Intelligent detection method of power grid frame and satellite image processing system

Abstract

The invention relates to the technical field of intelligent operation inspection of power transmission lines, in particular to an intelligent detection method of a power grid frame and a satellite image processing system.The method comprises the steps that the power grid frame is detected according to collected multi-source monitoring information, information of a plurality of power grid structures is obtained, and a power grid topology model is established; performing connection analysis on the power grid topology model to obtain a power grid topology analysis result, and performing spatial distribution analysis on the power grid frame by collecting coordinate information of the power grid frame and the power grid topology analysis result to obtain power grid coverage data. And performing data visualization on the plurality of pieces of power grid structure information, the power grid topology analysis result and the power grid coverage data to generate a power grid frame detection result. According to the invention, the method effectively solves a problem that the weak edge region is liable to leak detection in the recognition of the power grid frame in a large-range and high-complexity scene, remarkably improves the recognition efficiency and accuracy of the power grid frame, and reduces the cost and time of manual inspection.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent operation and inspection of power transmission lines, and in particular to an intelligent detection method for a power grid and a satellite image processing system. Background Art

[0002] With the continuous expansion of the scale of power systems and the continuous improvement of their intelligence levels, the monitoring and efficient management of power grids have become the core needs of the development of the power industry. Satellite remote sensing technology has gradually become an important tool for monitoring and analyzing power grids due to its wide coverage, short cycle for acquiring multiple types of data, and high spatial resolution. Especially in complex terrains and remote areas, satellite remote sensing technology can provide large-scale, high-precision information on power grid infrastructure, providing strong data support for power grid planning, operation and maintenance, and disaster warning.

[0003] However, the existing intelligent detection methods for power grids still have significant deficiencies in multi-source data fusion, transmission line continuity optimization, substation layout dynamic evaluation, and tower feature extraction. Traditional solutions mostly rely on a single remote sensing data source, which makes it difficult to fully capture the multi-dimensional characteristics of the power grid, resulting in limited recognition accuracy. In addition, the existing methods fail to effectively integrate multi-source remote sensing data and power grid operation historical data in power grid topology modeling and spatial distribution analysis, resulting in a low match between the model and the actual power grid structure, especially in key links such as transmission line continuity optimization, substation layout evaluation, and tower feature extraction.

[0004] The information disclosed in this background technology section is only intended to deepen the understanding of the overall background technology of the present disclosure, and should not be regarded as acknowledging or suggesting in any form that the information constitutes the prior art known to those skilled in the art. Summary of the invention

[0005] The present invention provides a cardiac stress response regulation method based on data analysis, which can effectively solve the problems in the background technology.

[0006] In order to achieve the above object, the technical solution adopted by the present invention is:

[0007] An intelligent detection method for a power grid, the method comprising:

[0008] Collect multi-source monitoring information, detect the power grid frame according to the multi-source monitoring information, and obtain some power grid structure information;

[0009] Establishing a power grid topology model according to the power grid structure information, and performing connection analysis on the power grid topology model to obtain a power grid topology analysis result;

[0010] Collecting grid coordinate information, performing spatial distribution analysis on the grid according to the grid coordinate information and the grid topology analysis result, and obtaining grid coverage data;

[0011] The grid structure information, grid topology analysis results and grid coverage data are visualized to generate grid frame detection results.

[0012] Further, performing spatial distribution analysis on the power grid frame according to the power grid frame coordinate information and the power grid topology analysis result includes:

[0013] According to the grid coordinate information, the geographical coordinates of the transmission line, the geographical coordinates of the tower and the geographical coordinates of the substation are respectively obtained;

[0014] Generate a transmission line spatial path according to the transmission line geographical coordinates by fitting, and calculate the tower distribution density according to the tower geographical coordinates;

[0015] Acquire regional load center data according to the geographic information system, perform spatial overlay analysis on the substation geographic coordinates and the regional load center data, and obtain substation layout assessment results;

[0016] Based on the transmission line spatial path, tower distribution density and substation layout assessment results, the spatial coverage of the power grid is drawn.

[0017] Furthermore, the substation geographic coordinates and regional load center data are spatially superimposed and analyzed to obtain substation layout assessment results, including:

[0018] Obtaining coordinates of load centers and load demand data of several regions according to the geographic information system;

[0019] According to each of the load center coordinates, the load-substation mapping relationship is matched with a plurality of the substation geographical coordinates to obtain the load-substation mapping relationship, and the spatial straight-line distance is calculated according to the load-substation mapping relationship, wherein the load-substation mapping relationship represents the substation geographical coordinates with the shortest distance to each of the load center coordinates;

[0020] The capacity of each substation is collected respectively, and the load demand data of the corresponding load center is determined according to the load-substation mapping relationship and the substation capacity, and the power station capacity matching degree is calculated;

[0021] The spatial threshold and the capacity threshold are set according to the historical power system operation data, and the substation layout is evaluated according to the matching degree between the spatial straight-line distance and the power station capacity to obtain the substation layout evaluation result.

