Data fusion method, system and equipment of power transmission and transformation equipment and medium

Abstract

The invention discloses a data fusion method, system and device for power transmission and transformation equipment and a medium, and the method comprises the steps: obtaining the data of the power transmission and transformation equipment through a sensor, and extracting a multi-scale visual feature set based on a Gaussian filter; fitting the laser point cloud data through a principal component analysis method to generate a multi-scale point cloud feature set, and performing feature matching through a nearest neighbor search algorithm; and converting the visual image and the laser point cloud data into a common coordinate system, and fusing the laser points and the pixel points according to a preset weight to generate multi-source fusion data. According to the invention, by fusing the visual image and the laser point cloud data, the multi-dimensional information of the power transmission and transformation equipment can be obtained, and a more comprehensive and more accurate digital twinborn model is constructed; the subtle change of the power transmission and transformation equipment can be more accurately captured through multi-source fusion data, and the dynamic sensing of the operation state of the equipment is realized; through the multi-scale fusion method, complementary information in multi-source data can be effectively utilized, and the precision of the digital twin model is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data fusion, and in particular to a data fusion method, system, equipment and medium for power transmission and transformation equipment. Background Art

[0002] In recent years, with the rapid development of technologies such as artificial intelligence and big data, the power system has an increasing demand for status assessment, fault prediction and maintenance of power transmission and transformation equipment. Digital twin technology, as an emerging technology, can build virtual models that match real-world equipment or systems, and realize real-time monitoring, prediction and control of equipment or system status. It has broad application prospects in the power system. To build a digital twin model of power transmission and transformation equipment, it is necessary to integrate information from multiple data sources, including visual images, laser point cloud data, sensor data, etc., to achieve comprehensive perception of multi-dimensional information such as equipment structure, status, and operating parameters.

[0003] However, the existing digital twin model construction methods have the following problems: the traditional single data source construction method is difficult to fully reflect the subtle changes in the structure of power transmission and transformation equipment, resulting in the data twin model's dynamic perception of the equipment's operating status being not accurate enough; the existing simple fusion method cannot effectively utilize the complementary information in multi-source data, resulting in the construction efficiency and accuracy of the digital twin model being limited.

[0004] Therefore, how to effectively integrate information from multiple data sources and build a digital twin model of power transmission and transformation equipment to achieve comprehensive and accurate perception of the equipment's operating status has become a technical problem that needs to be urgently solved in the current power system field. Summary of the Invention

[0005] In view of the above existing problems, the present invention is proposed.

[0006] Therefore, the present invention provides a data fusion method and system for power transmission and transformation equipment to solve the problems of existing digital twin model construction methods, such as single data, incomplete information, insufficient model accuracy, and how to effectively fuse multi-source data.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0008] In a first aspect, the present invention provides a data fusion method for power transmission and transformation equipment, including: acquiring visual images and laser point cloud data of the power transmission and transformation equipment through sensors, and extracting a multi-scale visual feature set based on a Gaussian filter; fitting the laser point cloud data through principal component analysis to generate a multi-scale point cloud feature set, and performing feature matching through a nearest neighbor search algorithm; converting the visual image and laser point cloud data into a common coordinate system, fusing laser points and pixel points according to preset weights, and generating multi-source fusion data; the multi-source fusion data is used to construct a digital twin model of the power transmission and transformation equipment.

[0009] As a preferred solution of the data fusion method for power transmission and transformation equipment described in the present invention, the extraction of a multi-scale visual feature set based on a Gaussian filter includes: performing convolution smoothing processing on a visual image through a Gaussian filter to obtain filtered images of different scales; detecting the filtered image through an edge detection algorithm to obtain edge features; performing texture extraction on the filtered image through a local binary pattern to obtain texture features; and using the edge features and texture features as a multi-scale visual feature set of the visual image.

