A method and system for three-dimensional reconstruction of transmission lines based on edge-side computation

Abstract

This application discloses a method and system for 3D reconstruction of transmission lines based on edge computing. The method includes: extracting features from acquired transmission line image data using pre-defined prior constraints and a pre-defined structural feature extraction model based on the edge environment to obtain key feature information; mapping 3D point clouds in the edge environment based on the key feature information to construct an initial 3D model; and transmitting the initial 3D model to the cloud for model optimization to obtain a 3D model of the transmission line. This application uses prior constraints related to the geometric structure of the transmission line to provide a more accurate and effective feature extraction strategy, and performs feature extraction and initial model construction directly in the edge environment, eliminating the need to transmit large amounts of data to the cloud for processing, reducing network bandwidth pressure and improving the speed of 3D model construction.

Description

Technical Field

[0001] This application belongs to the field of spatial data processing technology for power transmission lines, specifically relating to a method and system for three-dimensional reconstruction of power transmission lines based on edge-side computation. Background Technology

[0002] With the development of smart grids and drone inspection technology, more accurate transmission line structure data collected by drones can be obtained to support the management and maintenance of power facilities in the power system. Among them, three-dimensional reconstruction of transmission lines is an important means of power facility management and maintenance. Based on the transmission line structure data, the construction of transmission line models can effectively identify nodes and lines in the transmission line that need optimization, promptly detect power facilities with fault risks or those that have already failed, and improve the safety and operational stability of the power system.

[0003] Traditional 3D reconstruction methods often face challenges in constructing large-scale, high-precision transmission line models, including high computational demands and long processing times. This is especially true when processing massive amounts of data centrally in the cloud, where network latency and bandwidth limitations further exacerbate the computational difficulties, leading to high data transmission pressure and slow response times. Furthermore, the lack of effective feature extraction strategies in centralized cloud computing before model construction results in a significant waste of data resources, further reducing the accuracy of the constructed model. Summary of the Invention

[0004] This application proposes a method and system for 3D reconstruction of transmission lines based on edge computing. It uses prior constraints related to the geometry of the transmission line to provide a more accurate and effective feature extraction strategy, and performs feature extraction and initial model construction directly in the edge environment. This eliminates the need to transmit a large amount of data to the cloud for processing, reducing network bandwidth pressure and improving the speed of 3D model construction.

[0005] The first aspect of this application provides a method for three-dimensional reconstruction of transmission lines based on edge-side computation, the method comprising:

[0006] Based on the edge environment, key feature information is obtained by extracting features from the collected transmission line image data through preset prior constraints and preset structural feature extraction models.

[0007] Based on key feature information, 3D point cloud mapping is performed in the edge environment to construct an initial 3D model;

[0008] The initial 3D model is transmitted to the cloud for model optimization to obtain a 3D model of the transmission line.

[0009] The above scheme uses prior constraints to introduce additional information into the structural feature extraction model for feature extraction, improving the model's generalization ability and performance. It can extract more accurate and relevant key feature information of the transmission line, which is then used to construct the initial 3D model. Because feature extraction and model construction are performed at the edge, there is no need to transmit data to the cloud for processing. This not only reduces reliance on cloud computing resources but also lowers network transmission pressure, preventing slow model construction due to network bandwidth limitations. Furthermore, cloud-based model optimization further improves the accuracy of the 3D transmission line model, making the reconstruction of transmission lines more standardized.

[0010] In one possible implementation of the first aspect, based on the edge-side environment, feature extraction is performed on the acquired transmission line image data through preset prior constraints and a preset structural feature extraction model to obtain key feature point information, specifically:

[0011] Load the prior constraints and the structural feature extraction model in the edge-side environment;

[0012] Based on the prior constraints, the feature extraction strategy of the structural feature extraction model is determined;

[0013] Based on the aforementioned feature extraction strategy, the structural feature extraction model is used to extract features from the transmission line image data to obtain key feature information.

