A method for optimizing power grid operation and maintenance using machine vision and predictive analytics

Abstract

The application provides a kind of method for optimizing power grid operation and maintenance using machine vision and predictive analysis, and relates to the technical field of image processing.The method for optimizing power grid operation and maintenance specifically includes: collecting multiple video images of a power transmission line to obtain a video set; performing image stitching on the collected video images to obtain a corresponding complete panoramic image; inputting the panoramic image into an ice coating detection model to obtain an ice coating detection result; and generating operation and maintenance recommendations based on the ice coating detection result.Through the above steps, the application can improve the response capability and operation efficiency of the power grid, reduce maintenance costs, and optimize resource allocation.

Description

Technical Field

[0001] This invention relates to the field of image processing, and more specifically to a method for optimizing power grid operation and maintenance using machine vision and predictive analytics. Background Technology

[0002] Transmission lines often face challenges from extreme weather conditions during operation, especially icy and snowy weather. Icing refers to the condensation and freezing of moisture on transmission lines in low-temperature environments. This not only increases the weight of the line but may also lead to line breakage or reduced power transmission efficiency.

[0003] Machine vision technology has been widely applied in various fields, including manufacturing, healthcare, and transportation. Through advanced image processing and pattern recognition technologies, machine vision systems can automatically detect and analyze various objects and scenes. For example, in the automotive industry, machine vision is used to detect defects in the manufacturing process; in the medical field, it is used to analyze medical images to assist in disease diagnosis.

[0004] While machine vision technology has been widely applied in many fields, its application in power grids, especially under extreme weather conditions, is still in its early stages. For example, the power sector often relies on data from meteorological departments to assess potential weather risks, but this data is typically provided for larger geographical areas rather than specific local areas of transmission lines. Therefore, this meteorological data often fails to accurately reflect the actual climate conditions at the location of a particular transmission line. Furthermore, traditional systems often lack the ability to adapt to constantly changing environmental conditions. For instance, in strong winds, light icing may naturally dissipate, but current technologies fail to take advantage of this environmental advantage, still relying on standard de-icing procedures, leading to unnecessary resource waste.

[0005] To improve the power grid's ability to cope with extreme weather, reduce maintenance costs, and enhance the overall reliability of the power grid, this patent proposes a method for optimizing power grid operation and maintenance using machine vision and predictive analytics. This method optimizes the operation and maintenance of the power grid under extreme weather conditions, thereby protecting power grid assets and ensuring public safety. Summary of the Invention

[0006] This invention provides a method for optimizing power grid operation and maintenance using machine vision and predictive analytics. The method specifically includes the following steps:

[0007] S1: Acquire multiple video images of the transmission line to obtain a video set;

[0008] S2: Stitch the acquired video images to obtain the corresponding complete panoramic image;

[0009] S3: Input the panoramic image into the icing detection model to obtain the icing detection results;

[0010] S4: Generate operation and maintenance recommendations based on the icing detection results.

[0011] This invention provides a system for optimizing power grid operation and maintenance using machine vision and predictive analytics. The system includes:

[0012] Image acquisition device: The image acquisition device acquires multiple video images of the transmission line to obtain a video set;

[0013] Panoramic image generation module: The panoramic image generation module stitches together the acquired video images to obtain the corresponding complete panoramic image;

[0014] Icing detection module: The icing detection module inputs the panoramic image into the icing detection model to obtain the icing detection result;

[0015] Operation and maintenance module: The operation and maintenance module generates operation and maintenance suggestions based on the icing detection results.

[0016] An electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the aforementioned method for optimizing power grid operation and maintenance using machine vision and predictive analytics.

[0017] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the aforementioned method for optimizing power grid operation and maintenance using machine vision and predictive analytics.

