A power transmission tower drawing generation method and related device

Abstract

The application provides a power transmission tower drawing generation method and related device, and belongs to the technical field of power transmission tower drawing generation. The application extracts structure surface point clouds with topological continuity from a neural radiation field model; converts the structure surface point clouds into a topological skeleton, and identifies node clusters and connecting rods based on the topological skeleton; intercepts a local cross-section slice from the topological skeleton, matches the local cross-section slice with a preset industrial steel specification library, and obtains standard physical specifications of each component; constructs an energy function based on the identified node clusters and connecting rods, and corrects collinearity, coplanarity and symmetry of a full tower topological structure based on the energy function, to obtain an optimized model, wherein the full tower topological structure comprises the node clusters and the connecting rods; and maps the optimized model to absolute physical scales, generates power transmission tower drawings in combination with the standard physical specifications of each component. The application solves the problem of low efficiency of power transmission tower drawing generation.

Description

Technical Field

[0001] This invention belongs to the field of power transmission tower drawing generation technology, specifically relating to a method and related apparatus for generating power transmission tower drawings. Background Technology

[0002] In recent years, with the sustained and rapid growth of energy consumption in my country, the scale and complexity of power transmission networks have significantly increased, highlighting the growing importance of operational safety and structural stability issues for transmission lines. Against this backdrop, the finite element model (FEM) possesses strong versatility and interpretability, making it particularly suitable for structural mechanical analysis, fatigue life prediction, and safety assessment under extreme loads. The FEM model eliminates the need for long-term deployment of numerous sensors and can simulate various operating conditions, thus offering significant advantages when real-time monitoring systems are lacking or when facing extreme operating condition simulations.

[0003] However, the construction of finite element models relies heavily on accurate tower structure drawings, which are often difficult to obtain in actual engineering projects. Many old transmission towers have been in service for decades, and their original structural drawings are mostly lost or incomplete. Without complete drawings, modeling work can only be carried out through manual on-site measurements, which is not only inefficient and costly, but also poses significant safety hazards due to high-altitude manual work. This industry pain point and technical bottleneck severely restricts the safety assessment and operation and maintenance optimization of aging transmission towers in service. Summary of the Invention

[0004] The purpose of this invention is to provide a method and related apparatus for generating transmission tower drawings, in order to solve the problem of low efficiency in generating transmission tower drawings in the prior art.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: In a first aspect, the present invention provides a method for generating transmission tower drawings, comprising the following steps: Acquire multi-scale image data covering the entire transmission tower and key structural nodes, and record it as the first data; Based on the first data, a point cloud focusing on the main body of the transmission tower is constructed; We use neural networks to learn the implicit density distribution of the transmission tower body space, construct a neural radiation field model, and extract structural surface point clouds with topological continuity from the neural radiation field model. The point cloud of the structural surface is converted into a topological skeleton, and node clusters and connecting rods are identified based on the topological skeleton. Extract local cross-sectional slices from the topological skeleton, match the local cross-sectional slices with the preset industrial steel specification library to obtain the standard physical specifications of each component; Based on the identified node clusters and connecting rods, an energy function is constructed, and the collinearity, coplanarity, and symmetry of the entire tower topology are corrected based on the energy function to obtain an optimized model. The entire tower topology includes node clusters and connecting rods. The optimized model is mapped to absolute physical scale, and the standard physical specifications of each component are combined to generate transmission tower drawings.

[0006] A further improvement of this invention is that the acquisition of multi-scale image data covering the entire transmission tower and key structural nodes specifically involves: The method of using drones to fly around the power transmission tower from multiple perspectives was adopted to acquire multi-scale image data covering the entire power transmission tower and key structural nodes.

[0007] A further improvement of the present invention is that, before constructing a point cloud focusing on the main body of the transmission tower based on the first data, the first data is preprocessed. The preprocessing includes deframe the first data, joint camera calibration, background interference elimination, and color constancy processing.

[0008] A further improvement of this invention is that the extraction of topologically continuous structural surface point clouds from the neural radiation field model specifically involves: We use isosurface sampling to extract topologically continuous structural surface point clouds from the neural radiation field model.

