Transmission conductor sag monitoring method and device based on digital twinning

Abstract

The invention provides a power transmission conductor sag monitoring method and device based on digital twinning, and relates to the technical field of sag monitoring, and the method comprises the steps: reading the design data of a power transmission line after obtaining the authorization of a user, and constructing a basic digital twinning model; carrying out sag detection key point identification, and configuring binocular acquisition parameters; power transmission line acquisition is executed, and a binocular image is established; selecting a front view image to carry out local feature analysis, and establishing an adaptive authentication window; pixel matching cost calculation is carried out, and sag detection three-dimensional point cloud reconstruction is carried out; performing depth data extraction, establishing fused depth data of sag detection key points, updating the basic digital twinborn model, and generating a corrected digital twinborn model; dynamic simulation is carried out, and sag abnormity early warning is generated. The method solves the technical problems that the existing sag monitoring method often cannot realize real-time monitoring and is characterized in that high accuracy is difficult to guarantee in a dynamic environment, so that sag abnormity cannot be found in time and early warning cannot be performed.

Description

Technical Field

[0001] This invention relates to the field of sag monitoring technology, specifically to a method and device for monitoring the sag of power transmission lines based on digital twins. Background Technology

[0002] Sag refers to the degree of sag of power transmission lines due to external forces such as gravity and wind loads. Abnormal sag can lead to contact or breakage between conductors or collisions with ground objects, and in severe cases, it can cause power accidents, even resulting in personal injury and property damage. Existing methods for monitoring power transmission line sag mainly rely on manual inspections, periodic checks, and traditional sensor systems. While manual inspections can detect some obvious sag problems, they suffer from high labor costs, low efficiency, and limited coverage, and cannot acquire data in real time. Traditional sensor systems can collect some data, but they are costly and difficult to deploy and maintain in large-scale power transmission networks. Furthermore, these methods are limited by external factors such as weather and terrain, significantly reducing the comprehensiveness and timeliness of monitoring. Summary of the Invention

[0003] This application provides a method and device for monitoring the sag of transmission lines based on digital twins, aiming to solve the technical problem that existing sag monitoring methods often cannot achieve real-time monitoring, and are characterized by difficulty in ensuring high accuracy in dynamic environments, resulting in the inability to detect sag anomalies and issue early warnings in a timely manner.

[0004] The first aspect disclosed in this application provides a method for monitoring the sag of transmission lines based on digital twins. The method includes: after obtaining user authorization, reading design data of the transmission line and constructing a basic digital twin model of the transmission line based on the design data; establishing sag detection key point identification results based on the basic digital twin model and configuring binocular acquisition parameters; controlling a binocular camera to acquire data from the transmission line using the binocular acquisition parameters to establish a binocular image; selecting any image from the binocular images as the main view image and performing local feature analysis to establish an adaptive authentication window; calculating the pixel matching cost of the preprocessed binocular image using the adaptive authentication window, and reconstructing a sag detection 3D point cloud based on the pixel matching cost calculation result; establishing fusion depth data of sag detection key points using the sag detection 3D point cloud, updating the basic digital twin model based on the fusion depth data, and generating a corrected digital twin model; performing dynamic simulation based on the corrected digital twin model, and generating a sag anomaly warning based on the dynamic simulation results.

[0005] The second aspect of this application discloses a digital twin-based transmission line sag monitoring device. The device is used in the aforementioned digital twin-based transmission line sag monitoring method. The device includes: a basic twin model construction module, used to read design data of the transmission line after obtaining user authorization and construct a basic digital twin model of the transmission line based on the design data; a binocular acquisition parameter configuration module, used to establish sag detection key point identification results based on the basic digital twin model and configure binocular acquisition parameters; a transmission line acquisition module, used to control a binocular camera to acquire transmission line data using the binocular acquisition parameters and establish a binocular image; and a local feature analysis module, used to select… Any image from the binocular images is used as the main view image, and local feature analysis is performed to establish an adaptive authentication window; a pixel matching cost calculation module is used to calculate the pixel matching cost of the preprocessed binocular images using the adaptive authentication window, and to reconstruct the 3D point cloud of sag detection based on the pixel matching cost calculation result; a corrected twin model generation module is used to establish fusion depth data of key points of sag detection using the 3D point cloud of sag detection, update the basic digital twin model based on the fusion depth data, and generate a corrected digital twin model; a sag anomaly warning generation module is used to perform dynamic simulation based on the corrected digital twin model, and generate a sag anomaly warning based on the dynamic simulation result.

