A GIS power grid topology anomaly identification method based on image recognition technology

Abstract

The application relates to the technical field of GIS power grid, and discloses a GIS power grid topology abnormality identification method based on image recognition technology, which comprises an operation simulation robot program, an online power grid portal system browser extension plug-in program and a GIS graph intelligent identification program; the GIS power grid topology abnormality identification method based on image recognition technology realizes the intelligentization and automation of online power grid system GIS graph identification management by designing three functional modules of the online power grid portal system chrome browser extension plug-in program, the operation simulation robot program and the GIS graph intelligent identification program and the organic cooperation combination among the three functional modules, and combining an online power grid GIS graph element image intelligent identification training model; the identification accuracy is improved; subjective judgment is converted into artificial intelligence quantitative judgment; multi-level full-coverage traversal intelligent analysis is realized; and the omission of manual analysis is avoided.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of GIS power grid, in particular to a GIS power grid topology anomaly identification method based on image recognition technology. BACKGROUND

[0002] The production systems in the power grid maintain GIS topology data of cables and conductors of different voltage levels, and the data is synchronized to the online power grid system to provide browsing and display based on GIS layers. Due to the requirements of core production data security management and information system security management, these data are encrypted and stored, and the personnel in the front line cannot directly obtain the decrypted data from the system or database. They can only browse the digital image information of the power grid topology GIS layer rendered by the browser GIS engine from the web portal page. For the error data in the massive GIS topology data (such as data entry errors, system import information missing, and missing lines), the personnel in the front line can only find the errors by browsing the page. In the interlaced and dense mesh GIS topology, the efficiency of manual error finding is very low. The GIS topology data covers a wide area, and the manual checking workload is huge and easy to miss. Therefore, it is necessary to invent a technical method to provide an automatic means for identifying GIS power grid topology anomalies. SUMMARY

[0003] In view of the deficiencies in the prior art, the present application provides a GIS power grid topology anomaly identification method based on image recognition technology, which has the advantages of developing an operation simulation robot program to simulate the operation of the maintenance personnel, realizing automatic scanning coverage of GIS power grid topology in the whole region, and introducing artificial intelligence image recognition technology to train the identification model, realizing automatic identification and information registration of the flying line phenomenon of GIS topology errors.

[0004] To achieve the above-mentioned purpose, the present application provides the following technical scheme: a GIS power grid topology anomaly identification method based on image recognition technology, comprising an operation simulation robot program, an online power grid portal system browser extension plug-in program, and a GIS graphics intelligent identification program, wherein:

[0005] The operation simulation robot program realizes the virtual mouse device function, provides mouse left button, right button, and scroll wheel operation events, and communicates with the online power grid portal system browser extension plug-in program and the GIS graphics intelligent identification program through the interface;

[0006] The online power grid portal system browser extension plug-in program is triggered to execute after the browser loads the online power grid portal system page, and the variable attribute value of the online power grid GIS graphics page is obtained;

[0007] The GIS graphic intelligent recognition program provides screenshot, reverse electrical element extraction and analysis of GIS graphic images, identification of flying wires and erroneous line segments, and acquisition and saving of parameter information of GIS graphic cable and conductor lines.

[0008] Furthermore, the browser extension plugin for the online power grid portal system includes the following steps:

[0009] Step 1-1: Load the online power grid portal system browser extension plugin into your browser's extensions. In the browser options, add the online power grid portal system URL to the permissions settings for "Allow this extension to read and change all data left on the websites you visit" so that the plugin can obtain the online power grid portal system page information.

[0010] Steps 1-2: Open the online power grid website in your browser, enter your account and password to log in, and then load the GIS graphic page;

[0011] Steps 1-3: After the online power grid URL is loaded, the browser extension program of the online power grid portal system can automatically recognize that the page has been loaded and can obtain the attribute variable parameter information of the online power grid GIS graphic page.

[0012] Furthermore, the operation simulation robot program includes the following steps:

[0013] Step 2-1: Open the robot simulation program, set the GIS Level start level parameters, end level parameters, level increment interval, GIS X / Y coordinate scanning range, and set the overlap coefficient between each frame screenshot and adjacent screenshots of the same scale in the four directions (up, down, left, and right).

[0014] Step 2-2: After the robot simulation program successfully loads the system driver and sets the parameters, it begins to simulate mouse control. The available operation commands include: moving and dragging by specified pixels up, down, left, and right, and zooming in and out using the mouse wheel.