[0009] A further improvement of this invention lies in the fact that the absolute physical scale mapping of the optimized model, combined with the standard physical specifications of each component, generates transmission tower drawings, specifically as follows: Spatial coordinate calculation and similarity transformation registration are performed on the optimized model, and transmission tower drawings are generated by combining the standard physical specifications of each component.

[0010] A further improvement of the present invention is that the transmission tower drawings include component cross-sectional parameters, actual cutting length and spatial angle.

[0011] A further improvement of this invention is that, after generating the transmission tower drawings, a summary table of engineering materials and a list of component cutting lengths are generated.

[0012] Secondly, the present invention provides a transmission tower drawing generation system, comprising: The data acquisition module is used to acquire multi-scale image data covering the entire transmission tower and key structural nodes, which is referred to as the first data. The dataset construction module is used to build a point cloud focusing on the main body of the transmission tower based on the first data; The point cloud extraction module is used to learn the implicit density distribution of the transmission tower body space using neural networks, construct a neural radiation field model, and extract structural surface point clouds with topological continuity from the neural radiation field model. The conversion module is used to convert the point cloud of the structural surface into a topological skeleton, and to identify the node clusters and connecting rods based on the topological skeleton; The extraction module is used to extract local cross-sectional slices from the topological skeleton, match the local cross-sectional slices with the preset industrial steel specification library, and obtain the standard physical specifications of each component. The model optimization module is used to construct an energy function based on the identified node clusters and connecting rods, and to perform collinearity, coplanarity and symmetry corrections on the entire tower topology based on the energy function to obtain an optimized model. The entire tower topology includes node clusters and connecting rods. The drawing generation module is used to perform absolute physical scale mapping on the optimized model and generate engineering drawings for transmission towers by combining the standard physical specifications of each component.

[0013] Thirdly, the present invention provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the transmission tower drawing generation method described above.

[0014] Fourthly, the present invention provides a storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the transmission tower drawing generation method described above.

[0015] Compared with the prior art, the present invention has the following beneficial effects: The proposed method for generating transmission tower drawings has two main aspects. First, it acquires multi-scale image data covering the entire transmission tower and key structural nodes. This operation replaces the traditional manual high-altitude climbing and point-by-point measurement method, avoiding the safety hazards of manual high-altitude operations and eliminating the tedious process of manual on-site measurement, recording, and data processing. This significantly shortens the initial data acquisition cycle and improves the efficiency of generating transmission tower drawings. Second, it utilizes neural networks to learn the implicit density distribution of the transmission tower's spatial structure, constructs a neural radiation field model, and extracts structural surface point clouds with topological continuity from the neural radiation field model. The structural surface point clouds are then transformed into a topological skeleton, and node clusters and connecting rods are identified based on the topological skeleton. As can be seen, this invention eliminates the need for manual outlining of the tower body and division of node clusters and connecting rods, significantly reducing the workload of manual modeling and thus improving the efficiency of generating transmission tower drawings.

[0016] Furthermore, this invention discloses a method of acquiring multi-scale image data covering the entire transmission tower and key structural nodes by using a drone for multi-view circling flight. This design is safe and efficient, and can avoid the extremely high risks and costs of traditional manual tower climbing measurement.

[0017] Furthermore, this invention discloses a method of extracting structural surface point clouds with topological continuity from a neural radiation field model using isosurface sampling, which can overcome the problem of reconstruction fracture in slender rods. Attached Figure Description

[0018] Figure 1 This is a flowchart of the method for generating transmission tower drawings according to the present invention; Figure 2 This is a schematic diagram of the transmission tower drawing generation system of the present invention; Figure 3 This is a flowchart of the method for generating transmission tower drawings in Embodiment 4 of the present invention; Figure 4 This is a schematic diagram of a sparse point cloud in Embodiment 4 of the present invention; Figure 5 This is a schematic diagram of a dense point cloud in Embodiment 4 of the present invention; Figure 6 This is a schematic diagram of the point cloud mask in Embodiment 4 of the present invention; Figure 7 This is a schematic diagram of the origin cloud map covering the mask in Embodiment 4 of the present invention; Figure 8 This is a schematic diagram of the point cloud on the surface of the structure in Embodiment 4 of the present invention; Figure 9 This is a schematic diagram of the drawing generation result in Embodiment 4 of the present invention; Figure 10 This is a schematic diagram of the structure of the electronic device of the present invention. Detailed Implementation

[0019] To further understand the content of this invention, the invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments are merely illustrative and not limiting of the invention.