[0006] One or more technical solutions provided in this application have at least the following beneficial effects: By reading the design data of transmission lines and establishing a basic digital twin model based on this data, an accurate virtual model is provided for subsequent monitoring and analysis. This lays a solid foundation for real-time monitoring of transmission line sag, enabling more precise reflection of the line's geometry and operating status. By identifying key points for sag detection on the digital twin model and configuring the binocular camera's acquisition parameters based on the identification results, the impact of sag changes on the transmission line can be accurately identified. Adjusting the binocular camera's acquisition parameters ensures the quality of image data and the rationality of the acquisition angle, thereby improving the accuracy of sag monitoring. Using a binocular camera for image acquisition and preprocessing effectively improves image quality, ensuring the accuracy of subsequent feature extraction and analysis. Furthermore, selecting any image from the preprocessed binocular images as the main view image allows for local feature analysis, providing a basis for subsequent... Pixel matching and sag detection provide clear and high-quality image data. An adaptive authentication window is established through local feature analysis, and the pixel matching cost of the binocular image is calculated based on this window to perform 3D point cloud reconstruction for sag detection. This method considers the changes in local image features and enhances the system's adaptability to image matching under different conditions. The depth data extracted from the key points of sag detection is used to update the basic digital twin model, thereby generating a corrected digital twin model. This model more accurately reflects the actual state of the transmission line, especially under dynamic environmental conditions, and can update the state in real time. Dynamic simulation based on the corrected digital twin model can simulate the sag changes of the conductor under different environmental and load conditions, and generate sag anomaly warnings based on the simulation results. This simulation can comprehensively consider the influence of different environmental factors, further improving the prediction and warning capabilities for sag anomalies.

[0007] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0008] Figure 1 A schematic diagram of the process for monitoring the sag of transmission lines based on digital twins, provided in an embodiment of this application.

[0009] Figure 2 A schematic diagram of the structure of a transmission line sag monitoring device based on digital twin provided in this application embodiment.

[0010] Figure labeling: Basic twin model construction module 10, binocular acquisition parameter configuration module 20, transmission line acquisition module 30, local feature analysis module 40, pixel matching cost calculation module 50, correction twin model generation module 60, sag anomaly early warning generation module 70. Detailed Implementation

[0011] This application provides a method and device for monitoring the sag of transmission lines based on digital twins, which solves the technical problem that existing sag monitoring methods often cannot achieve real-time monitoring, and are characterized by difficulty in ensuring high accuracy in dynamic environments, resulting in the inability to detect sag anomalies and issue early warnings in a timely manner.

[0012] After introducing the basic principles of this application, various non-limiting embodiments of this application will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are merely illustrative of this application and are not intended to limit this application.

[0013] Example 1, as Figure 1 As shown in the figure, this application provides a method for monitoring the sag of transmission lines based on digital twins, the method comprising: After obtaining user authorization, the design data of the transmission line is read, and a basic digital twin model of the transmission line is constructed based on the design data.

[0014] Before executing the monitoring task, user authorization is first obtained. This can be done through user login, authorization confirmation, etc. User authorization is the first step to ensure the legality of the operation and compliance with privacy requirements. After obtaining user authorization, the design data of the transmission line is read. This design data contains detailed information about the transmission line, such as conductor layout, suspension points, conductor materials, line geometry, span, conductor height, and sag design values. Based on the read design data, a basic model of the transmission line is built using digital twin technology. A digital twin is a virtual model that can accurately replicate the physical characteristics and structure of the transmission line. In this step, the basic digital twin model includes the physical properties of the conductors, transmission status, and other engineering data for subsequent analysis and monitoring.

[0015] Based on the aforementioned basic digital twin model, the key point recognition results for sag detection are established, and the binocular acquisition parameters are configured.

[0016] By using a basic digital twin model, the changes in sag can be accurately calculated, and key points for sag detection in transmission lines can be identified. These key points refer to the places where the conductor bends under gravity, especially at the bearing point and suspension point of the conductor, and are used to monitor the deformation, tension and possible anomalies of the conductor.

[0017] Based on the identified sag detection key points, the binocular camera acquisition parameters are configured. The binocular camera acquires images from different angles using two cameras and obtains depth information by calculating parallax. The specific binocular acquisition parameter configuration includes viewing angle settings, focal length and exposure settings, and resolution settings. Among them, the viewing angle setting is used to ensure that the viewing angle of the binocular camera can cover the identified sag detection key points. The focal length and exposure settings are used to adjust parameters such as the camera's focal length and exposure time to ensure image quality, especially under different lighting and environmental conditions. The resolution configuration is used to set the image resolution to ensure that sufficient detailed information can be obtained in subsequent processing and analysis.

[0018] The binocular acquisition parameters are used to control the binocular camera to acquire data from the power transmission line and establish a binocular image.

[0019] Before data acquisition, the binocular camera is initialized using the configured binocular acquisition parameters. Initialization steps include calibrating the camera and adjusting its internal and external parameters to ensure that the camera's viewing angle, focal length, exposure, and other configurations conform to the aforementioned settings. After initialization, the binocular camera simultaneously captures images of the transmission line from two angles. Since the shooting positions of the two cameras are known, images of the same object can be obtained from different perspectives. These images exhibit parallax, which is used for subsequent depth analysis. The two images acquired by the binocular camera constitute a binocular image. This image not only provides a 2D image of the transmission line but also contains depth information. This is because the binocular camera estimates the distance or depth of the object through parallax calculation (the difference between the two viewing angles). The image reveals the specific location, shape, sag, and other features of the transmission conductor, providing an image foundation for subsequent data analysis.

[0020] Select any image from the binocular images as the main view image, perform local feature analysis, and establish an adaptive authentication window.