[0015] Steps 2-3: The operation simulation robot program calls the interface of the browser extension plugin program of the online power grid portal system to obtain the initial parameter information of the current graphic page, including the GIS Level layer variable, the coordinate information of the four points of the GIS graphic window (top, bottom, left, and right), the GIS graphic window page resolution, etc., and then starts to adjust the browser GIS graphic window to start analyzing and displaying the GIS graphic content.

[0016] Steps 2-4: Operate the simulated robot program to zoom in and out of the GIS Level variables using the mouse wheel, and adjust the map coordinate range of the GIS graphics window using simulated directional drag commands;

[0017] Steps 2-5: The robot operation program simulates mouse operations to adjust the GIS graphic level and window GIS coordinate range according to the set input parameters. After each operation, the subsequent operation steps are paused, and the GIS graphic intelligent recognition program is notified through the interface to enter step 3-2 to perform the current GIS graphic window analysis operation.

[0018] Steps 2-6: During the pause of the robot simulation program, the system will listen for notifications indicating that the GIS graphic intelligent recognition program has completed its analysis task. Then, repeat steps 2-5 until all set levels and scanning ranges have completed their analysis operations.

[0019] Furthermore, the GIS graphic intelligent recognition program includes the following steps:

[0020] Step 3-1: After the GIS graphic intelligent recognition program is opened, wait for the operation simulation robot program interface call notification screenshot and image analysis and other subsequent work.

[0021] Step 3-2: After receiving the notification that the graphical window adjustment of the operation simulation robot program interface is complete, the GIS graphic intelligent recognition program obtains the Level parameter and GIS coordinate range parameter information of the current graphical window from the operation simulation robot program through the interface, and performs distance conversion of the window height and width.

[0022] Step 3-3: The GIS graphic intelligent recognition program performs a screenshot operation to obtain the height and width pixels of the GIS graphic window screenshot. Based on the window height and width distance calculated in Step 3-2, the height and width length represented by each pixel of the screenshot image are calculated.

[0023] Steps 3-4: The GIS graphic intelligent recognition program performs GIS cable and wire path identification and analysis on the screenshot image. For the specific analysis process, please refer to steps 4-1 to 4-4.

[0024] Step 3-5: If no abnormalities are found in the image identification and analysis in step 3-4, proceed to step 3-9.

[0025] Steps 3-6: If step 3-4 finds suspicious GIS electrical wiring or abnormal line segments in the image, calculate the reasonable pixel position of the line segment on the computer screen and notify the operation simulation robot program of the relevant screen pixel position information through the interface.

[0026] Steps 3-7: After receiving the relevant information, the simulated robot program simulates mouse operation to click on the corresponding line segment, triggering a pop-up window. The pop-up window contains the device parameter information of the line segment. After the pop-up window is successfully displayed, the simulated robot program interface notifies the GIS graphic intelligent recognition program that the currently selected device information has been displayed in the pop-up window.

[0027] Steps 3-8: The GIS graphic intelligent identification program takes a screenshot and uses image text recognition technology to identify the device name and device code information in the screenshot image. The GIS graphic intelligent identification program uses the device name and device code information identified by the image to query the actual length of the specific wire or cable segment from the system equipment ledger interface or database, and then compares it with the length analyzed by the image recognition. When the difference is large, it can be determined that there is a problem with the GIS information of the device, such as a flying wire or error. The program registers and archives the problem information for subsequent rectification and management.

[0028] Steps 3-9: Once the image analysis in the current GIS graphic window is complete, the GIS graphic intelligent recognition program calls the operation simulation robot program's listening interface to notify it that it can repeat steps 2-5.

[0029] Furthermore, the GIS graphic intelligent recognition program's startup steps for identifying and analyzing GIS cable and conductor paths in the image include:

[0030] Step 4-1: The GIS graphic intelligent identification program starts the GIS cable and conductor path identification and analysis;

[0031] Step 4-2: The GIS graphic intelligent identification program intelligently identifies and extracts symbol elements of equipment such as line segments and stations, and identifies and analyzes each cable and conductor segment based on the pre-completed image recognition training model.

[0032] Step 4-3: If step 4-2 finds that the line segments in the image exceed the image border, then perform multiple image overlap and stitching according to the direction of the line segments exceeding the border, and then repeat the analysis in step 4-2.

[0033] Step 4-4: Based on the pixel width and height calculation results from Step 3-3, the GIS graphic intelligent identification program calculates the length of cable segments and conductor segments. For those exceeding the set length range, they are identified as suspicious GIS electrical wiring jumpers or abnormal segments.

