Multi-mode data fusion power transmission engineering geographic information basic data processing system

Abstract

The application discloses a power transmission and transformation engineering geographic information basic data processing system based on multi-mode data fusion, comprising: a power grid multi-mode data acquisition module, a multi-source geographic data receiving module, an ICP registration calculation module, a quality perception fusion module, a topological graph modeling module and a geographic data processing module. The system obtains data such as buildings, terrains, bridges, power lines and towers through the power grid multi-mode data acquisition module, and obtains data such as unmanned aerial vehicle aerial images and laser radars through the multi-source geographic data receiving module; the ICP registration calculation module calibrates multi-source geographic data to improve coordinate consistency; the quality perception fusion module adjusts weight according to data reliability to improve fusion accuracy; the topological graph modeling module constructs device node correlation based on fused data to improve model comprehensiveness; and the geographic data processing module integrates data and topological parameters to generate comprehensive data sets. The system improves data integrity and provides reliable data support for power transmission and transformation engineering survey and design and later operation and maintenance.

Description

Technical Field

[0001] This invention relates to the field of geographic information processing technology for power transmission and transformation projects, and in particular to a basic data processing system for geographic information of power transmission and transformation projects based on multi-mode data fusion. Existing technology

[0002] In the construction and operation and maintenance of power transmission and transformation projects, a large amount of geographic information data is required to support engineering design, equipment deployment, and fault diagnosis. Currently, the sources of geographic information data for power transmission and transformation projects are diverse, encompassing multi-mode data generated by power grid operation (such as conductor sag, tower coordinates, and insulator leakage current data) and multi-source geographic data (such as satellite remote sensing imagery, UAV aerial photography, and ground-measured elevation data). However, the data from different sources have significantly different formats, inconsistent coordinate systems, and weak correlations, making it difficult to form a complete and accurate geographic information dataset. Furthermore, with the expansion of power transmission and transformation projects and the increasing demand for intelligent operation and maintenance, traditional data processing methods are insufficient to efficiently integrate multiple types of data and cannot provide comprehensive data support for engineering decisions. Therefore, there is an urgent need to build a system capable of multi-mode data fusion and efficient processing to improve the usability and application value of geographic information data for power transmission and transformation projects.

[0003] Existing technologies for processing geographic information foundation data in power transmission and transformation projects have two significant drawbacks: First, the registration accuracy of multi-source geographic data is insufficient and the data fusion effect is poor. Traditional registration methods do not fully consider the differences in data characteristics and are unable to eliminate coordinate deviations from geographic data from different sources. This results in weak correlation between the registered data and the multi-mode data of the power grid. Furthermore, the data weights are not dynamically adjusted during the fusion process, failing to highlight the role of high-reliability data and causing a decrease in the accuracy of the fused data. Second, equipment topology modeling is disconnected from geographic data processing. Existing technologies do not deeply integrate the multi-dimensional data after fusion when constructing equipment topology relationships, relying only on single equipment parameters. This fails to comprehensively reflect the spatial location and attribute relationships of equipment. Moreover, geographic data processing does not effectively integrate topology model parameters, resulting in a geographic information foundation dataset lacking equipment topology information support, which is insufficient to meet the comprehensive data requirements of power transmission and transformation projects. Summary of the Invention

[0004] In order to overcome the shortcomings and deficiencies of existing technologies, this invention provides a multi-mode data fusion geographic information basic data processing system for power transmission and transformation projects.

[0005] The technical solution adopted in this invention is a multi-mode data fusion geographic information basic data processing system for power transmission and transformation projects, comprising: a power grid multi-mode data acquisition module, a multi-source geographic data receiving module, an ICP registration calculation module, a quality perception fusion module, a topology map modeling module, and a geographic data processing module; the power grid multi-mode data acquisition module acquires conductor sag data, tower coordinate data, and insulator leakage current data in the power transmission and transformation project and transmits them to the quality perception fusion module; the multi-source geographic data receiving module receives satellite remote sensing image data, UAV aerial photography data, and ground measured elevation data and transmits them to the ICP registration module. The calculation module includes an ICP registration calculation module that performs coordinate transformation calculations on the received multi-source geographic data and then transmits it to a quality perception fusion module. The quality perception fusion module performs feature layer fusion calculations on the power grid multi-mode data and the registered geographic data and then transmits them to a topology map modeling module and a geographic data processing module, respectively. The topology map modeling module constructs a node association model of power transmission and transformation equipment based on the fused data and transmits the model parameters to the geographic data processing module. The geographic data processing module performs spatial coordinate mapping calculations on the fused data and the topology model parameters to generate a basic dataset of geographic information for power transmission and transformation projects.

[0006] Furthermore, the quality perception fusion module adopts a power grid multi-mode quality perception fusion model, the model expression of which is: Among them, F fusion ω represents the feature value of the fused data. i Let D be the weighting coefficient for the i-th type of power grid multi-mode data. i This is the i-th type of power grid multi-mode data (conductor sag / tower coordinates / insulator leakage current). Let be the variance of the i-th type of power grid multimode data. The variance of the corresponding geographic data after registration is given by λ, where λ is the covariance adjustment coefficient, and cov(D) is the variance of the geographic data after registration. i G i Let be the covariance between the i-th type of power grid multi-mode data and the registered geographic data; the ICP registration calculation module adopts the multi-source geographic data ICP registration algorithm, the algorithm expression is: Among them, T opt For the optimal registration transformation matrix, SE(3) is the three-dimensional Euclidean transformation group, and G is the transformation matrix. j Let G' be the coordinates of the j-th feature point in the satellite remote sensing image data. j Let be the coordinates of the j-th feature point in the drone aerial data, δ be the distance attenuation coefficient, and m be the total number of feature points.

[0007] Furthermore, the ICP registration calculation module also includes a registration accuracy correction submodule, with the correction formula being: Where ΔT is the corrected registration transformation matrix, I is the identity matrix, α is the correction step size coefficient, and E is the registration error function. For the registration error function T opt The partial derivative of the transpose matrix; the quality-sensing fusion module also includes a fusion weight update formula: in, The weight coefficients for the (k+1)th iteration are... The weight coefficients for the k-th iteration are... The signal-to-noise ratio of the fused data in the (k+1)th iteration is given. Let be the signal-to-noise ratio of the fused data in the k-th iteration.

[0008] Furthermore, the topology graph modeling module constructs a device topology relationship model, the model expression of which is: Among them, M topo Let A be the device topology matrix. p,r Let N be the adjacency coefficient between the p-th device node and the r-th device node. p Let N be the attribute vector (tower height / conductor type / insulator type) of the p-th device node. r Let be the attribute vector of the r-th device node. For tensor product operations, diag(L) p ,L r Let L be the spatial distance between the p-th device node and the r-th device node. p ,L r The matrix is ​​a diagonal matrix with diagonal elements, where q is the total number of device nodes; the geographic data processing module uses the data spatial mapping formula: P geo =proj(F fusion M topo )·R utm +T offset Among them, P geo Let R be the final geographic coordinates, proj(·) be the projection transformation function, and R be the coordinates. utm T is the transformation matrix for the UTM coordinate system. offset This represents the coordinate offset.

[0009] Furthermore, the power grid multi-mode data acquisition module adopts a data acquisition frequency control formula: Among them, f sample f0 is the actual sampling frequency, and var(D) is the reference sampling frequency. real ) represents the variance of the actual power grid multi-mode data, var(D pre ) represents the variance of the preset multi-mode power grid data, and β is the frequency adjustment coefficient; the multi-source geographic data receiving module uses the data receiving priority calculation formula: Among them, P prio Q represents the receiving priority of the j-th type of geographic data. j R is the resolution of the j-th type of geographic data.j Let τ be the update rate of the j-th type of geographic data. j Let the transmission delay of the j-th type of geographic data be denoted by loss. j Let be the packet loss rate of the j-th type of geographic data.