[0018] Compared with existing technologies, this invention aims to solve the problem of the lack of flexible adjustment of maintenance strategies in the process of power grid operation optimization and maintenance. This invention can calculate the curvature of the current line by stitching together panoramic images, and then infer the local wind conditions of the current transmission line. It does not rely on meteorological data provided by meteorological departments, and obtains more accurate meteorological information for predictive analysis to optimize operating parameters and maintenance plans based on the actual location of the transmission line. In addition, this application also constructs an icing detection model to detect the amount and location of ice in the current line. The icing detection model integrates multi-layer features and enhances features to improve the recognition ability of small targets. Combining the icing detection results with the current wind speed, it provides comprehensive suggestions for power grid operation parameter optimization and de-icing maintenance plans, thereby improving the reliability of power grid operation. Attached Figure Description

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

[0020] Figure 1 This is a flowchart of the proposed optimization of operation and maintenance in this application;

[0021] Figure 2 This is a structural diagram of the icing detection model in this application. Detailed Implementation

[0022] The embodiments of this application will now be described in detail with reference to the accompanying drawings.

[0023] The following specific examples illustrate the implementation of this application. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. This application can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. It should be noted that, in the absence of conflict, the following embodiments and features in the embodiments can be combined with each other. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0024] It should be noted that various aspects of embodiments within the scope of the appended claims are described below. It will be apparent that the aspects described herein can be embodied in a wide variety of forms, and any particular structure and / or function described herein is merely illustrative. Based on this application, those skilled in the art will understand that one aspect described herein can be implemented independently of any other aspect, and two or more of these aspects can be combined in various ways. For example, any number and aspects set forth herein can be used to implement the device and / or practice the method. Additionally, this device and / or method can be implemented using structures and / or functionalities other than one or more of the aspects set forth herein.

[0025] Additionally, specific details are provided in the following description to facilitate a thorough understanding of the examples. However, those skilled in the art will understand that practice can be carried out without these specific details.

[0026] This specification presents an embodiment of a method for optimizing power grid operation and maintenance using machine vision and predictive analytics. The method specifically includes the following steps:

[0027] S1: Acquire multiple video images of the transmission line to obtain a video set;

[0028] S2: Stitch the acquired video images to obtain the corresponding panoramic image;

[0029] S3: Input the panoramic image into the icing detection model to obtain the icing detection result;

[0030] S4: Generate operation and maintenance recommendations based on the icing detection results.

[0031] This invention utilizes unmanned aerial vehicle (UAV) equipment to capture video images of overhead power transmission lines. Specifically, the UAV equipment can be selected from DJI's Matrice 300RTK, Inspire series UAVs, or other UAV equipment with long endurance, stable flight performance, high-quality shooting effects, and good environmental adaptability. Preferably, the UAV should be equipped with redundant systems, such as dual GPS and dual IMUs, to enhance safety and reliability in complex environments.

[0032] Before using drones to capture video images of overhead power transmission lines, it is necessary to first obtain information on the power grid construction in the area to be surveyed, including the number and location of transmission towers, and the number, height, and route of overhead transmission lines. Based on the obtained information on the power grid construction in the area to be surveyed, a corresponding drone data collection plan is formulated. This plan includes setting the drone's flight altitude, speed, flight path, take-off and landing locations, and contingency plans.

[0033] Specifically, for example, the power grid construction information for a certain overhead transmission area is retrieved in advance, revealing a total transmission line length of approximately 10 kilometers, encompassing about 40 transmission towers. ArcGIS or QGIS is used to analyze the geographic data of the transmission line, including the precise location, altitude, and surrounding terrain of each tower. The flight altitude is ensured to be at least 20 meters above the highest transmission line, with a planned flight altitude between 120-140 meters to avoid any collision risks. A straight-line flight is planned from each tower to the next, with each segment averaging 250 meters. The drone's built-in GPS navigation system is used to maintain a precise flight path. The drone's flight speed is set at 5 meters per second to balance shooting stability and mission efficiency. Each 250-meter straight-line flight segment takes approximately 50 seconds, and the entire 10-kilometer line will be divided into 40 segments, requiring a total flight time of approximately 33 minutes. The drone model (such as the DJI Matrice 300 RTK, with approximately 55 minutes of flight endurance) is selected based on the flight time and distance. A ground station is set up every 5 kilometers along the drone's flight path to monitor its flight status and ensure the mission proceeds smoothly. Each ground station is equipped with at least two backup batteries to ensure the mission can be completed or any emergency can be dealt with.

