Power transmission line wire hidden danger identification method, device, equipment, medium and product

Abstract

The invention relates to a transmission line wire hidden danger identification method and device, computer equipment, a computer readable storage medium and a computer program product. The method comprises the following steps: acquiring section point cloud data of a plurality of target traverse sections, and establishing a cylinder envelope model for each target traverse section; dividing each cylinder envelope model into a plurality of local cylinder sections; calculating the section point cloud data corresponding to each local cylindrical section to obtain a position parameter of a local cylindrical axis; extracting a plurality of section areas along the axes of the local cylinders, and projecting area point cloud data of the section areas to a plane perpendicular to the corresponding axes of the local cylinders to obtain section projection data; and conducting wire hidden danger identification according to the profile projection data. The method can improve the recognition accuracy of the hidden danger problem of the wire in the power transmission line.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and in particular to a method, device, computer equipment, computer-readable storage medium, and computer program product for identifying potential hazards in power transmission line conductors. Background Technology

[0002] With the widespread adoption of airborne and ground-based lidar and UAV photogrammetry technologies, it is possible to acquire large-scale, high-density 3D point clouds of railway corridors in a short time, providing a data foundation for conductor inspection and spatial risk assessment.

[0003] In the conductor extraction stage, relevant technologies employ methods such as threshold-based height or density screening, connectivity-based clustering, and geometric model fitting. Thresholding methods eliminate ground features and features by setting a pre-defined height or density range, then use morphological or clustering algorithms to find possible linear structures. While these methods have achieved some success in controlled scenarios or with good point cloud quality, they still face significant challenges in actual power transmission line engineering environments. Due to the slender nature of the conductors and the significant impact of viewing angle, distance, and occlusion on point cloud acquisition, the point cloud density is sparse and varies significantly across different spans. Furthermore, external interference such as tree branches, bird nests, temporary supports, and adjacent conductors or tower components, which are spatially similar to the conductors, often causes false detections or misclassifications, resulting in low accuracy in identifying potential problems with conductors in power transmission lines. Summary of the Invention

[0004] Therefore, it is necessary to provide a method, device, computer equipment, computer-readable storage medium, and computer program product for identifying potential problems in power transmission line conductors, which can improve the accuracy of identifying potential problems in conductors of power transmission lines.

[0005] Firstly, this application provides a method for identifying potential hazards in transmission line conductors, including:

[0006] Acquire segment point cloud data of multiple target conductor segments, and establish a cylindrical envelope model for each target conductor segment;

[0007] Each of the cylindrical envelope models is divided into multiple local cylindrical segments;

[0008] The position parameters of the local cylindrical axis are obtained by calculating the point cloud data of each local cylindrical segment.

[0009] Multiple profile regions are extracted along each of the local cylindrical axes, and the regional point cloud data of the profile regions are projected onto a plane perpendicular to the corresponding local cylindrical axis to obtain profile projection data.

[0010] The potential hazards of the conductor are identified based on the cross-sectional projection data.

[0011] In one embodiment, acquiring segment point cloud data of multiple target conductor segments includes:

[0012] Acquire the line point cloud data of the target transmission line;

[0013] The locations of multiple conductor suspension points are determined based on the line point cloud data, and the coordinates of the suspension points are determined based on the locations of the conductor suspension points.

[0014] Based on the coordinates of the suspension point, the target transmission line is divided into multiple target conductor segments, and the segment point cloud data of each target conductor segment is determined.

[0015] In one embodiment, dividing each of the cylindrical envelope models into multiple local cylindrical segments includes:

[0016] The direction of the section axis is determined based on the coordinates of the two suspension points adjacent to the cylindrical envelope model.

[0017] The length of the local segment is determined based on the point cloud density of the segment point cloud data;

[0018] Along the axis of the section, the cylindrical envelope model is divided into multiple local cylindrical sections according to the length of the local section.

[0019] In one embodiment, the step of calculating the position parameters of the local cylindrical axis from the point cloud data corresponding to each local cylindrical segment includes:

[0020] For each of the local cylindrical segments, a point cloud covariance matrix is ​​established for the corresponding segment point cloud data;

[0021] Based on the point cloud covariance matrix, the principal component direction vector is determined, and the direction parameter of the local cylindrical axis is obtained.

[0022] Based on the point cloud data of the section, the centroid of the local point cloud is selected as the reference position coordinate of the local cylindrical axis;

[0023] The position parameters of the local cylindrical axis are determined based on the reference position coordinates and the direction parameters.

[0024] In one embodiment, after determining the principal component direction vector based on the point cloud covariance matrix to obtain the direction parameters of the local cylindrical axis, the method further includes:

[0025] For the current local cylindrical axis, calculate the angle between the current local cylindrical axis and the adjacent local cylindrical axis;

[0026] If the included angle exceeds the correction threshold, the direction parameter of the current local cylindrical axis is corrected according to the direction parameter of the adjacent local cylindrical axis.

[0027] If the included angle exceeds the abnormal threshold, an abnormal alert message will be output.

[0028] In one embodiment, the extraction of multiple cross-sectional regions along the axes of each of the local cylinders includes:

[0029] The sampling interval is determined based on the angle between the local cylindrical axis and the adjacent local cylindrical axis.

[0030] Along the axis of each local cylinder, a sampling point is selected as the profile center according to the sampling interval;

[0031] Using the center of the profile as the base point, extract the profile region.

[0032] Secondly, this application also provides a device for identifying potential hazards in transmission line conductors, comprising:

[0033] The data acquisition module is used to acquire segment point cloud data of multiple target traverse segments and to establish a cylindrical envelope model for each target traverse segment.