[0010] Furthermore, the quality-aware fusion module also includes an abnormal data removal formula: in, For the i-th type of power grid multi-mode data after removing anomalies, H(·) is a step function. Let be the mean of the i-th type of power grid multi-mode data. The threshold for identifying abnormal data; the topology graph modeling module uses a node importance evaluation formula: Among them, I p Let deg(N) be the importance index of the p-th device node. p ) represents the degree of the p-th device node, deg(N) r ) represents the degree of the r-th device node, ∈ is used to avoid the minimum value where the denominator is zero, and γ is the importance adjustment coefficient.

[0011] Furthermore, the topology graph modeling module includes a device node extraction unit, a node attribute quantification unit, an adjacency relationship construction unit, and a topology matrix generation unit. The device node extraction unit performs feature point detection on the fused data transmitted by the quality perception fusion module, filters out the tower center points, conductor connection points, and insulator installation points in the power transmission and transformation project, distinguishes nodes corresponding to different equipment types through coordinate clustering, and establishes preliminary node identification information. The node attribute quantification unit performs numerical conversion on the physical parameters corresponding to the extracted device nodes, converting non-numerical attributes such as tower height, conductor cross-sectional area, and insulator insulation level into standardized values, and constructs a node feature vector containing spatial location and physical attributes by combining node coordinate information. The adjacency relationship construction unit calculates the spatial distance and electrical connection relationship between each device node, sets a distance threshold and an electrical connectivity judgment condition, and marks nodes as adjacent nodes and assigns an adjacency coefficient when the spatial distance between nodes is less than the threshold and an electrical connection path exists. The topology matrix generation unit constructs a topology relationship matrix containing node attributes and adjacency information based on the adjacency relationship and node feature vector, and stores the correspondence between matrix elements and device node identifiers.

[0012] Furthermore, the geographic data processing module includes a spatial coordinate transformation unit, a data layer overlay unit, a geographic attribute association unit, and a basic dataset generation unit. The spatial coordinate transformation unit receives fused data transmitted from the quality-aware fusion module and topological model parameters transmitted from the topological mapping modeling module. It reads the original coordinate system information contained in the data, calls a coordinate transformation algorithm to uniformly transform data from different sources to the target coordinate system, calculates the coordinate deviation during the transformation process, and records the deviation value. The data layer overlay unit divides the transformed fused data into different layers according to data type, including a conductor data layer, a tower data layer, and a geographic background map. The system employs a layer alignment algorithm to adjust the spatial position of each layer, ensuring that the overlap of corresponding feature points between layers meets set requirements. The geographic attribute association unit extracts attribute information from each layer's data, establishes attribute association rules, associates conductor sag data with elevation data of the corresponding geographic region, and associates tower coordinate data with geological data of the corresponding region, generating a multi-attribute associated data table. The basic dataset generation unit standardizes the format of the associated data, organizes the data structure according to the specifications for geographic information data of power transmission and transformation projects, stores the data in blocks, and establishes an index, forming a directly accessible basic dataset for geographic information of power transmission and transformation projects.

[0013] Furthermore, the quality-aware fusion module includes a multi-mode data preprocessing unit, a data feature extraction unit, a fusion weight calculation unit, and a feature layer fusion unit. The multi-mode data preprocessing unit receives power grid multi-mode data transmitted from the power grid multi-mode data acquisition module and registered geographic data transmitted from the ICP registration calculation module. It performs format conversion and missing value imputation on the data, eliminates high-frequency noise in the data using a data smoothing algorithm, and retains effective feature information in the data. The data feature extraction unit performs feature extraction on the preprocessed power grid multi-mode data and registered geographic data respectively, and uses a feature detection algorithm to extract trend features and abrupt changes in the power grid multi-mode data. The system transforms features by extracting spatial texture and contour features from the registered geographic data and converting the extracted features into feature vectors. The fusion weight calculation unit calculates the fusion weights of various types of data based on the variance and signal-to-noise ratio parameters of the data feature vectors using a weight allocation algorithm. It dynamically adjusts the weight coefficients according to the data credibility, so that data with high credibility receives higher weights. The feature layer fusion unit uses a feature layer fusion algorithm to weight and combine the feature vectors of various types of data according to the calculated fusion weights. It optimizes the fusion result by combining the covariance information between the data, generates a fused comprehensive feature vector, and transmits it to the topology map modeling module and the geographic data processing module.

[0014] The multi-mode data fusion geographic information basic data processing system for power transmission and transformation projects operates in the following steps: First, the power grid multi-mode data acquisition module periodically collects conductor sag data, tower coordinate data, and insulator leakage current data from the power transmission and transformation project. Simultaneously, the multi-source geographic data receiving module receives satellite remote sensing imagery, UAV aerial photography data, and measured ground elevation data. Second, the multi-source geographic data received by the multi-source geographic data receiving module is transmitted to the ICP registration calculation module. The ICP registration algorithm performs coordinate transformation and registration calculations on the multi-source geographic data to eliminate coordinate deviations between different sources. Third, the multi-mode data collected by the power grid multi-mode data acquisition module and the registered geographic data output by the ICP registration calculation module are transmitted to the quality sensing system. The first step involves the knowledge fusion module, which uses a multi-mode quality perception fusion model for power grids to perform feature layer fusion calculations on two types of data to generate fused data. The second step involves transmitting the fused data to the topology map modeling module, which constructs a node association model for power transmission and transformation equipment based on the equipment parameters and spatial location information in the fused data, generating an equipment topology relationship matrix. The third step involves transmitting the fused data and the topology relationship matrix to the geographic data processing module, which performs spatial coordinate mapping calculations on the fused data, associates the topology relationship matrix with geographic data, and generates comprehensive data containing equipment topology information and geographic attributes. The fourth step involves standardizing the format of the comprehensive data output by the geographic data processing module, storing it in blocks according to data type and spatial region, establishing a data index, and forming a basic dataset of geographic information for power transmission and transformation projects.

[0015] Beneficial Effects: This invention proposes a multi-mode data fusion-based geographic information basic data processing system for power transmission and transformation projects. This system accurately captures power grid operation data such as conductor sag and tower coordinates through a multi-mode power grid data acquisition module. It also acquires geographic data from satellite remote sensing and UAV aerial photography through a multi-source geographic data receiving module. An ICP registration calculation module calibrates the coordinates of the multi-source geographic data, significantly improving the coordinate consistency of geographic data from different sources. A quality-aware fusion module dynamically adjusts weights based on data reliability, significantly improving the accuracy of multi-mode data feature layer fusion. A topology mapping modeling module constructs equipment node relationships based on the fused data, effectively improving the comprehensiveness of the equipment topology model. Finally, a geographic data processing module integrates the fused data and topology parameters to generate a comprehensive geographic data model. The information foundation dataset significantly improves the completeness, accuracy, and application value of geographic information data for power transmission and transformation projects, providing more reliable data support for engineering design, equipment operation and maintenance, and fault diagnosis. Simultaneously, the system addresses the shortcomings of existing technologies. On one hand, it leverages the ICP registration algorithm to improve the registration accuracy of multi-source geographic data and strengthens the role of high-reliability data through a quality-aware fusion model, resolving the problems of large registration deviations and poor fusion effects in existing technologies. On the other hand, the topology mapping module deeply integrates fused data, improving the adaptability of equipment topology relationships to actual engineering scenarios. The geographic data processing module synchronously integrates topology information and geographic data, enhancing the comprehensiveness of the dataset and resolving the issues of disconnect between topology modeling and geographic data processing, and insufficient data usability in existing technologies. Attached Figure Description

[0016] Figure 1 This is a diagram showing the system module composition of the present invention;

[0017] Figure 2 This is a flowchart of the system operation steps of the present invention. Detailed Implementation

[0018] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. The application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0019] like Figure 1 As shown, the multi-mode data fusion power transmission and transformation engineering geographic information basic data processing system includes: a power grid multi-mode data acquisition module, a multi-source geographic data receiving module, an ICP registration calculation module, a quality perception fusion module, a topology map modeling module, and a geographic data processing module.