[0034] Furthermore, the distance conversion of the window height and width in step 3-2 is calculated using the distance between points A and B on the GIS sphere:

[0035] (1) C=sin(LatA)*sin(LatB)*cos(LonA-LonB)+cos(LatA)*cos(LatB)

[0036] (2) Distance=R*Arccos(C)*Pi / 180

[0037] Among them: (1) The latitude and longitude coordinates of point A: (LonA, LatA), and the latitude and longitude coordinates of point B: (LonB, LatB);

[0038] (2) R = 6371.004 km, Distance is the distance between two points, in km.

[0039] Furthermore, step 3-3 calculates the height and width length represented by each pixel in the screenshot image. The height and width length represented by each pixel are calculated as follows:

[0040] (1) Pixel height: Height = Window height / Number of pixels in the vertical direction of the window;

[0041] (2) Pixel width: Width = window width / number of pixels in the horizontal direction of the window.

[0042] The present invention has the following beneficial effects:

[0043] 1. This GIS power grid topology anomaly identification method based on image recognition technology, through the design of three functional modules—a Chrome browser extension plugin for the online power grid portal system, an operation simulation robot program, and a GIS graphic intelligent identification program—and their organic synergistic combination, combined with an intelligent image recognition training model for online power grid GIS graphic elements, realizes the intelligent and automated management of GIS graphic identification in the online power grid system, improves the accuracy of identification, transforms subjective judgment into quantitative judgment by artificial intelligence, and performs multi-level, full-coverage intelligent analysis, avoiding omissions in manual analysis.

[0044] 2. This GIS power grid topology anomaly identification method based on image recognition technology simulates the operation of maintenance personnel by operating a robot program to achieve automatic scanning coverage of the GIS power grid topology across the entire region. It also introduces artificial intelligence image recognition technology to train the identification model, enabling automatic identification and registration of flying wire phenomena in GIS topology errors. Compared with the traditional manual methods of graphic browsing, GIS power grid topology anomaly location, and problem recording and organization, it greatly saves human resources, requires almost no manual intervention, and can be repeated periodically. Attached Figure Description

[0045] Figure 1 This is a schematic diagram of the operation simulation robot program, the online power grid portal system browser extension plugin program, and the GIS graphic intelligent recognition program module of the present invention.

[0046] Figure 2 This is a schematic diagram illustrating an example of the image training part of the present invention. Detailed Implementation

[0047] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. Example

[0048] Please see Figure 1 A GIS-based power grid topology anomaly identification method based on image recognition technology includes an operation simulation robot program, an online power grid portal system browser extension plugin program, and a GIS graphic intelligent identification program, wherein:

[0049] Operating the robot simulation program: The operating robot simulation program implements the function of a virtual mouse device, provides left mouse button, right mouse button and scroll wheel operation events, and communicates with the online power grid portal system browser extension plug-in program and GIS graphic intelligent recognition program through interface communication.

[0050] The browser extension plugin for the online power grid portal system is triggered when the browser loads the online power grid portal system page to obtain the variable attribute values ​​of the online power grid GIS graphic page.

[0051] The GIS graphic intelligent recognition program provides screenshot, reverse electrical element extraction and analysis of GIS graphic images, identification of flying wires and erroneous line segments, and acquisition and saving of parameter information of GIS graphic cable and conductor lines. Example

[0052] Based on Embodiment 1, the distinguishing feature is that the online power grid portal system browser extension plugin program includes the following steps:

[0053] Step 1-1: Load the online power grid portal system browser extension plugin into your browser's extensions. In the browser options, add the online power grid portal system URL to the permissions settings for "Allow this extension to read and change all data left on the websites you visit" so that the plugin can obtain the online power grid portal system page information.

[0054] Steps 1-2: Open the online power grid website in your browser, enter your account and password to log in, and then load the GIS graphic page;

[0055] Steps 1-3: After the online power grid URL is loaded, the browser extension program of the online power grid portal system can automatically recognize that the page has been loaded and can obtain the attribute variable parameter information of the online power grid GIS graphic page.

[0056] Furthermore, the operation simulation robot program includes the following steps:

[0057] Step 2-1: Open the robot simulation program, set the GIS Level start level parameters and end level parameters (parameter range 1-22), set the level increment interval, set the GIS X / Y coordinate scanning range, and set the overlap coefficient between each frame screenshot and the adjacent same-scale screenshots in the four directions (up, down, left, and right).