[0034] After ensuring all flight parameters match the preset parameters, the filming task is executed. During flight, the operator monitors the camera output in real time to ensure video quality meets requirements. For each power transmission line segment, when flying from one tower to the next, the operator must control the drone to maintain stability, ensuring the continuity and clarity of the power transmission line in the video. At each new tower, the start and end points of the video are marked for subsequent processing.

[0035] Preferably, during the shooting process, the drone's real-time transmission function is used to transmit data back to the ground station in real time. The data is backed up at the ground station to ensure that no data is lost even in the event of equipment failure.

[0036] The acquired continuous video was segmented according to the intervals of the transmission towers, and necessary metadata, such as transmission tower number, date and time, and flight parameters, was added to each video segment to facilitate subsequent analysis and retrieval.

[0037] The process of stitching together the acquired video images to obtain the corresponding complete panoramic image specifically includes:

[0038] S21: Based on the drone's flight speed, keyframes are extracted from the video clips to obtain a set of keyframe images {D}. k |k∈[1,n]},D k It is the kth keyframe image, n is the number of keyframes extracted, and D1 is the first panoramic image. The extracted adjacent keyframe images have some overlapping areas along the shooting direction.

[0039] S22: Identify stable SIFT feature points in the t-th panoramic image and the (t+1)-th frame image, where 1≤t≤n-1;

[0040] S23: Determine the matching feature points between the t-th panoramic image and the (t+1)-th frame image using the k-nearest neighbor algorithm and Euclidean distance, and obtain the set of matching feature points in the t-th panoramic image. The set of matching feature points in the (t+1)th keyframe Where N represents the number of matching points. and These represent the i-th matching feature point in the t-th panoramic image and the t+1 keyframe image, respectively;

[0041] S24: Obtain the (t+1)th panoramic image I according to the image transformation formula. t+1 ;

[0042]

[0043] Where tl and tr are the location information of the transition region, H is the location information of the edge, γ is the balance weight information, and I is the location information of the transition region. t (x, y) represents the image information of the t-th panoramic image at coordinates (x, y), D t+1 (x,y) represents the image information of the (t+1)th keyframe image at coordinates (x,y);

[0044] S25: Iterate through steps S22-S24 and generate the final panoramic image I. n-1 As a complete panoramic image I panorama .

[0045] The specific methods for calculating the transition region location information tl and tr include:

[0046] Matching feature point set and We obtain a first transformation matrix T and a second transformation matrix T', where the first transformation matrix T is used to transform pixels in the t-th panoramic image to the (t+1)-th keyframe, and the second transformation matrix T' is used to transform pixels in the (t+1)-th keyframe to the t-th panoramic image; the first boundary B1 of the t-th panoramic image is {(x b ,y b The first transformation is performed to obtain the result in keyframe D. t+1 The first mapping boundary set E1 = {(x` e ,y` e The first conversion calculation formula is:

[0047] (x` e ,y` e )=T(x b ,y b )

[0048] For keyframe D t+1 The second boundary B2 = {(x` b ,y` b The second transformation is performed to obtain the result in keyframe D. t The second mapping boundary set E2 = {(x e ,y e The second conversion calculation formula is:

[0049] (x e ,y e )=T`(x` b ,y` b )

[0050] Define the first mapping boundary set E1 = {(x` e ,y` e x` in )} eThe maximum value is tr, and the second mapping boundary set is set as E2 = {(x e ,y e The minimum value in )} is tl. The first boundary is the image boundary corresponding to the drone's flight direction, and the second boundary is the image boundary in the opposite direction to the drone's flight direction.

[0051] This invention achieves efficient acquisition of panoramic images of power transmission lines by introducing image stitching technology. In step S1, the acquisition device is controlled to fly in a straight line along the transmission line, thereby systematically collecting image data along the route. This method allows for continuous capture of the transmission line's condition, and the image stitching technology combines these continuous images into a single panoramic image. This not only significantly improves the integrity of the image data but also ensures the continuity and accuracy of the image information.