[0014] The above solution processes the acquired transmission line image data directly and in real time in the edge environment, saving the time of uploading the data to the cloud; moreover, it uses prior constraints to obtain a more accurate feature extraction strategy, and performs feature localization and extraction on the transmission line image data more quickly, further saving data processing time.

[0015] In one possible implementation of the first aspect, the prior constraints are specifically:

[0016] Based on the preset geometric design of the transmission line, the geometric relationship of the power equipment is detected by using the preset prior model on the historical structural data of the transmission line, and several key geometric points of the transmission line are obtained.

[0017] The prior constraints are generated based on the geometric key points of the transmission line.

[0018] The above scheme is based on the geometric design of transmission lines. First, the geometric key points of specified power equipment are identified from the historical structural data of the transmission lines. Then, the prior constraints are generated based on the geometric key points of the transmission lines, so that the target features can be quickly identified and located from the data according to the prior constraints in the subsequent feature extraction, thereby improving the efficiency of feature extraction.

[0019] In one possible implementation of the first aspect, the prior model is specifically:

[0020] The specific formula for the prior model is as follows:

[0021] P = F prior (I)→(x i ,y i ,z i ,n i ,θ i );

[0022] In the formula, P is the set of key points of the transmission line extracted from the historical structural data I of the transmission line, and F prior (I) is the deep neural network on which the prior model is based, (x i ,y i ,z i ) represents the spatial coordinates of the geometric key points of the transmission line, and n is the coordinates of the key points. i Let θ be the normal vector of the geometric key point of the transmission line. i Let be the angle parameters of the geometric key points of the transmission line, and i be the i-th data in the historical structural data I of the transmission line.

[0023] In one possible implementation of the first aspect, a 3D point cloud is mapped in the edge-side environment based on key feature information to construct an initial 3D model, specifically as follows:

[0024] Based on a preset spatial perspective, the key feature point set of the current transmission line is extracted from the key feature information;

[0025] The key feature point set is optimized and mapped to three-dimensional space using a multi-view stereo matching algorithm to obtain an initial three-dimensional model.

[0026] The above scheme performs preliminary 3D model construction in the edge environment; firstly, it extracts key feature point set from key feature information, then optimizes the points in the point set through multi-view stereo matching algorithm to reduce noise in the point set and improve the accuracy of the data as much as possible, and then maps the optimized point set into 3D space to construct an initial 3D model.

[0027] In one possible implementation of the first aspect, the initial 3D model is transmitted to the cloud for model optimization to obtain a 3D model of the transmission line, specifically as follows:

[0028] After the initial 3D model is transmitted to the cloud, it is locally optimized and globally fused according to the preset transmission line structure standards to obtain the transmission line 3D model.

[0029] The above solution performs local optimization and global fusion of the initial 3D model in the cloud, correcting errors and refining model details. Because the cloud only needs to optimize the model and does not need to process large amounts of transmission line image data, this greatly reduces the computational load on the cloud and accelerates the entire transmission line reconstruction process.

[0030] The second aspect of this application provides an error correction system for an open-air gap current transformer, the system comprising: a feature extraction module, a model building module, and a model optimization module;

[0031] The feature extraction module is used to extract key feature information from the collected transmission line image data based on the edge environment, through preset prior constraints and preset structural feature extraction models.

[0032] The model building module is used to perform 3D point cloud mapping in the edge environment based on key feature information to build an initial 3D model;

[0033] The model optimization module is used to transmit the initial 3D model to the cloud for model optimization, resulting in a 3D model of the transmission line.

[0034] In one possible implementation of the second aspect, the feature extraction module includes: a key feature information extraction unit;

[0035] The key feature information extraction unit is used to load the prior constraints and the structural feature extraction model in the edge-side environment; determine the feature extraction strategy of the structural feature extraction model according to the prior constraints; and extract key feature information from the transmission line image data through the structural feature extraction model based on the feature extraction strategy.