[0022] Further, the power grid is detected according to the multi-source monitoring information, including:

[0023] Performing grayscale conversion on the remote sensing image in the multi-source monitoring information to obtain a grayscale image, and performing noise reduction processing on the grayscale image;

[0024] Calculating the gradient magnitude and gradient direction of the image according to the denoised grayscale image, and identifying potential edge areas according to the gradient magnitude and gradient direction;

[0025] Screening pixels along the gradient direction based on the potential edge region to obtain a refined edge region;

[0026] A gradient amplitude threshold is set based on historical power system operation data, and threshold segmentation is performed according to the gradient amplitude threshold to obtain a strong edge area and a weak edge area;

[0027] It is detected whether the pixel points of the weak edge area are connected to the strong edge area. If yes, the weak edge area is retained; otherwise, the weak edge area is removed, and the weak edge area is connected to the strong edge area to obtain a continuous transmission line.

[0028] Further, obtaining a continuous transmission line and optimizing the generation of the continuous transmission line includes:

[0029] Perform binary segmentation on the continuous power transmission line based on a deep learning algorithm to obtain a binary segmentation image;

[0030] extracting historical transmission line characteristics according to the historical power system operation data, and constructing a linear kernel according to the historical transmission line characteristics;

[0031] The binary segmented image is edge-filled and denoised according to the linear kernel, and linear fitting is performed on the binary segmented image after the edge-filled and denoised process to optimize the continuous power transmission line.

[0032] Further, the power grid is detected according to the multi-source monitoring information, including:

[0033] Extracting the edge contour of the tower according to the multi-source monitoring information, and calculating the geometric characteristics of the tower;

[0034] Determine local feature points of the contour according to the edge contour of the pole tower, smooth and correct the edge contour of the pole tower based on the local feature points of the contour, and obtain the contour of the pole tower area;

[0035] According to the outline of the tower area, the spatial distribution of pixel gray values ​​is obtained, and several local tower texture features are extracted respectively;

[0036] Performing feature enhancement on a number of local tower texture features to generate a texture response map;

[0037] The tower information is obtained according to the tower geometric features, local tower texture features and texture response map.

[0038] Further, the power grid is detected according to the multi-source monitoring information, including:

[0039] Extracting a substation area image according to the multi-source monitoring information, locating substation equipment according to the substation area image, and obtaining a number of substation boundary boxes;

[0040] Extracting and classifying the overall features of the substation from the substation area image to generate a probability distribution map of the functional areas of the substation;

[0041] Performing pixel-level classification on the substation area image to obtain mask images of several power station equipment;

[0042] The boundary range of the power station is determined according to the substation boundary box, the functional area probability distribution map and the power station equipment mask map.

[0043] Furthermore, a power grid topology model is established according to the power grid structure information, including:

[0044] Extracting transmission line information and grid facility information based on some of the grid structure information;

[0045] Determine the grid nodes and grid paths respectively according to the grid facility information and the transmission line information, and create an adjacency matrix according to the number of the grid nodes, wherein each row and column of the adjacency matrix corresponds to one grid node;

[0046] Traversing the power grid path corresponding to the transmission line information, determining the power grid nodes connected to the power grid path, representing the correspondingly connected power grid nodes as 1, and the rest as 0, and filling in the adjacency matrix;

[0047] The connection relationship of the power grid facilities is determined according to the adjacency matrix, and the power grid configuration attributes are added respectively to construct a power grid topology model.

[0048] A satellite image processing system, the system comprising:

[0049] A power grid structure extraction module collects multi-source monitoring information, detects the power grid frame according to the multi-source monitoring information, and obtains some power grid structure information;

[0050] A power grid topology analysis module, which establishes a power grid topology model according to the power grid structure information, and performs connection analysis on the power grid topology model to obtain a power grid topology analysis result;

[0051] A power grid spatial analysis module collects power grid frame coordinate information, performs spatial distribution analysis on the power grid frame according to the power grid frame coordinate information and the power grid topology analysis result, and obtains power grid coverage data;

[0052] The power grid visualization display module visualizes some of the power grid structure information, power grid topology analysis results and power grid coverage data to generate power grid frame detection results.

[0053] Furthermore, the power grid space analysis module includes:

[0054] A power grid coordinate extraction unit, which respectively obtains the geographical coordinates of the transmission line, the geographical coordinates of the tower and the geographical coordinates of the substation according to the power grid grid coordinate information;

[0055] A transmission path modeling unit is configured to generate a transmission line spatial path according to the geographical coordinates of the transmission line and calculate the distribution density of the towers according to the geographical coordinates of the towers;

[0056] The substation layout analysis unit obtains regional load center data according to the geographic information system, performs spatial superposition analysis on the substation geographic coordinates and the regional load center data, and obtains a substation layout evaluation result;

[0057] The power grid coverage analysis unit draws the spatial coverage of the power grid frame based on the spatial path of the transmission line, the distribution density of the towers and the substation layout evaluation results.

