Ground wire feature extraction method and system for complex environment

Abstract

The invention relates to the technical field of ground wire feature extraction. The invention relates to a ground wire feature extraction method and system for a complex environment. The method comprises the following steps: S1, acquiring remote sensing data of a multi-source satellite, positioning the position of a power transmission tower, and identifying attitude data of the power transmission tower in combination with the remote sensing data; s2, determining a ground wire suspension coordinate according to the attitude data, framing a region of interest of the ground wire by combining the ground wire suspension coordinate with a catenary equation, and performing same-position screening on the remote sensing data according to the region of interest; through collaborative design of multi-source remote sensing data fusion and power transmission tower attitude constraint, the problem of fuzzy positioning of a ground wire range in a complex environment is effectively solved, power transmission tower position data of a power transmission line management end is taken as a basis, power transmission tower attitude parameters are identified in combination with the multi-source remote sensing data, and the positioning accuracy of the ground wire range is improved. The earth wire interest area is accurately framed through suspension point coordinates and a catenary equation, the calculation range is greatly reduced, and irrelevant background interference is eliminated.

Description

Technical Field

[0001] This invention relates to the field of conductor and ground wire feature extraction technology, and more specifically, to a method and system for extracting conductor and ground wire features in complex environments. Background Technology

[0002] In the field of power transmission line inspection technology, conductors and ground wires, as the core carriers of power transmission, have geometric and state characteristics that are key to assessing line operation safety and predicting potential faults.

[0003] In complex environments such as heavy snow and ice, existing technologies have many unavoidable defects that seriously affect the accuracy and reliability of feature extraction. In complex environments, conductors and ground wires are easily covered by snow and ice, resulting in blurred boundaries between the conductors and ground wires and the background. Traditional visible light-based identification methods cannot effectively distinguish the conductor and ground wire body from the snow and ice. Single-mode remote sensing data cannot balance penetration and resolution, causing feature localization errors. Secondly, existing technologies mostly focus only on the impact of snow and do not fully consider the coupling effect of snow and ice. The calculation of cold shrinkage does not incorporate the thermal resistance correction of the ice layer, and the load calculation ignores the weight of the ice layer, resulting in large errors in the calculation of state features. Therefore, a feature extraction method for conductors and ground wires in complex environments is proposed. Summary of the Invention

[0004] The purpose of this invention is to provide a method for extracting conductor features in complex environments, so as to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, a method for extracting conductor and ground wire features in complex environments is provided, comprising the following steps: S1. Acquire remote sensing data from multiple satellite sources, locate the position of the power transmission tower, and identify the attitude data of the power transmission tower by combining the remote sensing data; S2. Determine the suspension coordinates of the conductor and ground wire based on the attitude data. Combine the suspension coordinates of the conductor and ground wire with the catenary equation to define the region of interest of the conductor and ground wire. Then, filter the remote sensing data according to the region of interest so that only the location of the region of interest is retained in the remote sensing data. S3. Based on the retained remote sensing data, analyze the snow cover parameters and icing parameters of the region of interest, divide the region of interest into sub-regions based on the snow cover parameters and icing parameters, incorporate the thermal resistance of the icing layer to correct the surface temperature of the conductor and ground wire, and calculate the cooling shrinkage of each region in combination with the conductor and ground wire equipment parameters. S4. Extract historical remote sensing data and combine it with the conductor and ground wire exposure ratio of the remote sensing data retained in S2 to set reference quality values ​​for different regions, and set dynamic reference thresholds to filter high-quality regions that meet the requirements. S5. Based on the snow cover parameters, icing parameters and remote sensing data of the high-quality area, the conductor and ground wire morphology is deduced for the conductor and ground wire characteristics of the intermediate area between adjacent high-quality areas, and then the conductor and ground wire characteristics of the intermediate area are selected according to the reference quality value. S6. Combine the characteristics of conductors and ground wires in all regions to perform feature conflict correction, and output the correction result as the final characteristics of conductors and ground wires.

[0006] As a further improvement to this technical solution, in S1, the remote sensing data from the multi-source satellites includes SAR fully polarimetric data, high-resolution multispectral data, InSAR data, and thermal infrared data. By connecting to the transmission line management terminal, the location data of the transmission towers is extracted from the transmission line management terminal, and then remote sensing data of the same location as the transmission tower location data is extracted. The attitude data of the transmission towers is identified based on the remote sensing data. Attitude data includes three-dimensional coordinates, tower tilt angle and crossarm azimuth angle, and crossarm length.

[0007] As a further improvement to this technical solution, the method of identifying the attitude data of transmission towers based on remote sensing data involves detecting candidate regions of transmission towers from multispectral data using the YOLOv8 algorithm, filtering real transmission towers by combining the metal backscattering characteristics of SAR fully polarized data, extracting key points of the transmission towers through SAR contours, calculating the angle between the tower's central axis and the vertical direction to obtain the tower's tilt angle, calculating the crossarm azimuth angle by combining the coordinates of the crossarm endpoints, and simultaneously obtaining the crossarm length.

