Electric power engineering construction quality evaluation method and device based on three-dimensional Gaussian modeling, equipment and medium

Abstract

The invention discloses an electric power engineering construction quality evaluation method and device based on three-dimensional Gaussian modeling, equipment and a medium. The method comprises the steps of abstracting a plurality of pieces of equipment in electric power engineering into a physical parameter group, and performing multi-view image fusion based on images of the electric power engineering; modeling is carried out based on a 3DGS three-dimensional scene, parameterized Gaussian distribution is used for representing the geometric morphology of the power equipment, and mapping of physical parameters and the three-dimensional scene is achieved; projecting the 3DGS model into a two-dimensional rendered image based on differentiable rendering, and realizing gradient propagation from pixels of the two-dimensional rendered image to physical parameters; constructing a loss function, and optimizing Gaussian function parameters and physical parameters based on the image of the electric power engineering and the two-dimensional rendering image; and after the optimization is completed, carrying out physical parameter extraction on the 3DGS model corresponding to the electric power engineering so as to realize quality evaluation, graded early warning or visual display. The invention belongs to the field of engineering evaluation. According to the invention, accurate evaluation of the construction quality can be realized.

Description

Technical Field

[0001] This invention relates to the field of engineering evaluation, and in particular to a method, apparatus, equipment, and medium for evaluating the construction quality of power engineering based on three-dimensional Gaussian modeling. Background Technology

[0002] In the field of power engineering construction quality assessment, traditional methods mainly rely on manual inspection, two-dimensional image comparison, or three-dimensional reconstruction techniques based on meshes / voxels. However, these methods suffer from drawbacks such as low efficiency and high subjectivity in manual inspection, high complexity in three-dimensional reconstruction techniques, and weak correlation between images and parameters.

[0003] With the development of computer vision and graphics technologies, 3D Gaussian sputtering (3DGS) has emerged as a new scene representation method, boasting strong explicit geometric modeling capabilities and differentiable rendering characteristics. However, existing research largely focuses on general scene modeling and has not yet systematically integrated methods to address the specific needs of power engineering. Therefore, there is an urgent need for a power engineering construction quality assessment method based on 3D Gaussian modeling to achieve accurate evaluation of power engineering construction quality. Summary of the Invention

[0004] This invention provides a method, apparatus, equipment, and medium for evaluating the construction quality of power engineering based on three-dimensional Gaussian modeling. It solves the technical problem of how to achieve engineering quality evaluation based on three-dimensional Gaussian sputtering in the prior art, and achieves the technical effect of accurate evaluation of engineering quality based on three-dimensional Gaussian sputtering.

[0005] In a first aspect, the present invention provides a method for evaluating the construction quality of power engineering based on three-dimensional Gaussian modeling, including: Several devices in power engineering are abstracted into physical parameter groups, and multi-view image fusion is performed based on the images of power engineering, where the physical parameters correspond one-to-one with the devices; Based on 3DGS three-dimensional scene modeling, the geometric shape of power equipment is represented by a parametric Gaussian distribution, and the mapping between physical parameters and three-dimensional scene is realized. Based on differentiable rendering, the 3DGS model is projected into a two-dimensional rendered image, and gradient propagation from the pixels of the two-dimensional rendered image to the physical parameters is realized. A loss function is constructed, and the Gaussian function parameters and physical parameters are optimized based on images of power engineering and two-dimensional rendered images; Once optimization is complete, physical parameters are extracted from the 3DGS model corresponding to the power engineering project to achieve quality assessment, graded early warning, or visualization.

[0006] Furthermore, based on differentiable rendering, the 3DGS model is projected into a 2D rendered image, and gradient propagation from pixels of the 2D rendered image to physical parameters is implemented, including: The 3DGS model is projected onto a two-dimensional plane to generate a two-dimensional rendered image, which corresponds one-to-one with the image of the power engineering project. The color value of each pixel in the 2D rendered image is obtained by integrating all Gaussian functions along the viewing direction, including:

[0007] in, For the first The opacity of a Gaussian function, For the first A Gaussian function color value, For the first The opacity of a Gaussian function, The color value after integration. The number of Gaussian functions.