[0021] Before selecting the main view image, the binocular images are preprocessed to improve image quality and accuracy. The preprocessing steps include denoising, contrast enhancement, and image correction. Denoising is used to filter out noise in the image so as to capture key features in the image more clearly. Contrast enhancement is used to improve the contrast of the image so that the difference between the lines and the background is more obvious, which facilitates feature extraction. Image correction includes perspective distortion correction to ensure that objects in the image are displayed in the correct proportions.

[0022] After preprocessing, any one of the binocular images is selected as the main view image. The main view image is usually the clearest or most representative one, with high resolution and less distortion. The main view image will become the basis for subsequent feature analysis, used to analyze and extract detailed information in the image.

[0023] Local feature analysis is performed on the selected front view image. Local features refer to regions in the image that have significant changes. These regions contain spatial information and can be used for object recognition and localization. Local feature analysis methods can include corner detection or edge detection. Corner detection is used to detect corners and points of interest in the image, while edge detection identifies the edges of conductors by recognizing regions in the image with significant brightness changes.

[0024] Based on the results of local feature analysis, an adaptive authentication window is established. The adaptive authentication window refers to a window area dynamically set in the image according to the features, which is used for subsequent pixel matching, disparity calculation and depth analysis. The window automatically adjusts its size and position according to changes in the image to ensure that each pixel in the image can be processed effectively. The key to the adaptiveness is that the size and shape of the window will change according to the characteristics of the image and the needs of different regions. For example, a larger window is used in complex regions, while a smaller window is used in simple regions to improve the accuracy and speed of matching.

[0025] The pixel matching cost of the preprocessed binocular image is calculated using the adaptive authentication window, and the sag detection 3D point cloud is reconstructed based on the pixel matching cost calculation result.

[0026] Pixel matching compares images from two different viewpoints pixel by pixel to find matching relationships between corresponding pixels, thereby obtaining depth information. An adaptive authentication window helps improve the accuracy of pixel matching because it dynamically adjusts the matching area based on the image content, avoiding the limitations of a fixed-size window. Pixel matching cost evaluates the matching situation between pixels in the image by calculating the similarity of each pair of pixels. Through cost calculation, a cost value is assigned to each pair of potentially matching pixels; a smaller cost value indicates a higher matching degree. Based on the matching results, key feature points in the image are found. These key points are located at locations such as suspension points of conductors and extreme points of sag, enabling sag detection and 3D point cloud reconstruction. Accurate pixel matching yields more 3D point cloud data, including sag detection key points, providing accurate locations for subsequent depth data extraction.

[0027] The fused depth data of key points for sag detection is established using the 3D point cloud of the sag detection, and the basic digital twin model is updated based on the fused depth data to generate a corrected digital twin model.

[0028] Using sag detection of 3D point clouds, depth data is extracted by calculating the disparity of binocular images. The depth data is obtained by calculating the disparity between matching pixels in two viewpoint images. The larger the disparity, the closer the object is to the camera; the smaller the disparity, the farther the object is from the camera. The depth data of each sag detection key point represents the position of that point in 3D space and is used to monitor the sag change of the conductor.

[0029] To improve the accuracy of depth data, the final depth data is constructed by fusing data from multiple sources, such as combining depth information obtained from images at different times and from different perspectives. The fusion process uses techniques such as weighted averaging and least squares to more accurately determine the depth position of each key point.

[0030] Depth data extracted and fused from binocular images is fed back into a basic digital twin model, which is a virtual representation of the conductor. This model allows for real-time monitoring of the conductor's shape and condition. The update process adjusts the positional information of each node in the digital twin model based on the actual measured depth data, making the model more closely resemble the actual situation. In this way, the gap between the virtual model and the real world gradually narrows. Based on the updated depth data, a corrected digital twin model is generated. This corrected model more accurately reflects the actual condition of the transmission line, especially in terms of sag. By comparing it with actual measured data, the corrected model can provide more accurate sag monitoring, fault warning, and other functions.

[0031] Dynamic simulation is performed based on the corrected digital twin model, and an early warning of sag anomalies is generated based on the dynamic simulation results.

[0032] Dynamic simulation is based on a calibrated digital twin model. This dynamic simulation divides the physical model of the transmission line into small discrete units; for example, each conductor segment can be considered an element. Mechanical analysis is then performed between these small units to simulate their response under external conditions (such as wind, temperature, and gravity). Through dynamic simulation, data such as sag deformation and tension distribution of the conductor under different loads and environments can be obtained. During the simulation, by calculating the sag changes of the conductor, any deviation from the normal range can be detected. If the simulation results show that the sag exceeds the allowable safety range, a sag anomaly warning can be triggered. This warning notifies maintenance personnel that the conductor may have overloaded, excessively wind-loaded, or abnormally temperatureed issues, allowing for timely inspection and repair. The sag anomaly warning system can prevent transmission line breakage, instability, and other safety accidents, ensuring the stable operation of the power grid.

[0033] Furthermore, the step of performing local feature analysis and establishing an adaptive authentication window includes: Configure multiple feature extraction channels, use the multiple feature extraction channels to extract features from the main view image, and establish a feature set, which includes texture features, color features, and geometric features; perform feature fusion based on an attention mechanism on the feature set, and establish a feature fusion result; after segmenting the main view image with the feature fusion result, establish an adaptive authentication window centered on each pixel.