[0020] Example 1: The flowchart of the method for generating transmission tower drawings in this invention is as follows: Figure 1 As shown, the method for generating transmission tower drawings according to the present invention includes the following steps: S1. Acquire multi-scale image data covering the entire transmission tower and key structural nodes, and record it as the first data; S2. Based on the first data, construct a point cloud focusing on the main body of the transmission tower; S3. Utilize neural networks to learn the implicit density distribution of the transmission tower body space, construct a neural radiation field model, and extract structural surface point clouds with topological continuity from the neural radiation field model; S4. Convert the point cloud of the structural surface into a topological skeleton, and identify the node clusters and connecting rods based on the topological skeleton; S5. Extract local cross-sectional slices from the topological skeleton, match the local cross-sectional slices with the preset industrial steel specification library to obtain the standard physical specifications of each component; S6. Based on the identified node clusters and connecting rods, an energy function is constructed, and the collinearity, coplanarity, and symmetry of the entire tower topology are corrected based on the energy function to obtain an optimized model. The entire tower topology includes node clusters and connecting rods. S7. Perform absolute physical scale mapping on the optimized model, and generate transmission tower drawings by combining the standard physical specifications of each component.

[0021] Example 2: A schematic diagram of the transmission tower drawing generation system of the present invention is shown below. Figure 2 As shown, the transmission tower drawing generation system of the present invention includes: The data acquisition module is used to acquire multi-scale image data covering the entire transmission tower and key structural nodes, which is referred to as the first data. The dataset construction module is used to build a point cloud focusing on the main body of the transmission tower based on the first data; The point cloud extraction module is used to learn the implicit density distribution of the transmission tower body space using neural networks, construct a neural radiation field model, and extract structural surface point clouds with topological continuity from the neural radiation field model. The conversion module is used to convert the point cloud of the structural surface into a topological skeleton, and to identify the node clusters and connecting rods based on the topological skeleton; The extraction module is used to extract local cross-sectional slices from the topological skeleton, match the local cross-sectional slices with the preset industrial steel specification library, and obtain the standard physical specifications of each component. The model optimization module is used to construct an energy function based on the identified node clusters and connecting rods, and to perform collinearity, coplanarity and symmetry corrections on the entire tower topology based on the energy function to obtain an optimized model. The entire tower topology includes node clusters and connecting rods. The drawing generation module is used to perform absolute physical scale mapping on the optimized model and generate engineering drawings for transmission towers by combining the standard physical specifications of each component.

[0022] Example 3: The method for generating transmission tower drawings according to the present invention includes the following steps: S1. Acquire multi-scale image data covering the entire transmission tower and key structural nodes, denoted as the first data.

[0023] This step involves acquiring multi-scale image data covering the entire transmission tower and key structural nodes, specifically: The method of using drones to fly around the power transmission tower from multiple perspectives was adopted to acquire multi-scale image data covering the entire power transmission tower and key structural nodes.

[0024] S2. Based on the first data, construct a point cloud focusing on the main body of the transmission tower.

[0025] Based on the first data, a point cloud focusing on the main body of the transmission tower is constructed (the point cloud includes sparse point clouds and dense point clouds, such as...). Figure 4 and Figure 5 express) Before constructing a point cloud focusing on the main body of the transmission tower based on the first data in this step, the first data is preprocessed. The preprocessing includes frame de-framing, camera joint calibration, background interference removal, and color constancy processing of the first data.

[0026] S3. Utilize neural networks to learn the implicit density distribution of the transmission tower body space, construct a neural radiation field model, and extract structural surface point clouds with topological continuity from the neural radiation field model.

[0027] This step involves extracting topologically continuous structural surface point clouds from the neural radiation field model, specifically: We use isosurface sampling to extract topologically continuous structural surface point clouds from the neural radiation field model.

[0028] S4. Convert the point cloud of the structural surface into a topological skeleton, and identify the node clusters and connecting rods based on the topological skeleton.

[0029] S5. Extract local cross-sectional slices from the topological skeleton, match the local cross-sectional slices with the preset industrial steel specification library to obtain the standard physical specifications of each component.