[0008] Furthermore, the loss function is constructed, including:

[0009] in, For the total loss, For loss of photometric uniformity, and All are weights. For geometric consistency loss, This represents the loss due to physical constraints.

[0010] Furthermore, the Gaussian function parameters and physical parameters are optimized based on images from power engineering and 2D rendered images, including: Based on the loss function, the total loss between the power engineering image and the corresponding two-dimensional rendered image is determined; If the total loss exceeds the preset loss threshold, the Gaussian function parameters and physical parameters are iteratively updated based on the adaptive gradient optimization method until convergence to the optimum.

[0011] Furthermore, it also includes: If the total loss between all power engineering images and their corresponding 2D rendered images is less than a preset loss threshold, then determine whether the sum of the total losses is less than the preset total loss threshold. If the value is less than 1, physical parameters are extracted from the 3DGS model corresponding to the power engineering project. If the total loss threshold is greater than or equal to the preset threshold, then all Gaussian function parameters and all physical parameters are iteratively updated based on the adaptive gradient optimization method until convergence to the optimum.

[0012] Furthermore, after optimization, physical parameters are extracted from the 3DGS model corresponding to the power engineering project to achieve quality assessment, graded early warning, or visualization, including: After extracting the physical parameters, the physical parameters are compared with the preset standard parameters to determine the deviation value corresponding to each physical parameter; The quality and early warning level of the power project are determined based on the relationship between the sum of the deviation values ​​and the preset deviation threshold.

[0013] Furthermore, based on 3DGS three-dimensional scene modeling, the geometric shape of the power equipment is represented by a parametric Gaussian distribution, and the mapping between physical parameters and the three-dimensional scene is realized, including: The geometric structure is abstracted as a Gaussian function, where the geometric structure of power equipment is represented by a Gaussian function or a superposition of Gaussian functions; Construct parameterized constraint relationships between the Gaussian function of power equipment and its physical parameters.

[0014] Secondly, the present invention provides a power engineering construction quality assessment device based on three-dimensional Gaussian modeling, comprising: The physical representation module is used to abstract several devices in power engineering into a group of physical parameters and perform multi-view image fusion based on the images of power engineering. The physical parameters correspond one-to-one with the devices. The Gaussian characterization module is used for 3D scene modeling based on 3DGS to characterize the geometric shape of power equipment with a parametric Gaussian distribution and realize the mapping between physical parameters and the 3D scene. The differential rendering module is used to project a 3DGS model into a 2D rendered image based on differentiable rendering, and to implement gradient propagation from the pixels of the 2D rendered image to the physical parameters. The loss optimization module is used to construct the loss function and optimize the Gaussian function parameters and physical parameters based on the power engineering images and 2D rendered images; The results evaluation module is used to extract physical parameters from the 3DGS model corresponding to the power engineering project after optimization, so as to achieve quality assessment, graded early warning or visualization.

[0015] Thirdly, the present invention provides an electronic device, comprising: processor; Memory used to store processor-executable instructions; The processor is configured to execute a power engineering construction quality assessment method based on three-dimensional Gaussian modeling, as provided in the first aspect.

[0016] Fourthly, the present invention provides a non-transitory computer-readable storage medium, wherein when the instructions in the non-transitory computer-readable storage medium are executed by a processor of an electronic device, the electronic device is able to execute the power engineering construction quality assessment method based on three-dimensional Gaussian modeling as provided in the first aspect.

[0017] One or more technical solutions provided in this invention have at least the following technical effects or advantages: This invention abstracts equipment in power engineering into a parameterized three-dimensional Gaussian model (3DGS model). It utilizes multi-view image fusion technology to accurately represent the geometric shape of the power equipment and map its physical parameters to the three-dimensional scene. Differentiable rendering technology projects the 3DGS model into a two-dimensional rendered image, achieving gradient propagation from image pixels to the equipment's physical parameters. This allows for the construction of a loss function to optimize the Gaussian function parameters and the physical parameters. This invention not only enables objective evaluation and graded early warning of power engineering construction quality based on the optimized 3DGS model, but also helps to quickly locate potential problems through visualization, greatly improving the efficiency and accuracy of construction quality inspection in power engineering and contributing to the prevention of potential risks.