[0034] When processing images, in order to obtain more comprehensive feature information, multiple feature extraction channels need to be configured to extract different types of image features. These channels use specific algorithms or models to process the image and extract corresponding features, including texture features, color features, and geometric features. Among them, the texture feature channel uses methods such as gray-level co-occurrence matrix to extract texture information in the image, capturing the local patterns and texture distribution of the image; the color feature channel extracts color information and identifies different color regions by analyzing the color distribution of the image, such as the values ​​of the RGB color space; the geometric feature channel extracts geometric shape information in the image, such as the outline and angle of the wire, through edge detection and other methods.

[0035] The front view image is processed using the above multi-feature extraction channels to obtain different types of features. The feature information generated by each channel is integrated to establish a feature set, which includes texture features, color features, geometric features, etc.

[0036] The attention mechanism is used to automatically focus on the most important parts of the input data. In this step, the attention mechanism is used to fuse features from multiple feature extraction channels. The basic idea of ​​the attention mechanism is to assign different weights to different features based on the importance of each feature. Specifically, the importance of each feature in the whole image is analyzed, a weight is given, and the contribution of each feature is adjusted according to the weight so that important features have a greater impact on subsequent analysis. Through feature fusion, a feature fusion result is established, which can more accurately represent the overall features of the image.

[0037] Image segmentation divides an image into multiple regions, each representing a part of the image with similar features. Based on the feature fusion result, different regions in the image are identified, such as lines, backgrounds, etc., and image segmentation is performed on the main view image. On the basis of image segmentation, an adaptive authentication window is established. Adaptive means that the size and shape of the window are adjusted according to the feature changes around each pixel. For example, the window is larger in areas with more complex textures to capture more information, while the window can be smaller in simple or smooth areas.

[0038] Furthermore, the calculation of the pixel matching cost of the preprocessed binocular image using the adaptive authentication window also includes: The binocular camera is used for camera acquisition and analysis to establish a preset deviation granularity; after the granularity of the preset deviation granularity is expanded, the preprocessed binocular image is segmented to establish a mapping segmentation block; all pixel similarity comparisons within the mapping segmentation block are performed to establish a disparity candidate set; the adaptive authentication window is called to perform matching cost analysis of the disparity candidate set and establish pixel matching cost calculation results.

[0039] Camera data acquisition and analysis using a stereo camera aims to identify deviations in the image data. These deviations may be caused by factors such as lighting conditions and inaccurate camera calibration. The analysis identifies key parameters in the image, such as noise level, contrast, and color deviation. Preset deviation granularity refers to the tolerable error range or degree of deviation when processing image data. Based on the acquisition and analysis results, an appropriate deviation granularity is set. A higher deviation granularity reduces false alarms in image processing but may also increase false negatives.

[0040] Granularity expansion is performed on the preset deviation granularity. Granularity expansion is achieved by adjusting the range or flexibility of the deviation granularity to adapt to areas in the image with large changes or high uncertainty. Granularity expansion can be achieved by increasing the error tolerance or adjusting the error evaluation algorithm, so that the system can better adapt to image regions with different complexities.

[0041] The preprocessed stereo image is segmented using an adjusted preset deviation granularity. Segmentation methods include threshold-based methods and clustering-based methods. This divides the image into multiple mapped segmentation blocks with similar features. These mapped segmentation blocks will serve as basic processing units in subsequent processing. The establishment of mapped segmentation blocks means that pixels within each block will be considered to have similar features or depth attributes, which helps reduce computational complexity and improve processing efficiency.

[0042] Pixels within each mapped segmentation block undergo similarity comparison. The purpose of similarity comparison is to determine the degree of matching between two pixels. The absolute difference method can be used to calculate the differences between two pixels in terms of grayscale, color, etc. Within each segmentation block, similarity calculations are performed to compare pixels at different locations. Based on the pixel similarity comparison results, a disparity candidate set is established. The disparity candidate set includes all corresponding pixel pairs from different viewpoints. A larger disparity value indicates that the object is closer to the camera, while a smaller disparity value indicates that the object is farther away. Through the disparity candidate set, more accurate depth information can be obtained.

[0043] The adaptive authentication window is invoked, which adjusts the matching weight of each pixel according to specific features to improve the accuracy of the disparity candidate set matching process, especially in complex areas. The final pixel matching cost calculation results provide high-precision matching data for subsequent depth calculation and 3D modeling.

[0044] Furthermore, the step of invoking the adaptive authentication window to perform disparity candidate set matching cost analysis and establishing pixel matching cost calculation results includes: Based on the adaptive authentication window, center weights, gradient weights, and color weights are created as follows: ; ; ; in, Characterizing the center weight, The position of the current pixel. The position of the center pixel of the window. Parameters characterizing the rate of weight decay. Characterizing gradient weights Characterization location Pixel gradient magnitude at that location This represents the maximum value of the gradient within the window. To adapt to the position of any pixel within the authentication window. Characterizes small positive numbers that avoid division by zero. Characterizing color weights, Characterization location The pixel color value at that location, The pixel color value representing the center position of the window. To control the parameters of color weight decay, the center weight, the gradient weight, and the color weight are weighted and fused, and the matching cost of the disparity candidate set is weighted and calculated based on the weighted fusion result to establish the pixel matching cost calculation result.