[0030] S6. Based on the identified node clusters and connecting rods, an energy function is constructed, and the collinearity, coplanarity, and symmetry of the entire tower topology are corrected based on the energy function to obtain an optimized model. The entire tower topology includes node clusters and connecting rods.

[0031] S7. Perform absolute physical scale mapping on the optimized model, and generate transmission tower drawings by combining the standard physical specifications of each component.

[0032] This step involves mapping the optimized model to absolute physical scales and combining this with the standard physical specifications of each component to generate transmission tower drawings. Specifically: Spatial coordinate calculation and similarity transformation registration are performed on the optimized model, and transmission tower drawings are generated by combining the standard physical specifications of each component.

[0033] In this step, the transmission tower drawings include component cross-sectional parameters, actual cutting length, and spatial angle.

[0034] After generating the transmission tower drawings in this embodiment, a summary table of engineering materials and a list of component cutting lengths are generated.

[0035] Example 4: The method of the present invention will be described in detail below. The flowchart of the method for generating transmission tower drawings of the present invention is as follows: Figure 3 As shown, the method for generating transmission tower drawings according to the present invention includes the following steps: S1: Obtain the original video Given the characteristics of transmission towers—slender, hollow, and situated in complex outdoor electromagnetic environments—this step aims to address the issues of data completeness and reconstruction robustness.

[0036] This step involves designing a multi-scale adaptive flight strategy that combines overall and local approaches to capture high-overlap image sequences that meet the requirements for consistency across multiple perspectives of the implicit neural radiation field (NeRF), ensuring comprehensive coverage of structural details without blind spots.

[0037] First, the drone used needs to be equipped with a visual sensor with a resolution of no less than 2K and a frame rate of 60FPS, and the camera exposure and white balance parameters need to be locked to reduce the abrupt changes in brightness during subsequent volume rendering.

[0038] During the actual filming, a multi-layered spiral flight strategy with a large backscatter was adopted. The UAV used the tower center as the center and set 3 to 5 horizontal reference planes at different heights. On each height plane, the UAV gimbal pitch angle was controlled (e.g., set to -30°, -45°, and -60° respectively), and the UAV flew in a constant orbit at a constant speed. This stage ensured that the directional overlap between adjacent frames was greater than 80%, and the lateral overlap was greater than 70%, forming a stable global closed-loop trajectory, providing basic data for delineating the spatial boundary of the global implicit field. For the crossarms, tower head insulator attachment points, and densely intersecting areas of the transverse web members of the transmission tower, close-range hovering and multi-directional pan-and-tilt photography were performed. While ensuring a safe distance, complex nodes were supplemented with gridded shots in four directions (up, down, left, and right), focusing on capturing the thickness information and connection relationships of the angle steel at the nodes to avoid missing key structures in the final drawings.

[0039] S2: Semantic-guided pose estimation and data preprocessing After the video data acquisition is completed, the background environment is decoupled (also known as background interference elimination) by introducing deep learning semantic mask and image enhancement algorithm, providing a pure subject-object mapping relationship for the implicit field, and calculating the high-precision camera spatial pose.

[0040] Combining IMU (Inertial Measurement Unit) data recorded by the UAV with visual odometry, adaptive frame extraction (also known as frame de-framing) is performed on the original high frame rate video. Keyframes are extracted when the field of view change between adjacent frames exceeds 5° or the translation distance exceeds 0.5 meters, and redundant data is removed. Subsequently, Zhang Zhengyou's calibration method combined with bundle adjustment (also known as camera joint calibration) is used to adjust the camera's intrinsic parameters such as focal length and principal point, as well as radial distortion. k 1 ,k 2 ,k 3) and tangential distortion ( p 1 ,p 2) Perform joint computation and image distortion correction. Input the distortion-corrected keyframes into a pre-trained deep learning instance segmentation network. This deep learning instance segmentation network is specifically designed for feature extraction from transmission towers and metal truss structures, outputting a binarized semantic mask specific to the tower structure. A schematic diagram of the point cloud mask is shown below. Figure 6 As shown, the schematic diagram of the origin cloud map covering the mask is as follows: Figure 7 As shown.