[0034] The segmentation module is used to divide each of the cylindrical envelope models into multiple local cylindrical segments;

[0035] The parameter calculation module is used to calculate the point cloud data of each local cylindrical segment to obtain the position parameters of the local cylindrical axis.

[0036] The data analysis module is used to extract multiple profile regions along each of the local cylindrical axes, and project the regional point cloud data of the profile regions onto a plane perpendicular to the corresponding local cylindrical axis to obtain profile projection data.

[0037] The hazard identification module is used to identify potential hazards in the conductor based on the profile projection data.

[0038] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0039] Acquire segment point cloud data of multiple target conductor segments, and establish a cylindrical envelope model for each target conductor segment;

[0040] Each of the cylindrical envelope models is divided into multiple local cylindrical segments;

[0041] The position parameters of the local cylindrical axis are obtained by calculating the point cloud data of each local cylindrical segment.

[0042] Multiple profile regions are extracted along each of the local cylindrical axes, and the regional point cloud data of the profile regions are projected onto a plane perpendicular to the corresponding local cylindrical axis to obtain profile projection data.

[0043] The potential hazards of the conductor are identified based on the cross-sectional projection data.

[0044] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps:

[0045] Acquire segment point cloud data of multiple target conductor segments, and establish a cylindrical envelope model for each target conductor segment;

[0046] Each of the cylindrical envelope models is divided into multiple local cylindrical segments;

[0047] The position parameters of the local cylindrical axis are obtained by calculating the point cloud data of each local cylindrical segment.

[0048] Multiple profile regions are extracted along each of the local cylindrical axes, and the regional point cloud data of the profile regions are projected onto a plane perpendicular to the corresponding local cylindrical axis to obtain profile projection data.

[0049] The potential hazards of the conductor are identified based on the cross-sectional projection data.

[0050] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, performs the following steps:

[0051] Acquire segment point cloud data of multiple target conductor segments, and establish a cylindrical envelope model for each target conductor segment;

[0052] Each of the cylindrical envelope models is divided into multiple local cylindrical segments;

[0053] The position parameters of the local cylindrical axis are obtained by calculating the point cloud data of each local cylindrical segment.

[0054] Multiple profile regions are extracted along each of the local cylindrical axes, and the regional point cloud data of the profile regions are projected onto a plane perpendicular to the corresponding local cylindrical axis to obtain profile projection data.

[0055] The potential hazards of the conductor are identified based on the cross-sectional projection data.

[0056] The aforementioned method, apparatus, computer equipment, computer-readable storage medium, and computer program product for identifying potential hazards in transmission line conductors acquire segment point cloud data of multiple target conductor segments and establish a cylindrical envelope model for each target conductor segment. By establishing a cylindrical envelope model for each target conductor segment and using it to limit the candidate point cloud space, a large number of background points and noise unrelated to the conductor in the original 3D point cloud can be removed from the processing area, thereby reducing sources of false detection. Each cylindrical envelope model is then divided into multiple local cylindrical segments, allowing the fitting of each segment to independently adapt to the sag, wind deflection, or local displacement of that segment. The segment point cloud data corresponding to each local cylindrical segment is calculated to obtain the position parameters of the local cylindrical axis. Multiple profile regions are extracted along each local cylindrical axis, and the regional point cloud data of the profile regions are projected onto a plane perpendicular to the corresponding local cylindrical axis to obtain profile projection data. Based on the profile projection data, conductor hazard identification is performed, improving the accuracy of identifying potential hazards in transmission line conductors. Attached Figure Description

[0057] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0058] Figure 1 This is an application environment diagram of the method for identifying potential hazards in transmission line conductors in one embodiment;

[0059] Figure 2 This is a flowchart illustrating a method for identifying potential hazards in transmission line conductors in one embodiment;

[0060] Figure 3 This is a flowchart illustrating step S206 in one embodiment;

[0061] Figure 4 This is a structural block diagram of a power transmission line conductor hazard identification device in one embodiment;

[0062] Figure 5 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0063] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0064] It should be noted that the terms "first," "second," etc., used in this application can be used to describe various elements, but these elements are not limited by these terms. These terms are only used to distinguish the first element from the second element. The terms "comprising" and "having," and any variations thereof, used in this application, are intended to cover non-exclusive inclusion. The term "multiple" used in this application refers to two or more. The term "and / or" used in this application refers to one of the embodiments, or any combination of multiple embodiments.

[0065] The method for identifying potential hazards in transmission line conductors provided in this application can be applied to, for example... Figure 1 In the application environment shown, terminal 102 communicates with server 104 via a network. Terminal 102 can acquire point cloud data and send it to server 104. A data storage system can store the data that server 104 needs to process. The data storage system can be integrated onto server 104 or located on the cloud or other network servers. The data storage system can be used to store line point cloud data and other data. Terminal 102 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, drones, low-altitude aircraft, IoT devices, and portable wearable devices. IoT devices can include smart speakers, smart TVs, smart air conditioners, smart vehicle devices, projection devices, etc. Portable wearable devices can include smartwatches, smart bracelets, head-mounted devices, etc. Head-mounted devices can be virtual reality (VR) devices, augmented reality (AR) devices, smart glasses, etc. Server 104 can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing cloud computing services.

[0066] In one exemplary embodiment, such as Figure 2 As shown, a method for identifying potential hazards in transmission line conductors is provided, and this method is applied to... Figure 1 Taking server 104 as an example, the explanation includes the following steps S202 to S210. Wherein:

[0067] Step S202: Obtain segment point cloud data of multiple target traverse segments, and establish a cylindrical envelope model for each target traverse segment.