[0008] As a further improvement to this technical solution, in step S2, the offset of the suspension point relative to the tower base is calculated based on the azimuth angle and length of the crossarm of the transmission tower, and the suspension coordinates are obtained by combining the three-dimensional coordinates of the transmission tower. The region of interest for the conductor and ground wire is defined by combining the suspended coordinates of the conductor and ground wire with the catenary equation. The region of interest is a buffer zone that extends to both sides along the catenary; Remote sensing data are filtered based on the location of the region of interest, retaining only the remote sensing data located at the same location as the region of interest.

[0009] As a further improvement to this technical solution, in step S3, snow cover parameters and icing parameters of the region of interest are analyzed based on the retained remote sensing data to obtain the snow cover parameters and icing parameters of the region of interest. Set differential snow cover parameters and differential icing parameters, and then combine the snow cover parameters and icing parameters of the region of interest with the differential snow cover parameters and differential icing parameters to divide the region of interest into multiple sub-regions; In this process, the snow accumulation parameters and icing parameters of each node in the region of interest are calculated. If the snow accumulation parameters and icing parameters between adjacent nodes do not exceed the difference snow accumulation parameters and difference icing parameters, the adjacent nodes are merged and then compared with the next node. Conversely, if the difference snow accumulation parameters or difference icing parameters between adjacent nodes exceed the difference snow accumulation parameters or difference icing parameters, it is determined that a sub-region is formed. A fixed length is set for each sub-region based on the length of the region of interest. When the fixed length is reached between adjacent nodes, a sub-region is formed.

[0010] As a further improvement to this technical solution, the snow accumulation parameters and icing parameters of the region of interest are obtained; The snow cover parameters include snow depth, which is obtained by calculating the Normalized Snow Index (NDSI) from high-resolution multispectral data. The icing parameters include the icing layer thickness, which is inverted by the proportion of surface scattering components in SAR fully polarized data.

[0011] As a further improvement to this technical solution, in step S4, historical remote sensing data of the region of interest is extracted, and conductor and ground wire features are extracted from the remote sensing data retained in step S2. The exposed ratio of conductor and ground wire is calculated based on the extracted conductor and ground wire features. Then, the exposed ratio of conductor and ground wire is combined with historical remote sensing data to set reference quality values ​​for sub-regions, thereby obtaining the reference quality values ​​corresponding to each sub-region. Reference thresholds are set based on the total snow cover parameters and ice parameters summarized from all sub-regions; If the reference quality value of a region exceeds the reference threshold, the region is determined to be a high-quality region. If the reference quality value of a sub-region does not exceed the reference threshold, the sub-region is determined to be a non-high-quality region, and the next sub-region is compared.

[0012] As a further improvement to this technical solution, in step S5, the characteristics of the conductor and ground wire in the intermediate region between adjacent high-quality regions are deduced based on the snow accumulation parameters, icing parameters, and conductor and ground wire morphology of the high-quality region and remote sensing data. The conductor morphology consists of snow, ice, and the conductor itself, and includes surface curvature and sag value. The deduction of the conductor and ground wire characteristics in the intermediate region includes geometric characteristic deduction and state characteristic deduction.

[0013] The conductor characteristics of the intermediate region are selected based on the reference quality value; When the reference quality value of the sub-region in the middle area is greater than 0.7, the derived features are used directly; When the reference quality value of the sub-region in the middle area is greater than 0.5 and less than 0.7, the inferred features are smoothed and corrected by combining the morphological trend of the nearest high-quality region. When the reference quality value of the sub-region in the middle region is less than 0.5, the inferred features are discarded and the average value of the features of the adjacent high-quality regions is used to fill the gap.

[0014] As a further improvement to this technical solution, in S6, the coordinate curve of the conflict area is fitted using the least squares method based on the three-dimensional coordinates of the high-quality area. When there is a sag conflict, the total load of snow and ice is recalculated by combining the catenary equation and the total load of snow and ice. When there is a shrinkage conflict, the surface temperature and shrinkage of the conflict area are recalculated based on the standard value of the thermal expansion and contraction coefficient of the conductor. When there is a snow and ice load conflict, historical load data from the same period are introduced for weighted average correction.