[0020] The power grid multi-mode data acquisition module collects conductor sag data, tower coordinate data, and insulator leakage current data in power transmission and transformation projects and transmits them to the quality perception fusion module.

[0021] Specifically, the power grid multi-mode data acquisition module is used to collect key multi-mode data during the operation of power transmission and transformation projects. Its acquisition frequency is set to 5 minutes per acquisition to ensure real-time data transmission. The acquisition range covers the entire length of the transmission line, including data related to core equipment such as towers, conductors, and insulators. During implementation, high-precision GNSS positioning equipment deployed on the top of the towers acquires tower coordinate data with a positioning accuracy of ±3cm. Simultaneously, conductor sag data is collected using conductor tension sensors, with a measurement range of 0-10m and an accuracy of ±0.05m. Additionally, leakage current data is collected through insulator leakage current monitoring devices, with a monitoring range of 0-1000μA and a resolution of 1μA. The module performs real-time data verification during acquisition. When data deviation exceeds set thresholds (e.g., coordinate deviation exceeding 5cm, sag deviation exceeding 0.1m, leakage current deviation exceeding 10μA), a re-acquisition mechanism is automatically triggered. The data collected by this module is directly transmitted to the quality perception fusion module, providing a foundation for subsequent data fusion. Through precise and high-frequency data collection, it ensures the accuracy and integrity of the multi-mode data of the power grid, providing reliable raw data support for the geographic information basic data processing of power transmission and transformation projects, and avoiding the impact of data loss or excessive errors on subsequent processing results.

[0022] The multi-source geographic data receiving module receives satellite remote sensing image data, UAV aerial photography data, and ground measured elevation data and transmits them to the ICP registration calculation module.

[0023] Specifically, the multi-source geographic data receiving module is responsible for receiving multi-source geographic data of the area where the power transmission and transformation project is located. The types of data received include satellite remote sensing imagery, UAV aerial photography data, and ground-measured elevation data. The resolution of the satellite remote sensing imagery is set to 0.5m, with an update cycle of 7 days. The resolution of the UAV aerial photography data is 0.1m, the flight altitude is controlled at 500m, and the coverage area of ​​a single aerial photography session is 10km. 2 The measurement accuracy of the ground-based measured elevation data is ±0.1m, and the distance between measurement points is 50m. During implementation, the module establishes connections with satellite data service providers, UAV control systems, and ground measurement equipment through a dedicated data receiving interface. Data transmission is performed using the TCP / IP protocol at a rate of no less than 10Mbps. Simultaneously, the received data is converted to GeoTIFF format for subsequent processing. The module also features data caching; when the network is interrupted, it can cache at least 24 hours of collected data, automatically resuming transmission upon network recovery. It also verifies the integrity of the received data; if the data loss rate exceeds 5%, a retransmission request is sent to the data sender. By receiving multi-type, high-resolution geographic data, this module provides rich geospatial information for the geographic information foundation data processing of power transmission and transformation projects, ensuring the accuracy of subsequent registration calculations and meeting the high-precision geographic data requirements of the project.

[0024] The ICP registration calculation module performs coordinate transformation calculations on the received multi-source geographic data and then transmits the data to the quality perception fusion module.

[0025] Specifically, the ICP registration calculation module performs registration calculations for multi-source geographic data. Its core function is to eliminate coordinate discrepancies between geographic data from different sources. The registration process employs an iterative nearest-point algorithm, with 20 iterations and a convergence threshold of 0.001m. During implementation, the module first acquires satellite remote sensing imagery, UAV aerial photography data, and measured ground elevation data from the multi-source geographic data receiving module. It then extracts feature points from each data set, with a feature point extraction density of 100m. 2 At least one feature point is used for registration. Then, the Euclidean distance between feature points in different data sets is calculated. Feature points with a distance less than 0.5m are used as matching pairs. The optimal registration transformation matrix is ​​then calculated iteratively to transform the coordinates of the UAV aerial data and the measured ground elevation data, unifying the coordinates of all three into the WGS-84 coordinate system. During registration, the module calculates the registration error in real time. When the error exceeds 0.1m, feature points are reselected for matching calculation. After registration, the registered geographic data is transmitted to the quality-aware fusion module. This module, through precise registration calculation, unifies the coordinate system of multi-source geographic data, reduces spatial deviations between data, provides a consistent spatial benchmark for subsequent multi-mode data fusion, avoids deviations in the fusion results due to coordinate inconsistencies, and ensures the spatial accuracy of geographic information data for power transmission and transformation projects.

[0026] The quality perception fusion module performs feature layer fusion calculations on the power grid multi-mode data and the registered geographic data, and then transmits them to the topology map modeling module and the geographic data processing module, respectively.

[0027] Specifically, the quality perception fusion module undertakes the task of fusing power grid multi-mode data with registered geographic data. During the fusion process, the input data is first preprocessed, including data denoising using a mean filtering algorithm with a 3×3 filter window. Then, data features are extracted. Features extracted from the power grid multi-mode data include data change trends and abrupt change points, while features extracted from the geographic data include terrain undulations and feature outlines. During implementation, the module assigns fusion weights based on data reliability. The weight coefficient for tower coordinate data is set to 0.3, for conductor sag data to 0.25, for insulator leakage current data to 0.2, for registered satellite remote sensing image data to 0.15, and for ground-measured elevation data to 0.1. The weights are dynamically adjusted based on the signal-to-noise ratio (SNR); when the SNR of a certain type of data is below 20dB, the weight is reduced by 10%. A feature-layer fusion method is used, combining the feature information of various data types according to their weights, while simultaneously calculating the covariance between data to optimize the fusion result. After fusion, the fused data is transmitted to the topology mapping module and the geographic data processing module, respectively. This module improves the accuracy of multi-mode data fusion by dynamically adjusting weights and optimizing the fusion algorithm, so that the fused data can not only reflect the operating status of power grid equipment, but also contain accurate geospatial information, providing a high-quality data foundation for subsequent topology modeling and geographic data processing.

[0028] The topology map modeling module constructs a node relationship model for power transmission and transformation equipment based on fused data and transmits the model parameters to the geographic data processing module.

[0029] Specifically, the topology mapping modeling module constructs a topology model of power transmission and transformation equipment based on fused data. During the modeling process, equipment node information is first extracted from the fused data, including tower nodes, conductor nodes, and insulator nodes. The node extraction accuracy is such that the identification error does not exceed 0.3m. Then, the node attributes are quantified. Tower node attributes include height (measurement range 5-100m, accuracy ±0.1m) and material (converted to numerical codes, e.g., steel is 1, concrete is 2). Conductor node attributes include cross-sectional area (range 50-630mm²). 2 Accuracy ±1mm 2The module calculates the spatial distance and electrical connection between each node. Nodes with a spatial distance of less than 10m and an electrical connection are marked as adjacent nodes and assigned an adjacency coefficient ranging from 0 to 1, with the coefficient increasing as the distance increases. A topology matrix is ​​then constructed, with dimensions matching the number of equipment nodes. Matrix elements represent the strength of the association between nodes. The topology model is validated; if the node association error rate exceeds 3%, node information is re-extracted and the model is rebuilt. After modeling, the topology model parameters are transmitted to the geographic data processing module. This module, by constructing an accurate equipment topology model, clearly reflects the spatial and electrical relationships between power transmission and transformation equipment, providing topological support for equipment management and fault location in power transmission and transformation projects. This ensures that geographic information data not only includes spatial location but also equipment association information, enhancing the application value of the data.

[0030] The geographic data processing module performs spatial coordinate mapping calculations on the fused data and topology model parameters to generate a basic dataset of geographic information for power transmission and transformation projects.