[0058] The technical solution of the present invention can achieve the following technical effects:

[0059] It effectively solves the problem that traditional methods are prone to missing detection in weak edge areas in power grid monitoring under large-scale and high-complexity scenarios. Through satellite remote sensing technology, it can quickly and widely obtain spatial distribution information of power grids. Combined with automatic recognition algorithms, it significantly improves the efficiency and accuracy of power grid recognition and reduces the cost and time of manual inspections. The identified power grid structure is checked with ledgers such as power line distribution maps. Through accurate recognition of power grids and spatial information management, it provides a data basis for the planning, construction and operation and maintenance of smart grids, and helps the digital transformation and intelligent upgrading of power systems.

[0060] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0062] Figure 1 A schematic diagram of a flow chart of an intelligent detection method for a power grid;

[0063] Figure 2 It is a schematic diagram of the process of spatial distribution analysis of power grid;

[0064] Figure 3 Schematic diagram of the process for obtaining substation layout assessment results;

[0065] Figure 4 A schematic diagram for the spatial distribution analysis of the power grid;

[0066] Figure 5 This is the architecture diagram of the satellite image processing system. DETAILED DESCRIPTION

[0067] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0068] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more related listed items.

[0069] Embodiment 1;

[0070] like Figure 1 As shown, the present application provides an intelligent detection method for a power grid, the method comprising:

[0071] S100: Collect multi-source monitoring information, detect the power grid according to the multi-source monitoring information, and obtain some power grid structure information;

[0072] S200: establishing a power grid topology model according to a number of power grid structure information, and performing connection analysis on the power grid topology model to obtain a power grid topology analysis result;

[0073] S300: Collecting grid coordinate information, performing spatial distribution analysis on the grid according to the grid coordinate information and grid topology analysis results, and obtaining grid coverage data;

[0074] S400: Visualize some grid structure information, grid topology analysis results and grid coverage data to generate grid frame detection results.

[0075] Specifically, first, by collecting multi-source monitoring information, such as power grid operation status information, equipment monitoring data, environmental and meteorological data, etc., to ensure comprehensive coverage of the power grid, then preprocess the collected multi-source monitoring information, including radiation correction, geometric correction and image enhancement; after preprocessing, the multi-source monitoring information is used to detect the power grid, which includes transmission lines, tower substations, power plants and substations, etc. For example, the following method can be used to extract and identify power plants: Use YOLOv5 to detect landmark structures such as chimneys, cooling towers, photovoltaic panels, wind turbines and dams to obtain key target information of power plants, and further use ResNet Extract target features and use classifiers to identify station types. The simultaneous appearance of chimneys and cooling towers indicates a thermal power plant, large photovoltaic panels indicate a photovoltaic power station, tall wind turbines indicate a wind power plant, and dams correspond to hydroelectric power plants. For areas with a variety of power generation equipment, comprehensively analyze target distribution and feature matching to identify hybrid energy stations. Then, based on the identified grid structure information, extract the geographic coordinates of transmission lines, towers, and substations, and build a grid topology model. Use topological analysis methods (such as short path algorithms and network centrality analysis) to evaluate the connection characteristics of the grid, identify possible isolated nodes or grid redundant areas, and optimize the grid layout. In the process of analyzing the spatial distribution of the grid, collect grid coordinate information and combine it with the results of grid topology analysis to evaluate and analyze the spatial coverage of the grid and obtain grid coverage data. The obtained grid structure information, topological analysis results and spatial coverage data are visualized to generate grid frame detection results. After that, the geographic information system (GIS) software can be used to visualize the grid frame data, including key structures such as transmission lines, towers, substations, power plants, etc., and the topological visualization method (such as force-directed graph layout) is used to display the connectivity of the grid. In addition, different grid structures are classified, including radial structure, ring structure, mesh structure, chain structure, hybrid structure, double-circuit structure, multi-terminal DC transmission structure, hierarchical partition structure and microgrid structure, intuitively presenting different types of grid topology. At the same time, the spatial coverage of the transmission line is drawn in combination with the geographic information system, and the power supply capacity and load coverage of the substation are evaluated through grid load matching analysis, providing scientific decision support for grid planning and management.

[0076] The technical solution of the present invention effectively solves the problem that weak edge areas are easily missed in the monitoring of power grids in large-scale and high-complexity scenarios by traditional methods. Satellite remote sensing technology can be used to quickly and widely acquire spatial distribution information of power grids. Combined with the automatic recognition algorithm, the efficiency and accuracy of power grid recognition are significantly improved, and the cost and time of manual inspections are reduced. The identified power grid structure is verified with the ledgers such as the power line distribution map. Through the accurate recognition of the power grid and the management of spatial information, a data basis is provided for the planning, construction and operation and maintenance of smart grids, which helps the digital transformation and intelligent upgrading of the power system.

[0077] Further, if Figure 2 and Figure 4 As shown, the spatial distribution analysis of the power grid is performed based on the power grid coordinate information and the power grid topology analysis results, including:

[0078] S310: Obtaining the geographical coordinates of the transmission line, the geographical coordinates of the tower and the geographical coordinates of the substation respectively according to the grid coordinate information;

[0079] S320: Generate a transmission line spatial path according to the geographic coordinates of the transmission line, and calculate the distribution density of the towers according to the geographic coordinates of the towers;

[0080] S330: Acquire regional load center data according to the geographic information system, perform spatial overlay analysis on the substation geographic coordinates and the regional load center data, and obtain substation layout assessment results;

[0081] S340: Based on the evaluation results of the spatial path of transmission lines, tower distribution density and substation layout, draw the spatial coverage of the power grid.

