference in equipment views and the unique properties of substation equipment.

8. A substation modeling system based on point cloud data, characterized in that: include: Substation equipment model library construction module, point cloud data conversion module, minimum repetitive structural unit extraction module and substation model construction module; The substation equipment model library construction module is used to construct a substation equipment model library based on the point cloud data of each device in the substation, combined with the image data and design standards of each device; The point cloud data conversion module is used to convert the point cloud data of each device into a two-dimensional plane bounding box topological structure of each device; The minimum repeating structural unit extraction module is used to perform geometric shape detection on the two-dimensional plane bounding box topological structure of each device to extract the minimum repeating structural unit; The substation model construction module is used to perform shape retrieval and template fitting on the minimum repeating structural unit and the substation equipment model library to obtain a three-dimensional point cloud model of the minimum repeating structural unit, and complete the rapid construction of the substation model based on the three-dimensional point cloud model of the minimum repeating structural unit.

9. The system according to claim 8, characterized in that The point cloud data conversion module comprises: A bounding box generation submodule is used to project the point cloud data of each device onto a two-dimensional coordinate plane to obtain cluster points corresponding to the point cloud data of each device, and generate a bounding box corresponding to each device according to the coordinates of the cluster points on the two-dimensional coordinate plane; The bounding box clustering submodule is used to cluster the bounding boxes of all devices using the K-means++ data clustering algorithm to obtain multiple bounding box cluster sets; The bounding box topology structure generation submodule is used to take the bounding boxes in each bounding box cluster set as nodes, determine the edges of the nodes according to the spatial relationship between the bounding boxes in each bounding box cluster set, and obtain the topological structure of each bounding box cluster set according to the relationship between the nodes and the edges as the two-dimensional plane bounding box topological structure of the corresponding device.

10. The system according to claim 9, characterized in that The minimum repeating structural unit extraction module is specifically used for: Using Hough transform, detect the geometric shapes existing in the topological structure of the two-dimensional plane bounding box of each device; matching the geometric shape with a predefined device pattern to determine a repeating structure of the geometric shape; Based on the repetitive structure, the two-dimensional plane bounding box topological structure of each device is compared to determine the two-dimensional plane bounding box topological structure of similar devices, and the minimum repetitive structure unit is extracted from the topological structure of the similar bounding box cluster set.

Images