[0036] In one possible implementation of the second aspect, the feature extraction module includes: a prior constraint construction unit;

[0037] The prior constraint construction unit is used to detect the geometric relationship of power equipment in the historical structural data of the transmission line based on the preset transmission line geometric design and a preset prior model, thereby obtaining several key geometric points of the transmission line; and to generate the prior constraints based on the key geometric points of the transmission line.

[0038] In one possible implementation of the second aspect, the model building module includes: an initial model building unit;

[0039] The initial model building unit is used to extract the key feature point set of the current transmission line from the key feature information based on a preset spatial perspective; and to map the key feature point set to three-dimensional space through a multi-view stereo matching algorithm to obtain an initial three-dimensional model.

[0040] In one possible implementation of the second aspect, the model optimization module includes: a model verification optimization unit;

[0041] The model verification and optimization unit is used to transmit the initial three-dimensional model to the cloud, and then perform local optimization and global fusion on the initial three-dimensional model according to the preset transmission line structure standard to obtain the three-dimensional model of the transmission line. Attached Figure Description

[0042] To more clearly illustrate the technical solution of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0043] Figure 1 This is a schematic diagram of a specific process of a three-dimensional reconstruction method for transmission lines based on edge-side calculation provided in a certain embodiment of this application;

[0044] Figure 2 This is a structural diagram of a three-dimensional reconstruction system for transmission lines based on edge-side computation provided in a certain embodiment of this application. Detailed Implementation

[0045] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0046] It should be understood that the step numbers used in the text are for ease of description only and are not intended to limit the order in which the steps are performed.

[0047] First Embodiment

[0048] To facilitate the monitoring of power transmission line operations and electrical equipment, it is necessary to reconstruct 3D models of the transmission lines to create a more detailed representation of the electrical equipment's working conditions. However, reconstructing transmission line models is generally a large undertaking, requiring the processing of massive amounts of data. Centralizing these data in the cloud would increase the computational burden on cloud computing. Furthermore, transmitting massive amounts of data to the cloud can lead to longer data transmission times due to bandwidth limitations and network latency, further reducing the efficiency of model building.

[0049] Therefore, instead of processing data in the cloud, processing it directly at the edge and constructing the initial 3D model of the transmission line can save time in transmitting massive amounts of data and alleviate the pressure on cloud computing. Furthermore, the accuracy of model reconstruction can be further improved by adjusting feature extraction strategies at the edge.

[0050] like Figure 1 As shown, Figure 1 This application provides a schematic flowchart of a three-dimensional reconstruction method for transmission lines based on edge-side computation according to a certain embodiment. The three-dimensional reconstruction method for transmission lines based on edge-side computation in this embodiment includes steps S1 to S3, which are detailed below:

[0051] Step S1: Based on the edge environment, feature extraction is performed on the collected transmission line image data through preset prior constraints and preset structural feature extraction models to obtain key feature information.

[0052] In this embodiment of the application, a large amount of historical structural data of transmission lines needs to be collected as samples. Then, a target detection network that can capture the features of transmission lines is constructed based on a convolutional neural network, which is the prior model. The historical structural data of transmission lines is then input into this prior model to extract geometric features and obtain prior constraints.

[0053] Furthermore, in the process of geometric feature extraction, based on the preset geometric design of the transmission line, the spatial relationships and dimensional patterns of the electrical facilities in the transmission line are extracted and processed to obtain the corresponding geometric key points of the transmission line. Each geometric key point of the transmission line includes spatial coordinates, a normal vector, and angular parameters. Based on these geometric key points of the transmission line, the key structures of the transmission line at branches and nodes can be determined, providing important data support for reconstructing the three-dimensional model of the transmission line.

[0054] Optionally, this application embodiment uses YOLOv5 as the target detection network, which can be used for real-time object detection and has excellent data processing speed and accuracy. Geometric feature extraction mainly targets important components of transmission lines such as towers, insulator strings, and conductors.