[0052] In the acquired panoramic images, the degree of curvature of the power transmission lines reflects the impact of wind on them. Specifically, when wind acts on power transmission lines, it causes them to bend or shift to a certain extent. This physical deformation can be accurately captured and quantified by analyzing the shape of the power transmission lines in the panoramic images. In-depth analysis of this image data can determine the current wind force experienced by the transmission lines, providing reliable weather and environmental information for subsequent predictive analysis.

[0053] The panoramic image is input into the icing detection model to obtain the icing detection results.

[0054] The icing detection model includes a feature extraction module, a feature enhancement module, a feature fusion module, and an attention guidance module. The feature extraction module uses the first five residual modules of ResNet50 as the backbone network for feature extraction. The input complete panoramic image is fed into two feature enhancement modules after passing through the backbone network for multi-scale feature fusion. At the same time, the attention guidance module is used for feature enhancement, and the multi-scale features are input into the classification module to obtain the icing detection result.

[0055] The feature enhancement module employs parallel sub-channel enhancement, with four sub-channels. The feature enhancement module enhances the output feature F of the previous layer. m-1 Channel enhancement is performed, and F is spliced ​​using a skip connection method. m-1 The enhanced feature F is obtained m :

[0056]

[0057] Where Conv represents convolution, Act represents the activation function, 1*1 and 3*3 represent convolution kernels of size 1 and 3 respectively, and d represents the dilation size. These represent the output features F of the previous layer, respectively. m-1The four sub-channels are sub-channel 1, sub-channel 2, sub-channel 3 and sub-channel 4; the number of each sub-channel is divided equally according to the total number of channels or according to a preset rule, and the feature maps in the sub-channels do not overlap with each other;

[0058] The feature fusion module employs a cross-layer fusion approach to fuse the output features. Specifically, the cross-layer fusion involves fusing the output features F from the previous layer. m-1 and the output features F of the next layer m+1 To merge:

[0059] F c =Concat(DwnS(C l (F m-1 UpS(C) h (F m+1 )))

[0060] Where F c For the output feature F of the previous layer m-1 and the output features F of the next layer m+1 The fused output features, where Concat represents concatenation by channel, DwnS and UpS represent downsampling and upsampling of the feature maps, respectively, to ensure consistent scale across different layers of feature maps, C l and C h This indicates that 1*1 dimensionality reduction and dimensionality increase are performed on low-level and high-level features to ensure consistency of channels in feature maps of different layers;

[0061] Specifically, this application performs cross-layer fusion of the output F3 of the third residual module and the output F5 of the fifth residual module in the backbone network to obtain fused features. The output of the fourth residual module F4 in the backbone network and the output of the first feature enhancement layer are combined. Cross-layer fusion is performed to obtain fusion features Fusion features and and the output of the second feature enhancement layer After attention guidance, the data is input into the classification module for target detection.

[0062] Furthermore, the icing detection model also includes an attention guidance module, which uses the output F3 of the third residual module to generate attention weights to guide the output of the first feature enhancement layer. The fourth residual module output F4 is used to generate attention weights to guide the output of the second feature enhancement layer.

[0063] The calculation method for the attention guidance module is as follows:

[0064]

[0065] in, and The output features are those after attention guidance, where σ represents the sigmoid activation function, and AVG and MAX represent average pooling and max pooling, respectively.

[0066] After inputting the panoramic image into the icing detection model to obtain the icing detection result, the calculation of wind speed is also included. The calculation of wind speed specifically includes:

[0067] S31: The complete panoramic image obtained in step S2 is converted to grayscale and Gaussian filtered, and edge detection is performed using the Canny operator to obtain the edge image I. edge The edge image I edge Including transmission line curves;

[0068] I edge =Canny(GaussianBlur(GrayScale(I panorama )))

[0069] Where Canny represents the Canny operator, GaussianBlur represents Gaussian filtering, and GrayScale represents grayscale processing;

[0070] S32: For edge image I edge The endpoints of the transmission line curve are identified to obtain the endpoint coordinates (x1, y1) and (x2, y2). Based on the endpoint coordinates, the coordinates of the midpoint (x1, y1) on the curve are obtained. m y m );

[0071] S33: Connect the endpoint coordinates (x1, y1) and the midpoint coordinates (x... m y m We obtain the first line segment L1, the coordinates of its endpoint (x2, y2), and the coordinates of its midpoint (x). m y m We obtain the second line segment L2;

[0072] S34: Calculate the angle θ between the first line segment L1 and the second line segment L2.