[0015] The second objective of this invention is to provide a conductor and ground wire feature extraction system for complex environments, including any one of the conductor and ground wire feature extraction methods for complex environments described above, comprising a data acquisition module, a region selection module, a region calculation module, a region quality screening module, and a feature output module. The data acquisition module is used to acquire remote sensing data from multiple satellites, locate the position of the transmission tower, and identify the attitude data of the transmission tower by combining the remote sensing data. The region selection module is used to determine the suspension coordinates of the conductor and ground wire based on the attitude data, combine the suspension coordinates of the conductor and ground wire with the catenary equation to define the region of interest of the conductor and ground wire, and filter the remote sensing data according to the region of interest so that only the location of the region of interest is retained in the remote sensing data. The region calculation module is used to analyze the snow cover parameters and icing parameters of the region of interest based on the retained remote sensing data, divide the region of interest into regions based on the snow cover parameters and icing parameters, incorporate the thermal resistance of the icing layer to correct the surface temperature of the conductor and ground wire, and calculate the cooling shrinkage of each region in combination with the conductor and ground wire equipment parameters. The regional quality screening module is used to extract historical remote sensing data and combine it with the conductor and ground wire exposure ratio of the retained remote sensing data to set reference quality values ​​for different regions, and to set dynamic reference thresholds to screen high-quality regions that meet the requirements. The feature output module is used to deduce the conductor and ground wire characteristics of the intermediate region between adjacent high-quality regions based on the snow accumulation parameters, icing parameters and conductor and ground wire morphology of the high-quality region and remote sensing data. Then, the conductor and ground wire characteristics of the intermediate region are selected according to the reference quality value, and feature conflict correction is performed by combining the conductor and ground wire characteristics of all regions. The correction result is output as the final feature of the conductor and ground wire.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This method and system for extracting conductor and ground wire features in complex environments effectively solves the problem of ambiguous positioning of conductor and ground wire range in complex environments by combining multi-source remote sensing data fusion with transmission tower attitude constraints. Based on the transmission tower location data at the transmission line management end, the system identifies the transmission tower attitude parameters by combining multi-source remote sensing data, and accurately defines the region of interest of conductor and ground wire by using the suspension point coordinates and catenary equation, thereby significantly reducing the calculation range and eliminating irrelevant background interference.

[0017] 2. This method and system for extracting conductor features in complex environments significantly improves the accuracy of feature parameter inversion in complex environments through the collaborative inversion of snow and ice parameters and refined region division. It uses high-resolution multispectral data to calculate the NDSI index to invert snow depth and combines the surface scattering component ratio of SAR fully polarimetric data to invert ice layer thickness, thus achieving accurate distinction between snow and ice. At the same time, based on the parameter difference threshold and the fixed-length dual-condition division rule, the region of interest is divided into sub-regions with uniform snow and ice states, avoiding the drawback of traditional fixed-length division ignoring parameter differences.

[0018] 3. In this method and system for extracting features of conductors and ground wires in complex environments, the data quality of each sub-region is quantified by weighted calculation of the exposure ratio of conductors and ground wires combined with the stability of historical data. High-quality regions are selected as the benchmark, and a differentiated feature selection strategy is adopted for intermediate regions of different quality levels to avoid the drag of low-quality data on the overall results and to achieve accurate inference of the geometric and state features of intermediate regions. Attached Figure Description

[0019] Figure 1 This is a flowchart illustrating a method for extracting conductor features in complex environments according to the present invention. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] Please see Figure 1 As shown, the purpose of this embodiment is to provide a method for extracting features of conductors and ground wires in complex environments, including the following steps: S1. Acquire remote sensing data from multiple satellite sources, locate the position of the power transmission tower, and identify the attitude data of the power transmission tower by combining the remote sensing data; In S1, the remote sensing data from multiple satellites includes SAR fully polarimetric data (which completes radiometric calibration, multi-view processing, and denoising by filtering to preserve the backscattering characteristics of metallic targets), high-resolution multispectral data (which performs atmospheric correction, geometric registration error ≤1, and enhances the contrast between power transmission towers and the background by individual pixels), InSAR data (which completes image registration and interferogram generation), and thermal infrared data. By connecting to the transmission line management terminal, the location data of the transmission towers is extracted from the transmission line management terminal, and then remote sensing data of the same location as the transmission tower location data is extracted. The attitude data of the transmission towers is identified based on the remote sensing data. Attitude data includes three-dimensional coordinates, tower tilt angle and crossarm azimuth angle, and crossarm length.