[0010] As a preferred solution of the data fusion method for power transmission and transformation equipment described in the present invention, the generating of a multi-scale point cloud feature set includes pre-processing laser point cloud data to obtain target point cloud data; performing cluster analysis on the target point cloud data by principal component analysis to obtain component point cloud features; obtaining connection point cloud features of the target point cloud data based on the distance relationship between point clouds; and using the component point cloud features and the connection point cloud features as a multi-scale point cloud feature set.

[0011] As a preferred solution of the data fusion method for power transmission and transformation equipment described in the present invention, the preprocessing of the laser point cloud data includes removing noise points in the laser point cloud data by statistical filtering to obtain denoised point cloud data; and using a random sampling consistency algorithm to remove ground points in the denoised point cloud data to obtain target point cloud data.

[0012] As a preferred solution of the data fusion method for power transmission and transformation equipment described in the present invention, the feature matching through the nearest neighbor search algorithm includes time synchronization of a multi-scale visual feature set and a multi-scale point cloud feature set; matching the feature points of the synchronized multi-scale visual feature set and the feature points of the synchronized multi-scale point cloud feature set through the nearest neighbor search algorithm to obtain feature points to be matched; and using a random sampling consistency algorithm to screen out matching feature point pairs of the feature points to be matched based on the geometric constraint relationship of the power transmission and transformation equipment.

[0013] As a preferred solution of the data fusion method for power transmission and transformation equipment described in the present invention, the time synchronization includes extracting visual timestamps and point cloud timestamps to obtain a time deviation between a multi-scale visual feature set and a multi-scale point cloud feature set; and performing time synchronization on the multi-scale visual feature set and the multi-scale point cloud feature set based on the time deviation.

[0014] As a preferred solution of the data fusion method for power transmission and transformation equipment described in the present invention, the generating of multi-source fusion data includes converting the visual image and the laser point cloud data into the same common coordinate system; obtaining the posture information of the matching feature point pairs according to the common coordinate system; and according to the posture information, fusing the laser points of the laser point cloud data and the pixel points of the visual image according to preset weights to obtain the multi-source fusion data.

[0015] In a second aspect, the present invention provides a data fusion system for power transmission and transformation equipment, comprising: a data processing module, a feature matching module, and a multi-source fusion module; the data processing module is used to obtain visual images and laser point cloud data of the power transmission and transformation equipment through sensors, and extract a multi-scale visual feature set based on a Gaussian filter; the feature matching module is used to fit the laser point cloud data through principal component analysis to generate a multi-scale point cloud feature set, and perform feature matching through a nearest neighbor search algorithm; the multi-source fusion module is used to convert visual images and laser point cloud data into a common coordinate system, fuse laser points and pixel points according to preset weights, and generate multi-source fusion data.

[0016] In a third aspect, the present invention provides an electronic device, comprising:

[0017] memory and processor;

[0018] The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the data fusion method for power transmission and transformation equipment are implemented.

[0019] In a fourth aspect, the present invention provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the data fusion method for power transmission and transformation equipment.

[0020] Compared with the existing technology, the beneficial effects of the present invention are as follows: by fusing visual images and laser point cloud data, the present invention can obtain multi-dimensional information such as the geometric shape, size, surface details, texture, etc. of the power transmission and transformation equipment, and build a more comprehensive and accurate digital twin model; through multi-source fusion data, it can more accurately capture subtle changes in the power transmission and transformation equipment and realize dynamic perception of the equipment operation status; through multi-scale fusion methods, it can effectively utilize the complementary information in multi-source data, avoid information redundancy and error accumulation, and improve the accuracy of the digital twin model; the present invention can automatically perform data fusion and improve the construction efficiency of the digital twin model. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0022] Figure 1 The figure is a schematic diagram of the overall process of a data fusion method for power transmission and transformation equipment according to an embodiment of the present invention.

[0023] Figure 2 The figure is a schematic diagram of the overall process of a data fusion system for power transmission and transformation equipment according to an embodiment of the present invention. DETAILED DESCRIPTION

[0024] To make the above-mentioned objects, features, and advantages of the present invention more clearly understood, the following detailed description of the specific embodiments of the present invention is given in conjunction with the accompanying drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in this field without creative work should fall within the scope of protection of the present invention.