[0031] Specifically, the geographic data processing module integrates the fused data and topology model parameters to generate a basic geographic information dataset for power transmission and transformation projects. The processing includes spatial coordinate mapping, further calibrating the coordinates of the fused data to a dedicated engineering coordinate system (such as the Beijing 54 coordinate system) with a mapping accuracy of ±0.05m. Then, data attribute association is performed, linking equipment association information in the topology model with terrain and feature information in the geographic data, with an association accuracy requirement of no less than 98%. During implementation, the module first standardizes the format of the fused data, uniformly converting it to SHP format. Then, the data is processed in segments according to the power transmission and transformation lines, with each 5km segment representing a data segment. The data for towers, conductors, and insulators in each segment is associated with the elevation, vegetation, and road data of the corresponding area, generating a comprehensive data record containing equipment parameters, spatial location, topological relationships, and geographic environment. Simultaneously, the data undergoes quality checks, including data completeness (missing rate less than 2%) and accuracy (error less than 0.1m). After passing these checks, the data is stored in blocks of 1GB each, with a spatial index established, and the index query response time is no more than 1 second. The resulting basic dataset can be directly used in the design, construction, operation and maintenance of power transmission and transformation projects. This module, through comprehensive data integration and processing, forms a complete and accurate basic dataset of geographic information for power transmission and transformation projects, providing data support for the entire project lifecycle, solving the problems of scattered and low-correlation traditional data, and improving project management efficiency.

[0032] Preferably, the quality perception fusion module adopts a power grid multi-mode quality perception fusion model, the model expression of which is: Among them, F fusion ω represents the feature value of the fused data. iLet D be the weighting coefficient for the i-th type of power grid multi-mode data. i This is the i-th type of power grid multi-mode data (conductor sag / tower coordinates / insulator leakage current). Let be the variance of the i-th type of power grid multimode data. The variance of the corresponding geographic data after registration is given by λ, where λ is the covariance adjustment coefficient, and cov(D) is the variance of the geographic data after registration. i G i Let be the covariance between the i-th type of power grid multi-mode data and the registered geographic data; the ICP registration calculation module adopts the multi-source geographic data ICP registration algorithm, the algorithm expression is: Among them, T opt For the optimal registration transformation matrix, SE(3) is the three-dimensional Euclidean transformation group, and G is the transformation matrix. j Let G' be the coordinates of the j-th feature point in the satellite remote sensing image data. j Let be the coordinates of the j-th feature point in the drone aerial data, δ be the distance attenuation coefficient, and m be the total number of feature points.

[0033] Specifically, the optimization of the power grid multi-mode quality perception fusion model in the quality perception fusion module and the multi-source geographic data ICP registration algorithm in the ICP registration calculation module is discussed. During implementation, the weighting coefficients for various data types in the fusion model are set based on data reliability. Tower coordinate data, with a positioning accuracy of ±3cm, has the highest reliability and a weighting coefficient of 0.3; conductor sag data, with a measurement accuracy of ±0.05m, has a weighting coefficient of 0.25; insulator leakage current data, with a resolution of 1μA, has a weighting coefficient of 0.2; registered satellite remote sensing image data, with a resolution of 0.5m, has a weighting coefficient of 0.15; and ground-measured elevation data, with an accuracy of ±0.1m, has a weighting coefficient of 0.1. The covariance adjustment coefficient is set to 0.8 based on data correlation. When the correlation between two types of data is high, this coefficient is increased to strengthen the influence of covariance on the fusion result. In the registration algorithm, the number of iterations for the 3D Euclidean transform group is set to 20 to ensure convergence of the transformation matrix calculation. The distance attenuation coefficient is set to 0.5m based on the feature point distribution density. When the distance between feature points exceeds this value, the attenuation factor decreases significantly, reducing interference from invalid matches. The total number of feature points is calculated per 100m. 2 Density calculations were performed on a single sample to ensure coverage of the entire power transmission and transformation project area. During implementation, a comprehensive feature value was first calculated using a fusion model, and then the optimal transformation matrix was obtained using a registration algorithm. This unified the coordinates of the multi-source data, controlling the registration error within 0.1m, significantly improving data fusion accuracy and registration precision, and providing a high-quality data foundation for subsequent topology modeling and geographic data processing.

[0034] Preferably, the ICP registration calculation module further includes a registration accuracy correction submodule, and the correction formula is: Where ΔT is the corrected registration transformation matrix, I is the identity matrix, α is the correction step size coefficient, and E is the registration error function. For the registration error function T opt The partial derivative of the transpose matrix; the quality-sensing fusion module also includes a fusion weight update formula: in, The weight coefficients for the (k+1)th iteration are... The weight coefficients for the k-th iteration are... The signal-to-noise ratio of the fused data in the (k+1)th iteration is given. Let be the signal-to-noise ratio of the fused data in the k-th iteration.

[0035] Specifically, the registration accuracy correction of the ICP registration calculation module and the fusion weight update of the quality perception fusion module are optimized. During registration accuracy correction, the identity matrix is ​​used as the reference matrix, and the correction step size coefficient is dynamically adjusted according to the magnitude of the registration error. When the registration error is 0.1-0.2m, the coefficient is set to 0.3; when the error is 0.05-0.1m, the coefficient is set to 0.1, avoiding over-correction that could lead to data distortion. The registration error function is obtained by calculating the sum of squares of the differences between the actual coordinates and the transformed coordinates of the feature points. Its partial derivative with respect to the transpose of the transformation matrix reflects the error change trend, thus guiding the correction direction. In the fusion weight update, the number of iterations is set according to the data acquisition cycle, with one iteration completed every 5 minutes. The signal-to-noise ratio (SNR) of the fused data is calculated during each iteration. When the SNR drops from 25dB to below 20dB, the corresponding data weight is reduced by 10%, ensuring that the weights match the data quality in real time. During implementation, the registered geographic data is first corrected for accuracy, keeping the error within 0.05m. Then, the fusion weights are updated based on the changes in the signal-to-noise ratio (SNR) of the data during the iteration process, ensuring that high-reliability data always dominates. Through this optimization, the registration accuracy is improved by 30% compared to traditional methods, and the SNR of the fused data is maintained above 20dB, effectively solving the problems of poor fusion results caused by unstable registration accuracy and fixed weights in traditional methods.

[0036] Preferably, the topology mapping module constructs a device topology relationship model, the model expression of which is: Among them, M topo Let A be the device topology matrix. p,r Let N be the adjacency coefficient between the p-th device node and the r-th device node. p Let N be the attribute vector (tower height / conductor type / insulator type) of the p-th device node. r Let be the attribute vector of the r-th device node. For tensor product operations, diag(L) p ,L r Let L be the spatial distance between the p-th device node and the r-th device node. p ,Lr The matrix is ​​a diagonal matrix with diagonal elements, where q is the total number of device nodes; the geographic data processing module uses the data spatial mapping formula: P geo =proj(F fusion M topo )·R utm +T offset Among them, P geo Let R be the final geographic coordinates, proj(·) be the projection transformation function, and R be the coordinates. utm T is the transformation matrix for the UTM coordinate system. offset This represents the coordinate offset.

[0037] Specifically, the topology mapping between the equipment topology modeling module and the geographic data processing module is optimized. When constructing the topology model, the total number of equipment nodes is determined based on the scale of the power transmission and transformation project. A 50km line typically includes approximately 200 tower nodes, 400 conductor nodes, and 800 insulator nodes. The adjacency coefficient is set based on the spatial distance between nodes, with a coefficient of 1 within 5m and decreasing to 0.1 inversely proportional to the distance between 5-10m. The node attribute vector includes parameters such as height, material, and cross-sectional area. Tower height measurements range from 5-100m, and conductor cross-sectional areas range from 50-630mm². 2 All data is quantized and encoded according to actual measured values. Tensor product operations are used to integrate the attribute associations between two nodes. The spatial distance in the diagonal matrix uses measured values ​​to ensure that the matrix elements accurately reflect the association strength. In the data spatial mapping, the projection transformation function is selected according to the characteristics of the target coordinate system, the UTM coordinate system transformation matrix is ​​set in conjunction with the regional latitude and longitude, and the coordinate offset is determined through on-site calibration and controlled within ±0.05m. During implementation, a topology relationship matrix is ​​first constructed to verify that the node association error rate is less than 3%. Then, the fused data is transformed to the engineering-specific coordinate system through spatial mapping to ensure that the geographic coordinates and topology information correspond accurately. This optimization achieves an accuracy rate of over 98% in associating the topology model with geographic data, providing accurate spatial-topology association data for power transmission and transformation equipment management and fault location.