[0022] The attitude data of transmission towers is identified based on remote sensing data. Specifically, candidate regions for transmission towers are detected from multispectral data using the YOLOv8 algorithm. Real transmission towers are then selected by combining the metallic backscattering characteristics of SAR fully polarimetric data. Key points of the transmission towers are extracted using SAR contours. The tower tilt angle is obtained by calculating the angle between the tower's central axis and the vertical direction. The azimuth angle of the crossarm is calculated by combining the coordinates of the crossarm endpoints, and the crossarm length is also obtained. The steps are as follows: The model is trained by inputting a set of labeled transmission tower samples (including tower images with different angles and snow cover). Multispectral data is input into the trained model, and the bounding rectangle of the candidate area of ​​the transmission tower is output. The full polarization data corresponding to the candidate area is extracted, and the following calculations are performed: VV is used, and the polarization backscattering coefficient is used to set a metal target screening threshold (-15dB). Candidate areas that meet the threshold condition are retained, which are the real transmission tower areas. Non-tower metal targets (such as billboards and metal buildings) are excluded. Then, the key feature points of the real transmission tower (tower base key points, tower top key points, crossarm end key points) are extracted using the contour extraction algorithm (Canny, through edge detection and contour tracking) using the following methods: InSAR is used, and the data is combined with the tower base key points using the following methods: DEM is used, and the data is used for elevation calibration to obtain the three-dimensional coordinates of the transmission tower base. The tower tilt angle is calculated by the angle between the line connecting the top of the tower and the center of the tower base (the central axis of the tower) and the vertical direction. The azimuth angle of the crossarm (the angle between the crossarm and the due north direction) is calculated by combining the coordinates of the end points of the crossarm. Then, the length of the crossarm is calculated by using the Euclidean distance formula based on the plane coordinates of the two ends of the crossarm.

[0023] S2. Determine the suspension coordinates of the conductor and ground wire based on the attitude data. Combine the suspension coordinates of the conductor and ground wire with the catenary equation to define the region of interest of the conductor and ground wire. Then, filter the remote sensing data according to the region of interest so that only the location of the region of interest is retained in the remote sensing data. In S2, the offset of the suspension point relative to the tower base is calculated based on the azimuth angle and length of the crossarm of the transmission tower, and the suspension coordinates are obtained by combining the three-dimensional coordinates of the transmission tower. The region of interest for the conductor and ground wire is defined by combining the suspended coordinates of the conductor and ground wire with the catenary equation. Among them, the region of interest is a buffer zone extending to both sides along the catenary; a strip-shaped region of interest with a width of 3m centered on the catenary is generated, covering the conductor and the surrounding possible snow and ice coverage area; Based on the location of the region of interest (ROI), remote sensing data is filtered to retain only those data located at the same location as the ROI. Spatial overlay analysis is used to retain the portions of multi-source remote sensing data that spatially overlap with the ROI, while irrelevant background data (such as trees and buildings) outside the ROI are removed to reduce subsequent computational load. The formula is as follows: ; in, Let x be the horizontal offset in the direction. The length of the crossarm. For the azimuth angle of the horizontal bearing ; in, Let y be the horizontal offset in the direction, which is... ; in, This is the elevation offset. Installation height of suspension point ; in, This refers to the conductor spacing (horizontal distance). , , , The plane coordinates of the suspension points of the two towers ; in, For the conductor elevation, The horizontal coordinates are along the span direction. The difference in elevation between the two suspension points. As the conductor is under no load tension, The weight per unit length of the conductor / ground wire, It is a hyperbolic cosine function.

[0024] S3. Based on the retained remote sensing data, analyze the snow cover parameters and icing parameters of the region of interest, divide the region of interest into sub-regions based on the snow cover parameters and icing parameters, incorporate the thermal resistance of the icing layer to correct the surface temperature of the conductor and ground wire, and calculate the cooling shrinkage of each region in combination with the conductor and ground wire equipment parameters. In S3, snow cover parameters and icing parameters of the region of interest are analyzed based on the retained remote sensing data to obtain the snow cover parameters and icing parameters of the region of interest. Obtain snow cover and icing parameters for the region of interest; Among them, snow cover parameters include snow depth, which is obtained by calculating the Normalized Difference Snow Index (NDSI) from high-resolution multispectral data, as shown in the following formula: ; in, Snow depth The normalized snow cover index is 0.8 and 0.02, which are model calibration coefficients (obtained by fitting actual snow cover samples). Icing parameters include ice layer thickness, which is inverted by the proportion of surface scattering components in SAR fully polarimetric data, as shown in the following formula:

[0025] in, For the thickness of the ice layer, The value represents the proportion of surface scattering components, and 0.8 and 0.02 are model calibration coefficients (obtained through fitting from indoor icing experiments). Set differential snow accumulation parameters (set as , passing through 2mm, threshold for the difference in snow depth between adjacent nodes) and differential icing parameters (set as 0.5mm, threshold for the difference in icing thickness between adjacent nodes). Then, combine the snow accumulation parameters and icing parameters of the region of interest with the differential snow accumulation parameters and differential icing parameters to divide the region of interest into multiple sub-regions. Nodes are evenly distributed within the region of interest along the conductor direction (catenary direction), with a node spacing of 1m. In this process, the snow accumulation parameters and icing parameters of each node in the region of interest are calculated. If the snow accumulation parameters and icing parameters between adjacent nodes do not exceed the difference snow accumulation parameters and difference icing parameters, the adjacent nodes are merged and then compared with the next node. Conversely, if the difference snow accumulation parameters or difference icing parameters between adjacent nodes exceed the difference snow accumulation parameters or difference icing parameters, it is determined that a sub-region is formed. A fixed length is set for each sub-region based on the length of the region of interest (where 5m is the maximum allowable length of a single sub-region). When the fixed length is reached between adjacent nodes, a sub-region is formed.