[0018] This invention utilizes differentiable rendering technology to project a 3DGS model containing multiple Gaussian functions describing its shape and appearance into a 2D rendered image. The color value of each pixel is calculated by integrating all relevant Gaussian functions along the viewing direction, simulating the effect of light passing through a semi-transparent layer. This not only enhances the realism of the rendering but also allows for precise tracking of the relationship between each pixel's error and the 3DGS model parameters. By constructing a loss function that comprehensively considers photometric consistency, geometric consistency, and physical constraints, this invention optimizes the Gaussian function parameters and physical parameters, ensuring a high degree of consistency between the rendered image and the actual power engineering image. Adaptive gradient optimization methods (such as the Adam optimizer) iteratively update parameters until convergence, simultaneously optimizing the parameters of thousands of Gaussian functions and equipment physical parameters to achieve global optimization. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 A flowchart illustrating the power engineering construction quality assessment method based on three-dimensional Gaussian modeling provided by this invention; Figure 2 A flowchart illustrating another method for evaluating the construction quality of power engineering based on three-dimensional Gaussian modeling provided by this invention. Detailed Implementation

[0021] This invention provides a method for evaluating the construction quality of power engineering based on three-dimensional Gaussian modeling, thus solving the technical problem of how to achieve engineering quality evaluation based on three-dimensional Gaussian sputtering in the prior art.

[0022] The technical solution of this invention is to solve the above-mentioned technical problems, and the overall idea is as follows: A method for assessing the construction quality of power engineering based on 3D Gaussian modeling includes: abstracting several devices in the power engineering project into a set of physical parameters and performing multi-view image fusion based on the images of the power engineering project, wherein the physical parameters correspond one-to-one with the devices; modeling a 3D scene based on 3DGS, using a parametric Gaussian distribution to represent the geometric shape of the power equipment, and realizing the mapping between physical parameters and the 3D scene; projecting the 3DGS model into a 2D rendered image based on differentiable rendering, and realizing gradient propagation from the pixels of the 2D rendered image to the physical parameters; constructing a loss function and optimizing the Gaussian function parameters and physical parameters based on the images of the power engineering project and the 2D rendered image; after optimization, extracting physical parameters from the 3DGS model corresponding to the power engineering project to achieve quality assessment, graded early warning, or visualization.

[0023] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.

[0024] First, it should be clarified that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0025] This invention provides, for example Figure 1 The power engineering construction quality assessment method based on 3D Gaussian modeling shown includes steps S11-S14: Step S11: Abstract several devices in the power engineering into a group of physical parameters and perform multi-view image fusion based on the image of the power engineering, wherein the physical parameters correspond one-to-one with the devices.

[0026] Specifically, equipment in power engineering (such as transmission towers, crossarms, conductors, insulator strings, etc.) can be abstracted into optimizable physical parameters, including the verticality angle θ of the tower, the horizontal deviation angle α of the crossarm, the sag coefficient f of the conductor, and the equipment spacing d. The types of equipment can be related to the construction quality; that is, any equipment related to the construction quality can be abstracted into a single physical parameter.

[0027] By using parametric geometric models, a direct mathematical mapping relationship is established between each physical parameter and the three-dimensional spatial morphology of the equipment. This allows information about these parameters to be extracted directly from images of power engineering projects without the need for the complex three-dimensional reconstruction process of traditional methods.

[0028] In terms of multi-view fusion processing, multiple images from different shooting angles can be processed simultaneously. The system establishes a unified world coordinate system to calibrate and align camera parameters from different viewpoints, ensuring the consistency of the 3DGS model in space. Multi-view constraints not only improve the accuracy of parameter estimation but also enhance the system's robustness under partial occlusion conditions. When a device is occluded from a single viewpoint, information from other viewpoints can supplement the constraints, ensuring the stability and accuracy of parameter optimization.

[0029] In other words, the purpose of step S11 is to provide multi-view images and optimizable physical parameters for each device for 3DGS 3D scene modeling.

[0030] Step S12: Based on 3DGS three-dimensional scene modeling, the geometric shape of the power equipment is represented by a parametric Gaussian distribution, and the mapping between physical parameters and the three-dimensional scene is realized.