[0055] The specific formula for the prior model is as follows:

[0056] P = F prior (I)→(x i ,y i ,z i ,n i ,θ i );

[0057] In the formula, P is the set of key points of the transmission line extracted from the historical structural data I of the transmission line, and Fprior (I) is the deep neural network on which the prior model is based, (x i ,y i ,z i ) represents the spatial coordinates of the geometric key points of the transmission line, and n is the coordinates of the key points. i Let θ be the normal vector of the geometric key point of the transmission line. i Let be the angle parameters of the geometric key points of the transmission line, and i be the i-th data in the historical structural data I of the transmission line.

[0058] Because these geometric key points of transmission lines can very typically represent the characteristics of transmission lines in specific physical structures, they can be used as prior constraints to identify and locate key features in other transmission line images. In this embodiment, the geometric key points of transmission lines are used as prior constraints for the structural feature extraction model to improve the model's generalization ability, impose certain restrictions on model parameters or prediction results, and reduce the model's prediction error. Secondly, prior constraints can reduce data processing time and computational resource consumption during the model feature extraction process.

[0059] First, after drones and other devices collect image data of power transmission lines, features are extracted from the image data in the field using prior constraints and structural feature extraction models loaded in the edge environment. This allows for feature extraction from massive amounts of data using edge computing, rather than transmitting the collected power transmission line image data to the cloud for processing. The advantage of using edge computing in this embodiment is that the application is initiated at the edge, resulting in faster network service response and direct local data processing without transmitting large amounts of data to the cloud, saving data transmission time, especially when the amount of data to be transmitted is very large.

[0060] Edge computing refers to providing the nearest end service on the side closest to the object or data source, enabling real-time data processing and reducing data transmission latency.

[0061] Optionally, in some embodiments, an open platform integrating core capabilities such as networking, computing, storage, and applications, such as an unmanned aerial vehicle (UAV) computer or a field intelligent terminal, can be used as the edge environment.

[0062] Before performing feature extraction, the feature extraction strategy of the structural feature extraction model is adjusted based on prior constraints, and data noise is reduced.

[0063] For example, the structural feature extraction model described in this application embodiment is generated based on ResNet50. ResNet50 is widely used in image classification, object detection, and image segmentation. Therefore, by using prior constraints that include geometric key points of the transmission line, geometric spatial knowledge about the transmission line is introduced into the structural feature extraction model, which can generate a feature extraction strategy that uses the prior constraints to make assumptions and constraints on the targets in the transmission line image data.

[0064] Based on this feature extraction strategy, if there are missing data in the transmission line image data, the missing values ​​can be filled according to the prior constraints, and the missing values ​​will not have an excessive impact on the results. With the help of prior constraints, the structural feature extraction model can quickly identify and locate the spatial information of the reconstructed transmission line model in the transmission line image data, obtain key feature information, and improve the efficiency of feature extraction.

[0065] In this embodiment of the application, using prior constraints is equivalent to performing a data preprocessing on the transmission line image data, which can avoid the waste of computing resources caused by having to standardize the data first during feature extraction.

[0066] Step S2: Based on key feature information, perform 3D point cloud mapping in the edge environment to construct an initial 3D model.

[0067] In this embodiment, the initial 3D model is still constructed based on key feature information in the original edge-side environment.

[0068] First, based on a preset spatial perspective, a set of spatially correlated key feature points of the current transmission line is extracted from images of different spatial perspectives. Then, a multi-view stereo matching algorithm is used to optimize and map the set of key feature points into three-dimensional space.

[0069] Optionally, to adapt to the computing power of the edge environment, the embodiments of this application adopt a lightweight 3D reconstruction algorithm, which is mainly a simplified multi-view stereo matching algorithm that simplifies the 3D reconstruction steps into two steps: 3D point set optimization and point cloud stitching and densification, thereby reducing the computing pressure on the edge.