[0038] Preferably, the power grid multi-mode data acquisition module adopts a data acquisition frequency control formula: Among them, f sample f0 is the actual sampling frequency, and var(D) is the reference sampling frequency. real ) represents the variance of the actual power grid multi-mode data, var(D pre ) represents the variance of the preset multi-mode power grid data, and β is the frequency adjustment coefficient; the multi-source geographic data receiving module uses the data receiving priority calculation formula: Among them, P prio Q represents the receiving priority of the j-th type of geographic data. j R is the resolution of the j-th type of geographic data. jLet τ be the update rate of the j-th type of geographic data. j Let the transmission delay of the j-th type of geographic data be denoted by loss. j Let be the packet loss rate of the j-th type of geographic data.

[0039] Specifically, the frequency control of the power grid multi-mode data acquisition module and the data reception priority calculation of the multi-source geographic data receiving module were optimized. In the frequency control, the baseline acquisition frequency was set to 5 minutes / time, and the actual acquisition frequency was adjusted according to changes in data variance. When the actual power grid data variance increased by 20% compared to the preset value, the frequency was increased to 3 minutes / time; when the variance decreased by 20%, the frequency was reduced to 10 minutes / time, ensuring that the data could reflect the equipment status in real time while avoiding redundant acquisition. The frequency adjustment coefficient was set according to the importance of the equipment: the coefficient for tower coordinate data was 1.2, and the coefficient for insulator leakage current data was 1.0, balancing the acquisition density of critical and general data. In the data reception priority calculation, satellite remote sensing image data had a resolution of 0.5m, an update cycle of 7 days, a transmission delay of 10s, and a packet loss rate of 2%; UAV aerial photography data had a resolution of 0.1m, an update cycle of 1 day, a transmission delay of 5s, and a packet loss rate of 1%; and ground measured elevation data had a resolution of 0.1m, an update cycle of 3 days, a transmission delay of 3s, and a packet loss rate of 0.5%. During implementation, the acquisition frequency is first dynamically adjusted based on the data variance, and then the receiving order is determined according to the priority calculation formula. UAV aerial photography data, with the highest priority, is received first, while satellite remote sensing image data, with the lowest priority, is received last. Through this optimization, data acquisition efficiency is improved by 25%, and the integrity of received data reaches 99%, meeting the project's requirements for data real-time performance and integrity.

[0040] Preferably, the quality perception fusion module further includes an abnormal data removal formula: in, For the i-th type of power grid multi-mode data after removing anomalies, H(·) is a step function. Let be the mean of the i-th type of power grid multi-mode data. The threshold for identifying abnormal data; the topology graph modeling module uses a node importance evaluation formula: Among them, I p Let deg(N) be the importance index of the p-th device node. p ) represents the degree of the p-th device node, deg(N) r ) represents the degree of the r-th device node, ∈ is used to avoid the minimum value where the denominator is zero, and γ is the importance adjustment coefficient.

[0041] Specifically, optimizations were made to the anomaly removal in the quality perception fusion module and the node importance assessment in the topology map modeling module. For anomaly removal, the anomaly threshold was set at 3 times the data standard deviation. For tower coordinate data, the standard deviation was 5cm, so the threshold was 15cm; for conductor sag data, the standard deviation was 0.05m, so the threshold was 0.15m. When data exceeded the corresponding threshold, a step function triggered the removal mechanism to filter out the anomaly. In node importance assessment, the node degree was calculated based on the number of connected devices. The node degree for main towers was approximately 10, and for branch towers, it was approximately 3. To avoid a minimum value of zero, the degree was set to 0.01. The importance adjustment coefficient was set based on the node's role in the power grid, with a node coefficient of 1.5 for main towers and 1.0 for branch towers. In implementation, anomaly removal was first performed on the fused data to ensure an anomaly rate below 1%. Then, the importance index of each device node was calculated, maintaining the main tower node importance index above 5 and the branch tower node importance index in the range of 2-3. This optimization can effectively remove outliers in the data, reduce interference with the fusion results, and identify key equipment nodes, providing a basis for the division of key areas for operation and maintenance of power transmission and transformation projects, thereby improving operation and maintenance efficiency by 20% and shortening fault location time by 30%.

[0042] Preferably, the topology graph modeling module includes a device node extraction unit, a node attribute quantification unit, an adjacency relationship construction unit, and a topology matrix generation unit. The device node extraction unit performs feature point detection on the fused data transmitted by the quality perception fusion module, filtering out tower center points, conductor connection points, and insulator installation points in the power transmission and transformation project. It distinguishes nodes corresponding to different equipment types through coordinate clustering and establishes preliminary node identification information. The node attribute quantification unit performs numerical conversion on the physical parameters corresponding to the extracted device nodes, converting non-numerical attributes such as tower height, conductor cross-sectional area, and insulator insulation level into standardized values. It then constructs a node feature vector containing spatial location and physical attributes, combining this with node coordinate information. The adjacency relationship construction unit calculates the spatial distance and electrical connection relationship between each device node, sets a distance threshold and electrical connectivity judgment condition, and marks nodes as adjacent nodes and assigns an adjacency coefficient when the spatial distance between nodes is less than the threshold and an electrical connection path exists. The topology matrix generation unit, based on the adjacency relationship and node feature vector, uses matrix operations to construct a topology relationship matrix containing node attributes and adjacency information, storing the correspondence between matrix elements and device node identifiers.

[0043] Specifically, the topology graph modeling module includes a device node extraction unit, a node attribute quantification unit, an adjacency relationship construction unit, and a topology matrix generation unit. When processing the fused data transmitted from the quality perception fusion module, the device node extraction unit uses a feature point detection algorithm with a detection accuracy set to a recognition deviation of no more than 0.3m. It distinguishes between tower center points, conductor connection points, and insulator installation points through coordinate clustering, with a clustering radius of 0.5m to avoid confusion between different device nodes. Simultaneously, it establishes preliminary node identification information, including node type and initial coordinates. The node attribute quantification unit measures tower height (measurement range 5-100m, accuracy ±0.1m) and conductor cross-sectional area (50-630mm²). 2 Accuracy ±1mm 2 Non-numerical attributes such as insulator insulation class are converted into standardized values, e.g., insulation class A corresponds to a value of 1, and class B corresponds to a value of 2. Feature vectors are constructed based on node coordinates, with a 6-dimensional vector to cover key attributes. The adjacency relationship construction unit calculates the spatial distance between nodes, with a distance threshold of 10m. Simultaneously, it determines electrical connection paths using a loop impedance detection method; an impedance below 5Ω is considered an electrical connection, and nodes meeting this condition are marked as adjacent. The adjacency coefficient is set according to distance: 1 for 0-5m and linearly decreasing to 0.1 for 5-10m. The topology matrix generation unit constructs a matrix based on adjacency relationships and feature vectors. The matrix dimension is consistent with the total number of nodes, and the element value is the product of the adjacency coefficient and the attribute association value. A mapping table between matrix elements and node identifiers is established during storage to ensure accurate matching in subsequent calls. Through the collaboration of these units, the node identification accuracy of the topology model reaches over 98%, providing a reliable foundation for equipment association analysis.

[0044] Preferably, the geographic data processing module includes a spatial coordinate transformation unit, a data layer overlay unit, a geographic attribute association unit, and a basic dataset generation unit. The spatial coordinate transformation unit receives fused data transmitted by the quality-aware fusion module and topological model parameters transmitted by the topological map modeling module, reads the original coordinate system information contained in the data, calls a coordinate transformation algorithm to uniformly transform data from different sources to the target coordinate system, calculates the coordinate deviation during the transformation process, and records the deviation value. The data layer overlay unit divides the transformed fused data into different layers according to data type, including a conductor data layer, a tower data layer, and a geographic background layer. The spatial position of each layer is adjusted through a layer alignment algorithm to ensure that the overlap of corresponding feature points between layers meets the set requirements. The geographic attribute association unit extracts attribute information from the data of each layer, establishes attribute association rules, associates conductor sag data with the elevation data of the corresponding geographic area, and associates tower coordinate data with the geological data of the corresponding area to generate an association data table containing multiple attributes. The basic dataset generation unit performs format standardization processing on the associated data, organizes the data structure according to the geographic information data specification for power transmission and transformation projects, stores the data in blocks and establishes an index to form a basic dataset of geographic information for power transmission and transformation projects that can be directly accessed.