[0082] As a preferred embodiment of the above, firstly, the grid coordinate information is obtained by extracting remote sensing images, and corrected in combination with historical grid data to ensure the accuracy of the path. The grid coordinate information includes the geographical coordinates of the transmission line, the geographical coordinates of the tower and the geographical coordinates of the substation; after obtaining the basic geographical data, the transmission line coordinates are analyzed: the affine transformation matrix can be used to convert the pixels of the transmission line extracted from the remote sensing image into geographical coordinates, and then the spline interpolation or Bezier curve method can be used to smoothly fit the path of the transmission line to eliminate the influence of remote sensing image noise and errors, and ensure the continuity and accuracy of the transmission line. In addition, the spatial distribution density of the geographical coordinates of the tower is calculated: preferably, the calculation is performed by dividing the fixed area Calculate the number of poles and towers per unit area, and detect the distribution pattern of poles and towers through spatial clustering algorithms (such as DBSCAN and K-Means) to identify dense areas, sparse areas and possible missing points of poles and towers; further, in order to optimize the spatial layout of the power grid, obtain regional load center data based on the geographic information system, and perform spatial superposition analysis on the geographic coordinates of the substation and the load center data to evaluate the rationality and power supply capacity of the substation. After completing the transmission line path fitting, pole tower distribution density calculation and substation layout evaluation, draw the spatial coverage of the power grid. This process comprehensively considers the spatial path of the transmission line, the pole tower distribution density and the substation layout evaluation results, and can use Thiessen polygons or triangulation to generate the coverage map of the power grid.

[0083] Further, if Figure 3 As shown in the figure, the geographical coordinates of the substation and the regional load center data are spatially superimposed and analyzed to obtain the substation layout assessment results, including:

[0084] S331: Obtaining coordinates of load centers and load demand data of several regions according to a geographic information system;

[0085] S332: Match each load center coordinate with a plurality of substation geographical coordinates to obtain a load-substation mapping relationship, and calculate the spatial straight-line distance according to the load-substation mapping relationship, wherein the load-substation mapping relationship represents the substation geographical coordinate with the shortest distance to each load center coordinate;

[0086] S333: Collect the capacity of each substation respectively, determine the load demand data of the corresponding load center according to the load-substation mapping relationship and the substation capacity, and calculate the power station capacity matching degree;

[0087] S334: Set a spatial threshold and a capacity threshold according to historical power system operation data, and evaluate the substation layout according to the spatial straight-line distance and the power station capacity matching degree to obtain the substation layout evaluation result.

[0088] In this embodiment, first, the geographic coordinates and corresponding load demand data of multiple regional load centers are extracted based on the geographic information system. The load center refers to an area with intensive electricity demand, such as a city core area, an industrial park, a large commercial area, etc. Its data can be obtained from channels such as the power grid management system, smart meter data, and an urban electricity demand forecasting system. In addition, combined with nighttime light remote sensing data, high-energy consumption areas can be further identified to enhance the accuracy of the data. Then, the geographic space matching method is used to match each load center with the substation to form a load-substation mapping relationship; specifically, the Euclidean distance between the load center and the substation is calculated, and the nearest substation is selected as its main power supply station based on the nearest neighbor algorithm to form a mapping relationship. After obtaining the load-substation mapping relationship, the power supply capacity of the substation is further calculated to ensure that its load distribution is reasonable: First, the rated capacity of each substation is extracted from the power system database, and the actual load of the substation is calculated in combination with the load-substation mapping relationship. According to the actual load of the power station, the power station capacity matching degree is calculated. The power station capacity matching degree calculation formula is: , is the capacity matching degree of substation j, , is the total power supply load of the substation, is the rated capacity of the substation, where the total power supply load is obtained based on the addition of all load demand data; when the power station capacity matching degree is greater than 1, it means that the substation is in an overloaded operation state, and it is necessary to adjust the power supply range or add a new substation; when the power station capacity matching degree is less than 0.5, it means that the substation has power supply redundancy, and the load distribution may need to be readjusted. Finally, based on the above calculation results, the rationality of the current substation layout is evaluated to obtain the substation layout evaluation result.

[0089] Furthermore, the power grid is tested based on multi-source monitoring information, including:

[0090] Convert the remote sensing images in the multi-source monitoring information into grayscale, obtain grayscale images, and perform noise elimination on the grayscale images;

[0091] The gradient magnitude and gradient direction of the image are calculated based on the denoised grayscale image, and the potential edge area is identified based on the gradient magnitude and gradient direction;

[0092] Filter pixels along the gradient direction based on the potential edge area to obtain the refined edge area;

[0093] A gradient amplitude threshold is set based on historical power system operation data, and threshold segmentation is performed according to the gradient amplitude threshold to obtain strong edge areas and weak edge areas;

[0094] Detect whether the pixel points in the weak edge area are connected to the strong edge area. If so, retain the weak edge area, otherwise remove the weak edge area, connect the weak edge area with the strong edge area, and obtain a continuous transmission line.