[0025] Example 1, reference Figure 1 , as one embodiment of the present invention, provides a data fusion method for power transmission and transformation equipment, comprising:

[0026] S1: Obtain visual images and laser point cloud data of power transmission and transformation equipment through sensors, and extract multi-scale visual feature sets based on Gaussian filters;

[0027] S2: Fit the laser point cloud data through principal component analysis to generate a multi-scale point cloud feature set, and perform feature matching through the nearest neighbor search algorithm;

[0028] S3: Convert the visual image and laser point cloud data into a common coordinate system, fuse the laser points and pixel points according to the preset weights, and generate multi-source fusion data.

[0029] It should be noted that multi-source fusion data is used to build digital twin models of power transmission and transformation equipment.

[0030] It should be noted that the sensors include visual sensors and laser scanning devices.

[0031] It should be noted that the digital twin model, as an important tool for the management and maintenance of power transmission and transformation equipment, can connect the physical entity of the equipment with the virtual model to achieve real-time monitoring of equipment status, fault prediction and performance optimization; therefore, it is very important to improve the dynamic perception capability of the digital twin model.

[0032] Therefore, in order to address the above-mentioned problems of single twin model data, incomplete information, and insufficient model accuracy, through steps S1-S3, a multi-scale visual feature set is extracted based on a Gaussian filter to achieve image smoothing, remove noise, and highlight the edge and corner features of the image; a multi-scale point cloud feature set is obtained through principal component analysis, and feature matching is performed through the nearest neighbor search algorithm to achieve dimensionality reduction and feature extraction of laser point cloud data, convert a large amount of high-dimensional point cloud data into low-dimensional features, and establish a connection between visual information and spatial structure; by converting the common coordinate system and presetting weights, multi-source fusion data is generated, and multi-scale fusion of visual images and laser point cloud data is achieved.

[0033] Example 2, reference Figure 1 , which is an embodiment of the present invention, provides a data fusion method for power transmission and transformation equipment based on the above embodiment.

[0034] In the embodiment of the present application, obtaining the visual image and laser point cloud data of the power transmission and transformation equipment by the sensor in step S1 includes converting optical signals such as the appearance, shape, color and other optical information of the power transmission and transformation equipment into electrical signals by the visual sensor, thereby generating a digital image, and reflecting the appearance characteristics of the power transmission and transformation equipment by the digital image;

[0035] A device that uses a laser scanning device to scan power transmission and transformation equipment using a laser beam and obtain surface spatial information. The laser point cloud data describes the surface shape of the power transmission and transformation equipment in the form of three-dimensional coordinates, which can accurately reflect the geometric structure data of the equipment;

[0036] Laser point cloud data includes but is not limited to the outline of the equipment, subtle changes in shape, and spatial relationships between components;

[0037] Specifically, a high-resolution camera is used to photograph the power transmission and transformation equipment, and visual images of the power transmission and transformation equipment under different viewing angles and lighting conditions are collected. The power transmission and transformation equipment is scanned using a laser scanning device to obtain laser point cloud data that can accurately reflect the geometric shape, size and other information of the power transmission and transformation equipment.

[0038] In the embodiment of the present application, extracting a multi-scale visual feature set based on a Gaussian filter in step S1 includes the following steps A1-A4:

[0039] A1: Perform convolution smoothing on the visual image using a Gaussian filter to obtain filtered images of different scales;

[0040] A2: Detect the filtered image using an edge detection algorithm to obtain edge features;

[0041] A3: Extract texture from the filtered image using local binary patterns to obtain texture features;

[0042] A4: Use edge features and texture features as a multi-scale visual feature set of visual images.

[0043] Specifically, in step A1, the Gaussian filter is provided with Gaussian kernels of different sizes. The Gaussian kernel is a matrix used for convolution operation of image pixels in the Gaussian filter. The visual image is filtered using matrices of different sizes.