[0070] Furthermore, 3D point set optimization involves globally optimizing the key feature point set through bundle adjustment. This constructs a Jacobian matrix for the camera parameters and key feature points, followed by optimization using the Gauss-Newton nonlinear optimization method. This iterative optimization of the key feature point set reduces data noise. The bundle adjustment method aims to minimize the coordinate error between the 3D point cloud coordinates obtained by the algorithm and traditional 3D point cloud reconstruction algorithms using the full image.

[0071] The iterative optimization formula used in the 3D point set optimization is as follows:

[0072]

[0073] In the formula, x represents the camera parameters and key feature points. The camera parameters are the equipment parameters of the camera that captures image data of the transmission line. J is the Jacobian matrix, r is the reprojection error, and Δx represents the camera parameters and key feature points with minimized residuals.

[0074] Furthermore, after completing the optimization of the 3D point set, the optimized key feature point set is stitched together and densified to map the key feature points into the 3D space and form an initial 3D model.

[0075] The specific formula for the initial three-dimensional model is as follows:

[0076] M edge =G(P1,P2,...,P) n );

[0077] In the formula, M edge For the initial 3D model constructed in the edge-side environment, P1, P2, ..., P n G represents the set of key feature points, and G is the simplified multi-view stereo matching algorithm used in the embodiments of this application.

[0078] Point cloud stitching and densification are performed using ICP, a point cloud registration method with minimal computational cost. Since the prior constraints and 3D point set optimization used earlier reduce data noise, ICP can be used to perform accurate and efficient 3D mapping on high-precision key feature point sets.

[0079] Step S3: The initial 3D model is transmitted to the cloud for model optimization to obtain the 3D model of the transmission line.

[0080] In this embodiment of the application, the constructed initial three-dimensional model is transmitted to the cloud, and then the initial three-dimensional model is locally optimized and globally fused according to the preset transmission line structure standard to correct errors in the model and improve model details.

[0081] Because the initial feature extraction and model building are performed at the edge, while the subsequent model optimization is done in the cloud, this division of labor and collaboration places the part that needs to process massive amounts of data at the edge and the part that needs to be refined in the cloud, greatly reducing the computing load on the cloud and accelerating the process of reconstructing the entire transmission line 3D model.

[0082] Implementing the embodiments of this application has the following beneficial effects:

[0083] This application's embodiments introduce additional information about the specific structure of the transmission line into the structural feature extraction model by using prior constraints for feature extraction. This improves the model's generalization ability and performance, enabling the extraction of more accurate and relevant key feature information of the transmission line. This key feature information is then used to construct the initial 3D model. Because feature extraction and model construction are performed at the edge, there is no need to transmit data to the cloud for processing. This not only reduces reliance on cloud computing resources but also lowers network transmission pressure, preventing slow model construction due to network bandwidth limitations. Furthermore, fine-tuning the model in the cloud further improves the accuracy of the 3D transmission line model, making the reconstruction of the transmission line more standardized.

[0084] Second Embodiment

[0085] Furthermore, in order to implement the edge-side computing-based three-dimensional reconstruction system for transmission lines corresponding to the above method embodiments, and to achieve the corresponding functional and technical effects, Figure 2 A structural diagram of a three-dimensional reconstruction system for transmission lines based on edge-side computation is provided. For ease of explanation, only the parts relevant to this embodiment are shown. The three-dimensional reconstruction system for transmission lines based on edge-side computation provided in this application embodiment includes:

[0086] The feature extraction module 201 is used to extract key feature information from the collected transmission line image data based on the edge environment, through preset prior constraints and preset structural feature extraction models.

[0087] In this embodiment of the application, a large amount of historical structural data of transmission lines needs to be collected as samples. Then, a target detection network that can capture the features of transmission lines is constructed based on a convolutional neural network, which is the prior model. The historical structural data of transmission lines is then input into this prior model to extract geometric features and obtain prior constraints.