[0073]

[0074] Where the vector α = (x1 - x) m y1-y m ), β=(x2-x m y2-y m ), · represents the dot product, |||| represents the vector magnitude; the wind speed v in the transmission line area is calculated based on the included angle value:

[0075]

[0076] Where v represents wind speed, C d The resistance coefficient of the transmission line is represented by ρ, the air density is represented by A, the diameter of the transmission line is represented by E, the elastic modulus of the transmission line material is represented by I, the moment of inertia of the cross section of the transmission line is represented by L, and the length of the transmission line is represented by L.

[0077] Operation and maintenance recommendations are generated based on the icing detection results.

[0078] The icing detection results include the amount of icing T in the complete panoramic image. n and icing location T c Based on the amount of ice T in the complete panoramic image n The number of drones to be dispatched for de-icing is determined by the wind speed (v) in the power transmission line area.

[0079]

[0080] Where T min V represents the minimum amount of ice buildup required to trigger the dispatch of a drone for de-icing. maxsafe This indicates the maximum wind speed required to ensure the drone's flight.

[0081] Based on the amount of ice T in the complete panoramic image n Optimize grid load parameters based on wind speed v in the transmission line area:

[0082]

[0083] Among them, S(T) n v) represents the amount of ice accumulation T n Recommended load parameters for wind speed v in the transmission line area, S max The maximum load capacity of the line, μ1 is the icing influence coefficient, μ2 is the wind speed influence coefficient, T max It is the preset icing threshold, V max The preset maximum safe wind speed threshold for power transmission lines.

[0084] This invention provides a system for optimizing power grid operation and maintenance using machine vision and predictive analytics. The system includes:

[0085] Image acquisition device: The image acquisition device acquires multiple video images of the transmission line to obtain a video set;

[0086] Panoramic image generation module: The panoramic image generation module stitches together the acquired video images to obtain the corresponding complete panoramic image;

[0087] Icing detection module: The icing detection module inputs the panoramic image into the icing detection model to obtain the icing detection result;

[0088] Operation and maintenance module: The operation and maintenance module generates operation and maintenance suggestions based on the icing detection results.

[0089] An electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the aforementioned method for optimizing power grid operation and maintenance using machine vision and predictive analytics.

[0090] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the aforementioned method for optimizing power grid operation and maintenance using machine vision and predictive analytics.

[0091] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0092] In this specification, the same or similar parts between the various embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the descriptions of the embodiments described later are relatively simple, and relevant parts can be referred to the descriptions of the foregoing embodiments.

[0093] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for optimizing power grid operation and maintenance using machine vision and predictive analytics, characterized in that, The method includes the following steps: S1: Acquire multiple video images of the transmission line to obtain a video set; S2: Stitch the acquired video images to obtain the corresponding complete panoramic image; S3: Input the panoramic image into the icing detection model to obtain the icing detection results; S4: Generate operation and maintenance recommendations based on icing detection results; The icing detection model includes a feature extraction module, a feature enhancement module, a feature fusion module, an attention guidance module, and a classification module; The feature fusion module fuses the output features in a cross-layer fusion manner, the cross-layer fusion fuses the output features F m-1 of the previous layer and the output features F m+1 of the next layer: F c = Concat(DwnS(C l (F m-1 )), UpS(C h (F m+1 ))) where F c is the fusion output feature of the previous layer output feature F m-1 and the next layer output feature F m+1 , m represents the current layer, Concat represents channel concatenation, DwnS and UpS represent down-sampling and up-sampling of the feature map respectively, C l and C h represent 1*1 dimension reduction and dimension increase of low-level features and high-level features; The feature enhancement module adopts sub-channel parallel enhancement, the number of sub-channels is 4, and the feature enhancement module enhances the output feature F m-1 Channel enhancement is performed, and F m-1 is obtained by using a jump connection method to splice F m ​ Where Conv represents convolution, Act represents the activation function, 1*1 and 3*3 represent convolution kernels of size 1 and 3 respectively, and d represents the dilation size. These represent the output features F of the previous layer, respectively. m-1 Subchannel 1, subchannel 2, subchannel 3 and subchannel 4 are the four subchannels.