[0031] Specifically, this includes: abstracting the geometric structure into a Gaussian function, where the geometric structure of the power equipment is represented by a Gaussian function or a superposition of Gaussian functions; and constructing a parameterized constraint relationship between the Gaussian function of the power equipment and the physical parameters.

[0032] Specifically, the geometric structure of the power equipment is first abstracted into a series of Gaussian functions or superpositions of functions. Each Gaussian function is defined by parameters such as position mean, covariance matrix, opacity, and color, thereby accurately representing the complex geometry of the equipment.

[0033] Establish parameterized constraint relationships between Gaussian functions and key physical parameters (such as tower verticality angle, crossarm horizontal deviation angle, conductor sag coefficient, and equipment spacing).

[0034] For example, the verticality of a tower directly affects the spatial distribution of its corresponding Gaussian function and the direction of its covariance matrix, while the sag of a conductor is accurately simulated by adjusting the vertical position of the Gaussian function along the conductor path.

[0035] Unlike traditional mesh or voxel representations, this invention enables efficient and accurate mapping directly from physical parameters to 3D scenes.

[0036] 3DGS is an abbreviation for 3D Gaussian Splatting. 3DGS is a 3D scene representation and rendering technology. Its core idea is to use a large number of 3D Gaussian functions with attributes such as position, shape, color, and transparency to represent and render complex scenes. Unlike traditional mesh or voxel modeling methods, 3DGS does not rely on explicit geometric structures. Instead, it accurately expresses the geometry and appearance of objects through the distribution, orientation, and color superposition of tens of thousands of Gaussian spheres in space. Gaussian functions can be efficiently projected onto a 2D image plane for differentiable rendering, thereby achieving high-quality new perspective compositing and supporting the back-optimization of scene parameters from image errors.

[0037] Step S13: Based on differentiable rendering, the 3DGS model is projected into a two-dimensional rendered image, and gradient propagation from the pixels of the two-dimensional rendered image to the physical parameters is realized.

[0038] Specifically, this includes: projecting the 3DGS model onto a two-dimensional plane to generate a two-dimensional rendered image, where each rendered image corresponds one-to-one with an image of the electrical engineering project; and for the color value of each pixel in the two-dimensional rendered image, integrating all Gaussian functions along the viewing direction, including:

[0039] in, For the first The opacity of a Gaussian function, For the first A Gaussian function color value, For the first The opacity of a Gaussian function, The color value after integration. The number of Gaussian functions.

[0040] The 3DGS model is a power engineering construction scene, which contains multiple Gaussian functions to describe its shape and appearance. Each Gaussian function has a color value and opacity.

[0041] For each pixel in the generated 2D rendered image, its color value is obtained by integrating over all relevant Gaussian functions.

[0042] The formula above simulates the effect of light passing through a series of semi-transparent layers, each of which contributes to the final color but also blocks light from the layers behind it.

[0043] After generating a 2D rendered image through the above process, the parameters of the 3DGS model can be optimized to make the rendered image as close as possible to the target image.

[0044] To this end, the gradient from image errors to 3DGS model parameters can be calculated, allowing these parameters to be adjusted in subsequent optimizations. Since the entire rendering process is differentiable, this means it's possible to precisely track how the error of each pixel depends on the parameter variations of the individual Gaussian functions in the 3DGS model.

[0045] Step S14: Construct a loss function and optimize the Gaussian function parameters and physical parameters based on the power engineering image and the two-dimensional rendered image.

[0046] Constructing the loss function includes:

[0047] in, For the total loss, For loss of photometric uniformity, and All are weights. For geometric consistency loss, This represents the loss due to physical constraints.

[0048] Photometric consistency loss measures the color difference between two images at the pixel level; geometric consistency loss measures the difference in the accuracy of the projection position of the geometric features of the device (such as edges, corners, etc.) in the image; physical constraint loss measures the difference between the physical laws and construction specifications of power engineering.