[0045] The center weight is The center weight is calculated based on the distance between the current pixel and the center pixel of the window. Pixels closer to the center of the window will be given a higher weight, while pixels farther away from the center will have a lower weight. This is done so that the pixels in the center have a greater impact on the final image processing, because the central region usually contains more important information.

[0046] Gradient weights are Gradient weights focus on regions with large gradients in an image, typically the edges or areas of significant change. The larger the gradient value, the more dramatic the change in the image, such as edges or contours. These regions are usually more important and therefore given higher weights. By calculating the ratio of pixel gradients to the maximum gradient within a window, edge information in the image can be highlighted.

[0047] Color weight is Color weights are calculated based on the similarity between the color of the current pixel and the color of the center of the window. Pixels with similar colors are given higher weights, so that areas with strong color consistency receive more attention. This helps to ensure that areas with similar colors are given priority during image matching, reducing interference caused by color differences.

[0048] Through this weighted mechanism, the adaptive authentication window can dynamically adjust its processing strategy according to different parts and features of the image, enabling tasks such as pixel matching and feature extraction to focus more accurately on important information, thereby improving the overall effect of image analysis.

[0049] The center weight, gradient weight, and color weight are weighted and fused to determine their influence on the final result. The final weights are obtained through weighted fusion and used to calculate the matching cost of the disparity candidate set to adjust the matching score of different pixel pairs. Specifically, pixel pairs with larger weights have a greater impact on the final matching result, thus affecting subsequent depth calculations. In the weighted fusion calculation, the final pixel matching cost result reflects the degree of matching of pixel pairs in the image. For example, in the process of stereo matching, the matching cost is calculated based on the differences in color, texture, gradient, etc. between pixels. The weighted fusion cost calculation helps to more accurately determine whether pixel pairs match, especially when there are complex backgrounds or noise in the image. The weighted calculation improves the robustness of the matching.

[0050] Furthermore, the generation of sag anomaly early warning based on dynamic simulation results includes: A static safety threshold is established based on the design data; environmental images of the transmission line are collected to establish an environmental dataset; the static safety threshold is corrected using the environmental dataset to establish a dynamic safety threshold; threshold trigger analysis of dynamic simulation results is performed based on the dynamic safety threshold to establish an early warning system for sag anomalies.

[0051] Static safety thresholds are established based on design data. These thresholds define the safe range of transmission lines under standard conditions, specifically including sag tolerance, current load limit, wind load, and temperature limits. The sag tolerance is the tolerable range of conductor sag deviation under normal operating conditions without external interference. The current load limit is the maximum current a conductor can carry under normal operating conditions. Wind load and temperature limits are the limits that a conductor can withstand under environmental conditions such as maximum wind speed and temperature, based on the transmission line's design conditions. These thresholds are static because they are calculated based on fixed design data that do not consider real-time environmental changes.

[0052] Using image acquisition devices, such as high-definition cameras, images are taken of the power transmission line and its surrounding environment to obtain an environmental dataset about the environmental condition. These images can provide information about environmental factors such as wind speed, climate change, plant growth, and snow cover. Furthermore, these environmental datasets are transformed into the same coordinate system as the key points for sag detection to construct a three-dimensional point cloud of the environment, thereby reflecting the current environmental condition of the power transmission line.

[0053] The static safety threshold is corrected based on the environmental dataset. For example, increased wind speed may lead to increased conductor sag, necessitating adjustments to the sag tolerance value. Temperature changes may cause conductor expansion or contraction, affecting the safe operating range. The corrected threshold reflects the actual safety status of the transmission line under real-time environmental conditions and can dynamically adapt to external factors such as weather changes and load fluctuations. By correcting the static safety threshold, a dynamic safety threshold is formed. This threshold adjusts according to real-time environmental changes, better reflecting the current safety status of the transmission line and helping the monitoring system detect potential risks in real time.

[0054] The simulation results are compared with the dynamic safety threshold, and threshold-triggered analysis is performed. The main purpose of threshold-triggered analysis is to monitor whether the simulation results exceed the dynamic safety threshold range. If the sag or other safety parameters in the simulation results exceed the dynamic safety threshold, it is determined that there is a potential safety risk. In this case, an abnormal sag warning is automatically triggered to report the risk to the operation and maintenance personnel.

[0055] Furthermore, the step of generating an anomaly warning based on dynamic simulation results also includes: Establish a time-series sag detection dataset for the transmission line; perform time-series state change analysis using the dynamic simulation results and the time-series sag detection dataset to establish time-series state change results; generate additional anomaly warnings using the time-series state change results and execute the reporting of additional anomaly warnings.

[0056] The time-series sag detection dataset was established by long-term monitoring and recording of sag changes in transmission lines. It covers sag data from different time periods. The establishment of the time-series sag dataset helps to analyze the trend of sag changes, especially to identify some gradually accumulating abnormal changes at an early stage.