[0068] The target conductor section can be a portion of the transmission line defined by two adjacent suspension points (such as insulator suspension points or conductor fixing points on tower crossarms). This section may have a certain sag shape and exhibit an approximately parabolic spatial curve due to the influence of wind load, temperature, and line tension. Section point cloud data refers to a set of spatial data points acquired through lidar, UAV aerial surveying, or other 3D scanning equipment. This point data is used to record the geometric morphology of the conductor surface and surrounding environmental features in 3D space, such as trees, ground, and tower structures. The cylindrical envelope model refers to a virtual cylindrical structure formed by expanding outwards with a preset radius using the conductor's central axis as the geometric reference. It can be used to construct an approximate 3D spatial range for the conductor, covering the conductor point cloud and incorporating potential interference objects into the spatial boundary for subsequent risk and hazard detection.

[0069] For example, server 104 can obtain line point cloud data of the target transmission line; locate multiple conductor suspension points based on the line point cloud data, and determine the suspension point coordinates based on the conductor suspension point positions; divide the target transmission line into multiple target conductor segments based on the suspension point coordinates, and determine the segment point cloud data of each target conductor segment.

[0070] When acquiring data, server 104 can access point cloud data sources, such as LiDAR scanners mounted on drones, laser point clouds collected by overhead line inspection robots, or 3D data generated by periodic scanning of fixed monitoring equipment. Server 104 can then preprocess these raw point clouds, such as removing isolated noise points, unifying coordinate systems, and timestamps, to ensure high data consistency and integrity.

[0071] For example, in a power transmission line that crosses a valley, the conductor suspension points are located on the transmission towers on both sides. The server 104 can identify the specific coordinates of the suspension points based on the point cloud data. For example, the suspension point on the east tower is located at (X1, Y1, Z1), and the suspension point on the west tower is located at (X2, Y2, Z2). The server 104 can then extract the data of the conductor section from the overall point cloud using these two points as boundaries.

[0072] After extracting the point cloud of each target conductor segment, the server 104 can establish a corresponding cylindrical envelope model for it. Specifically, the server 104 can establish the spatial orientation of the conductor segment with the coordinates of the suspension point as the endpoint, and generate a cylindrical geometric skeleton on the initial approximate curve of this orientation. Then, the radius of the cylinder is determined according to the common radius of the conductor and a certain safety margin, for example, using two or three times the actual conductor radius as the model radius, so as to ensure that the point cloud of the conductor itself is completely contained within it and can cover the potential range of foreign object interference.

[0073] Furthermore, when constructing the cylindrical envelope model, the server 104 can dynamically adjust the envelope shape according to the suspension characteristics of the conductor. Since the conductor is not perfectly straight in its natural state, but has sag and lateral sway, the server 104 can perform preliminary fitting with the conductor point cloud to obtain an approximate curved axis, and then generate a dynamically fitted cylindrical envelope along this curve. The model generated in this way can more accurately cover the conductor shape. For example, in mountainous transmission lines, the conductor sag may be significantly asymmetrical due to the influence of terrain and wind deflection. In this case, using a straight envelope is prone to deviation, while the cylinder formed by fitting the curve through the point cloud can well adapt to the actual shape.

[0074] For example, when processing a three-span continuous transmission line, server 104 can first acquire the point cloud data of the entire line, identify the locations of each suspension point through an algorithm, and assuming five suspension points are identified, four target conductor segments are naturally formed. Server 104 can extract the point cloud of each of these four segments using the endpoints as references, and establish a separate cylindrical envelope model for each segment. For example, the first span conductor segment consists of suspension points from tower A to tower B, and the extracted point cloud can contain approximately 100,000 laser points. Server 104 can confine these points within the cylindrical envelope, thereby accurately defining the analysis range of this span conductor in space.

[0075] Step S204: Divide each cylindrical envelope model into multiple local cylindrical segments.

[0076] The cylindrical envelope model itself establishes a spatial range for the entire target conductor segment, and its length can cover the entire conductor span. If processed as a whole, it is easy to encounter the problem of local details being averaged out. Therefore, server 104 can further divide the overall model into multiple local cylindrical segments. A local cylindrical segment refers to a sub-segment cylinder cut from the overall conductor segment at a certain length. Each sub-segment has an independent spatial range and a corresponding point cloud set, facilitating subsequent more refined fitting and profile analysis.

[0077] During the partitioning process, server 104 can first determine the direction of the segment axis. For example, server 104 can determine the segment axis direction based on the coordinates of two suspension points adjacent to the cylindrical envelope model; determine the length of a local segment based on the point cloud density of the segment point cloud data; and divide the cylindrical envelope model into multiple local cylindrical segments along the segment axis direction, based on the length of the local segments. Server 104 can determine the approximate spatial orientation of the segment based on the coordinates of two adjacent suspension points and use it as the central axis of the cylinder. For example, if the coordinates of suspension point A on tower A are (X1, Y1, Z1) and the coordinates of suspension point B on tower B are (X2, Y2, Z2), server 104 can use the line connecting these two points as the initial axial reference, and then further correct the axis direction by combining it with the guide curve fitted from the point cloud, so that the partitioning of the cylinder can conform to the actual guide shape.

[0078] Point cloud density refers to the number of points collected within a unit length of conductor. Its value is affected by the scanning equipment resolution, flight altitude, conductor reflection characteristics, and environmental interference. Areas with high point cloud density indicate rich data detail and potentially more hidden hazards, such as tree branches near the conductor or foreign objects attached. Using excessively long segmentation lengths in these areas would result in too much data in the profile, masking subtle risk features. Conversely, in areas with low point cloud density, overly fine segmentation would lead to sparse local point clouds and unstable fitting. Therefore, server 104 can dynamically adjust the segmentation length: shorter local segments for high-density areas and longer local segments for low-density areas to maintain the balance of point cloud data across sub-segments.