2. The method for optimizing power grid operation and maintenance using machine vision and predictive analysis according to claim 1, characterized in that, The process of stitching together the acquired video images to obtain the corresponding complete panoramic image specifically includes: S21: According to the flight speed of the unmanned aerial vehicle, key frame extraction is performed on the video segment to obtain a key frame image set {D k |k∈[1,n]},D k is the kth key frame image, n is the number of extracted key frames, and the first preset D1 is the first panoramic image. S22: Identify SIFT feature points in the t-th panoramic image and the (t+1)-th frame image, where 1≤t≤n-1; S23: Determine the matching feature points between the t-th panoramic image and the (t+1)-th frame image using the k-nearest neighbor algorithm and Euclidean distance, and obtain the set of matching feature points in the t-th panoramic image. The set of matching feature points in the (t+1)th keyframe Where N represents the number of matching points. and These represent the i-th matching feature point in the t-th panoramic image and the t+1 keyframe image, respectively; S24: Obtain the t+1th panoramic image I according to the image conversion formula t+1 ; wherein tl and tr are transition region position information, H is edge position information, γ is balance weight information, I t (x,y) is image information of the tth panoramic image at coordinate (x, y), D t+1 (x,y) is image information of the t+1th key frame image at coordinate (x, y). S25: iteratively performing steps S22-S24, resulting in a final panorama image I n-1 as a complete panorama image I panorama .

3. The method for optimizing power grid operation and maintenance using machine vision and predictive analysis according to claim 2, characterized in that, The method for calculating the transition region location information tl and tr is as follows: Based on the set of matching feature points and A first transformation matrix T and a second transformation matrix T' are obtained. The first transformation matrix T is used to transform the pixels in the t-th panoramic image to the (t+1)-th keyframe, and the second transformation matrix T' is used to transform the pixels in the (t+1)-th keyframe to the t-th panoramic image. The first boundary B1 of the tth panoramic image {(x b ,y b )} is first converted to obtain a first mapping boundary set E1 in the key frame D t+1 {(x` e ,y` e )}, and the first conversion formula is: (x e ,y e ) = T(x b ,y b ) a second boundary B2 of the key frame D t+1 = {(x b ,y b )} is converted to obtain a second mapping boundary set E2 = {(x t ,y e )} in the key frame D e , and the second conversion is calculated by the following formula: (x e ,y e )=T`(x` b ,y` b ) Set the maximum value of x e in the first mapping boundary set E1 = {(x e ,y e} as tr, and set the minimum value in the second mapping boundary set E2 = {(x e ,y e} as tl.

4. The method for optimizing power grid operation and maintenance using machine vision and predictive analysis according to claim 1, characterized in that: The calculation method for the attention guidance module is as follows: in, and The output features are those after attention guidance, where σ represents the sigmoid activation function, and AVG and MAX represent average pooling and max pooling, respectively.