[0088] Furthermore, in the process of geometric feature extraction, based on the preset geometric design of the transmission line, the spatial relationships and dimensional patterns of the electrical facilities in the transmission line are extracted and processed to obtain the corresponding geometric key points of the transmission line. Each geometric key point of the transmission line includes spatial coordinates, a normal vector, and angular parameters. Based on these geometric key points of the transmission line, the key structures of the transmission line at branches and nodes can be determined, providing important data support for reconstructing the three-dimensional model of the transmission line.

[0089] Optionally, this application embodiment uses YOLOv5 as the target detection network, which can be used for real-time object detection and has excellent data processing speed and accuracy. Geometric feature extraction mainly targets important components of transmission lines such as towers, insulator strings, and conductors.

[0090] Then, before performing feature extraction, the feature extraction strategy of the structural feature extraction model is adjusted according to prior constraints, and data noise is reduced.

[0091] For example, the structural feature extraction model described in this application embodiment is generated based on ResNet50. ResNet50 is widely used in image classification, object detection, and image segmentation. Therefore, by using prior constraints that include geometric key points of the transmission line, geometric spatial knowledge about the transmission line is introduced into the structural feature extraction model, which can generate a feature extraction strategy that uses the prior constraints to make assumptions and constraints on the targets in the transmission line image data.

[0092] Finally, based on the aforementioned feature extraction strategy, the structural feature extraction model is used to extract features from the transmission line image data to obtain key feature information.

[0093] The model building module 202 is used to perform 3D point cloud mapping in the edge environment based on key feature information to build an initial 3D model.

[0094] In this embodiment, within the existing edge-side environment, firstly, based on a preset spatial perspective, a set of spatially correlated key feature points of the current transmission line, extracted from images of different spatial perspectives, is obtained from the key feature information. Then, a multi-view stereo matching algorithm is used to optimize and map the set of key feature points into three-dimensional space to construct an initial three-dimensional model.

[0095] The model optimization module 203 is used to transmit the initial 3D model to the cloud for model optimization to obtain the 3D model of the transmission line.

[0096] In this embodiment of the application, the constructed initial three-dimensional model is transmitted to the cloud, and then the initial three-dimensional model is locally optimized and globally fused according to the preset transmission line structure standard to correct errors in the model and improve model details.

[0097] Because the initial feature extraction and model building are performed at the edge, while the subsequent model optimization is done in the cloud, this division of labor and collaboration places the part that needs to process massive amounts of data at the edge and the part that needs to be refined in the cloud, greatly reducing the computing load on the cloud and accelerating the process of reconstructing the entire transmission line 3D model.

[0098] In some embodiments, the feature extraction module 201 further includes:

[0099] A key feature information extraction unit is used to load the prior constraints and the structural feature extraction model in the edge-side environment; determine the feature extraction strategy of the structural feature extraction model according to the prior constraints; and extract features from the transmission line image data through the structural feature extraction model based on the feature extraction strategy to obtain key feature information.

[0100] The prior constraint construction unit is used to detect the geometric relationship of power equipment on the historical structural data of the transmission line according to the preset transmission line geometric design and the preset prior model, so as to obtain a number of transmission line geometric key points; and generate the prior constraints based on the transmission line geometric key points.

[0101] The specific formula for the prior model is as follows:

[0102] P = F prior (I)→(x i ,y i ,z i ,n i ,θ i );

[0103] In the formula, P is the set of key points of the transmission line extracted from the historical structural data I of the transmission line, and F prior (I) is the deep neural network on which the prior model is based, (x i ,y i ,z i ) represents the spatial coordinates of the geometric key points of the transmission line, and n is the coordinates of the key points. i Let θ be the normal vector of the geometric key point of the transmission line. i Let be the angle parameters of the geometric key points of the transmission line, and i be the i-th data in the historical structural data I of the transmission line.

[0104] Because these geometric key points of transmission lines can very typically represent the characteristics of transmission lines in specific physical structures, they can be used as prior constraints to identify and locate key features in other transmission line images. In this embodiment, the geometric key points of transmission lines are used as prior constraints for the structural feature extraction model to improve the model's generalization ability, impose certain restrictions on model parameters or prediction results, and reduce the model's prediction error. Secondly, prior constraints can reduce data processing time and computational resource consumption during the model feature extraction process.