[0049] Optimization of Gaussian function parameters and physical parameters based on power engineering images and 2D rendered images, including: Based on the loss function, the total loss between the power engineering image and the corresponding 2D rendered image is determined; if the total loss is greater than the preset loss threshold, the Gaussian function parameters and physical parameters are iteratively updated based on the adaptive gradient optimization method until convergence to the optimum.

[0050] The gradient of the loss function with respect to each physical parameter can be calculated using the backpropagation algorithm, and the parameter values ​​can be iteratively updated using an adaptive gradient optimization method (such as the Adam optimizer) until the optimal solution is converged.

[0051] In this process, the parameters of thousands of Gaussian functions and the physical parameters of the device can be optimized simultaneously to achieve global optimization.

[0052] Understandably, in the above process, only the total loss between a pair of images and the corresponding 2D rendered image was calculated, and optimizations were performed based on this.

[0053] In addition to directly extracting the physical parameters of the device from the optimized pair of images and the corresponding 2D rendered image (which can be directly extracted if the loss is less than a preset threshold), to improve overall accuracy, the following steps are also included: if the total loss between all power engineering images and their corresponding 2D rendered images is less than a preset threshold, then determine whether the sum of the total losses is less than a preset total loss threshold; if it is less, then extract the physical parameters of the 3DGS model corresponding to the power engineering; if it is greater than or equal to the preset total loss threshold, then iteratively update all Gaussian function parameters and all physical parameters based on the adaptive gradient optimization method until convergence to the optimum.

[0054] Both the preset loss threshold and the preset total loss threshold can be set according to the actual situation, and this invention does not impose any restrictions.

[0055] Step S15: After optimization is completed, physical parameters are extracted from the 3DGS model corresponding to the power engineering project to achieve quality assessment, graded early warning or visualization.

[0056] This includes: after extracting physical parameters, comparing the physical parameters with preset standard parameters to determine the deviation value corresponding to each physical parameter; and determining the quality and early warning level of the power project based on the relationship between the sum of the deviation values ​​and the preset deviation threshold.

[0057] To further improve the system's practicality and accuracy, after parameter optimization, the back-derived equipment parameters are compared with the preset standard parameters in the BIM model to calculate the deviation values ​​corresponding to each physical parameter.

[0058] The convergence threshold and constraint weights of the optimization algorithm can be adjusted according to the accuracy requirements of different power equipment. Simultaneously, an intelligent early warning mechanism is established. When the detected parameter deviation exceeds a preset threshold, a quality problem report can be generated, marking the specific problem location, deviation value, and determining the corresponding early warning level based on the deviation value (the higher the deviation value, the higher the early warning level), providing intuitive and visual feedback for construction quality control.

[0059] In practical applications, a dynamic weight adjustment strategy can be adopted to address the complex environmental conditions at power construction sites. This strategy adjusts the weight coefficients of each loss item in real time based on factors such as image quality, lighting conditions, weather conditions, and degree of occlusion.

[0060] For example, increasing the weight of geometric constraints under strong light conditions. Increase the weight of physical constraints in cloudy or smoggy weather. This ensures that reliable test results can be obtained under various environmental conditions.

[0061] Furthermore, coarse-grained global parameters can be optimized first, followed by progressively refining local parameters, ensuring both optimization efficiency and improved final accuracy. The entire optimization process employs a multi-scale pyramid structure, gradually increasing from low-resolution images to high-resolution ones, effectively avoiding local optima and improving the stability and accuracy of parameter convergence.

[0062] For more information on the method and flow provided by this invention, please refer to [link / reference]. Figure 2 .

[0063] In summary, this invention abstracts equipment in power engineering into a parameterized three-dimensional Gaussian model (3DGS model), and utilizes multi-view image fusion technology to achieve accurate representation of the geometric shape of power equipment and mapping of physical parameters to the three-dimensional scene. Differentiable rendering technology is used to project the 3DGS model into a two-dimensional rendered image, realizing gradient propagation from image pixels to equipment physical parameters, thereby constructing a loss function to optimize the Gaussian function parameters and physical parameters. This invention not only enables objective evaluation and graded early warning of power engineering construction quality based on the optimized 3DGS model, but also helps to quickly locate potential problems through visualization, greatly improving the efficiency and accuracy of construction quality inspection in power engineering, while also helping to prevent potential risks.