[0057] By combining dynamic simulation results with a time-series sag detection dataset, the simulation results provide predicted sag changes based on environmental and design parameters, while the time-series sag detection dataset records actual sag change data. By combining these two types of data, the deviation between actual sag changes and theoretical predictions can be analyzed, thereby identifying potential anomalies. Time-series state change analysis mainly detects the changing trend of sag values ​​in different time periods, especially whether there are abnormal fluctuations or trends. Through this analysis, the periodic patterns of sag changes and sudden anomalies can be determined.

[0058] When the time sequence state change analysis shows abnormalities, additional anomaly warnings are generated based on these anomalies. These additional anomaly warnings target potential risks that have not been directly detected, or anomalies that have gradually accumulated and have not yet reached a severe level. For example, sag changes may exceed the normal fluctuation range of the long-term trend, sag values ​​may show a continuous increasing trend, and the deviation between the actual sag and the simulation results may exceed the set tolerance threshold. By issuing additional anomaly warnings, maintenance personnel can be alerted in advance to these potential risks and prevent problems from developing into more serious accidents.

[0059] Furthermore, the process of establishing a binocular image includes: Perform data stability analysis on the binocular images and establish data stability analysis results; if the data stability analysis results cannot meet the preset stability threshold, generate additional acquisition instructions; call the fusion sensor to perform additional data acquisition according to the additional acquisition instructions; perform binocular image compensation based on the additional data acquisition results.

[0060] Data stability analysis is performed on binocular images. For example, temporal stability analysis compares whether image data changes too rapidly or exhibits sudden jumps within consecutive time periods, which are caused by camera vibration or other external factors. Spatial stability analysis examines whether changes in different regions of the image are uniform and whether any areas show abnormal deformation or blurring, which are caused by changes in lighting, object movement, or image alignment issues. By analyzing these stability indicators, a data stability analysis result is established. This result can be expressed numerically or as a score, reflecting the image's stability in both time and space.

[0061] A preset stability threshold is used to determine whether the binocular images meet the required stability standard. This threshold is set based on empirical data or system requirements; for example, image displacement error, contrast variation, or noise level can all be included as part of the stability threshold. If the data stability analysis results show that the image stability does not meet the preset threshold, additional acquisition instructions are generated to correct the data instability and ensure that the next image acquisition can provide higher quality data.

[0062] A fusion sensor is a multi-sensor system that can acquire data from different types of sensors and fuse this data to provide more accurate and comprehensive information, such as accelerometers, gyroscopes, and lidar, to assist in image acquisition and compensation. Additional data includes information acquired using the aforementioned sensors to supplement image data and compensate for defects in previous images.

[0063] Binocular image compensation is performed based on additional data acquisition results. For example, based on sensor data, the angle, position, or other parameters of the image are adjusted so that the images acquired multiple times can be accurately aligned, reducing errors and further optimizing image quality. This provides high-quality data support for subsequent depth calculation and sag monitoring.

[0064] In summary, the digital twin-based transmission line sag monitoring method provided in this application has the following technical effects: By reading the design data of transmission lines and establishing a basic digital twin model based on this data, an accurate virtual model is provided for subsequent monitoring and analysis. This lays a solid foundation for real-time monitoring of transmission line sag, enabling more precise reflection of the line's geometry and operating status. By identifying key points for sag detection on the digital twin model and configuring the binocular camera's acquisition parameters based on the identification results, the impact of sag changes on the transmission line can be accurately identified. Adjusting the binocular camera's acquisition parameters ensures the quality of image data and the rationality of the acquisition angle, thereby improving the accuracy of sag monitoring. Using a binocular camera for image acquisition and preprocessing effectively improves image quality, ensuring the accuracy of subsequent feature extraction and analysis. Furthermore, selecting any image from the preprocessed binocular images as the main view image allows for local feature analysis, providing a basis for subsequent... Pixel matching and sag detection provide clear and high-quality image data. An adaptive authentication window is established through local feature analysis, and the pixel matching cost of the binocular image is calculated based on this window to perform 3D point cloud reconstruction for sag detection. This method considers the changes in local image features and enhances the system's adaptability to image matching under different conditions. The depth data extracted from the key points of sag detection is used to update the basic digital twin model, thereby generating a corrected digital twin model. This model more accurately reflects the actual state of the transmission line, especially under dynamic environmental conditions, and can update the state in real time. Dynamic simulation based on the corrected digital twin model can simulate the sag changes of the conductor under different environmental and load conditions, and generate sag anomaly warnings based on the simulation results. This simulation can comprehensively consider the influence of different environmental factors, further improving the prediction and warning capabilities for sag anomalies.