[0058] Step 2-2: After the robot simulation program successfully loads the system driver and sets the parameters, it begins to simulate mouse control. The available operation commands include: moving and dragging by specified pixels up, down, left, and right, and zooming in and out using the mouse wheel.

[0059] Steps 2-3: The operation simulation robot program calls the interface of the browser extension plugin program of the online power grid portal system to obtain the initial parameter information of the current graphic page, including the GIS Level layer variable, the coordinate information of the four points of the GIS graphic window (top, bottom, left, and right), the GIS graphic window page resolution, etc., and then starts to adjust the browser GIS graphic window to start analyzing and displaying the GIS graphic content.

[0060] Steps 2-4: Operate the simulated robot program to zoom in and out of the GIS Level variables using the mouse wheel, and adjust the map coordinate range of the GIS graphics window using simulated directional drag commands;

[0061] Steps 2-5: The robot operation program simulates mouse operations to adjust the GIS graphic level and window GIS coordinate range according to the set input parameters. After each operation, the subsequent operation steps are paused, and the GIS graphic intelligent recognition program is notified through the interface to enter step 3-2 to perform the current GIS graphic window analysis operation.

[0062] Steps 2-6: During the pause of the robot simulation program, the system will listen for notifications indicating that the GIS graphic intelligent recognition program has completed its analysis task. Then, repeat steps 2-5 until all set levels and scanning ranges have completed their analysis operations. Example

[0063] Based on Embodiment 2, the distinguishing feature is that the GIS graphic intelligent recognition program includes the following steps:

[0064] Step 3-1: After the GIS graphic intelligent recognition program is opened, wait for the operation simulation robot program interface call notification screenshot and image analysis and other subsequent work.

[0065] Step 3-2: After receiving the notification that the graphical window adjustment of the operation simulation robot program interface is complete, the GIS graphic intelligent recognition program obtains the Level parameter and GIS coordinate range parameter information of the current graphical window from the operation simulation robot program through the interface, and performs distance conversion of the window height and width.

[0066] The conversion of window height and width is calculated using the distance between two points A and B on the GIS sphere:

[0067] (1) C=sin(LatA)*sin(LatB)*cos(LonA-LonB)+cos(LatA)*cos(LatB)

[0068] (2) Distance=R*Arccos(C)*Pi / 180

[0069] Among them: (1) The latitude and longitude coordinates of point A: (LonA, LatA), and the latitude and longitude coordinates of point B: (LonB, LatB);

[0070] (2) R = 6371.004 km, Distance is the distance between two points, in km.

[0071] Step 3-3: The GIS graphic intelligent recognition program performs a screenshot operation to obtain the height and width pixels of the GIS graphic window screenshot. Based on the window height and width distance calculated in Step 3-2, the height and width length represented by each pixel of the screenshot image are calculated.

[0072] This involves calculating the height and width represented by each pixel in the screenshot image. The height and width represented by each pixel are calculated as follows:

[0073] (1) Pixel height: Height = Window height / Number of pixels in the vertical direction of the window;

[0074] (2) Pixel width: Width = window width / number of pixels in the horizontal direction of the window.

[0075] Steps 3-4: The GIS graphic intelligent recognition program performs GIS cable and wire path identification and analysis on the screenshot image. For the specific analysis process, please refer to steps 4-1 to 4-4.

[0076] Step 3-5: If no abnormalities are found in the image identification and analysis in step 3-4, proceed to step 3-9.

[0077] Steps 3-6: If step 3-4 finds suspicious GIS electrical wiring or abnormal line segments in the image, calculate the reasonable pixel position of the line segment on the computer screen and notify the operation simulation robot program of the relevant screen pixel position information through the interface.

[0078] Steps 3-7: After receiving the relevant information, the simulated robot program simulates mouse operation to click on the corresponding line segment, triggering a pop-up window. The pop-up window contains the device parameter information of the line segment. After the pop-up window is successfully displayed, the simulated robot program interface notifies the GIS graphic intelligent recognition program that the currently selected device information has been displayed in the pop-up window.

[0079] Steps 3-8: The GIS graphic intelligent identification program takes a screenshot and uses image text recognition technology to identify the device name and device code information in the screenshot image. The GIS graphic intelligent identification program uses the device name and device code information identified by the image to query the actual length of the specific wire or cable segment from the system equipment ledger interface or database, and then compares it with the length analyzed by the image recognition. When the difference is large, it can be determined that there is a problem with the GIS information of the device, such as a flying wire or error. The program registers and archives the problem information for subsequent rectification and management.