[0079] For example, server 104 can establish the correspondence between point cloud density and local segment length through a preset mapping table. For instance, server 104 can establish a rule: when the number of points per unit length exceeds 1000, the length of each local segment is 1 meter; when the number of points per unit length is between 500 and 1000, the length of each segment is 2 meters; and when the number of points is less than 500, the length of each segment can be increased to more than 3 meters. This mapping relationship ensures that in areas with abundant information, the segmentation is finer to highlight details, while in areas with insufficient information, the segment length can be increased to maintain the continuity and availability of the overall data. Server 104 can also automatically generate mappings based on other algorithms instead of a fixed table, such as using a sliding window to statistically analyze point cloud distribution density and adjusting the segment length of each segment in real time, thereby achieving dynamic adaptive segmentation.

[0080] For example, when processing a traverse span in a mountainous area, the point cloud is very dense near the forest, with 1200 points per meter of traverse length. In this case, server 104 can divide the area into small cylindrical segments of 1 meter each according to the mapping relationship to ensure fine-grained detection of potential tree obstacles. In areas crossing river valleys, the point cloud is sparse, with only about 400 points per meter of traverse length. Server 104 can automatically extend the length of local segments to 3 meters, ensuring that each segment still has a sufficient number of point clouds for stable fitting and analysis.

[0081] Step S206: Calculate the point cloud data corresponding to each local cylindrical segment to obtain the position parameters of the local cylindrical axis.

[0082] For example, server 104 can preprocess the segment point cloud data of each local cylindrical segment to ensure the robustness of subsequent statistical estimation. During the preprocessing stage, server 104 can remove obvious isolated points and try to remove point cloud remnants belonging to towers, ground, or large areas of vegetation. In segments with sparse point clouds or severe noise interference, server 104 can assign confidence weights to points based on neighborhood distance statistics or point intensity information for use in subsequent calculations.

[0083] After preprocessing, server 104 can characterize the distribution features of the point cloud in three orthogonal directions based on the point cloud covariance matrix of the segment point cloud. Server 104 can obtain several eigenvectors and corresponding eigenvalues ​​by performing eigenvalue decomposition on the covariance matrix. The eigenvector corresponding to the largest eigenvalue in the covariance matrix is ​​taken as the principal component direction vector of the segment point cloud, and it is used as the initial direction parameter of the local cylindrical axis. Server 104 can use the centroid of the calculated segment point cloud as the reference position coordinate of the axis, and use distance-weighted average or median position to improve robustness to outliers.

[0084] After obtaining the direction parameters and reference position, the server 104 can output the position parameters of the local cylindrical axis as the combination of the reference position and the direction vector, and at the same time record the point cloud coverage, point cloud density, radial distance distribution statistics and estimated confidence of the segment.

[0085] For example, when faced with complex situations where local point clouds are disturbed by tree branches, bird nests, or parallel wires, server 104 can employ a robust estimation strategy to enhance the accuracy of axis estimation. For example, server 104 can first identify candidate interior points in the segment point cloud and use a robust model fitting method to reduce the impact of outliers on direction estimation. Server 104 can select a set of points that best represent the slender wire structure using methods such as random sampling consensus or weighted principal component analysis and use it to obtain the principal direction. Next, server 104 can estimate the local cylinder radius based on the radial distance distribution from the point to this direction and use the median or truncated mean to avoid extreme value interference. Server 104 can also reduce the weight of radius estimation and expand the neighborhood joint estimation to improve stability.

[0086] Step S208: Extract multiple profile regions along each local cylindrical axis, and project the regional point cloud data of the profile regions onto a plane perpendicular to the corresponding local cylindrical axis to obtain profile projection data.

[0087] A profile region refers to a finite area defined in a plane perpendicular to a local cylindrical axis, centered on a sampling point. This area can be circular or rectangular and is used to extract a local slice from the point cloud to analyze the cross-sectional morphology of the conductor at that location. The regional point cloud data is the collection of all point clouds located within this profile region, reflecting the three-dimensional structural information of that local location.

[0088] For example, server 104 can determine the sampling interval based on the angle between the local cylindrical axis and the adjacent local cylindrical axis; select sampling points as the profile center along each local cylindrical axis according to the sampling interval; and extract the profile region with the profile center as the base point.

[0089] Among them, profile projection data refers to the set of two-dimensional coordinate points obtained by projecting the regional point cloud data along a plane perpendicular to the local cylindrical axis.

[0090] For example, server 104 can traverse each local cylindrical axis and select a series of profile center points on that axis according to a preset sampling interval or an adaptive sampling strategy. Server 104 can comprehensively consider factors such as point cloud density, the angle between the axis and adjacent axes, and the sag characteristics of the conductor to determine the optimal sampling interval. For example, in areas where the conductor undergoes large sag changes or is shifted by wind, the sampling interval can be reduced to extract profiles more densely, ensuring the integrity of local geometric details; while in areas where the conductor tends to be straight and the point cloud density is high, the sampling interval can be appropriately increased to reduce redundant calculations. Server 104 can also construct a tangent plane perpendicular to the local cylindrical axis at each sampling point, and delineate a profile region with a fixed radius or side length within the plane with that point as the geometric center.

[0091] Next, server 104 can extract the point cloud data falling within the profile area from the original point cloud data. Then, server 104 can project this regional point cloud data onto the tangent plane along the direction of the local cylindrical axis, thereby obtaining profile projection data on a two-dimensional plane. This projection data is equivalent to compressing the three-dimensional structure onto a two-dimensional plane, which can intuitively display the cross-sectional distribution of the conductor and any possible foreign objects or interference. Server 104 can also record information such as the spatial location, axial direction, and radius of the profile.