[0064] This invention utilizes differentiable rendering technology to project a 3DGS model containing multiple Gaussian functions describing its shape and appearance into a 2D rendered image. The color value of each pixel is calculated by integrating all relevant Gaussian functions along the viewing direction, simulating the effect of light passing through a semi-transparent layer. This not only enhances the realism of the rendering but also allows for precise tracking of the relationship between each pixel's error and the 3DGS model parameters. By constructing a loss function that comprehensively considers photometric consistency, geometric consistency, and physical constraints, this invention optimizes the Gaussian function parameters and physical parameters, ensuring a high degree of consistency between the rendered image and the actual power engineering image. Adaptive gradient optimization methods (such as the Adam optimizer) iteratively update parameters until convergence, simultaneously optimizing the parameters of thousands of Gaussian functions and equipment physical parameters to achieve global optimization.

[0065] Based on the same inventive concept, this invention provides a power engineering construction quality assessment device based on three-dimensional Gaussian modeling, comprising: The physical representation module is used to abstract several devices in power engineering into a group of physical parameters and perform multi-view image fusion based on the images of power engineering. The physical parameters correspond one-to-one with the devices. The Gaussian characterization module is used for 3D scene modeling based on 3DGS to characterize the geometric shape of power equipment with a parametric Gaussian distribution and realize the mapping between physical parameters and the 3D scene. The differential rendering module is used to project a 3DGS model into a 2D rendered image based on differentiable rendering, and to implement gradient propagation from the pixels of the 2D rendered image to the physical parameters. The loss optimization module is used to construct the loss function and optimize the Gaussian function parameters and physical parameters based on the power engineering images and 2D rendered images; The results evaluation module is used to extract physical parameters from the 3DGS model corresponding to the power engineering project after optimization, so as to achieve quality assessment, graded early warning or visualization.

[0066] Based on the same inventive concept, the present invention also provides an electronic device, comprising: processor; Memory used to store processor-executable instructions; The processor is configured to execute a power engineering construction quality assessment method based on 3D Gaussian modeling, as described above.

[0067] Based on the same inventive concept, the present invention also provides a non-transitory computer-readable storage medium, which, when the instructions in the storage medium are executed by the processor of an electronic device, enables the electronic device to execute the power engineering construction quality assessment method based on three-dimensional Gaussian modeling as described above.

[0068] Since the electronic device described in this embodiment is an electronic device used to implement the information processing method in the embodiments of the present invention, those skilled in the art can understand the specific implementation methods and various variations of the electronic device in this embodiment based on the information processing method described in the embodiments of the present invention. Therefore, how the electronic device implements the method in the embodiments of the present invention will not be described in detail here. Any electronic device used by those skilled in the art to implement the information processing method in the embodiments of the present invention falls within the scope of protection of the present invention.

[0069] 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.

[0070] 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.

[0071] 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.

[0072] 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.

[0073] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0074] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for evaluating the construction quality of power engineering based on three-dimensional Gaussian modeling, characterized in that, include: Several devices in power engineering are abstracted into physical parameter groups, and multi-view image fusion is performed based on the images of power engineering, where the physical parameters correspond one-to-one with the devices; Based on 3DGS three-dimensional scene modeling, the geometric shape of power equipment is represented by a parametric Gaussian distribution, and the mapping between physical parameters and three-dimensional scene is realized. Based on differentiable rendering, the 3DGS model is projected into a two-dimensional rendered image, and gradient propagation from the pixels of the two-dimensional rendered image to the physical parameters is realized. A loss function is constructed, and the Gaussian function parameters and physical parameters are optimized based on the power engineering image and the two-dimensional rendered image. Once optimization is complete, physical parameters are extracted from the 3DGS model corresponding to the power engineering project to achieve quality assessment, graded early warning, or visualization.