[0080] Steps 3-9: Once the image analysis in the current GIS graphic window is complete, the GIS graphic intelligent recognition program calls the operation simulation robot program's listening interface to notify it that it can repeat steps 2-5. Example

[0081] refer to Figure 2 Based on Embodiment 3, the distinguishing feature is that the GIS graphic intelligent recognition program's startup image GIS cable and conductor path identification and analysis steps include:

[0082] Step 4-1: The GIS graphic intelligent identification program starts the identification and analysis of GIS cable and wire paths in the image;

[0083] Step 4-2: The GIS graphic intelligent recognition program intelligently identifies and extracts line segments and equipment symbol elements such as stations in the image, and identifies and analyzes each cable and conductor segment based on the pre-completed image recognition training model.

[0084] Image training example: For the scenes in (1), (2), and (3) below, they can be identified as a single device, while (4), (5), and (6) can be identified as multiple devices (Note: In actual GIS images, there are no line distinctions; the line differences here are just for illustrative purposes).

[0085] Step 4-3 If step 4-2 found Figure 2 If the middle line segment extends beyond the image border, multiple images are overlapped and stitched together according to the direction of the extension beyond the border, and then step 4-2 is re-analyzed.

[0086] Step 4-4: The GIS graphic intelligent identification program calculates the length of cable segments and conductor segments based on the pixel width and height calculation results of Step 3-3. For segments that exceed the set length range, they are identified as suspicious GIS electrical wiring jumpers or abnormal segments.