[0105] In some embodiments, the model building module 202 further includes:

[0106] The initial model building unit is used to extract the key feature point set of the current transmission line from the key feature information based on a preset spatial perspective; and to map the key feature point set to three-dimensional space through a multi-view stereo matching algorithm to obtain an initial three-dimensional model.

[0107] In this embodiment of the application, in order to adapt to the computing power of the edge environment, a lightweight 3D reconstruction algorithm is adopted. It is mainly a simplified multi-view stereo matching algorithm that simplifies the 3D reconstruction steps into two steps: 3D point set optimization, point cloud stitching and densification, thereby reducing the computing pressure on the edge.

[0108] Furthermore, 3D point set optimization involves globally optimizing the key feature point set through bundle adjustment. This constructs a Jacobian matrix for the camera parameters and key feature points, followed by optimization using the Gauss-Newton nonlinear optimization method. This iterative optimization of the key feature point set reduces data noise. The bundle adjustment method aims to minimize the coordinate error between the 3D point cloud coordinates obtained by the algorithm and traditional 3D point cloud reconstruction algorithms using the full image.

[0109] The iterative optimization formula used in the 3D point set optimization is as follows:

[0110]

[0111] In the formula, x represents the camera parameters and key feature points. The camera parameters are the equipment parameters of the camera that captures image data of the transmission line. J is the Jacobian matrix, r is the reprojection error, and Δx represents the camera parameters and key feature points with minimized residuals.

[0112] Furthermore, after completing the optimization of the 3D point set, the optimized key feature point set is stitched together and densified to map the key feature points into the 3D space and form an initial 3D model.

[0113] The specific formula for the initial three-dimensional model is as follows:

[0114] M edge =G(P1,P2,...,P) n );

[0115] In the formula, M edge For the initial 3D model constructed in the edge-side environment, P1, P2, ..., P n G represents the set of key feature points, and G is the simplified multi-view stereo matching algorithm used in the embodiments of this application.

[0116] In some embodiments, the model optimization module 203 further includes:

[0117] The model verification and optimization unit is used to transmit the initial three-dimensional model to the cloud, and then perform local optimization and global fusion on the initial three-dimensional model according to the preset transmission line structure standard to obtain the transmission line three-dimensional model.

[0118] Implementing the embodiments of this application has the following beneficial effects:

[0119] This application's embodiments introduce prior knowledge into the structural feature extraction model using prior constraints related to the geometric structure of transmission lines. This provides a more accurate and effective feature extraction strategy, enabling rapid identification and location of key feature points in transmission line image data, thus improving data processing efficiency. Then, a simplified multi-view stereo matching algorithm simplifies the 3D reconstruction process into two steps: 3D point set optimization and point cloud stitching and densification. This reduces computational pressure on the edge side, optimizing and mapping key feature points into 3D space to obtain an initial 3D model. Feature extraction and initial model construction are performed directly on the edge side, eliminating the need to transmit large amounts of data to the cloud for processing, reducing network bandwidth pressure and increasing the speed of 3D model construction. Finally, the model is refined and optimized in the cloud, correcting errors and improving model details.

[0120] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of this application. It should be understood that the above descriptions are merely specific embodiments of this application and are not intended to limit the scope of protection of this application. In particular, it should be noted that any modifications, equivalent substitutions, or improvements made by those skilled in the art within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for three-dimensional reconstruction of transmission lines based on edge-side computation, characterized in that, include: Based on the edge-side environment, key feature information is obtained by extracting features from the collected transmission line image data through preset prior constraints and a preset structural feature extraction model. Specifically, the prior constraints and the structural feature extraction model are loaded into the edge-side environment; the feature extraction strategy of the structural feature extraction model is determined according to the prior constraints; and the key feature information is obtained by extracting features from the transmission line image data through the structural feature extraction model based on the feature extraction strategy. Based on key feature information, perform 3D point cloud mapping in the edge environment to construct an initial 3D model; The initial 3D model is transmitted to the cloud for model optimization to obtain a 3D model of the transmission line.