[0092] Furthermore, server 104 can extract features from the distribution of points. Server 104 can detect whether the point cloud within the profile exhibits an approximately circular distribution to verify whether the conductor cross-section conforms to the cylindrical assumption. If the distribution of the projected points in the profile deviates significantly from a circle or exhibits a double clustering pattern, server 104 can mark the profile as an anomaly, indicating the possible presence of multiple entangled conductors, external attachments, bird nests, or interference from tree branches. For example, in sections near forests, the profile projection data may show some discrete point clusters outside the circular distribution of the conductors. Server 104 can determine that these points may originate from tree branch intrusion and will focus on analyzing their distance and coverage in subsequent processes.

[0093] For example, when the included angle between adjacent local cylindrical axes is small, server 104 can use a larger sampling interval (e.g., 1 to 2 meters) to reduce the generation of duplicate profiles; while in areas with larger included angles and significant bending of the conductor, server 104 can automatically reduce the sampling interval (e.g., 0.2 to 0.5 meters) to ensure that the profile closely tracks the curvature changes of the conductor, thereby avoiding the risk of profile skipping. Server 104 can evaluate the actual point cloud coverage of each profile center. If the point cloud of a certain profile area is insufficient to form an effective projection, server 104 can automatically expand the profile radius to the adjacent segments in front and behind or merge adjacent profiles to ensure data integrity.

[0094] For example, server 104 can also set the profile radius according to the physical diameter of the wire. For instance, when the wire diameter is 30 mm, server 104 can set the profile area radius to 50 mm to 100 mm to ensure that the profile coverage area can encompass the main body of the wire and also capture potential interference objects at close range. After obtaining the profile projection data, server 104 can further generate a profile projection image, presenting the two-dimensional projection points in a pixelated form as input for subsequent image recognition algorithms.

[0095] Step S210: Identify potential hazards in the conductor based on the profile projection data.

[0096] Among these, potential hazards to conductors may include, but are not limited to, attachment of external foreign objects (such as bird nests, kites, and plastic bags), vegetation intrusion (such as tree branches entering the safe distance), entanglement of multiple conductors, conductor deformation (ellipticization, wear leading to changes in radius), and point cloud loss caused by obstruction by foreign objects.

[0097] For example, server 104 can fit an ideal circular or elliptical contour based on the distribution of the profile projection data and compare it with the standard radius of the traverse. If the profile contour closely matches the ideal circular contour and the radius is within the expected range, server 104 can determine that the profile area is normal; if the fitting result deviates too much, or the profile exhibits a non-circular, bimodal, or multi-clustered distribution, server 104 can mark it as having potential hazards. Server 104 can calculate multiple statistical features, such as the mean, variance, fitting residuals, and density distribution of profile points.

[0098] Furthermore, the server 104 can also classify the profile into types such as normal profile, profile with attached foreign objects, profile with vegetation invasion, or profile with conductor deformation based on the morphological characteristics and point cloud distribution of the profile projection data, combined with the historical sample library and pattern recognition model. For example, when the profile projection point cloud forms a sparse distribution on the outer ring in addition to the central circular distribution, and these points are distributed in a single direction, the server 104 can determine that it is a potential hazard caused by tree branch invasion; when the profile point cloud forms irregular blocky clusters outside the circular distribution, the server 104 can infer that there may be foreign objects such as bird nests or kites attached; if the profile point cloud is elliptical and the main axis direction continues to deviate, the server 104 can determine that the conductor cross-section has been deformed, and there may be wear or uneven stress.

[0099] In some embodiments, the server 104 can also make a judgment based on the continuity of multiple profiles. The server 104 can check whether there are similar anomalies in the projection data of adjacent profiles. If a certain hazard feature appears in consecutive profiles, the server 104 can confirm the anomaly as a real hazard; if the anomaly only appears in a few profiles and the data of adjacent profiles is normal, the server 104 can mark it as noise or acquisition error, thereby avoiding false alarms. For example, when a conductor is briefly obstructed (such as by a bird flying by), an anomaly may appear in a profile, but this situation will not occur continuously in the preceding and following profiles. Based on this, the server 104 can determine that the anomaly is not a hazard.

[0100] For example, when an anomaly is identified in the cross-section of a conductor, server 104 can generate a hazard label for the profile, along with records of the spatial distribution, degree of anomaly, and category of the abnormal point clusters. For instance, if the profile projection data shows a cluster of tree branches in a certain direction, server 104 can generate a "vegetation intrusion hazard" label and record the difference between the azimuth angle of the cluster and the conductor radius; if a large area of ​​non-circular distribution appears in the profile projection data, server 104 can generate a "conductor deformation hazard" label and record the difference between the actual fitted radius and the standard radius. Server 104 can write these identification results into a hazard database and generate alarm information for maintenance personnel to view in inspection reports.

[0101] In the aforementioned method for identifying potential hazards in transmission line conductors, point cloud data of multiple target conductor segments are acquired, and a cylindrical envelope model is established for each target conductor segment. By establishing a cylindrical envelope model for each target conductor segment and using it to limit the candidate point cloud space, a large number of background points and noise unrelated to the conductor in the original 3D point cloud can be removed from the processing area, thereby reducing the sources of false detection. Each cylindrical envelope model is then divided into multiple local cylindrical segments, allowing the fitting of each segment to independently adapt to the sag, wind deflection, or local displacement of that segment. The point cloud data corresponding to each local cylindrical segment is calculated to obtain the position parameters of the local cylindrical axis. Multiple profile regions are extracted along each local cylindrical axis, and the regional point cloud data of the profile regions are projected onto a plane perpendicular to the corresponding local cylindrical axis to obtain profile projection data. Based on the profile projection data, conductor hazard identification is performed, improving the accuracy of identifying potential problems in transmission line conductors.