5. The method for optimizing power grid operation and maintenance using machine vision and predictive analysis according to claim 1, characterized in that: After inputting the panoramic image into the icing detection model to obtain the icing detection result, the calculation of wind speed is also included. The calculation of wind speed specifically includes: S31: The complete panoramic image obtained in step S2 is subjected to grayscale processing and Gaussian filtering, and an edge image I is obtained by edge detection using a Canny operator edge , the edge image I edge includes a power transmission line curve; I edge = Canny(GaussianBlur(GrayScale(I panorama ))) Where Canny represents the Canny operator, GaussianBlur represents Gaussian filtering, and GrayScale represents grayscale processing; S32: edge image I edge The midpoint coordinates (x, y) on the curve are obtained based on the endpoint coordinates. m , y m ) S33: connecting the end point coordinate (x1, y1) and the midpoint coordinate (x m , y m ) to obtain a first line segment L1, and connecting the end point coordinate (x2, y2) and the midpoint coordinate (x m , y m ) to obtain a second line segment L2; S34: Calculate the angle θ between the first line segment L1 and the second line segment L2. where the vector a = (x1-x m , y1-y m ), β = (x2-x m , y2-y m ), · denotes the dot product, and || || denotes the vector norm; the wind speed v of the transmission line region is calculated according to the included angle value: where v represents the wind speed, C d represents the drag coefficient of the transmission line, p is the air density, A is the diameter of the transmission line, E represents the elastic modulus of the transmission line material, I describes the cross-sectional moment of inertia of the transmission line, and L represents the length of the transmission line.

6. The method for optimizing power grid operation and maintenance using machine vision and predictive analysis according to claim 1, characterized in that: The feature fusion module performs cross-layer fusion of the output F3 of the third residual module and the output F5 of the fifth residual module in the backbone network to obtain the fused feature. The output of the fourth residual module F4 in the backbone network and the output of the first feature enhancement layer are combined. Cross-layer fusion is performed to obtain fusion features Fusion features and and the output of the second feature enhancement layer After attention guidance, the data is input into the classification module for target detection.

7. The method for optimizing power grid operation and maintenance using machine vision and predictive analysis according to claim 1, characterized in that, The specific provisions for generating operation and maintenance recommendations based on icing detection results include: The icing detection result includes an icing quantity T in the complete panoramic image n and an icing position T c , the number of dispatched drones for deicing is determined according to the icing quantity T in the complete panoramic image n and the wind speed v of the power transmission line area; Where T min V represents the minimum amount of ice buildup required to trigger the dispatch of a drone for de-icing. maxsafe This indicates the maximum wind speed required to ensure the drone's flight. This indicates rounding up to the nearest integer.

8. A system for optimizing power grid operation and maintenance using machine vision and predictive analytics, characterized in that... The system includes: Image acquisition device: The image acquisition device acquires multiple video images of the transmission line to obtain a video set; Panoramic image generation module: The panoramic image generation module stitches together the acquired video images to obtain the corresponding complete panoramic image; Icing detection module: The icing detection module inputs the panoramic image into the icing detection model to obtain the icing detection result; Operation and maintenance module: The operation and maintenance module generates operation and maintenance suggestions based on the icing detection results. The icing detection model includes a feature extraction module, a feature enhancement module, a feature fusion module, and an attention guidance module; The feature fusion module fuses the output features in a cross-layer fusion manner, the cross-layer fusion fuses the output features F m-1 of the previous layer and the output features F m+1 of the next layer: F c =Concat(DwnS(C l (F m-1 )),UpS(C h (F m+1 ))) Where F c For the output feature F of the previous layer m-1 and the next layer output feature F m+1 The fused output features, where m represents the current layer, Concat means concatenation by channel, DwnS and UpS represent downsampling and upsampling of the feature map, respectively, and C l and C h This indicates 1x1 dimensionality reduction and dimensionality increase for low-level and high-level features; The feature enhancement module employs parallel sub-channel enhancement, with four sub-channels. The feature enhancement module enhances the output feature F of the previous layer. m-1 Channel enhancement is performed, and F is spliced ​​using a skip connection method. m-1 The enhanced feature F is obtained m : Where Conv represents convolution, Act represents the activation function, 1*1 and 3*3 represent convolution kernels of size 1 and 3 respectively, and d represents the dilation size. These represent the output features F of the previous layer, respectively. m-1 Subchannel 1, subchannel 2, subchannel 3 and subchannel 4 are the four subchannels.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the method for optimizing power grid operation and maintenance using machine vision and predictive analytics as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program that, when executed by a processor, implements the method for optimizing power grid operation and maintenance using machine vision and predictive analytics as described in any one of claims 1 to 7.

Images