[0045] Specifically, the implementation process of the spatial coordinate transformation unit, data layer overlay unit, geographic attribute association unit, and basic dataset generation unit in the geographic data processing module is as follows: After receiving the fused data and topology model parameters, the spatial coordinate transformation unit first reads the original coordinate system information. If it is the WGS-84 coordinate system, it is converted to the Beijing 54 coordinate system. The conversion process uses a seven-parameter method, with parameter errors controlled within ±0.001. The coordinate deviation is calculated and recorded; if the deviation exceeds 0.05m, the conversion is repeated. The data layer overlay unit divides the fused data into traverse, tower, and geographic background layers. The layer resolution is uniformly set to 0.1m. A feature point-based alignment algorithm is used, selecting no fewer than 20 evenly distributed feature points. After alignment, the feature point overlap must reach over 95% to ensure consistent spatial positioning of the layers. The geographic attribute association unit extracts attributes from each layer, such as conductor sag (range 0-10m, accuracy ±0.05m), tower coordinates (accuracy ±3cm), and geographic area elevation (accuracy ±0.1m). Association rules are established, such as binding conductor sag data with corresponding elevation data within a 100m range, generating an association data table. Each record in the table contains at least 5 association attributes, and the association error rate is controlled within 2%. The basic dataset generation unit organizes the structure according to the geographic information data specifications for power transmission and transformation projects, using block storage, with each block being 1GB in size. Data within a block is divided by spatial region, and an R-tree index is built. The index query response time is no more than 1 second. The final dataset must pass an integrity check, with a missing data rate of less than 1%, ensuring it can be directly used for engineering design and operation and maintenance, improving data application efficiency.

[0046] Preferably, the quality-aware fusion module includes a multi-mode data preprocessing unit, a data feature extraction unit, a fusion weight calculation unit, and a feature layer fusion unit. The multi-mode data preprocessing unit receives multi-mode power grid data transmitted from the power grid multi-mode data acquisition module and registered geographic data transmitted from the ICP registration calculation module. It performs format conversion and missing value imputation on the data, eliminates high-frequency noise in the data using a data smoothing algorithm, and retains effective feature information in the data. The data feature extraction unit performs feature extraction on the preprocessed multi-mode power grid data and registered geographic data respectively, and uses a feature detection algorithm to extract trend features and abrupt changes in the multi-mode power grid data. The system extracts spatial texture and contour features from the registered geographic data and converts these features into feature vectors. The fusion weight calculation unit calculates the fusion weights for various data types based on the variance and signal-to-noise ratio parameters of the data feature vectors using a weight allocation algorithm. It dynamically adjusts the weight coefficients according to data credibility, giving higher weights to data with higher credibility. The feature layer fusion unit uses a feature layer fusion algorithm to weight and combine the feature vectors of various data types according to the calculated fusion weights. It then optimizes the fusion result by combining the covariance information between data types, generating a fused comprehensive feature vector, which is then transmitted to the topology mapping module and the geographic data processing module.

[0047] Specifically, the quality perception fusion module's multi-mode data preprocessing unit, data feature extraction unit, fusion weight calculation unit, and feature layer fusion unit are implemented in detail. The multi-mode data preprocessing unit receives power grid multi-mode data and registered geographic data, converts the format to CSV, imputes missing values ​​using the mean impute method with an impute error not exceeding 10% of the data standard deviation, and smooths the data using a 3×3 window mean filtering algorithm, reducing high-frequency noise by more than 30% while retaining effective features. The data feature extraction unit extracts features from the preprocessed data. For power grid multi-mode data, a trend analysis algorithm is used to extract trend features, with a window size set to 10 acquisition cycles. Abrupt features are identified using the difference method; a difference exceeding twice the mean is considered an abrupt change. For geographic data, a texture analysis algorithm is used to extract spatial texture features, with a texture window of 5×5 pixels. Contour features are extracted using an edge detection algorithm with an edge extraction accuracy of 0.1m. All features are converted into 12-dimensional feature vectors. The fusion weight calculation unit calculates weights based on feature vector variance (variance of power grid multi-mode data controlled between 0.01 and 0.1, and variance of geographic data between 0.05 and 0.2) and signal-to-noise ratio (≥20dB). It uses the entropy weighting method for weight allocation, with an error of less than 5%. Weights are dynamically adjusted based on data reliability; for every 10% decrease in reliability, the weight is reduced by 8%. The feature layer fusion unit uses a weighted combination algorithm to fuse feature vectors according to the calculated weights. It then optimizes the results by combining the covariance (absolute value of covariance ≤0.5). The similarity of the fused feature vectors must be ≥90%. Finally, the fusion result is transmitted to subsequent modules, improving the accuracy of the fused data by 25% and providing high-quality input for topology modeling and geographic data processing.

[0048] The power grid multi-mode quality-aware fusion model in this invention is a core model for integrating power grid multi-mode data with registered geographic data, aiming to solve the problems of weak correlation and low fusion accuracy of data from different sources. Its implementation relies on four units of the quality-aware fusion module: the multi-mode data preprocessing unit first performs format unification (converting to CSV format), missing value imputation (mean imputation method, error not exceeding 10% of data standard deviation), and smoothing and denoising (3×3 window mean filtering, reducing high-frequency noise by more than 30%) on the power grid multi-mode data (conductor sag, tower coordinates, etc.) and registered geographic data; the data feature extraction unit uses a trend analysis algorithm (10 acquisition cycles per window) to extract the trend and abrupt change features of the power grid data, and performs texture analysis (5×5 pixel window), edge detection, and other techniques. The model extracts spatial texture and contour features from geographic data using measurements (accuracy 0.1m), converting all features into 12-dimensional vectors. A fusion weight calculation unit allocates weights based on feature vector variance (0.01-0.1 for power grid data, 0.05-0.2 for geographic data) and signal-to-noise ratio (≥20dB) using entropy weighting (error less than 5%), and dynamically adjusts the weights according to confidence level (weights are reduced by 8% if confidence level drops by 10%). A feature layer fusion unit fuses the vectors using a weighted combination algorithm, optimizing the results by combining covariance (absolute value ≤0.5) to ensure a similarity of ≥90% for the fused vectors. This model improves the accuracy of multi-mode data fusion (by 25% compared to traditional methods), enabling the fused data to simultaneously include the operating status of power grid equipment and geospatial information. It provides a high-quality data foundation for subsequent topology modeling and geographic data processing, preventing data quality issues from affecting the accuracy of power transmission and transformation engineering design and operation and maintenance decisions.

[0049] The multi-source geographic data ICP registration algorithm in this invention is used to eliminate coordinate deviations in multi-source geographic data such as satellite remote sensing imagery, UAV aerial photography, and measured ground elevation. Its core objective is to achieve coordinate unification across multi-source geographic data. This is achieved through an ICP registration calculation module: first, it receives various types of geographic data transmitted from the multi-source geographic data receiving module, and then registers the data in increments of 100m. 2Extract the density of one feature point and select the feature points for registration; then calculate the Euclidean distance between different data feature points, and define the feature points with a distance of less than 0.5m as matching pairs. Iterate the optimal registration transformation matrix based on the three-dimensional Euclidean transformation group (SE(3)), set the number of iterations to 20, and the convergence threshold to 0.001m. If the registration error exceeds 0.1m, reselect the matching pair; the subsequent optimization can also be achieved through the registration accuracy correction submodule. Based on the identity matrix, dynamically adjust the correction step size coefficient according to the registration error (coefficient 0.3 when the error is 0.1-0.2m, 0.1 when the error is 0.05-0.1m). Combine the partial derivative of the error function with respect to the transpose of the transformation matrix to correct the transformation matrix so that the final registration error is controlled within 0.05m. The algorithm aims to unify the coordinate system of multi-source geographic data (such as converting various types of data to the WGS-84 coordinate system), reduce spatial deviation, solve the problems of low accuracy and poor data compatibility of traditional registration methods, provide a consistent spatial benchmark for the effective integration of multi-mode data of power grid and geographic data, and ensure the spatial accuracy of geographic information data of power transmission and transformation projects.