[0102] In one exemplary embodiment, such as Figure 3 As shown, step S206 includes steps S302 to S308. Wherein:

[0103] Step S302: For each local cylindrical segment, establish the point cloud covariance matrix for the corresponding segment point cloud data.

[0104] Among them, the point cloud covariance matrix is ​​used to characterize the distribution width and interrelationship of point clouds in three-dimensional space.

[0105] For example, server 104 can perform structured statistical characterization of the point cloud data corresponding to each local cylindrical segment for subsequent orientation and position estimation. Server 104 can spatially register the point cloud from the segment and transform the point cloud to a reference frame based on the local coordinate system of the segment. In the preprocessing stage, server 104 can perform outlier removal and voxel downsampling to obtain an approximately uniform point density, and assign weights to each point based on the point's echo intensity, distance, or confidence information from the laser. Subsequently, server 104 can calculate the distribution characteristics of the point cloud in three orthogonal directions according to the three-dimensional offset relationship of the points relative to the centroid of the segment, and establish a point cloud covariance matrix. Server 104 can use this matrix as a statistical measure describing the shape and unfolding direction of the point cloud in local space, thereby quantifying the variance and interrelationship of the point cloud in each direction. Server 104 can consider weighting strategies when constructing the covariance matrix to reduce the impact of noise points, and use robust statistics (such as median centering or truncation) to replace the simple mean when there are insufficient points or obvious outliers, so as to ensure that the covariance matrix can represent the slender characteristics of the conductor sufficiently stably and reliably.

[0106] Furthermore, when the point cloud within a segment exhibits multiple clusters or contains significant occlusions, server 104 can first perform clustering and segmentation on the segment's point cloud to distinguish the main clusters representing the conductors, and then establish a covariance matrix for the main clusters. Next, server 104 can use a density-based algorithm to identify different point groups. Server 104 can only perform covariance calculation when the number of points within a cluster meets a minimum threshold (e.g., several dozen points); otherwise, server 104 can attempt to merge point clouds with adjacent segments to supplement the sample size. During the merging process, the source of the merge can be recorded, and a marker indicating a decrease in confidence can be added to the subsequent output.

[0107] Step S304: Determine the principal component direction vector based on the point cloud covariance matrix to obtain the direction parameters of the local cylindrical axis.

[0108] Here, the principal component direction vector refers to the unit vector indicating the direction of maximum variation in the point cloud obtained after performing eigenvalue decomposition on the covariance matrix.

[0109] For example, server 104 can identify three principal directions by eigenvalue decomposition of the covariance matrix and use the eigenvectors of the directions with the largest variance as the principal component direction vectors of the segment point cloud. Next, server 104 can regard the principal component direction vectors as an estimate of the axial direction of a conductor in the local segment. Since the conductor is spatially represented as a slender structure stretched along the axial direction, server 104 can process the normalization and sign consistency of the vectors, aligning the direction of the current segment with the direction of the adjacent segments to avoid discontinuities caused by a 180-degree flip in the direction of adjacent segments. At the same time, a direction confidence index is calculated, which can reflect whether the point cloud exhibits a significant linear structure based on the ratio of the principal eigenvalue to the secondary eigenvalue, so that low-confidence directions can be specially processed in subsequent processing.

[0110] For example, after determining the principal component direction vector based on the point cloud covariance matrix and obtaining the direction parameters of the local cylindrical axis, the server 104 can further: calculate the angle between the current local cylindrical axis and the adjacent local cylindrical axes; if the angle exceeds a correction threshold, correct the direction parameters of the current local cylindrical axis based on the direction parameters of the adjacent local cylindrical axes; if the angle exceeds an abnormal threshold, output an abnormal reminder message.

[0111] For example, server 104 can calculate the angle between the current local cylindrical axis and the adjacent local cylindrical axis and use it to quantify the change in direction. When the angle is within the allowable jitter range (e.g., less than about 10 degrees), server 104 can perform weighted smoothing based on the direction of adjacent segments to suppress short-period fluctuations caused by noise or uneven sampling. When the angle exceeds a preset abnormal threshold (e.g., more than about 30 degrees), server 104 can mark the segment as abnormal and output abnormal reminder information including abnormal location, angle size, point cloud density, and suggested review operation.

[0112] Step S306: Based on the segment point cloud data, select the local point cloud centroid as the reference position coordinate of the local cylindrical axis.

[0113] Among them, the centroid of the point cloud refers to the three-dimensional average position of the point cloud in this section and is used as the initial selection for the reference position of the axis; the position parameters of the local cylindrical axis are a set of parameters that can uniquely describe the position and direction of the local axis in three-dimensional space.

[0114] For example, server 104 can calculate the center position in three dimensions of the segment point cloud and use a weighted average or median method to reduce the influence of outliers on the centroid. Subsequently, server 104 can geometrically project the centroid onto the line defined by the direction vector, thereby obtaining a reference point that is both related to the point cloud center and located on the axis. Server 104 can use this projected point as the reference position coordinates of the local cylindrical axis to ensure that the position and direction parameters of the axis are geometrically consistent. When the point cloud samples are sparse or severely offset by external objects, server 104 can fall back to using the centroids of neighboring segments to estimate the reference position of the current segment through interpolation or fitting, and annotate the source and confidence information of the reference position in the output.