[0065] Example 2, based on the same inventive concept as the digital twin-based transmission line sag monitoring method in the foregoing examples, such as... Figure 2 As shown in the figure, this application provides a transmission line sag monitoring device based on digital twin, the device comprising: The basic digital twin model construction module 10 is used to read the design data of the transmission line after obtaining user authorization, and construct a basic digital twin model of the transmission line based on the design data; the binocular acquisition parameter configuration module 20 is used to establish the sag detection key point recognition result based on the basic digital twin model, and configure the binocular acquisition parameters; the transmission line acquisition module 30 is used to control the binocular camera to acquire transmission line data using the binocular acquisition parameters, and establish binocular images; the local feature analysis module 40 is used to select any image in the binocular images as the main view image, and perform local feature analysis to establish an adaptive... The system includes: an authentication window; a pixel matching cost calculation module 50, used to calculate the pixel matching cost of the preprocessed binocular image using the adaptive authentication window, and to reconstruct the 3D point cloud for sag detection based on the pixel matching cost calculation result; a calibration twin model generation module 60, used to establish fusion depth data of key points for sag detection using the 3D point cloud for sag detection, update the basic digital twin model based on the fusion depth data, and generate a calibration digital twin model; and a sag anomaly warning generation module 70, used to perform dynamic simulation based on the calibration digital twin model, and generate a sag anomaly warning based on the dynamic simulation result.

[0066] Furthermore, the local feature analysis module 40 includes: The feature extraction unit is used to configure multiple feature extraction channels, extract features from the main view image using the multiple feature extraction channels, and establish a feature set, which includes texture features, color features, and geometric features; the feature fusion unit is used to perform feature fusion based on an attention mechanism on the feature set and establish a feature fusion result; the authentication window establishment unit is used to segment the main view image with the feature fusion result and establish an adaptive authentication window centered on each pixel.

[0067] Furthermore, the pixel matching cost calculation module 50 includes: The camera acquisition and analysis unit is used to perform camera acquisition and analysis on the stereo camera and establish a preset deviation granularity; the stereo image segmentation unit is used to perform granularity expansion of the preset deviation granularity and segment the preprocessed stereo image to establish a mapping segmentation block; the similarity comparison unit is used to perform all pixel similarity comparisons within the mapping segmentation block and establish a disparity candidate set; the matching cost analysis unit is used to call the adaptive authentication window to perform matching cost analysis on the disparity candidate set and establish pixel matching cost calculation results.

[0068] Furthermore, the pixel matching cost calculation module 50 includes: The weight creation unit is used to create center weights, gradient weights, and color weights based on the adaptive authentication window, as follows: ; ; ; in, Characterizing the center weight, The position of the current pixel. The position of the center pixel of the window. Parameters characterizing the rate of weight decay. Characterizing gradient weights Characterization location Pixel gradient magnitude at that location This represents the maximum value of the gradient within the window. To adapt to the position of any pixel within the authentication window. Characterizes small positive numbers that avoid division by zero. Characterizing color weights, Characterization location The pixel color value at that location, The pixel color value representing the center position of the window. The parameters for controlling color weight decay are: a weighted calculation unit, which performs weighted fusion of the center weight, the gradient weight, and the color weight, and then performs weighted calculation of the matching cost of the disparity candidate set based on the weighted fusion result to establish the pixel matching cost calculation result.

[0069] Furthermore, the sag anomaly early warning generation module 70 includes: The system includes a static safety threshold establishment unit for establishing a static safety threshold based on the design data; an environmental image acquisition unit for acquiring environmental images of the transmission line and establishing an environmental dataset; a static safety threshold correction unit for correcting the static safety threshold using the environmental dataset and establishing a dynamic safety threshold; and a threshold trigger analysis unit for performing threshold trigger analysis on the dynamic simulation results based on the dynamic safety threshold and establishing an abnormal sag warning.

[0070] Furthermore, the sag anomaly early warning generation module 70 includes: The detection dataset establishment unit is used to establish a time-series sag detection dataset for the transmission line; the change analysis unit is used to perform time-series state change analysis using the dynamic simulation results and the time-series sag detection dataset, and establish time-series state change results; the anomaly warning reporting unit is used to generate additional anomaly warnings using the time-series state change results, and execute the reporting of additional anomaly warnings.

[0071] Furthermore, the transmission line acquisition module 30 includes: The data stability analysis unit is used to perform data stability analysis on the binocular images and establish data stability analysis results; the instruction generation unit is used to generate additional acquisition instructions if the data stability analysis results cannot meet the preset stability threshold; the additional data acquisition unit is used to call the fusion sensor to perform additional data acquisition according to the additional acquisition instructions; and the binocular image compensation unit is used to perform binocular image compensation according to the additional data acquisition results.

[0072] Through the foregoing detailed description of the transmission conductor sag monitoring method based on digital twins, those skilled in the art can clearly understand the transmission conductor sag monitoring device based on digital twins in this embodiment. Since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and relevant parts can be referred to the method section.

[0073] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for monitoring the sag of transmission conductors based on digital twins, characterized in that, The method includes: After obtaining user authorization, the design data of the transmission line is read, and a basic digital twin model of the transmission line is constructed based on the design data; Based on the aforementioned basic digital twin model, the key point identification results for sag detection are established, and the binocular acquisition parameters are configured. The binocular acquisition parameters are used to control the binocular camera to acquire data from the power transmission line and establish a binocular image; Select any image from the binocular images as the main view image, perform local feature analysis, and establish an adaptive authentication window; The pixel matching cost of the preprocessed binocular image is calculated using the adaptive authentication window, and the sag detection 3D point cloud reconstruction is performed based on the pixel matching cost calculation result. The fusion depth data of key points for sag detection is established using the 3D point cloud of the sag detection, and the basic digital twin model is updated based on the fusion depth data to generate a corrected digital twin model. Dynamic simulation is performed based on the corrected digital twin model, and an early warning of sag anomalies is generated based on the dynamic simulation results.