[0026] S4. Extract historical remote sensing data and combine it with the conductor and ground wire exposure ratio of the remote sensing data retained in S2 to set reference quality values ​​for different regions, and set dynamic reference thresholds to filter high-quality regions that meet the requirements. In S4, historical remote sensing data of the region of interest is extracted. Simultaneously, conductor wire features are extracted from the remote sensing data retained in S2. The exposed conductor wire ratio is calculated based on the extracted features. Then, the exposed conductor wire ratio is combined with historical remote sensing data to set reference quality values ​​for each sub-region, thereby obtaining the reference quality value corresponding to each sub-region. The steps are as follows: Through the remote sensing data management platform, historical remote sensing data (data from the same period and region as the current region of interest, including data from the same season and similar weather conditions over the past 3 years) are extracted to ensure that the spatial range and resolution of the historical data are consistent with the current data. Then, using the metal backscattering characteristics of the fully polarized data, the conductor and ground wire body region is identified. Combined with snow accumulation parameters and icing parameters, the areas covered by snow and ice are marked, and the remaining areas are the exposed areas of the conductor and ground wire. Measure the total length of the conductor and ground wire in each sub-region (the sub-region length along the direction of the conductor and ground wire), measure the length of the exposed area of ​​the conductor and ground wire in the sub-region, and then calculate the ratio of the two to obtain the exposed proportion of the conductor and ground wire in that sub-region. For the extracted historical remote sensing data, the standard deviation of snow cover parameters and icing parameters at corresponding locations in each sub-region is calculated to assess the stability of historical data (the smaller the standard deviation, the higher the stability). The reference quality value is calculated by weighting the conductor exposure ratio with the stability of historical data. The exposure ratio of the conductor accounts for 60% of the weight; the higher the exposure ratio, the better the data quality. Historical data stability accounts for 40% of the weight; the higher the stability, the better the data quality. Reference thresholds are set based on the total snow cover parameters and ice parameters summarized from all sub-regions; If the total snow accumulation parameter is <5mm and the total ice accumulation parameter is <1mm, then the reference threshold is ≥0.8. If 5mm ≤ total snow accumulation parameter < 15mm, and 1mm ≤ total ice accumulation parameter < 3mm, then the reference threshold is ≥ 0.7. If the total snow accumulation parameter is ≥15mm, and the ice accumulation parameter is ≥3mm, then the reference threshold is ≥0.6. If the reference quality value of a region exceeds the reference threshold, the region is determined to be a high-quality region. If the reference quality value of a sub-region does not exceed the reference threshold, the sub-region is determined not to be a high-quality region, and the next sub-region is compared. S5. Based on the snow cover parameters, icing parameters and remote sensing data of the high-quality area, the conductor and ground wire morphology is deduced for the conductor and ground wire characteristics of the intermediate area between adjacent high-quality areas, and then the conductor and ground wire characteristics of the intermediate area are selected according to the reference quality value. In S5, based on the snow cover parameters, icing parameters and the conductor morphology of remote sensing data in high-quality areas, the conductor characteristics of the intermediate area between adjacent high-quality areas are deduced. The conductor morphology consists of snow, ice, and the conductor itself, and includes surface curvature and sag value. The derivation of the conductor and ground wire characteristics in the intermediate region includes geometric feature derivation and state feature derivation, and the steps are as follows: Using the set of three-dimensional coordinate points of adjacent high-quality areas as control points, and employing spline curve interpolation, a three-dimensional coordinate curve of the intermediate region along the conductor line is generated. Then, the second derivative of the interpolated three-dimensional coordinate curve is calculated to determine the surface curvature of each coordinate point in the intermediate region. Simultaneously, considering the snow depth and ice layer thickness in the intermediate region, the total snow and ice load is calculated. This load is then substituted into the catenary equation to correct the sag value, as shown in the following formula: ; in, The coordinates of the interpolation points in the intermediate region. To control the number of points, Let B be the spline order. Let B be the value obtained through spline basis functions. For high-quality regional control point coordinates, For parameter variables; ; in, For surface curvature, The first derivative of the three-dimensional coordinate curve. The second derivative of the three-dimensional coordinate curve; ; in, The total load from snow and ice accumulation. snow density, The density of ice, This represents the snow depth in the intermediate area. The thickness of the ice layer in the middle region. Gravitational acceleration, The cross-sectional area of ​​the conductor / ground wire; Based on the snow depth and ice thickness of adjacent high-quality regions, weights are assigned according to the distance from the intermediate region's sub-nodes to the two high-quality regions (the closer the distance, the greater the weight). The weighted calculation yields the snow depth and ice thickness of each sub-node in the intermediate region. Then, combined with the surface temperature of the conductor in the intermediate region (derived from thermal infrared data), the state characteristics are calculated using the cold shrinkage formula, as follows: ; in, For the interpolation parameter values ​​in the intermediate region, Distance weights and These are parameter values ​​for adjacent high-quality regions. and This represents the distance from the intermediate node to the high-quality region. ; in, This refers to the amount of cold shrinkage in the middle area. The length of the middle region. The coefficient of thermal expansion and contraction of the conductor material is taken as 23.1 × 10⁻⁶ for aluminum stranded wire. -6 ), The design reference temperature is 25°C. The surface temperature of the intermediate region (obtained by inversion from thermal infrared data); The conductor characteristics of the intermediate region are selected based on the reference quality value; When the reference quality value of the sub-region in the middle area is greater than 0.7, the inferred features are used directly (high data reliability). When the reference quality value of the sub-region in the middle region is greater than 0.5 and less than 0.7, the inferred characteristics are smoothed and corrected by combining the morphological trend of the nearest high-quality region (such as the rate of change of curvature and the increasing sag). When the reference quality value of the sub-region in the middle area is less than 0.5, the inferred features are discarded and the average value of the features of the adjacent high-quality regions is used to fill the gap (such as the average snow depth and the average curvature).