[0115] Furthermore, in certain complex environments, server 104 can apply additional rules to the selection of reference positions to avoid misselection. For example, when the centroid is detected to be severely offset by tree branches, large attachments, or parallel wire clusters, the point closest to the direction line can be selected as the reference position, or the interpolation point of the center trajectory calculated by fitting multiple neighboring segments through splines or straight lines at the current segment can be used as the reference position. Server 104 can record the reasons for the substitution and the scope of impact in these alternative strategies, and highlight them in the subsequent visualization and manual review interface for maintenance personnel to make judgments.

[0116] Step S308: Determine the position parameters of the local cylinder axis based on the reference position coordinates and direction parameters.

[0117] For example, server 104 can construct a geometric description of the axis in the form of reference point coordinates and normalized direction vectors, and add several auxiliary attributes to the position parameter, including segment identifier, point cloud coverage, point cloud density, fitting confidence, preliminary value of estimated radius, and timestamp and other meta-information. Server 104 can also record the position parameter to the segment information table through a data structure for subsequent profile extraction, contour fitting and hazard identification modules to call.

[0118] Furthermore, when the confidence level of the position parameters is low or discontinuities are detected, the server 104 can perform curve fitting (e.g., cubic spline fitting) based on multiple position parameters to obtain a more reasonable continuous axis model, and use this global model to update the position and orientation parameters of a single segment. The server 104 can retain the original estimate as a record and write the difference before and after the update and its possible impact (e.g., the offset of the profile interception position) into a report for maintenance decision-making.

[0119] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages in other steps. It is understood that the steps in different embodiments can be freely combined as needed, and all non-contradictory solutions formed by such combinations are within the scope of protection of this application.

[0120] Based on the same inventive concept, this application also provides a transmission line conductor hazard identification device for implementing the above-mentioned method for identifying hidden dangers in transmission line conductors. The solution provided by this device is similar to the solution described in the above method. Therefore, the specific limitations of one or more embodiments of the transmission line conductor hazard identification device provided below can be found in the limitations of the transmission line conductor hazard identification method described above, and will not be repeated here.

[0121] In one exemplary embodiment, such as Figure 4 As shown, a power transmission line conductor hazard identification device is provided, comprising: a data acquisition module 402, a section division module 404, a parameter calculation module 406, a data analysis module 408, and a hazard identification module 410, wherein:

[0122] The data acquisition module 402 is used to acquire segment point cloud data of multiple target traverse segments and to establish a cylindrical envelope model for each target traverse segment.

[0123] The segmentation module 404 is used to divide each cylindrical envelope model into multiple local cylindrical segments;

[0124] The parameter calculation module 406 is used to calculate the point cloud data corresponding to each local cylindrical segment to obtain the position parameters of the local cylindrical axis.

[0125] The data analysis module 408 is used to extract multiple profile regions along each local cylindrical axis and project the regional point cloud data of the profile regions onto a plane perpendicular to the corresponding local cylindrical axis to obtain profile projection data.

[0126] The hazard identification module 410 is used to identify potential hazards in the conductor based on the profile projection data.

[0127] In one embodiment, the data acquisition module 402 is specifically used to: acquire line point cloud data of the target transmission line; locate multiple conductor suspension points based on the line point cloud data, and determine the suspension point coordinates based on the conductor suspension point positions; divide the target transmission line into multiple target conductor segments based on the suspension point coordinates, and determine the segment point cloud data of each target conductor segment.

[0128] In one embodiment, the segment division module 404 is specifically used to: determine the direction of the segment axis based on the coordinates of two suspension points adjacent to the cylindrical envelope model; determine the length of the local segment based on the point cloud density of the segment point cloud data; and divide the cylindrical envelope model into multiple local cylindrical segments along the segment axis direction based on the length of the local segment.

[0129] In one embodiment, the parameter calculation module 406 is specifically used to: establish a point cloud covariance matrix for each local cylindrical segment and the corresponding segment point cloud data; determine the principal component direction vector based on the point cloud covariance matrix to obtain the direction parameter of the local cylindrical axis; select the local point cloud centroid as the reference position coordinate of the local cylindrical axis based on the segment point cloud data; and determine the position parameter of the local cylindrical axis based on the reference position coordinate and the direction parameter.

[0130] In one embodiment, the parameter calculation module 406 is further configured to: calculate the angle between the current local cylindrical axis and the adjacent local cylindrical axis for the current local cylindrical axis; if the angle exceeds a correction threshold, correct the direction parameter of the current local cylindrical axis according to the direction parameter of the adjacent local cylindrical axis; and if the angle exceeds an abnormal threshold, output an abnormal reminder message.

[0131] In one embodiment, the data analysis module 408 is specifically used to: determine the sampling interval based on the angle between the local cylindrical axis and the adjacent local cylindrical axis; select sampling points as the profile center along each local cylindrical axis according to the sampling interval; and extract the profile area with the profile center as the base point.

[0132] Each module in the aforementioned power transmission line conductor hazard identification device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of a computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0133] In one exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 5As shown, the computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media. The database stores data such as line point cloud data. The I / O interfaces are used for information exchange between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a method for identifying potential hazards in power transmission line conductors.

[0134] Those skilled in the art will understand that Figure 5 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0135] In an exemplary embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to perform the following steps: acquiring segment point cloud data of multiple target conductor segments, and establishing a cylindrical envelope model for each target conductor segment; dividing each cylindrical envelope model into multiple local cylindrical segments; calculating the position parameters of the local cylindrical axis based on the segment point cloud data corresponding to each local cylindrical segment; extracting multiple profile regions along each local cylindrical axis, and projecting the regional point cloud data of the profile regions onto a plane perpendicular to the corresponding local cylindrical axis to obtain profile projection data; and identifying potential conductor hazards based on the profile projection data.