[0044] Among them, the small-sized Gaussian kernel acts on a small local area of ​​the visual image, capturing detailed information in the visual image during convolution, such as some small textures or fine edge parts in the image; the large-sized Gaussian kernel covers a larger area in the visual image, and extracts the overall structure of the image and larger-sized features when extracting features, such as the approximate outline of power transmission and transformation equipment.

[0045] In step A2, the edge detection algorithm used is the Canny edge detection algorithm, which performs Gaussian filtering on the visual image to reduce noise through the Canny edge detection algorithm; the gradient amplitude and direction of the visual image are calculated, where the gradient amplitude represents the intensity of the grayscale change, and the direction represents the direction of the fastest grayscale change; non-maximum suppression is performed on the view image, and the pixels with non-maximum gradient amplitude in the local range are checked, and the points with maximum gradient amplitude are retained to refine the edges; a high threshold and a low threshold are set, and pixels above the high threshold are edge points, and pixels below the low threshold are non-edge points. Pixels between the high threshold and the low threshold are judged as edge points if they are connected to edge points. Continuous edge features are obtained by dual-threshold detection and edge connection.

[0046] In step A3, through the local binary pattern, for each pixel in the filtered image, each pixel is taken as the center pixel, and the neighborhood pixels around the center pixel, such as a 3*3 neighborhood, are selected. The grayscale difference between the center pixel and the neighborhood pixels is calculated, and the neighborhood pixels are marked as 0 or 1 according to the grayscale difference. If the grayscale value of the neighborhood pixel is greater than or equal to the center pixel, it is marked as 1, otherwise it is marked as 0. The marked 0 and 1 are combined in a fixed order to form a binary number. After converting the binary number into a decimal number, the local binary pattern value corresponding to the center pixel is obtained, which represents the texture feature of the image.

[0047] In step A4, edge features and texture features are used as a multi-scale visual feature set of the visual image.

[0048] In the embodiment of the present application, generating a multi-scale point cloud feature set in step S2 includes the following steps B1-B4:

[0049] B1: Preprocess the laser point cloud data to obtain target point cloud data;

[0050] B2: Perform cluster analysis on the target point cloud data using principal component analysis to obtain the component point cloud features;

[0051] B3: Obtain the connection point cloud features of the target point cloud data based on the distance relationship between the point clouds;

[0052] B4: Use component point cloud features and connection point cloud features as multi-scale point cloud feature sets.

[0053] It should be noted that the target point cloud data is the data obtained after denoising of the laser point cloud data; the component point cloud features are the features that can characterize the various components of the power transmission and transformation equipment, which are extracted from the target point cloud data after cluster analysis using the principal component analysis method. The power transmission and transformation equipment is composed of multiple components, each of which has a unique geometric shape and spatial position relationship; the component point cloud features include but are not limited to the shape characteristics, spatial distribution characteristics, etc. of the components.

[0054] Specifically, in step B1, the target point cloud data is obtained by performing denoising preprocessing on the laser point cloud data, and the target point cloud data is clustered by principal component analysis. The component point cloud features representing the various components of the power transmission and transformation equipment are extracted from the target point cloud data. Based on the distance relationship between a large number of discrete point clouds, the connection point cloud features reflecting the connectivity of the point cloud data in space are extracted from the target point cloud data, and the component point cloud features and the connection point cloud features are used as a multi-scale point cloud feature set.

[0055] It should be noted that a point cloud is a geometry composed of a large number of discrete points. These points have their own coordinates in three-dimensional space. The distance relationship between point clouds refers to the spatial distance measurement between each point cloud data point. The connected point cloud feature is a feature extracted from the target point cloud data based on the distance relationship between point clouds, reflecting the connectivity of the point cloud data in space.

[0056] It should be noted that preprocessing the laser point cloud data to obtain the target point cloud data includes removing noise points in the laser point cloud data through statistical filtering to obtain denoised point cloud data; and using a random sampling consistency algorithm to remove ground points in the denoised point cloud data to obtain the target point cloud data.