[0027] S6. Combine the characteristics of conductors and ground wires in all regions to perform feature conflict correction, and output the correction result as the final characteristics of the conductors and ground wires. The feature conflict conditions are as follows: Three-dimensional coordinate conflict, with a continuous difference in three-dimensional coordinates at the junction of adjacent regions > 0.2m; Sag conflict, with sag values ​​in adjacent areas differing by more than 0.3m; Cold shrinkage conflict, the difference in cold shrinkage between adjacent areas is >5%; The loads from snow accumulation and icing conflict, with load deviations between adjacent areas exceeding 10%.

[0028] In S6, the coordinates of the high-quality region are used as the reference, and the least squares method is used to fit the coordinate curve of the conflict region so that the continuous difference between the corrected curve and the adjacent region is ≤0.2m. Extract the coordinates of the conflict area and two high-quality areas on each side as fitting control points, substitute them into the least squares formula to fit a new coordinate curve, and then verify the continuity difference between the corrected curve and the adjacent area to ensure that it is ≤0.2m. If the standard is not met, readjust the range of control points for fitting. When there is a sag conflict, the calculation should be recalculated by combining the catenary equation and the total load of snow and ice to ensure that the sag deviation is ≤0.3m; Extract the snow depth and ice layer thickness of the conflict area, calculate the total load, substitute the total load into the catenary equation, resolve the sag value, and then compare the deviation of the corrected sag with that of the adjacent area to ensure it is ≤0.3m. The formula is as follows:

[0029]

[0030] in, To correct the sag value, To take into account the effective tension after loading; When there is a conflict in the amount of thermal expansion and contraction, the surface temperature and amount of thermal expansion and contraction in the conflict area are recalculated based on the standard value of the thermal expansion and contraction coefficient of the conductor. Retrieve the standard coefficient of thermal expansion and contraction from the conductor and ground wire equipment manual, recalculate the surface temperature of the conductor and ground wire in the conflict area (based on thermal infrared data), and then substitute it into the shrinkage formula to calculate the corrected shrinkage, ensuring that the difference with the adjacent area is ≤5%; When snow accumulation and icing loads conflict, historical load data from the same period are used for weighted average correction to ensure that the load deviation is ≤5%. Extract historical data of snow cover and icing loads from the same period in the conflict area over the past three years. Calculate the weighted average load by weighting the current load at 60% and the historical average load at 40%. Verify the deviation between the corrected load and the adjacent areas, ensuring that it is ≤5%.

[0031] After all conflicts are corrected, feature continuity is checked again across the entire area to confirm that there are no uncorrected conflicts. The final features are then output, including the corrected 3D coordinate vector file, a quantitative report of sag / cold shrinkage / load for each area, and a feature overlay remote sensing image.