[0050] The equipment topology knowledge graph modeling platform in this invention is a platform built upon the topology graph modeling module to present the node relationships of power transmission and transformation equipment, aiming to clearly reflect the spatial and electrical connections of the equipment. Its implementation relies on four modules: an equipment node extraction unit performs feature point detection on the fused data (deviation ≤ 0.3m), and distinguishes nodes such as tower centers, conductor connections, and insulator installations through coordinate clustering with a 0.5m radius, establishing identification information including type and initial coordinates; a node attribute quantification unit quantifies tower height (5-100m, accuracy ±0.1m), conductor cross-sectional area (50-630mm²), etc. 2 Accuracy ±1mm 2 The platform standardizes non-numerical attributes such as insulator insulation class (A-level = 1, B-level = 2, etc.) and constructs a 6-dimensional feature vector based on coordinates. The adjacency relationship construction unit calculates the spatial distance between nodes (threshold 10m), uses loop impedance detection (impedance < 5Ω indicates continuity) to determine electrical connections, sets adjacency coefficients based on distance (0-5m = 1, 5-10m linearly decreases to 0.1), and marks adjacent nodes. The topology matrix generation unit constructs a matrix (dimension = total number of nodes) based on adjacency relationships and feature vectors, with elements being the product of adjacency coefficients and attribute association values, and establishes a mapping table between matrix elements and node identifiers. The platform's function is to generate topology models with node identification accuracy ≥ 98%, provide equipment association data, solve the problem of disconnect between traditional topology modeling and geographic data, provide topology support for power transmission and transformation engineering equipment management and fault location, and improve engineering operation and maintenance efficiency and fault diagnosis accuracy.

[0051] like Figure 2As shown, the multi-mode data fusion geographic information basic data processing system for power transmission and transformation projects includes the following steps: First, the power grid multi-mode data acquisition module periodically collects conductor sag data, tower coordinate data, and insulator leakage current data from the power transmission and transformation project. Simultaneously, the multi-source geographic data receiving module receives satellite remote sensing imagery, UAV aerial photography data, and ground-measured elevation data. Second, the multi-source geographic data received by the multi-source geographic data receiving module is transmitted to the ICP registration calculation module. The ICP registration algorithm performs coordinate transformation and registration calculations on the multi-source geographic data to eliminate coordinate deviations between different sources. Third, the multi-mode data collected by the power grid multi-mode data acquisition module and the registered geographic data output by the ICP registration calculation module are transmitted to a quality control system. The first step involves a quality sensing fusion module that uses a multi-mode power grid quality sensing fusion model to perform feature layer fusion calculations on two types of data to generate fused data. The second step involves transmitting the fused data to a topology mapping module. Based on the equipment parameters and spatial location information in the fused data, a node association model for power transmission and transformation equipment is constructed, generating an equipment topology relationship matrix. The third step involves transmitting the fused data and the topology relationship matrix to a geographic data processing module. This module performs spatial coordinate mapping calculations on the fused data and associates the topology relationship matrix with geographic data to generate comprehensive data containing equipment topology information and geographic attributes. The fourth step involves standardizing the format of the comprehensive data output by the geographic data processing module, storing it in blocks according to data type and spatial region, establishing a data index, and forming a basic dataset of geographic information for power transmission and transformation projects.

[0052] This multi-mode data fusion geographic information basic data processing system for power transmission and transformation projects boasts numerous significant advantages, comprehensively enhancing the efficiency of geographic information basic data processing for these projects. It accurately captures key data such as conductor sag and tower coordinates through a multi-mode power grid data acquisition module, coupled with a multi-source geographic data receiving module to acquire geographic data from satellite remote sensing and UAV aerial photography, achieving comprehensive data acquisition. An ICP registration calculation module calibrates the coordinates of the multi-source geographic data, significantly improving the consistency of coordinates across different sources. A quality-aware fusion module dynamically adjusts weights based on data reliability, significantly improving the accuracy of multi-mode data feature layer fusion. A topology mapping modeling module constructs equipment node relationships based on the fused data, effectively improving the comprehensiveness and accuracy of the equipment topology model. Finally, a geographic data processing module integrates the fused data and topology parameters to generate a comprehensive geographic information basic dataset, significantly improving the completeness, accuracy, and application value of geographic information data for power transmission and transformation projects, providing more reliable data support for engineering design, equipment operation and maintenance, and fault diagnosis.

[0053] This system effectively overcomes the shortcomings of existing technologies and addresses the deficiencies of traditional data processing. Addressing the issues of large registration deviations and poor fusion effects in existing multi-source geographic data, the system utilizes the ICP registration algorithm to perform precise coordinate calibration of multi-source geographic data, improving registration accuracy. Simultaneously, it employs a quality-aware fusion model to dynamically optimize weights based on data credibility, strengthening the role of high-credibility data and significantly enhancing the multi-model data fusion effect. Furthermore, addressing the disconnect between equipment topology modeling and geographic data processing, resulting in insufficient data usability in existing technologies, the system deeply integrates fused data through a topology mapping module. This fully incorporates power grid operation data and geographic data characteristics, improving the adaptability of equipment topology relationships to actual engineering scenarios. The geographic data processing module then synchronously integrates topology information and geographic data, enhancing the dataset's comprehensiveness. This ensures that the generated geographic information dataset possesses both geographic attributes and topological relationships, significantly improving data usability.

[0054] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "set," "install," "connect," "link," and "fix" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal communication between two components. Those skilled in the art will understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0055] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various equivalent changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A multi-mode data fusion geographic information basic data processing system for power transmission and transformation projects, characterized in that, include: The module includes a power grid multi-mode data acquisition module, a multi-source geographic data receiving module, an ICP registration calculation module, a quality perception fusion module, a topology map modeling module, and a geographic data processing module. The power grid multi-mode data acquisition module collects conductor sag data, tower coordinate data, and insulator leakage current data in the power transmission and transformation project and transmits them to the quality perception fusion module. The multi-source geographic data receiving module receives satellite remote sensing image data, UAV aerial photography data, and ground measured elevation data and transmits them to the ICP registration calculation module. The ICP registration calculation module performs coordinate transformation calculations on the received multi-source geographic data and then transmits it to the quality perception fusion module. The quality perception fusion module performs feature layer fusion calculations on the power grid multi-mode data and the registered geographic data and then transmits them to the topology map modeling module and the geographic data processing module, respectively. The topology map modeling module constructs a node association model of power transmission and transformation equipment based on the fused data and transmits the model parameters to the geographic data processing module. The geographic data processing module performs spatial coordinate mapping calculations on the fused data and the topology model parameters to generate a basic dataset of geographic information for the power transmission and transformation project. The quality perception fusion module adopts a power grid multi-mode quality perception fusion model, the model expression of which is: ;in, The feature values ​​of the fused data For the first Weighting coefficients for multi-mode data of power grids For the first Multi-mode data of power grids For the first Variance of multi-mode data for power grids The variance of the corresponding geographic data after registration. This is the covariance adjustment coefficient. For the first The covariance between multi-mode data of the power grid and the registered geographic data; the ICP registration calculation module adopts the multi-source geographic data ICP registration algorithm, the algorithm expression is: The optimal registration transformation matrix is... It is a three-dimensional Euclidean transformation group. For the first satellite remote sensing image data Coordinates of feature points The first drone aerial data Coordinates of feature points This is the distance attenuation coefficient. This represents the total number of feature points; The ICP registration calculation module also includes a registration accuracy correction submodule, and the correction formula is: ;in, The corrected registration transformation matrix, It is the identity matrix. To correct the step size factor, The registration error function is... For the registration error function pair The partial derivative of the transpose matrix; the quality-sensing fusion module also includes a fusion weight update formula: ,in, For the first Weight coefficients for the next iteration For the first Weight coefficients for the next iteration For the first Signal-to-noise ratio of fused data in the next iteration For the first Signal-to-noise ratio of fused data in the next iteration.