[0057] Statistical filtering is a method of denoising based on the spatial distribution characteristics of each point in the laser point cloud data. Noise points are abnormal points caused by factors such as the errors of the laser radar equipment itself and external interference. These points do not represent the actual surface information of the object and interfere with the subsequent analysis and processing of the laser point cloud data.

[0058] In an optional embodiment, the target point cloud data can also be obtained by randomly sampling data points to construct a hypothetical model, such as a plane model to fit the ground, calculating the number of points in the data set that conform to the hypothetical model, and finding the model containing the most inliers through iteration. Points that do not conform to the model are regarded as outliers, thereby removing the ground points in the denoised point cloud data to obtain the target point cloud data.

[0059] In steps B2-B3, after cluster analysis of the component point cloud features using principal component analysis, features representing each component of the power transmission and transformation equipment are extracted from the target point cloud data;

[0060] Perform cluster analysis on the target point cloud data using principal component analysis, calculate the eigenvalues ​​and eigenvectors of the point cloud covariance matrix, arrange the eigenvalues ​​in descending order, and use the eigenvector corresponding to the largest eigenvalue as the main direction of the component. Determine the component category based on the eigenvalue ratio and extract the component point cloud features.

[0061] At the same time, based on the Euclidean distance between point clouds, if the distance between two points is less than or equal to 0.1m, they are determined to be connection points and the connection point cloud features are extracted.

[0062] In an optional embodiment, principal component analysis is a method for dimensionality reduction and feature extraction of laser point cloud data. The principal component analysis is used to perform data fitting on laser point cloud data roughly distributed on a plane to extract features such as the planar and curved surface structures of power transmission and transformation equipment.

[0063] In step B4, the component point cloud features and the connection point cloud features are used as a multi-scale point cloud feature set.

[0064] In an embodiment of the present application, in step S2, the feature matching by the nearest neighbor search algorithm includes the following steps C1-C3:

[0065] C1: Temporally synchronize the multi-scale visual feature set and the multi-scale point cloud feature set;

[0066] C2: Match the feature points of the synchronized multi-scale visual feature set with the feature points of the synchronized multi-scale point cloud feature set through the nearest neighbor search algorithm to obtain the feature points to be matched;

[0067] C3: Using the random sampling consistency algorithm, the matching feature point pairs to be matched are screened out according to the geometric constraints of the power transmission and transformation equipment.

[0068] Specifically, in step C1, considering that the visual image and laser point cloud data are collected at different time points, there are temporal differences between the corresponding multi-scale visual feature set and multi-scale point cloud feature set. A time synchronization operation is performed to ensure that the multi-scale visual feature set and the multi-scale point cloud feature set are at the same time.

[0069] By extracting visual timestamps and point cloud timestamps, the time deviation between the multi-scale visual feature set and the multi-scale point cloud feature set is obtained; according to the time deviation, the multi-scale visual feature set and the multi-scale point cloud feature set are time-synchronized;

[0070] The time deviation can be the difference between the time information of the visual timestamp and the point cloud timestamp. Assume that the time mark of the feature point in the visual data is t v , the time mark of the corresponding feature point in the laser data is t l , then the time deviation Deltat is expressed as:

[0071] Deltal=t v -t l

[0072] During the data acquisition process, visual images and laser point cloud data are time-stamped.

[0073] It should be noted that the visual timestamp is used to record the time information mark of the visual image acquisition related to the multi-scale visual feature set; the point cloud timestamp is used to record the time information mark of the laser point cloud data acquisition related to the multi-scale point cloud feature set.

[0074] In step C2, the points to be matched are preliminary feature points obtained by matching the feature points in the visual data with the feature points in the laser data through the nearest neighbor search algorithm. The preliminary feature points are screened for errors, and false matches are removed to obtain the feature points to be matched.

[0075] In step C3, based on the geometric relationship of the power transmission and transformation equipment, a random sampling consistency algorithm is used for iterative screening to obtain matching feature point pairs.