[0041] In the subsequent neural field construction, this embodiment will only perform ray casting and color integration on the pixels within the mask, directly shielding the interference of dynamic backgrounds such as the sky and weeds at the mathematical level (also known as background interference elimination), thus significantly reducing computational resource consumption. Finally, to address the issues of reflection on the metal angle steel surface and high contrast on the backlit side, adaptive histogram equalization (CLAHE) and color constancy processing are performed on the effective mask area of ​​the keyframes to restore the texture in the dark areas. At the same time, the sharpness of each image is evaluated using the Laplacian variance. Once the variance value is lower than the preset sharpness threshold, the blurred frame is automatically discarded, ensuring that every ray input to the neural radiation field has a real and effective photometric constraint.

[0042] S3: Implicit Radiation Field Construction and Surface Extraction After completing the image dataset construction and initial camera pose estimation, the three-dimensional neural radiation field (NeRF) of the tower space is first constructed to achieve continuous reconstruction of the geometric surface. Specifically, in this embodiment, the pixels of the two-dimensional image are first back-projected into three-dimensional rays in space. ,in, It is a three-dimensional ray. t Let $o$ be the starting point of the 3D ray, $d$ be the direction vector of the 3D ray, and a multilayer perceptron (MLP) predict the volume density $σ$ and color $c$ at that point using the coordinates of sampling points on the ray and the viewpoint direction. Based on the predicted values, the desired color along the ray is... The desired color along this ray is calculated using a discretized volume rendering integral formula. The calculation formula is:

[0043] in, The cumulative transmittance represents the probability that light is not blocked. For the first i The color of each sampling point For the first i Volume density at each sampling point The step size of the sampling step. For all points on the 3D ray.

[0044] To enable the neural network to accurately represent real-world scenes, this embodiment constructs realistic image colors. With predicted color Photometric residual loss function ,in, Let the photometric residual loss function be . Three-dimensional ray The set, The network is trained using three-dimensional rays and backpropagation iterations. After the network converges and the spatial density field stabilizes, this embodiment sets a reasonable volume density threshold and uses the Moving Cubes algorithm to extract the spatial isosurfaces of the tower structure. Then, high-density uniform spatial sampling is performed on these isosurfaces, thereby transforming the implicit continuous field into a structural surface point cloud with topological continuity. A schematic diagram of the structural surface point cloud is shown below. Figure 8 As shown. This process fundamentally overcomes the problem of reconstruction fracture caused by the loss of local features at the small, hollowed-out web members.

[0045] S4: Component axis extraction and parametric identification After successfully acquiring continuous surface point clouds, PCA (Principal Component Analysis) is used to extract the principal axes of the topological skeleton and perform parametric identification for industrial specifications. This embodiment first locks the local point cloud set of a single member by selecting its start and end points. P sub And by calculating this point set (local point cloud set) P sub The covariance matrix of ) Eigenvalue decomposition is performed, and the eigenvector corresponding to the largest eigenvalue is established as the local principal axis direction of the member, where, For local point cloud collections P sub The covariance matrix, The number of points in the neighborhood. For sampling points within the neighborhood, This is the arithmetic mean of the coordinates of all points within this local neighborhood. Then, a two-dimensional point set is extracted along the normal plane perpendicular to this principal axis. Q The parameters are then compared with a pre-defined library of industrial standard steel specifications. This comparison process is transformed into a cross-section matching objective function based on Iterative Closest Point (ICP), and the expression for the cross-section matching objective function is as follows:

[0046] in, In order to target the k The minimum matching energy value (error) for a certain type of cross section (e.g., angle steel of a specific specification). It is a two-dimensional rotation matrix. It is a two-dimensional translation vector. Q For the set of points to be matched, this embodiment Q Two-dimensional points extracted from the surface of the reconstructed tower. This embodiment uses a standard template point set. Geometric contour points constructed for standard angle steel, For regularization terms, The optimization process is constrained based on prior knowledge of the cross-section to prevent the model from overfitting noise. This is the regularization coefficient, used to balance the weight between matching accuracy and constraint strength.

[0047] This embodiment traverses the specification library to find the template index that globally minimizes the objective function. k * This automatically establishes the precise physical model, cross-sectional shape, and thickness parameters of the current member.