[0095] Specifically, the remote sensing images in the collected multi-source monitoring information are first converted to grayscale, that is, the multi-band optical images are converted into single-channel grayscale images to reduce the computational complexity and enhance the edge information of the images. Gaussian filtering or bilateral filtering can be used for denoising. After denoising, the edge information of the images is extracted based on the gradient calculation method. The method is as follows: First, the Sobel or Prewitt operator can be used to perform convolution calculation on the denoised grayscale image, and the gradients in the x-axis and y-axis directions are calculated respectively. The gradient amplitude and gradient direction are calculated according to the gradients in the x-axis and y-axis directions. The gradient amplitude calculation formula is as follows: , Represents pixel The gradient strength at represents the gradient in the x direction, Represents the gradient in the y direction; the gradient direction calculation formula is as follows: , is the gradient direction, represents the gradient in the x direction, represents the gradient in the y direction; according to the gradient direction, potential edge regions are extracted, and possible transmission line regions are preliminarily identified. On this basis, non-maximum suppression can be used to screen pixels along the gradient direction, eliminate noise edges, and obtain refined edge regions; then the historical power system operation data is extracted and the gradient amplitude threshold is set. According to the set gradient amplitude threshold, the edge detection results are threshold segmented to obtain strong edge regions and weak edge regions, respectively. Among them, the strong edge region corresponds to an obvious transmission line, while the weak edge region may contain some breakpoints or low-contrast regions of the transmission line. Then, the connectivity analysis of the pixels in the weak edge region is performed to detect whether they are connected to the strong edge region. If there is a pixel-level connection between the weak edge region and the strong edge region, the weak edge region is retained, otherwise it is removed to reduce false detection. In the final optimization stage, the weak edge region is connected to the strong edge region, and the missing part is filled by morphological operations (such as closing operations) to ensure the integrity and continuity of the transmission line, thereby obtaining the final continuous transmission line.

[0096] Specifically, obtaining a continuous transmission line and optimizing the generation of the continuous transmission line include:

[0097] Perform binary segmentation on continuous transmission lines based on deep learning algorithms to obtain binary segmentation images;

[0098] Extract historical transmission line characteristics based on historical power system operation data, and construct a linear kernel based on the historical transmission line characteristics;

[0099] The binary segmentation image is processed with edge filling and noise reduction according to the linear kernel, and the binary segmentation image after edge filling and noise reduction is linearly fitted to optimize the continuous transmission line.

[0100] As a preferred embodiment of the above embodiment, the image of the continuous transmission line obtained is used as input, and the U-Net model is used for pixel-level segmentation in the following manner: based on the multi-source monitoring information obtained; and a training data set is established through manual annotation or automatic edge detection method, and cross entropy loss and optimization are used, and then the trained U-Net model can automatically detect the transmission line in the image and generate a binary segmentation map, in which the white pixel 1 represents the transmission line and the black pixel 0 represents the background; after the U-Net segmentation, the historical characteristics of the transmission line are extracted based on the historical power system operation data, including line direction characteristics, spatial distribution characteristics and topological consistency characteristics, and a linear kernel is constructed based on these characteristics; after the binary segmentation and linear kernel construction are completed, edge filling and noise reduction processing are further performed in the following manner: first, the morphological closing operation is used to connect the broken transmission line area to avoid line loss due to noise interference. At the same time, morphological opening operation is used to remove isolated noise points, reduce false detection, and improve the stability of recognition. In addition, in order to ensure the consistency of the direction of the transmission line, the directional gradient analysis is combined to optimize the filled transmission line to keep it consistent with the direction of the original line to avoid misconnection. Then, in this embodiment, the Hough transform is used to fit the optimized transmission line. The fitting process is as follows: extract preliminary edge information based on the continuous transmission line, and input it into the Hough transform for straight line detection. By constructing the Hough accumulator matrix, traverse all edge pixel points, calculate the polar coordinate parameters at different angles, and vote in the accumulator. Subsequently, the straight line with the highest number of votes is selected as the main direction of the transmission line. On this basis, the least squares method is used to calculate the fitting equation of the transmission line to obtain the optimized continuous transmission line.

[0101] Furthermore, the power grid is tested based on multi-source monitoring information, including:

[0102] Extract the edge contour of the tower based on multi-source monitoring information and calculate the geometric characteristics of the tower;

[0103] Determine local feature points of the contour according to the edge contour of the tower, smooth and correct the edge contour of the tower based on the local feature points of the contour, and obtain the contour of the tower area;

[0104] According to the outline of the tower area, the spatial distribution of pixel gray values ​​is obtained, and several local tower texture features are extracted respectively;

[0105] Enhance the texture features of some local towers and generate texture response maps;

[0106] The tower information is obtained based on the tower geometric characteristics, local tower texture characteristics and texture response diagram.