[0032] The second objective of this invention is to provide a conductor and ground wire feature extraction system for complex environments, including any one of the above-mentioned conductor and ground wire feature extraction methods for complex environments, comprising a data acquisition module, a region selection module, a region calculation module, a region quality screening module, and a feature output module. The data acquisition module is used to acquire remote sensing data from multiple satellites, locate the position of the power transmission tower, and identify the attitude data of the power transmission tower by combining the remote sensing data. The region selection module is used to determine the suspension coordinates of the conductor and ground wire based on the attitude data, combine the suspension coordinates of the conductor and ground wire with the catenary equation to define the region of interest of the conductor and ground wire, and filter the remote sensing data according to the region of interest so that only the location of the region of interest is retained in the remote sensing data. The region calculation module is used to analyze snow cover and icing parameters of the region of interest based on the retained remote sensing data, divide the region of interest into regions based on the snow cover and icing parameters, incorporate the thermal resistance of the icing layer to correct the surface temperature of the conductor and ground wire, and calculate the cooling shrinkage of each region in combination with the conductor and ground wire equipment parameters. The regional quality screening module is used to extract historical remote sensing data and combine it with the conductor and ground wire exposure ratio of the retained remote sensing data to set reference quality values ​​for different regions, and set dynamic reference thresholds to screen high-quality regions that meet the requirements. The feature output module is used to deduce the conductor and ground wire characteristics of the intermediate region between adjacent high-quality regions based on the snow cover parameters, icing parameters and remote sensing data of the high-quality region. Then, the conductor and ground wire characteristics of the intermediate region are selected according to the reference quality value. The feature conflict correction is performed by combining the conductor and ground wire characteristics of all regions, and the correction result is output as the final feature of the conductor and ground wire.

[0033] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for extracting features of conductors and ground wires in complex environments, characterized in that: Includes the following steps: S1. Acquire remote sensing data from multiple satellite sources, locate the position of the power transmission tower, and identify the attitude data of the power transmission tower by combining the remote sensing data; S2. Determine the suspension coordinates of the conductor and ground wire based on the attitude data. Combine the suspension coordinates of the conductor and ground wire with the catenary equation to define the region of interest of the conductor and ground wire. Then, filter the remote sensing data according to the region of interest so that only the location of the region of interest is retained in the remote sensing data. S3. Based on the retained remote sensing data, analyze the snow cover parameters and icing parameters of the region of interest, divide the region of interest into sub-regions based on the snow cover parameters and icing parameters, incorporate the thermal resistance of the icing layer to correct the surface temperature of the conductor and ground wire, and calculate the cooling shrinkage of each region in combination with the conductor and ground wire equipment parameters. S4. Extract historical remote sensing data and combine it with the conductor and ground wire exposure ratio of the remote sensing data retained in S2 to set reference quality values ​​for different regions, and set dynamic reference thresholds to filter high-quality regions that meet the requirements. S5. Based on the snow cover parameters, icing parameters and remote sensing data of the high-quality area, the conductor and ground wire morphology is deduced for the conductor and ground wire characteristics of the intermediate area between adjacent high-quality areas, and then the conductor and ground wire characteristics of the intermediate area are selected according to the reference quality value. S6. Combine the characteristics of conductors and ground wires in all regions to perform feature conflict correction, and output the correction result as the final characteristics of conductors and ground wires.

2. The method for extracting conductor and ground wire features in complex environments according to claim 1, characterized in that: In S1, the remote sensing data from the multi-source satellites includes SAR fully polarimetric data, high-resolution multispectral data, InSAR data, and thermal infrared data. By connecting to the transmission line management terminal, the location data of the transmission towers is extracted from the transmission line management terminal, and then remote sensing data of the same location as the transmission tower location data is extracted. The attitude data of the transmission towers is identified based on the remote sensing data. Attitude data includes three-dimensional coordinates, tower tilt angle and crossarm azimuth angle, and crossarm length.

3. The method for extracting conductor and ground wire features in complex environments according to claim 2, characterized in that: The method of identifying the attitude data of transmission towers based on remote sensing data involves using the YOLOv8 algorithm to detect candidate regions of transmission towers from multispectral data, combining the metal backscattering characteristics of SAR fully polarimetric data to screen real transmission towers, extracting key points of the transmission towers through SAR contours, calculating the angle between the tower's central axis and the vertical direction to obtain the tower's tilt angle, and calculating the crossarm azimuth angle and crossarm length by combining the coordinates of the crossarm endpoints.

4. The method for extracting conductor and ground wire features in complex environments according to claim 1, characterized in that: In S2, the offset of the suspension point relative to the tower base is calculated based on the azimuth angle and length of the crossarm of the transmission tower, and the suspension coordinates are obtained by combining the three-dimensional coordinates of the transmission tower. The region of interest for the conductor and ground wire is defined by combining the suspended coordinates of the conductor and ground wire with the catenary equation. The region of interest is a buffer zone that extends to both sides along the catenary; Remote sensing data are filtered based on the location of the region of interest, retaining only the remote sensing data located at the same location as the region of interest.

5. The method for extracting conductor and ground wire features in complex environments according to claim 1, characterized in that: In step S3, the snow cover parameters and icing parameters of the region of interest are analyzed based on the retained remote sensing data to obtain the snow cover parameters and icing parameters of the region of interest. Set differential snow cover parameters and differential icing parameters, and then combine the snow cover parameters and icing parameters of the region of interest with the differential snow cover parameters and differential icing parameters to divide the region of interest into multiple sub-regions; In this process, the snow accumulation parameters and icing parameters of each node in the region of interest are calculated. If the snow accumulation parameters and icing parameters between adjacent nodes do not exceed the difference snow accumulation parameters and difference icing parameters, the adjacent nodes are merged and then compared with the next node. Conversely, if the difference snow accumulation parameters or difference icing parameters between adjacent nodes exceed the difference snow accumulation parameters or difference icing parameters, it is determined that a sub-region is formed. A fixed length is set for each sub-region based on the length of the region of interest. When the fixed length is reached between adjacent nodes, a sub-region is formed.