[0048] S5: Topology Consistency Optimization After completing the parametric identification, in order to eliminate spatial defects caused by local point cloud noise, this embodiment abstracts the transmission tower as a spatial undirected graph containing nodes and connections. Spatial density clustering (DBSCAN) is performed on the pole endpoints extracted in step S4 to merge adjacent endpoints into a unified structural node. Subsequently, a global graph optimization energy function jointly driven by data approximation terms, rigid constraint terms, and engineering prior regularization terms is constructed. The expression of the global graph optimization energy function is as follows:

[0049] Among them, the data approximation term introduces the Huber robust kernel function. Constrain the nodes to not deviate from their initial spatial positions. ,Right now The rigid constraint term will match the standard steel section length obtained in step S4. Introducing the Laplace matrix forces the optimized member lengths to conform to the specification: Engineering Prior Regularization Terms Including collinearity constraints on the main materials, the bending of the main leg nodes is penalized through vector cross product: Furthermore, a quadruple rotational symmetry penalty is applied to the tower body. In this embodiment, the Jacobian matrix is ​​calculated and the Levenberg-Marquardt algorithm is used to iteratively solve the highly nonlinear energy function, ultimately outputting a high-precision three-dimensional topological skeleton that fully satisfies rigid mechanical constraints and symmetry specifications.

[0050] S6: Parametric Drawing Generation For a 3D topological skeleton that only possesses relative virtual space coordinates after global topology optimization, it is accurately mapped to the real physical space during the drawing generation stage. This embodiment introduces a reference distance as a scaling prior for the absolute scale, and performs a similarity transformation registration across the entire spatial scale on the optimized skeleton model, thereby establishing an unbiased absolute engineering coordinate system.

[0051] After the physical benchmark is established, this embodiment accurately calculates the core geometric parameters such as the actual axial length, spatial assembly angle, and connection hole distance of each major component of the tower body based on the converged three-dimensional spatial coordinates of the nodes.

[0052] Subsequently, this embodiment deeply integrates these precise geometric data with the standard steel physical specifications assigned by the cross-section matching in the previous stage. Based on both geometric and semantic attributes, this embodiment applies the principle of spatial geometric orthographic projection transformation to automatically map and generate standardized industrial CAD drawings containing a front view, side view, top view, and key local sectional views. A schematic diagram of the generated drawings is shown below. Figure 9 As shown. In addition, in order to directly connect with actual construction operations and subsequent finite element analysis (FEA), this embodiment simultaneously extracts and exports a precise summary table of engineering materials and a list of component cutting lengths while outputting graphic files, thus completely opening up a closed-loop technology process from UAV generalized image acquisition to the automatic generation of high-precision, quantifiable industrial-grade drawings.

[0053] Example 5: Please see Figure 10 As shown, the present invention also provides an electronic device 100 for a method of generating transmission tower drawings; the electronic device 100 includes a memory 101, at least one processor 102, a computer program 103 stored in the memory 101 and executable on the at least one processor 102, and at least one communication bus 104.

[0054] The memory 101 can be used to store the computer program 103. The processor 102 implements the steps of the transmission tower drawing generation method described in Embodiment 1 by running or executing the computer program stored in the memory 101 and calling the data stored in the memory 101. The memory 101 may mainly include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the electronic device 100 (such as audio data), etc. In addition, the memory 101 may include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other non-volatile solid-state storage device.

[0055] The at least one processor 102 may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The processor 102 may be a microprocessor or any conventional processor. The processor 102 is the control center of the electronic device 100, connecting various parts of the electronic device 100 via various interfaces and lines.

[0056] The memory 101 in the electronic device 100 stores multiple instructions to implement a method for generating transmission tower drawings, and the processor 102 can execute the multiple instructions to achieve the following: Acquire multi-scale image data covering the entire transmission tower and key structural nodes, and record it as the first data; Based on the first data, a point cloud focusing on the main body of the transmission tower is constructed; We use neural networks to learn the implicit density distribution of the transmission tower body space, construct a neural radiation field model, and extract structural surface point clouds with topological continuity from the neural radiation field model. The point cloud of the structural surface is converted into a topological skeleton, and node clusters and connecting rods are identified based on the topological skeleton. Extract local cross-sectional slices from the topological skeleton, match the local cross-sectional slices with the preset industrial steel specification library to obtain the standard physical specifications of each component; Based on the identified node clusters and connecting rods, an energy function is constructed, and the collinearity, coplanarity, and symmetry of the entire tower topology are corrected based on the energy function to obtain an optimized model. The entire tower topology includes node clusters and connecting rods. The optimized model is mapped to absolute physical scale, and the standard physical specifications of each component are combined to generate transmission tower drawings.