[0087] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various 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 method for identifying topological anomalies in GIS power grids based on image recognition technology, characterized in that: The identification method is based on an operation simulation robot program, an online power grid portal system browser extension plugin program, and a GIS graphic intelligent identification program, wherein: Operating the robot simulation program: The operating robot simulation program implements the function of a virtual mouse device, provides left mouse button, right mouse button and scroll wheel operation events, and communicates with the online power grid portal system browser extension plug-in program and GIS graphic intelligent recognition program through interface communication. The browser extension plugin for the online power grid portal system is triggered when the browser loads the online power grid portal system page to obtain the variable attribute values ​​of the online power grid GIS graphic page. The GIS graphic intelligent recognition program provides screenshot, reverse electrical element extraction and analysis of GIS graphic images, identification of flying wires and erroneous line segments, and acquisition and saving of parameter information of GIS graphic cables and conductors. The browser extension plugin for the online power grid portal system includes the following steps: Step 1-1: Load the Internet Power Grid Portal System browser extension plugin into your browser's extensions. In the browser options, under "Allow this extension to read and change all data left on the websites you visit," add the Internet Power Grid Portal System URL to the permissions settings. Steps 1-2: Open the online power grid website in your browser, enter your account and password to log in, and then load the GIS graphic page; Steps 1-3: After the online power grid website is loaded, the browser extension plugin of the online power grid portal system will automatically recognize that the page has been loaded and obtain the attribute variable parameter information of the online power grid GIS graphic page. The operation simulation robot program includes the following steps: Step 2-1: Open the robot simulation program, set the GISLevel start level parameter, end level parameter, level increment interval, GIS X / Y coordinate scanning range, and set the overlap coefficient between each frame screenshot and the adjacent same-scale screenshots in the four directions (up, down, left, and right). Step 2-2: After successfully loading the system driver and setting the parameters in the robot simulation program, start simulating mouse control; Steps 2-3: The operation simulation robot program calls the interface of the browser extension plugin program of the online power grid portal system to obtain the initial parameter information of the current graphic page, including the GIS Level layer variable, the coordinate information of the four points of the GIS graphic window (top, bottom, left, and right), and the page resolution of the GIS graphic window. Then, it starts to adjust the browser GIS graphic window to start analyzing and displaying the GIS graphic content. Steps 2-4: Operate the simulated robot program to zoom in and out of the GISLevel layer variable using the mouse wheel command, and adjust the map coordinate range of the GIS graphics window using simulated directional drag commands; Steps 2-5: The robot operation program simulates mouse operations to adjust the GIS graphic level and window GIS coordinate range according to the set input parameters. After each operation, the subsequent operation steps are paused, and the GIS graphic intelligent recognition program is notified through the interface to enter step 3-2 to perform the current GIS graphic window analysis operation. Steps 2-6: During the pause of the robot simulation program, the system will listen for notifications that the GIS graphic intelligent recognition program has completed its analysis task. Then, repeat steps 2-5 until all set levels and scanning ranges have completed their analysis operations. The GIS graphic intelligent recognition program includes the following steps: Step 3-1: After the GIS graphic intelligent recognition program is opened, wait for the operation simulation robot program interface call notification screenshot and image analysis follow-up work; Step 3-2: After receiving the notification that the graphical window adjustment of the operation simulation robot program interface is complete, the GIS graphic intelligent recognition program obtains the Level parameter and GIS coordinate range parameter information of the current graphical window from the operation simulation robot program through the interface, and performs distance conversion of the window height and width. Step 3-3: The GIS graphic intelligent recognition program performs a screenshot operation to obtain the height and width pixels of the GIS graphic window screenshot. Based on the window height and width distance calculated in Step 3-2, the height and width length represented by each pixel of the screenshot image are calculated. Steps 3-4: The GIS graphic intelligent recognition program performs GIS cable and wire path identification and analysis on the screenshot image. For the specific analysis process, please refer to steps 4-1 to 4-4. Step 3-5: If no abnormalities are found in the image identification and analysis in step 3-4, proceed to step 3-9. Steps 3-6: If step 3-4 finds that there are GIS electrical wiring fly wires or abnormal line segments in the image, calculate the reasonable pixel position of the line segment on the computer display screen, and notify the operation simulation robot program of the relevant screen pixel position information through the interface. Steps 3-7: After receiving the relevant information, the simulated robot program simulates mouse operation to click on the corresponding line segment, triggering a pop-up window. The pop-up window contains the device parameter information of the line segment. After the pop-up window is successfully displayed, the simulated robot program interface notifies the GIS graphic intelligent recognition program that the currently selected device information has been displayed in the pop-up window. Steps 3-8: The GIS graphic intelligent identification program takes a screenshot and uses image text recognition technology to identify the device name and device code information in the screenshot image. The GIS graphic intelligent identification program uses the device name and device code information identified by the image to query the actual length of the specific wire or cable segment through the system equipment ledger interface or database, and then compares it with the length analyzed by the image recognition. When there is a difference, it is determined that there is a problem with the GIS information of the device, such as a flying wire or error. The program registers and archives the problem information for subsequent rectification and management. Steps 3-9: Once the image analysis in the current GIS graphic window is complete, the GIS graphic intelligent recognition program calls the operation simulation robot program's listening interface to notify it to repeat steps 2-5. The steps for identifying and analyzing the GIS cable and conductor paths in the image initiated by the GIS graphic intelligent recognition program include: Step 4-1: The GIS graphic intelligent identification program starts the GIS cable and conductor path identification and analysis; Step 4-2: The GIS graphic intelligent identification program intelligently identifies and extracts line segments and station equipment symbol elements, and identifies and analyzes each cable and conductor segment based on the pre-completed image recognition training model. Step 4-3: If step 4-2 finds that the line segments in the image exceed the image border, then perform multiple image overlap and stitching according to the direction of the line segments exceeding the border, and then repeat the analysis in step 4-2. Step 4-4: Based on the pixel width and height calculation results from Step 3-3, the GIS graphic intelligent identification program calculates the length of cable segments and conductor segments. For segments exceeding the set length range, they are identified as flying wires or abnormal segments in GIS electrical circuits.

2. The GIS power grid topology anomaly identification method based on image recognition technology according to claim 1, characterized in that: The distance conversion of the window height and width in step 3-2 is calculated using the distance between two points A and B on the GIS sphere: (1) C=sin(LatA)*sin(LatB)*cos(LonA-LonB)+cos(LatA)*cos(LatB) (2) Distance=R*Arccos(C)*Pi / 180 Among them: (1) The latitude and longitude coordinates of point A: (LonA, LatA), and the latitude and longitude coordinates of point B: (LonB, LatB); (2) R = 6371.004 km, Distance is the distance between two points, in km.

3. The GIS power grid topology anomaly identification method based on image recognition technology according to claim 1, characterized in that: Step 3-3 calculates the height and width length represented by each pixel in the screenshot image. The height and width length represented by each pixel are calculated as follows: (1) Pixel height: Height = Window height / Number of pixels in the vertical direction of the window; (2) Pixel width: Width = window width / number of pixels in the horizontal direction of the window.

Images