6. The method for extracting conductor and ground wire features in complex environments according to claim 5, characterized in that: The snow cover parameters and icing parameters of the region of interest are obtained; The snow cover parameters include snow depth, which is obtained by calculating the Normalized Snow Index (NDSI) from high-resolution multispectral data. The icing parameters include the icing layer thickness, which is inverted by the proportion of surface scattering components in SAR fully polarized data.

7. The method for extracting conductor and ground wire features in complex environments according to claim 1, characterized in that: In step S4, historical remote sensing data of the region of interest is extracted, and ground wire features are extracted from the remote sensing data retained in step S2. The exposed ratio of ground wire is calculated based on the extracted ground wire features. Then, the exposed ratio of ground wire is combined with historical remote sensing data to set reference quality values ​​for each region, thereby obtaining the reference quality values ​​corresponding to each region. Reference thresholds are set based on the total snow cover parameters and ice parameters summarized from all sub-regions; If the reference quality value of a region exceeds the reference threshold, the region is determined to be a high-quality region. If the reference quality value of a sub-region does not exceed the reference threshold, the sub-region is determined to be a non-high-quality region, and the next sub-region is compared.

8. The method for extracting conductor and ground wire features in complex environments according to claim 1, characterized in that: In S5, based on the snow accumulation parameters, icing parameters and the conductor morphology of remote sensing data in the high-quality area, the conductor characteristics of the intermediate area between adjacent high-quality areas are deduced. The conductor morphology consists of snow, ice, and the conductor itself, and includes surface curvature and sag value. The deduction of the conductor and ground wire characteristics in the intermediate region includes geometric characteristic deduction and state characteristic deduction; The conductor characteristics of the intermediate region are selected based on the reference quality value; When the reference quality value of the sub-region in the middle area is greater than 0.7, the derived features are used directly; When the reference quality value of the sub-region in the middle area is greater than 0.5 and less than 0.7, the inferred features are smoothed and corrected by combining the morphological trend of the nearest high-quality region. When the reference quality value of the sub-region in the middle region is less than 0.5, the inferred features are discarded and the average value of the features of the adjacent high-quality regions is used to fill the gap.

9. The method for extracting conductor and ground wire features in complex environments according to claim 1, characterized in that: In S6, the coordinate curves of the conflict area are fitted using the least squares method based on the three-dimensional coordinates of the high-quality area. When there is a sag conflict, the total load of snow and ice is recalculated by combining the catenary equation and the total load of snow and ice. When there is a shrinkage conflict, the surface temperature and shrinkage of the conflict area are recalculated based on the standard value of the thermal expansion and contraction coefficient of the conductor. When there is a snow and ice load conflict, historical load data from the same period are introduced for weighted average correction.

10. A conductor / ground wire feature extraction system for complex environments, used to implement the conductor / ground wire feature extraction method for complex environments as described in any one of claims 1-9, characterized in that: It includes a data acquisition module, a region selection module, a region calculation module, a region quality screening module, and a feature output module; The data acquisition module is used to acquire remote sensing data from multiple satellites, locate the position of the transmission tower, and identify the attitude data of the transmission tower by combining the remote sensing data. The region selection module is used to determine the suspension coordinates of the conductor and ground wire based on the attitude data, combine the suspension coordinates of the conductor and ground wire with the catenary equation to define the region of interest of the conductor and ground wire, and filter the remote sensing data according to the region of interest so that only the location of the region of interest is retained in the remote sensing data. The region calculation module is used to analyze snow cover parameters and icing parameters of the region of interest based on the retained remote sensing data, divide the region of interest into regions based on the snow cover parameters and icing parameters, incorporate the thermal resistance of the icing layer to correct the surface temperature of the conductor and ground wire, and calculate the cooling shrinkage of each region in combination with the conductor and ground wire equipment parameters. The regional quality screening module is used to extract historical remote sensing data and combine it with the conductor and ground wire exposure ratio of the retained remote sensing data to set reference quality values ​​for different regions, and to set dynamic reference thresholds to screen high-quality regions that meet the requirements. The feature output module is used to deduce the conductor and ground wire characteristics of the intermediate region between adjacent high-quality regions based on the snow accumulation parameters, icing parameters and remote sensing data of the high-quality region. Then, the conductor and ground wire characteristics of the intermediate region are selected according to the reference quality value. The feature conflict is corrected by combining the conductor and ground wire characteristics of all regions, and the correction result is output as the final feature of the conductor and ground wire.

Images