[0107] In this embodiment, the edge contour of the pole tower is first extracted according to the multi-source monitoring information, and the Canny edge detection algorithm is used to perform gradient calculation to identify the area with large gradient changes and highlight the contour structure of the pole tower. Subsequently, the geometric features of the pole tower, including area, perimeter, aspect ratio, rectangularity, etc., are calculated to characterize the spatial form of the pole tower; wherein the aspect ratio measures the structural characteristics of the pole tower by calculating the aspect ratio of the minimum circumscribed rectangle of the pole tower, and the rectangularity is used to evaluate the regularity of the tower structure; in this embodiment, the Arris corner detection algorithm is used to extract the key points of the pole tower, and the extraction method is as follows: the Sobel operator is used to calculate the gradients in the x and y directions, and a second-order gradient matrix is ​​constructed to calculate the corner point response value, and possible key points of the pole tower are screened out; after obtaining the edge contour and geometric features of the pole tower, the local texture features of the pole tower are further extracted, the spatial grayscale distribution of pixels in the tower area is analyzed by calculating the grayscale co-occurrence matrix, and texture features such as contrast, correlation, energy, and homogeneity are extracted. At the same time, the local binary pattern is used to calculate the detailed texture information of the pole tower to distinguish different types of pole towers; after extracting the local texture features of the pole tower, the Gabor filter is used to extract the multi-scale and multi-directional texture features of the pole tower to enhance the edge information and structural information of the tower area. Finally, the geometric characteristics of the pole tower, the local pole tower texture features and the texture response map are integrated to generate complete pole tower recognition information.

[0108] Furthermore, the power grid is tested based on multi-source monitoring information, including:

[0109] Extract substation area images based on multi-source monitoring information, locate substation equipment based on the substation area images, and obtain several substation boundary boxes;

[0110] Extract and classify the overall features of the substation area image to generate a probability distribution map of the functional area of ​​the substation;

[0111] Perform pixel-level classification on substation area images to obtain mask images of several power station equipment;

[0112] The boundary range of the power station is determined based on the substation boundary box, the functional area probability distribution map and the power station equipment mask map.

[0113] Specifically, the substation area image is extracted from the multi-source monitoring information. The YOLO target detection model can be used to detect the landmark equipment inside the substation (such as transformers, switchgear, busbars, etc.) to obtain the location information and bounding box of the equipment. The YOLO model can extract the key features of the substation area based on the convolutional neural network, and predict the location of each substation equipment through bounding box regression; after the target detection result is output, it is further combined with the geographic information system data to perform spatial calibration on the location of the substation to ensure that the substation equipment identification result is consistent with the actual power grid layout. After completing the target detection of the substation equipment, the substation area is classified as a whole. The ResNet deep learning model can be used to segment the remote sensing image of the substation: First, the remote sensing The image is preprocessed, including pixel normalization, size adjustment and data enhancement, to ensure the consistency of input data. Then the ResNet model extracts the spatial features of the substation through convolutional layers and residual blocks, uses global average pooling for feature dimensionality reduction, and calculates the classification probability of the substation area through the fully connected layer to generate a probability heat map of the core area (such as transformers, switchgear) and the peripheral area of ​​the substation. Different areas correspond to different functions of the substation, such as the main transformer area, high-voltage switch area, busbar area, control room area, etc. This distribution map can provide prior information for subsequent pixel-level classification. After obtaining the overall classification information of the substation, the U-Net deep learning model can be further used for pixel-level classification to obtain several power station equipment mask maps. Finally, based on the substation boundary box, the functional area probability distribution map and the power station equipment mask map, the spatial boundary range of the substation is optimized.

[0114] Furthermore, a power grid topology model is established based on several power grid structure information, including:

[0115] Extracting transmission line information and grid facility information based on certain grid structure information;

[0116] According to the grid facility information and the transmission line information, the grid nodes and the grid paths are respectively determined, and an adjacency matrix is ​​created according to the number of grid nodes, where each row and column of the adjacency matrix corresponds to a grid node;

[0117] Traverse the power grid path corresponding to the transmission line information, determine the power grid nodes connected to the power grid path, represent the corresponding connected power grid nodes as 1, and the rest as 0, and fill in the adjacency matrix;

[0118] The connection relationship of power grid facilities is determined according to the adjacency matrix, and the power grid configuration attributes are added respectively to build a power grid topology model.

[0119] As a preferred embodiment of the above, firstly, a target detection algorithm combined with an edge detection algorithm can be used to extract transmission line information and grid facility information based on a number of grid structure information, wherein the transmission line information includes the spatial path of the transmission line, the geographical coordinates of the tower, and the voltage level of the line; the grid facility information includes the location, capacity, and power supply range of the substation; after the grid structure information is extracted, the grid nodes and grid pathways are further determined, and an adjacency matrix is ​​created based on the number of nodes to represent the connection relationship between the grid facilities. The grid nodes include key facilities such as substations, towers, and load centers, and the grid pathway refers to the transmission lines connecting these nodes; wherein if there is a transmission line directly between two grid nodes, it is represented as 1, and if there is no transmission line, it is represented as 0. is shown as 0; traverse the transmission line information, determine the grid nodes connected to each transmission line, and fill in the adjacency matrix. The traversal can be carried out in the following ways: based on the geographic coordinate data of the tower and the transmission line, calculate the start and end points of the transmission line, and determine the grid facilities connected to it. Use the shortest path algorithm or breadth-first search to calculate the connectivity between the grid nodes, and assign weights to transmission lines of different voltage levels in the adjacency matrix. The weight value corresponds to the voltage level of the transmission line and can be used for grid load flow calculation; after constructing the adjacency matrix, further determine the connection relationship of the grid facilities, and add grid configuration attributes to each grid node, including node type, voltage level, power supply capacity, etc. Based on the adjacency matrix and grid configuration data, finally construct a grid topology model.