2. The method for three-dimensional reconstruction of transmission lines based on edge-side computation according to claim 1, characterized in that, The prior constraints are specifically as follows: Based on the preset geometric design of the transmission line, the geometric relationship of the power equipment is detected by using the preset prior model on the historical structural data of the transmission line, and several key geometric points of the transmission line are obtained. The prior constraints are generated based on the geometric key points of the transmission line.

3. The method for three-dimensional reconstruction of transmission lines based on edge-side computation according to claim 2, characterized in that, The prior model is specifically as follows: The specific formula for the prior model is as follows: ； In the formula, P is the set of key points of the transmission line extracted from the historical structural data I of the transmission line. The prior model is based on a deep neural network. The spatial coordinates of the geometric key points of the transmission line. Here is the normal vector of the geometric key point of the transmission line. Let be the angle parameters of the geometric key points of the transmission line, and i be the i-th data in the historical structural data I of the transmission line.

4. The method for three-dimensional reconstruction of transmission lines based on edge-side computation according to claim 1, characterized in that, The step of mapping 3D point clouds in the edge environment based on key feature information to construct an initial 3D model is as follows: Based on a preset spatial perspective, the key feature point set of the current transmission line is extracted from the key feature information; The key feature point set is optimized and mapped to three-dimensional space using a multi-view stereo matching algorithm to obtain an initial three-dimensional model.

5. The method for three-dimensional reconstruction of transmission lines based on edge-side computation according to claim 1, characterized in that, The process of transmitting the initial 3D model to the cloud for model optimization to obtain a 3D model of the transmission line is as follows: After the initial 3D model is transmitted to the cloud, it is locally optimized and globally fused according to the preset transmission line structure standards to obtain the 3D model of the transmission line.

6. A three-dimensional reconstruction system for transmission lines based on edge-side computation, characterized in that, include: Feature extraction module, model building module, and model optimization module; The feature extraction module is used to extract features from the collected transmission line image data based on the edge-side environment, using preset prior constraints and a preset structural feature extraction model, to obtain key feature information. Specifically, it loads the prior constraints and the structural feature extraction model into the edge-side environment; determines the feature extraction strategy of the structural feature extraction model based on the prior constraints; and extracts features from the transmission line image data using the structural feature extraction model based on the feature extraction strategy to obtain key feature information. The model building module is used to perform 3D point cloud mapping in the edge environment based on key feature information to build an initial 3D model; The model optimization module is used to transmit the initial 3D model to the cloud for model optimization, thereby obtaining a 3D model of the transmission line.

7. The three-dimensional reconstruction system for transmission lines based on edge-side computation according to claim 6, characterized in that, The feature extraction module includes: a prior constraint construction unit; The prior constraint construction unit is used to detect the geometric relationship of power equipment in the historical structural data of the transmission line based on the preset transmission line geometric design and a preset prior model, thereby obtaining several key geometric points of the transmission line; and to generate the prior constraints based on the key geometric points of the transmission line.

8. The three-dimensional reconstruction system for transmission lines based on edge-side computation according to claim 6, characterized in that, The model building module includes: an initial model building unit; The initial model building unit is used to extract the key feature point set of the current transmission line from the key feature information based on a preset spatial perspective; and to map the key feature point set to three-dimensional space through a multi-view stereo matching algorithm to obtain an initial three-dimensional model.

9. The three-dimensional reconstruction system for transmission lines based on edge-side computation according to claim 6, characterized in that, The model optimization module includes: a model validation optimization unit; The model verification and optimization unit is used to transmit the initial three-dimensional model to the cloud, and then perform local optimization and global fusion on the initial three-dimensional model according to the preset transmission line structure standard to obtain the three-dimensional model of the transmission line.

Images