[0057] Example 6: If the modules / units integrated in the electronic device 100 are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, and a read-only memory (ROM).

[0058] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0059] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0060] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0061] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0062] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for generating transmission tower drawings, characterized in that, Includes the following steps: Acquire multi-scale image data covering the entire transmission tower and key structural nodes, and record it as the first data; Based on the first data, a point cloud focusing on the main body of the transmission tower is constructed; We use neural networks to learn the implicit density distribution of the transmission tower body space, construct a neural radiation field model, and extract structural surface point clouds with topological continuity from the neural radiation field model. The point cloud of the structural surface is converted into a topological skeleton, and node clusters and connecting rods are identified based on the topological skeleton. Extract local cross-sectional slices from the topological skeleton, match the local cross-sectional slices with the preset industrial steel specification library to obtain the standard physical specifications of each component; Based on the identified node clusters and connecting rods, an energy function is constructed, and the collinearity, coplanarity, and symmetry of the entire tower topology are corrected based on the energy function to obtain an optimized model. The entire tower topology includes node clusters and connecting rods. The optimized model is mapped to absolute physical scale, and the standard physical specifications of each component are combined to generate transmission tower drawings.

2. The method for generating transmission tower drawings according to claim 1, characterized in that, The acquisition of multi-scale image data covering the entire transmission tower and key structural nodes specifically includes: The method of using drones to fly around the power transmission tower from multiple perspectives was adopted to acquire multi-scale image data covering the entire power transmission tower and key structural nodes.

3. The method for generating transmission tower drawings according to claim 1, characterized in that, Before constructing a point cloud focusing on the main body of the transmission tower based on the first data, the first data is preprocessed. The preprocessing includes frame de-framing, camera joint calibration, background interference removal, and color constancy processing of the first data.

4. The method for generating transmission tower drawings according to claim 1, characterized in that, The extraction of topologically continuous structural surface point clouds from the neural radiation field model specifically involves: We use isosurface sampling to extract topologically continuous structural surface point clouds from the neural radiation field model.

5. The method for generating transmission tower drawings according to claim 1, characterized in that, The process involves mapping the optimized model to absolute physical scales and combining this with the standard physical specifications of each component to generate transmission tower drawings. Specifically: Spatial coordinate calculation and similarity transformation registration are performed on the optimized model, and transmission tower drawings are generated by combining the standard physical specifications of each component.

6. The method for generating transmission tower drawings according to claim 1, characterized in that, The transmission tower drawings include component cross-sectional parameters, actual cutting length, and spatial angle.

7. The method for generating transmission tower drawings according to claim 1, characterized in that, After generating the transmission tower drawings, a summary table of engineering materials and a list of component cutting lengths are generated.

8. A system for generating drawings of transmission towers, characterized in that, include: The data acquisition module is used to acquire multi-scale image data covering the entire transmission tower and key structural nodes, which is referred to as the first data. The dataset construction module is used to build a point cloud focusing on the main body of the transmission tower based on the first data; The point cloud extraction module is used to learn the implicit density distribution of the transmission tower body space using neural networks, construct a neural radiation field model, and extract structural surface point clouds with topological continuity from the neural radiation field model. The conversion module is used to convert the point cloud of the structural surface into a topological skeleton, and to identify the node clusters and connecting rods based on the topological skeleton; The extraction module is used to extract local cross-sectional slices from the topological skeleton, match the local cross-sectional slices with the preset industrial steel specification library, and obtain the standard physical specifications of each component. The model optimization module is used to construct an energy function based on the identified node clusters and connecting rods, and to perform collinearity, coplanarity and symmetry corrections on the entire tower topology based on the energy function to obtain an optimized model. The entire tower topology includes node clusters and connecting rods. The drawing generation module is used to perform absolute physical scale mapping on the optimized model and generate engineering drawings for transmission towers by combining the standard physical specifications of each component.

9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method for generating transmission tower drawings according to any one of claims 1 to 7.

10. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the transmission tower drawing generation method according to any one of claims 1 to 7.

Images