2. The method for monitoring the sag of transmission lines based on digital twins as described in claim 1, characterized in that, The process of performing local feature analysis and establishing an adaptive authentication window includes: Configure multiple feature extraction channels, use the multiple feature extraction channels to extract features from the front view image, and establish a feature set, which includes texture features, color features, and geometric features; The feature set is subjected to feature fusion based on an attention mechanism to establish the feature fusion result; After segmenting the main view image using the feature fusion result, an adaptive authentication window centered on each pixel is established.

3. The method for monitoring the sag of transmission lines based on digital twins as described in claim 2, characterized in that, The calculation of the pixel matching cost of the preprocessed binocular image using the adaptive authentication window further includes: The binocular camera is used to perform camera acquisition and analysis to establish a preset deviation granularity; After performing the granularity expansion with the preset deviation granularity, the preprocessed binocular image is segmented to establish a mapped segmentation block; Perform a similarity comparison of all pixels within the mapped segmentation block to establish a disparity candidate set; The adaptive authentication window is invoked to perform a matching cost analysis of the disparity candidate set, and a pixel matching cost calculation result is established.

4. The method for monitoring the sag of transmission lines based on digital twins as described in claim 3, characterized in that, The step of calling the adaptive authentication window to perform disparity candidate set matching cost analysis and establishing pixel matching cost calculation results includes: Based on the adaptive authentication window, center weights, gradient weights, and color weights are created as follows: ； ； ； in, Characterizing the center weight, The position of the current pixel. The position of the center pixel of the window. Parameters characterizing the rate of weight decay. Characterizing gradient weights Characterization location Pixel gradient magnitude at that location This represents the maximum value of the gradient within the window. To adapt to the position of any pixel within the authentication window. Characterizes small positive numbers that avoid division by zero. Characterizing color weights, Characterization location The pixel color value at that location, The pixel color value representing the center position of the window. Parameters for controlling color weight decay; After weighted fusion of the center weight, the gradient weight, and the color weight, the matching cost of the disparity candidate set is weighted based on the weighted fusion result to establish the pixel matching cost calculation result.

5. The method for monitoring the sag of transmission lines based on digital twins as described in claim 1, characterized in that, The generation of sag anomaly early warning based on dynamic simulation results includes: Establish a static safety threshold based on the design data; Environmental images of the transmission line are acquired to create an environmental dataset; The static security threshold is corrected using the environmental dataset to establish a dynamic security threshold; Based on the dynamic safety threshold, threshold trigger analysis of the dynamic simulation results is performed to establish an early warning system for sag anomalies.

6. The method for monitoring the sag of transmission lines based on digital twins as described in claim 1, characterized in that, The method of generating sag anomaly early warning based on dynamic simulation results also includes: Establish a time-series sag detection dataset for the transmission line; Using the dynamic simulation results and the time-series sag detection dataset, a time-series state change analysis is performed to establish the time-series state change results. Additional anomaly warnings are generated using the time-series state change results, and the additional anomaly warnings are reported.

7. The method for monitoring the sag of transmission lines based on digital twins as described in claim 1, characterized in that, The process of creating a binocular image includes: Data stability analysis is performed on the binocular images to establish data stability analysis results; If the data stability analysis results do not meet the preset stability threshold, an additional acquisition instruction will be generated. The additional acquisition command is used to invoke the fusion sensor to acquire additional data. Binocular image compensation is performed based on the additional data acquisition results.

8. A transmission line sag monitoring device based on digital twin, characterized in that, For implementing the digital twin-based transmission conductor sag monitoring method according to any one of claims 1-7, the apparatus comprises: The basic digital twin model construction module is used to read the design data of the transmission line after obtaining user authorization, and construct the basic digital twin model of the transmission line based on the design data. The binocular acquisition parameter configuration module is used to establish the key point recognition results of sag detection based on the basic digital twin model, and to configure the binocular acquisition parameters. The transmission line acquisition module is used to control the binocular camera to acquire data of the transmission line using the binocular acquisition parameters and to establish a binocular image; The local feature analysis module is used to select any image from the binocular images as the main view image, perform local feature analysis, and establish an adaptive authentication window; The pixel matching cost calculation module is used to calculate the pixel matching cost of the preprocessed binocular image using the adaptive authentication window, and to perform sag detection and 3D point cloud reconstruction based on the pixel matching cost calculation result. The calibration twin model generation module is used to establish fusion depth data of key points of sag detection using the 3D point cloud of sag detection, update the basic digital twin model based on the fusion depth data, and generate a calibration digital twin model. The sag anomaly early warning generation module is used to perform dynamic simulation based on the corrected digital twin model and generate sag anomaly early warning based on the dynamic simulation results.

Images