2. The multi-mode data fusion geographic information basic data processing system for power transmission and transformation projects according to claim 1, characterized in that, The topology mapping modeling module constructs a device topology relationship model, and the model expression is: ;in, This is a device topology matrix. For the first The device node and the first The adjacency coefficient of each device node. For the first The attribute vector of each device node. For the first The attribute vector of each device node. For tensor product operations, For the first The device node and the first Spatial distance of each device node A diagonal matrix with diagonal elements. The total number of device nodes; the geographic data processing module uses the data spatial mapping formula: ;in, For the final geographic information coordinates, Let be the projection transformation function. This is the transformation matrix for the UTM coordinate system. This represents the coordinate offset.

3. The multi-mode data fusion geographic information basic data processing system for power transmission and transformation projects according to claim 1, characterized in that, The power grid multi-mode data acquisition module adopts a data acquisition frequency control formula: ;in, This is the actual sampling frequency. As the reference sampling frequency, The variance of the actual power grid multi-mode data. To preset the variance of the multi-mode power grid data, The frequency adjustment coefficient is used; the multi-source geographic data receiving module uses the data receiving priority calculation formula: ;in, For the first Priority for receiving geographic data For the first Resolution of geographic data For the first The update rate of geographic data For the first Transmission latency of geographic data For the first Packet loss rate in the transmission of geographic data.

4. The multi-mode data fusion geographic information basic data processing system for power transmission and transformation projects according to claim 2, characterized in that, The quality perception fusion module also includes an abnormal data removal formula: ;in, For the first time after removing anomalies Multi-mode data of power grids It is a step function. For the first The mean of multi-mode data for power grids, The threshold for identifying abnormal data; the topology graph modeling module uses a node importance evaluation formula: ;in, For the first The importance index of each device node. For the first The degree of each device node For the first The degree of each device node To avoid the minimum value where the denominator is zero, This is the importance adjustment coefficient.

5. The multi-mode data fusion geographic information basic data processing system for power transmission and transformation projects according to claim 1, characterized in that, The topology graph modeling module includes a device node extraction unit, a node attribute quantification unit, an adjacency relationship construction unit, and a topology matrix generation unit. The device node extraction unit performs feature point detection on the fused data transmitted by the quality perception fusion module, filters out the tower center points, conductor connection points, and insulator installation points in the power transmission and transformation project, distinguishes nodes corresponding to different equipment types through coordinate clustering, and establishes preliminary node identification information. The node attribute quantification unit performs numerical conversion on the physical parameters corresponding to the extracted device nodes, converts non-numerical attributes such as tower height, conductor cross-sectional area, and insulator insulation level into standardized values, and constructs a node feature vector containing spatial location and physical attributes by combining node coordinate information. The adjacency relationship construction unit calculates the spatial distance and electrical connection relationship between each device node, sets a distance threshold and an electrical connectivity judgment condition, and marks the node as an adjacency node and assigns an adjacency coefficient when the spatial distance between nodes is less than the threshold and an electrical connection path exists. The topology matrix generation unit constructs a topology matrix containing node attributes and adjacency information based on adjacency relationships and node feature vectors, and stores the correspondence between matrix elements and device node identifiers.

6. The multi-mode data fusion geographic information basic data processing system for power transmission and transformation projects according to claim 1, characterized in that, The geographic data processing module includes a spatial coordinate transformation unit, a data layer overlay unit, a geographic attribute association unit, and a basic dataset generation unit. The spatial coordinate transformation unit receives fused data transmitted by the quality-aware fusion module and topological model parameters transmitted by the topological map modeling module, reads the original coordinate system information contained in the data, calls a coordinate transformation algorithm to uniformly transform data from different sources to the target coordinate system, calculates the coordinate deviation during the transformation process, and records the deviation value. The data layer overlay unit divides the transformed fused data into different layers according to data type, including a conductor data layer, a tower data layer, and a geographic background layer. It adjusts the spatial position of each layer through a layer alignment algorithm to ensure that the overlap of corresponding feature points between layers meets the set requirements. The geographic attribute association unit extracts attribute information from the data of each layer, establishes attribute association rules, associates the sag data of the conductor with the elevation data of the corresponding geographic area, and associates the tower coordinate data with the geological data of the corresponding area, generating an association data table containing multiple attributes. The basic dataset generation unit performs format standardization processing on the associated data, organizes the data structure according to the data specifications for geographic information of power transmission and transformation projects, stores the data in blocks and establishes an index to form a basic dataset for geographic information of power transmission and transformation projects that can be directly accessed.

7. The multi-mode data fusion geographic information basic data processing system for power transmission and transformation projects according to claim 1, characterized in that, The quality-aware fusion module includes a multi-mode data preprocessing unit, a data feature extraction unit, a fusion weight calculation unit, and a feature layer fusion unit. The multi-mode data preprocessing unit receives multi-mode data from the power grid multi-mode data acquisition module and registered geographic data from the ICP registration calculation module. It performs format conversion and missing value imputation on the data, eliminates high-frequency noise through a data smoothing algorithm, and retains effective feature information. The data feature extraction unit extracts features from the preprocessed multi-mode data and registered geographic data respectively. It uses a feature detection algorithm to extract trend and abrupt change features from the multi-mode data and spatial texture and contour features from the registered geographic data, converting the extracted features into feature vectors. The fusion weight calculation unit calculates the fusion weights for various data types based on the variance and signal-to-noise ratio parameters of the data feature vectors, dynamically adjusting the weight coefficients according to data credibility, so that data with high credibility receives higher weights. The feature layer fusion unit adopts a feature layer fusion algorithm to weight and combine the feature vectors of various types of data according to the calculated fusion weights, optimize the fusion result by combining the covariance information between data, generate the fused comprehensive feature vector, and transmit it to the topology map modeling module and the geographic data processing module.

8. A geographic information basic data processing system for power transmission and transformation projects based on multi-mode data fusion according to any one of claims 1-7, characterized in that, The system operation includes the following steps: First, the multi-mode data acquisition module periodically collects conductor sag data, tower coordinate data, and insulator leakage current data from power transmission and transformation projects. Simultaneously, the multi-source geographic data receiving module receives satellite remote sensing imagery, UAV aerial photography data, and measured ground elevation data. Second, the multi-source geographic data received by the multi-source geographic data receiving module is transmitted to the ICP registration calculation module. The ICP registration algorithm performs coordinate transformation and registration calculations on the multi-source geographic data to eliminate coordinate deviations between different sources. Third, the multi-mode data acquired by the multi-mode data acquisition module and the registered geographic data output by the ICP registration calculation module are transmitted to the quality perception fusion module. The multi-mode data is then processed using the multi-mode data acquisition module. The quality-aware fusion model performs feature-layer fusion calculations on the two types of data to generate fused data. The fourth step involves transmitting the fused data to the topology mapping module. Based on the equipment parameters and spatial location information in the fused data, a node association model for power transmission and transformation equipment is constructed, generating an equipment topology matrix. The fifth step involves transmitting the fused data and the topology matrix to the geographic data processing module. Spatial coordinate mapping calculations are performed on the fused data, and attribute associations are established between the topology matrix and geographic data to generate comprehensive data containing equipment topology information and geographic attributes. The sixth step involves standardizing the format of the comprehensive data output from the geographic data processing module, storing it in blocks according to data type and spatial region, establishing a data index, and forming a basic dataset of geographic information for power transmission and transformation projects.

Images