[0120] Embodiment 2:

[0121] Based on the same inventive concept as the intelligent detection method for power grid in the aforementioned embodiment, the present invention also provides a satellite image processing system, such as Figure 5 As shown, the system includes:

[0122] The power grid structure extraction module collects multi-source monitoring information, detects the power grid frame based on the multi-source monitoring information, and obtains some power grid structure information;

[0123] The power grid topology analysis module establishes a power grid topology model based on a number of power grid structure information, performs connection analysis on the power grid topology model, and obtains the power grid topology analysis results;

[0124] The power grid spatial analysis module collects the grid coordinate information, performs spatial distribution analysis on the grid according to the grid coordinate information and the grid topology analysis results, and obtains the grid coverage data;

[0125] The power grid visualization display module visualizes several power grid structure information, power grid topology analysis results and power grid coverage data to generate power grid frame detection results.

[0126] The above-mentioned adjustment system in the present invention can effectively realize an intelligent detection method of a power grid frame, and the technical effects that can be achieved are as described in the above-mentioned embodiments, which will not be repeated here.

[0127] Furthermore, the power grid spatial analysis module includes:

[0128] A power grid coordinate extraction unit, which obtains the geographical coordinates of the transmission line, the geographical coordinates of the tower and the geographical coordinates of the substation according to the grid coordinate information;

[0129] The transmission path modeling unit generates the transmission line spatial path according to the geographical coordinates of the transmission line and calculates the tower distribution density according to the geographical coordinates of the tower;

[0130] The substation layout analysis unit obtains regional load center data according to the geographic information system, performs spatial overlay analysis on the substation geographic coordinates and the regional load center data, and obtains the substation layout assessment result;

[0131] The grid coverage analysis unit draws the spatial coverage of the power grid based on the evaluation results of the spatial path of the transmission lines, the distribution density of the towers and the layout of the substations.

[0132] Similarly, the above-mentioned optimization schemes for the system can also respectively achieve the optimization effects corresponding to the method in Example 1, which will not be repeated here.

[0133] Although the present application has been described in conjunction with specific features and embodiments thereof, it is obvious that various modifications and combinations may be made thereto without departing from the spirit and scope of the present application. Accordingly, this specification and the accompanying drawings are merely exemplary illustrations of the present application as defined therein, and are deemed to have covered any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application is intended to include these modifications and variations.

Claims

1. An intelligent detection method for a power grid, characterized in that: The method comprises: Collect multi-source monitoring information, detect the power grid frame according to the multi-source monitoring information, and obtain some power grid structure information; Establishing a power grid topology model according to the power grid structure information, and performing connection analysis on the power grid topology model to obtain a power grid topology analysis result; Collecting grid coordinate information, performing spatial distribution analysis on the grid according to the grid coordinate information and the grid topology analysis result, and obtaining grid coverage data; The grid structure information, grid topology analysis results and grid coverage data are visualized to generate grid frame detection results.

2. The intelligent detection method of the power grid according to claim 1 is characterized in that: Performing spatial distribution analysis on the power grid frame according to the power grid frame coordinate information and the power grid topology analysis result includes: According to the grid coordinate information, the geographical coordinates of the transmission line, the geographical coordinates of the tower and the geographical coordinates of the substation are respectively obtained; Generate a transmission line spatial path according to the transmission line geographical coordinates by fitting, and calculate the tower distribution density according to the tower geographical coordinates; Acquire regional load center data according to the geographic information system, perform spatial overlay analysis on the substation geographic coordinates and the regional load center data, and obtain substation layout assessment results; Based on the transmission line spatial path, tower distribution density and substation layout assessment results, the spatial coverage of the power grid is drawn.

3. The intelligent detection method of the power grid according to claim 2 is characterized in that: The substation geographic coordinates and regional load center data are spatially superimposed and analyzed to obtain substation layout assessment results, including: Obtaining coordinates of load centers and load demand data of several regions according to the geographic information system; According to each of the load center coordinates, the load-substation mapping relationship is matched with a plurality of the substation geographical coordinates to obtain the load-substation mapping relationship, and the spatial straight-line distance is calculated according to the load-substation mapping relationship, wherein the load-substation mapping relationship represents the substation geographical coordinates with the shortest distance to each of the load center coordinates; The capacity of each substation is collected respectively, and the load demand data of the corresponding load center is determined according to the load-substation mapping relationship and the substation capacity, and the power station capacity matching degree is calculated; The spatial threshold and the capacity threshold are set according to the historical power system operation data, and the substation layout is evaluated according to the matching degree between the spatial straight-line distance and the power station capacity to obtain the substation layout evaluation result.