[0076] In an optional embodiment, an example of obtaining a matching feature point pair is to randomly select 4 pairs of matching points to fit the geometric transformation model, calculate the projection error of all matching points, and if the projection error is less than or equal to 0.01m, the matching point is determined to be an inlier; when the inlier point data exceeds 90% of the total matching points or the number of iterations reaches 1000, it is terminated and the inlier point is retained as the final matching feature point pair.

[0077] In an embodiment of the present application, generating multi-source fusion data in step S3 includes converting the visual image and the laser point cloud data into the same common coordinate system; obtaining the posture information of the matching feature point pairs based on the common coordinate system; and according to the posture information, fusing the laser points of the laser point cloud data and the pixel points of the visual image according to preset weights to obtain multi-source fusion data.

[0078] Specifically, the common coordinate system is a coordinate system in which the visual image and the laser point cloud data use the same coordinate rules to represent the position of elements;

[0079] After converting the visual image and laser point cloud data into the same common coordinate system, it is convenient to accurately locate and analyze objects in the visual image and the laser point cloud data;

[0080] Pose information is the data of the position of the matching feature point pairs relative to the common coordinate system;

[0081] The preset weight is a pre-set value used to determine the respective weights of the laser points of the laser point cloud data and the pixel points of the visual image during the fusion process.

[0082] In an optional embodiment, an example of preset weights is to set the weight of laser point cloud data to 0.6 and the weight of visual image to 0.4. During fusion, data fusion focuses on the information of laser point cloud data, and the specific weights are set according to actual conditions.

[0083] It should be noted that the digital twin model is a virtual digital representation of power transmission and transformation equipment, which can simulate the various characteristics, behaviors and states of real power transmission and transformation equipment in a virtual environment. Since the construction of a digital twin model requires a large amount of information about power transmission and transformation equipment, multi-source fusion data can provide a comprehensive data foundation for the construction of the digital twin model of power transmission and transformation equipment. Comprehensive information enables the digital twin model to more accurately reflect the actual state of power transmission and transformation equipment in reality, making the digital twin model more practical.

[0084] Example 3, reference Figure 2 The above is a schematic diagram of a data fusion method for power transmission and transformation equipment. It should be noted that the technical solution of the data fusion system for power transmission and transformation equipment and the technical solution of the data fusion method for power transmission and transformation equipment described above are based on the same concept. For details not described in detail in the technical solution of the data fusion system for power transmission and transformation equipment in this embodiment, please refer to the description of the technical solution of the data fusion method for power transmission and transformation equipment described above.

[0085] This embodiment further provides a data fusion system for power transmission and transformation equipment, including: a data processing module, a feature matching module, and a multi-source fusion module.

[0086] Among them, the data processing module is used to obtain visual images and laser point cloud data of power transmission and transformation equipment through sensors, and extract multi-scale visual feature sets based on Gaussian filters; the feature matching module is used to fit laser point cloud data through principal component analysis, generate multi-scale point cloud feature sets, and perform feature matching through the nearest neighbor search algorithm; the multi-source fusion module is used to convert visual images and laser point cloud data into a common coordinate system, fuse laser points and pixel points according to preset weights, and generate multi-source fusion data.

[0087] This embodiment also provides an electronic device suitable for data fusion of power transmission and transformation equipment, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement the data fusion method for power transmission and transformation equipment proposed in the above embodiment.

[0088] This embodiment further provides a storage medium storing a computer program. When the program is executed by a processor, the method for realizing data fusion of power transmission and transformation equipment proposed in the above embodiment is implemented.

[0089] The storage medium proposed in this embodiment and the method for realizing data fusion of power transmission and transformation equipment proposed in the above embodiment belong to the same inventive concept. For technical details not fully described in this embodiment, please refer to the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0090] Through the above description of the implementation methods, those skilled in the art can clearly understand that the present invention can be implemented with the help of software and necessary general hardware, and of course can also be implemented by hardware. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory (FLASH), hard disk or optical disk, etc., including a number of instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute the methods of various embodiments of the present invention.