2. The method for evaluating the construction quality of power engineering based on three-dimensional Gaussian modeling as described in claim 1, characterized in that, Based on differentiable rendering, a 3DGS model is projected into a 2D rendered image, and gradient propagation from pixels of the 2D rendered image to physical parameters is implemented, including: The 3DGS model is projected onto a two-dimensional plane to generate a two-dimensional rendered image, which corresponds one-to-one with the image of the power engineering project. The color value of each pixel in the 2D rendered image is obtained by integrating all Gaussian functions along the viewing direction, including: in, For the first The opacity of a Gaussian function, For the first A Gaussian function color value, For the first The opacity of a Gaussian function, The color value after integration. The number of Gaussian functions.

3. The method for evaluating the construction quality of power engineering based on three-dimensional Gaussian modeling as described in claim 1, characterized in that, Constructing the loss function includes: in, For the total loss, For loss of photometric uniformity, and All are weights. For geometric consistency loss, This represents the loss due to physical constraints.

4. The power engineering construction quality assessment method based on three-dimensional Gaussian modeling as described in claim 3, characterized in that, The Gaussian function parameters and physical parameters are optimized based on the image from the power engineering project and the two-dimensional rendered image, including: Based on the loss function, the total loss between the power engineering image and the corresponding two-dimensional rendered image is determined; If the total loss exceeds the preset loss threshold, the Gaussian function parameters and physical parameters are iteratively updated based on the adaptive gradient optimization method until convergence to the optimum.

5. The power engineering construction quality assessment method based on three-dimensional Gaussian modeling as described in claim 4, characterized in that, Also includes: If the total loss between all power engineering images and their corresponding 2D rendered images is less than a preset loss threshold, then determine whether the sum of the total losses is less than the preset total loss threshold. If the value is less than 1, physical parameters are extracted from the 3DGS model corresponding to the power engineering project. If the total loss threshold is greater than or equal to the preset threshold, then all Gaussian function parameters and all physical parameters are iteratively updated based on the adaptive gradient optimization method until convergence to the optimum.

6. The method for evaluating the construction quality of power engineering based on three-dimensional Gaussian modeling as described in claim 1, characterized in that, Once optimization is complete, physical parameters are extracted from the 3DGS model corresponding to the power engineering project to achieve quality assessment, graded early warning, or visualization, including: After extracting the physical parameters, the physical parameters are compared with the preset standard parameters to determine the deviation value corresponding to each physical parameter; The quality and early warning level of the power project are determined based on the relationship between the sum of the deviation values ​​and the preset deviation threshold.

7. The method for evaluating the construction quality of power engineering based on three-dimensional Gaussian modeling as described in claim 1, characterized in that, Based on 3DGS 3D scene modeling, the geometric shape of power equipment is represented by a parametric Gaussian distribution, and the mapping between physical parameters and the 3D scene is realized, including: The geometric structure is abstracted as a Gaussian function, where the geometric structure of power equipment is represented by a Gaussian function or a superposition of Gaussian functions; Construct parameterized constraint relationships between the Gaussian function of power equipment and its physical parameters.

8. A power engineering construction quality assessment device based on three-dimensional Gaussian modeling, characterized in that, include: The physical representation module is used to abstract several devices in power engineering into a group of physical parameters and perform multi-view image fusion based on the images of power engineering. The physical parameters correspond one-to-one with the devices. The Gaussian characterization module is used for 3D scene modeling based on 3DGS to characterize the geometric shape of power equipment with a parametric Gaussian distribution and realize the mapping between physical parameters and the 3D scene. The differential rendering module is used to project a 3DGS model into a 2D rendered image based on differentiable rendering, and to implement gradient propagation from the pixels of the 2D rendered image to the physical parameters. The loss optimization module is used to construct a loss function and optimize the Gaussian function parameters and physical parameters based on the power engineering image and the two-dimensional rendered image. The results evaluation module is used to extract physical parameters from the 3DGS model corresponding to the power engineering project after optimization, so as to achieve quality assessment, graded early warning or visualization.

9. An electronic device, characterized in that, include: processor; Memory used to store the processor's executable instructions; The processor is configured to execute the power engineering construction quality assessment method based on three-dimensional Gaussian modeling as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium, characterized in that, When the instructions in the non-transitory computer-readable storage medium are executed by the processor of the electronic device, the electronic device is able to perform the power engineering construction quality assessment method based on three-dimensional Gaussian modeling as described in any one of claims 1 to 7.

Images