4. The intelligent detection method of the power grid according to claim 1 is characterized in that: Detecting the power grid according to the multi-source monitoring information includes: Performing grayscale conversion on the remote sensing image in the multi-source monitoring information to obtain a grayscale image, and performing noise reduction processing on the grayscale image; Calculating the gradient magnitude and gradient direction of the image according to the denoised grayscale image, and identifying potential edge areas according to the gradient magnitude and gradient direction; Screening pixels along the gradient direction based on the potential edge region to obtain a refined edge region; A gradient amplitude threshold is set based on historical power system operation data, and threshold segmentation is performed according to the gradient amplitude threshold to obtain a strong edge area and a weak edge area; It is detected whether the pixel points of the weak edge area are connected to the strong edge area. If yes, the weak edge area is retained; otherwise, the weak edge area is removed, and the weak edge area is connected to the strong edge area to obtain a continuous power transmission line.

5. The intelligent detection method of the power grid according to claim 4 is characterized in that: Obtaining a continuous transmission line and optimizing the generation of the continuous transmission line includes: Perform binary segmentation on the continuous power transmission line based on a deep learning algorithm to obtain a binary segmentation image; extracting historical transmission line characteristics according to the historical power system operation data, and constructing a linear kernel according to the historical transmission line characteristics; The binary segmented image is edge-filled and denoised according to the linear kernel, and linear fitting is performed on the binary segmented image after the edge-filled and denoised process to optimize the continuous power transmission line.

6. The intelligent detection method of power grid according to claim 1, characterized in that: Detecting the power grid according to the multi-source monitoring information includes: Extracting the edge contour of the tower according to the multi-source monitoring information, and calculating the geometric characteristics of the tower; Determine local feature points of the contour according to the edge contour of the pole tower, smooth and correct the edge contour of the pole tower based on the local feature points of the contour, and obtain the contour of the pole tower area; According to the outline of the tower area, the spatial distribution of pixel gray values ​​is obtained, and several local tower texture features are extracted respectively; Performing feature enhancement on a number of local tower texture features to generate a texture response map; The tower information is obtained according to the tower geometric features, local tower texture features and texture response map.

7. The intelligent detection method of power grid according to claim 1, characterized in that: Detecting the power grid according to the multi-source monitoring information includes: Extracting a substation area image according to the multi-source monitoring information, locating substation equipment according to the substation area image, and obtaining a number of substation boundary boxes; Extracting and classifying the overall features of the substation from the substation area image to generate a probability distribution map of the functional areas of the substation; Performing pixel-level classification on the substation area image to obtain mask images of several power station equipment; The boundary range of the power station is determined according to the substation boundary box, the functional area probability distribution map and the power station equipment mask map.

8. The intelligent detection method of power grid according to claim 1, characterized in that: A power grid topology model is established according to the power grid structure information, including: Extracting transmission line information and grid facility information based on some of the grid structure information; Determine the grid nodes and grid paths respectively according to the grid facility information and the transmission line information, and create an adjacency matrix according to the number of the grid nodes, wherein each row and column of the adjacency matrix corresponds to one grid node; Traversing the power grid path corresponding to the transmission line information, determining the power grid nodes connected to the power grid path, representing the correspondingly connected power grid nodes as 1, and the rest as 0, and filling in the adjacency matrix; The connection relationship of the power grid facilities is determined according to the adjacency matrix, and the power grid configuration attributes are added respectively to construct a power grid topology model.

9. A satellite image processing system, characterized in that: The system comprises: A power grid structure extraction module collects multi-source monitoring information, detects the power grid frame according to the multi-source monitoring information, and obtains some power grid structure information; A power grid topology analysis module, which establishes a power grid topology model according to the power grid structure information, and performs connection analysis on the power grid topology model to obtain a power grid topology analysis result; A power grid spatial analysis module collects power grid frame coordinate information, performs spatial distribution analysis on the power grid frame according to the power grid frame coordinate information and the power grid topology analysis result, and obtains power grid coverage data; The power grid visualization display module visualizes some of the power grid structure information, power grid topology analysis results and power grid coverage data to generate power grid frame detection results.

10. The satellite image processing system according to claim 9, characterized in that: The power grid space analysis module includes: A power grid coordinate extraction unit, which respectively obtains the geographical coordinates of the transmission line, the geographical coordinates of the tower and the geographical coordinates of the substation according to the power grid grid coordinate information; A transmission path modeling unit is configured to generate a transmission line spatial path according to the geographical coordinates of the transmission line and calculate the distribution density of the towers according to the geographical coordinates of the towers; The substation layout analysis unit obtains regional load center data according to the geographic information system, performs spatial overlay analysis on the substation geographic coordinates and the regional load center data, and obtains a substation layout evaluation result; The power grid coverage analysis unit draws the spatial coverage of the power grid frame based on the evaluation results of the spatial path of the transmission line, the distribution density of the towers and the layout of the substations.

Images