[0091] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A data fusion method for power transmission and transformation equipment, characterized in that: include: The visual images and laser point cloud data of power transmission and transformation equipment are acquired through sensors, and multi-scale visual feature sets are extracted based on Gaussian filters. The laser point cloud data is fitted by principal component analysis to generate a multi-scale point cloud feature set, and feature matching is performed using the nearest neighbor search algorithm. Convert visual images and laser point cloud data into a common coordinate system, fuse laser points and pixel points according to preset weights, and generate multi-source fusion data; The multi-source fusion data is used to construct a digital twin model of power transmission and transformation equipment.

2. The data fusion method for power transmission and transformation equipment according to claim 1, characterized in that: The extracting of the multi-scale visual feature set based on the Gaussian filter includes performing convolution smoothing processing on the visual image through the Gaussian filter to obtain filtered images of different scales; The filtered image is detected by edge detection algorithm to obtain edge features; The texture of the filtered image is extracted by using local binary patterns to obtain texture features; Edge features and texture features are used as a multi-scale visual feature set of visual images.

3. The data fusion method for power transmission and transformation equipment according to claim 2, characterized in that: Generating a multi-scale point cloud feature set includes pre-processing the laser point cloud data to obtain target point cloud data; Cluster analysis is performed on the target point cloud data using principal component analysis to obtain the point cloud features of the components; According to the distance relationship between point clouds, the connection point cloud features of the target point cloud data are obtained; The component point cloud features and connection point cloud features are used as a multi-scale point cloud feature set.

4. The data fusion method for power transmission and transformation equipment according to claim 3, characterized in that: The preprocessing of the laser point cloud data includes removing noise points of the laser point cloud data by statistical filtering to obtain denoised point cloud data; The random sampling consensus algorithm is used to remove the ground points in the denoised point cloud data to obtain the target point cloud data.

5. The data fusion method for power transmission and transformation equipment according to claim 4, characterized in that: The feature matching by the nearest neighbor search algorithm includes time synchronization of the multi-scale visual feature set and the multi-scale point cloud feature set; The feature points of the synchronized multi-scale visual feature set and the feature points of the synchronized multi-scale point cloud feature set are matched by a nearest neighbor search algorithm to obtain feature points to be matched; The random sampling consistency algorithm is used to select the matching feature point pairs according to the geometric constraints of the power transmission and transformation equipment.

6. The data fusion method for power transmission and transformation equipment according to claim 5, characterized in that: The time synchronization includes extracting visual timestamps and point cloud timestamps to obtain a time deviation between a multi-scale visual feature set and a multi-scale point cloud feature set; According to the time deviation, the multi-scale visual feature set and the multi-scale point cloud feature set are temporally synchronized.

7. The data fusion method for power transmission and transformation equipment according to claim 6, characterized in that: Generating multi-source fusion data includes converting the visual image and the laser point cloud data into the same common coordinate system; According to the common coordinate system, the pose information of the matching feature point pairs is obtained; According to the pose information, the laser points of the laser point cloud data and the pixel points of the visual image are fused according to the preset weights to obtain multi-source fusion data.

8. A data fusion system for power transmission and transformation equipment, applying the method according to any one of claims 1 to 7, characterized in that: include: Data processing module, feature matching module, multi-source fusion module The data processing module is used to obtain visual images and laser point cloud data of power transmission and transformation equipment through sensors, and extract multi-scale visual feature sets based on Gaussian filters; The feature matching module is used to fit the laser point cloud data through principal component analysis to generate a multi-scale point cloud feature set and perform feature matching through a nearest neighbor search algorithm; The multi-source fusion module is used to convert visual images and laser point cloud data into a common coordinate system, fuse laser points and pixel points according to preset weights, and generate multi-source fusion data.

9. An electronic device comprising: memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the data fusion method for power transmission and transformation equipment according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing computer-executable instructions, wherein the computer-executable instructions, when executed by a processor, implement the steps of the data fusion method for power transmission and transformation equipment according to any one of claims 1 to 7.

Images