[0136] In one embodiment, when the processor executes the computer program, it further performs the following steps: acquiring line point cloud data of the target transmission line; locating multiple conductor suspension points based on the line point cloud data, and determining the suspension point coordinates based on the conductor suspension point locations; dividing the target transmission line into multiple target conductor segments based on the suspension point coordinates, and determining the segment point cloud data of each target conductor segment.

[0137] In one embodiment, when the processor executes the computer program, it further performs the following steps: determining the direction of the segment axis based on the coordinates of two suspension points adjacent to the cylindrical envelope model; determining the length of the local segment based on the point cloud density of the segment point cloud data; and dividing the cylindrical envelope model into multiple local cylindrical segments along the segment axis direction based on the length of the local segment.

[0138] In one embodiment, when the processor executes the computer program, it further implements the following steps: for each local cylindrical segment, for the corresponding segment point cloud data, establish a point cloud covariance matrix; determine the principal component direction vector based on the point cloud covariance matrix to obtain the direction parameter of the local cylindrical axis; select the local point cloud centroid as the reference position coordinate of the local cylindrical axis based on the segment point cloud data; and determine the position parameter of the local cylindrical axis based on the reference position coordinate and the direction parameter.

[0139] In one embodiment, when the processor executes the computer program, it further performs the following steps: for the current local cylindrical axis, calculate the angle between the current local cylindrical axis and the adjacent local cylindrical axis; if the angle exceeds a correction threshold, correct the direction parameter of the current local cylindrical axis according to the direction parameter of the adjacent local cylindrical axis; if the angle exceeds an abnormal threshold, output an abnormal reminder message.

[0140] In one embodiment, when the processor executes the computer program, it further performs the following steps: determining the sampling interval based on the angle between the local cylindrical axis and the adjacent local cylindrical axis; selecting sampling points as profile centers along each local cylindrical axis according to the sampling interval; and extracting the profile region with the profile center as the base point.

[0141] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above method embodiments.

[0142] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0143] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0144] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0145] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0146] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for identifying a conductor hazard of a power transmission line, characterized by, The method comprises: obtaining section point cloud data of a plurality of target conductor sections, and establishing a cylindrical envelope model for each target conductor section; dividing each cylindrical envelope model into a plurality of local cylindrical sections; calculating the section point cloud data corresponding to each local cylindrical section to obtain position parameters of a local cylindrical axis; extracting a plurality of cross-section regions along each local cylindrical axis, and projecting region point cloud data of the cross-section regions onto a plane perpendicular to the corresponding local cylindrical axis to obtain cross-section projection data; conducting conductor hidden danger identification according to the cross-section projection data.

2. The method of claim 1, wherein, The method comprises: obtaining line point cloud data of a target power transmission line; locating a plurality of conductor suspension point positions according to the line point cloud data, and determining suspension point coordinates according to the conductor suspension point positions; dividing the target power transmission line into a plurality of target conductor sections with the suspension point coordinates as a reference, and determining section point cloud data of each target conductor section.

3. The method of claim 2, wherein, The method comprises: determining a section axis direction according to two suspension point coordinates adjacent to the cylindrical envelope model; determining a local section length according to point cloud density of the section point cloud data; dividing the cylindrical envelope model into a plurality of local cylindrical sections according to the local section length along the section axis direction.

4. The method of claim 1, wherein, The method comprises: for each local cylindrical section, establishing a point cloud covariance matrix for the corresponding section point cloud data; determining a principal component direction vector according to the point cloud covariance matrix to obtain direction parameters of a local cylindrical axis; selecting a local point cloud centroid as a reference position coordinate of the local cylindrical axis according to the section point cloud data; determining position parameters of the local cylindrical axis according to the reference position coordinate and the direction parameters.

5. The method of claim 4, wherein, After the step of determining a principal component direction vector according to the point cloud covariance matrix to obtain direction parameters of a local cylindrical axis, the method further comprises: calculating an included angle between a current local cylindrical axis and an adjacent local cylindrical axis for the current local cylindrical axis; in a case where the included angle exceeds a correction threshold, correcting the direction parameters of the current local cylindrical axis according to direction parameters of the adjacent local cylindrical axis; in a case where the included angle exceeds an abnormality threshold, outputting an abnormality reminding information.

6. The method according to any one of claims 1 to 5, characterized in that, The method comprises: determining a sampling interval according to an included angle between the local cylindrical axis and an adjacent local cylindrical axis; selecting a sampling point as a cross-section center according to the sampling interval along each local cylindrical axis; extracting a cross-section region with the cross-section center as a reference point.

7. A power line conductor hazard identification device, characterized by, The device comprises: a data acquisition module configured to obtain section point cloud data of a plurality of target conductor sections, and establish a cylindrical envelope model for each target conductor section; a segment division module, configured to divide each of the cylindrical envelope models into a plurality of local cylindrical segments; a parameter calculation module, configured to calculate the segment point cloud data corresponding to each of the local cylindrical segments to obtain a position parameter of a local cylindrical axis; a data analysis module, configured to extract a plurality of cross-section regions along each of the local cylindrical axes, and project region point cloud data of the cross-section regions onto a plane perpendicular to the corresponding local cylindrical axis to obtain cross-section projection data; a hidden danger identification module, configured to identify a conductor hidden danger according to the cross-section projection data.

8. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that, The processor executes the computer program to implement the steps of the method in any one of claims 1 to 6.

9. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the method in any one of claims 1 to 6.

10. A computer program product comprising a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the method in any one of claims 1 to 6.

Images