Power transmission corridor area construction supervision method and system based on unmanned aerial vehicle group collaborative operation

Abstract

The invention relates to the technical field of power construction supervision, in particular to a power transmission corridor area construction supervision method and system based on cooperative operation of an unmanned aerial vehicle group. According to the method, a supervision area is linearly segmented through the geographic direction of a power transmission line corridor, line segments are divided according to the supervision area, sub-areas are generated, an unmanned aerial vehicle is used for carrying out image collection on a construction area, difference comparison is carried out on formed marks and historical images, changes are recognized through pixel-level comparison, and a change layer containing position and type information is output. A power transmission corridor area construction knowledge graph is constructed, abnormity judgment is carried out on a change area, a rule reasoning mechanism is triggered to judge whether violation happens or not, rectification notifications are automatically generated and pushed, cruise is rechecked after rectification to achieve a closed loop, and meanwhile according to violation recognition density and rectification progress, flight task priority and frequency are dynamically adjusted, and the safety of a power transmission corridor area is improved. Differentiated high-frequency patrol and iterative monitoring of a high-risk area are realized, and the intelligent level and the monitoring efficiency of construction supervision of a strip corridor area are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of power construction supervision technology, and in particular to a method and system for supervising construction in power transmission corridor areas based on collaborative operation of unmanned aerial vehicle (UAV) swarms. Background Technology

[0002] Currently, the supervision of engineering construction areas mainly relies on manual inspections or phased reviews based on remote sensing images, which suffers from problems such as high personnel input, low monitoring accuracy, and delayed response. Especially in power engineering, transmission lines and their towers are usually distributed in a "linear strip" pattern, traversing various terrain environments. The terrain is complex and the distribution is wide, making manual monitoring difficult to cover and inefficient. Traditional remote sensing methods are also difficult to effectively identify and track in real time.

[0003] In recent years, high-resolution imagery-based inspection methods using drones have been widely applied in engineering supervision. While these methods can significantly improve dynamic monitoring capabilities, most existing solutions rely on single-drone flights and lack swarm collaboration mechanisms, resulting in low flight efficiency, uneven task scheduling, and overlapping inspections. Furthermore, the large volume of drone images needs to be uploaded to the cloud for processing, leading to bandwidth constraints and response latency, making real-time violation detection difficult.

[0004] Furthermore, existing image-based violation identification technologies mostly employ rule templates or simple image differencing methods, which can only detect "changes" but cannot understand whether the changed behavior is compliant. They lack the ability to integrate and judge semantic information such as land use, approval history, and construction stage. Knowledge graph technology, as a structured semantic modeling and reasoning technology, has demonstrated powerful capabilities in fields such as intelligent question answering, risk control and compliance, and urban planning in recent years. By constructing a triplet relationship graph between land parcels and their use, approvals, and planning controls, and combining it with image recognition tags, cross-modal decision support can be achieved from image "change detection" to "behavioral compliance reasoning," significantly improving the accuracy and intelligence level of construction violation identification. However, there is currently no mature method to introduce knowledge graphs into drone construction supervision systems, especially in linear areas such as power corridors, where its application remains a blank. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides a method for construction supervision in power transmission corridor areas based on unmanned aerial vehicle (UAV) swarm collaborative operations, comprising the following steps:

[0006] S1. Obtain GIS data of transmission lines and tower coordinate data, construct a construction monitoring buffer zone for the transmission corridor based on the centerline of the transmission line, and dynamically adjust the construction monitoring buffer zone of the transmission corridor using a digital elevation model combined with the topographic data of the transmission line to obtain the construction monitoring area of ​​the transmission corridor.

[0007] S2. Based on tower location coordinate data, the construction monitoring area of ​​the power transmission corridor is divided into multiple construction monitoring sub-areas. With the operating parameters of the UAV as constraints, the UAV monitoring path optimization model of each sub-area is constructed. With the minimum flight time of the UAV and the maximum monitoring coverage of the UAV as objectives, the UAV monitoring path of each sub-area is calculated.

[0008] S3. Based on the drone monitoring paths of each sub-region, and using the drone swarm collaborative patrol mechanism, image data are collected from each construction monitoring sub-region by drone swarms to construct a monitoring image dataset for the power transmission corridor area.

[0009] S4. Based on the monitoring image dataset of the power transmission corridor area, semantic feature annotation and feature recognition are performed using a lightweight deep learning algorithm, and the features are compared with historical image data of the power transmission corridor area to obtain an abnormal feature dataset of the power transmission corridor area.

[0010] S5. Obtain the preset land parcel data and standard operating condition data of the power transmission corridor area, construct the construction knowledge graph of the power transmission corridor area, and perform feature comparison based on the rule reasoning algorithm and the abnormal feature dataset to obtain the abnormal construction dataset of the power transmission corridor area, generate the construction rectification report of the power transmission corridor area and push it to the construction unit for rectification.

[0011] S6. Obtain rectification frequency data, rectification progress data, and construction area sensitivity data from each construction unit, construct a construction violation risk value calculation model, calculate the construction violation risk level for each construction monitoring sub-area, and conduct drone swarm collaborative monitoring of each construction monitoring sub-area based on the construction violation risk level calculation until the construction rectification is completed.

[0012] This invention also provides a construction monitoring system for power transmission corridor areas based on UAV swarm collaborative operation. This system is based on the aforementioned method for construction monitoring of power transmission corridor areas based on UAV swarm collaborative operation. It includes a multi-source data acquisition and fusion module for acquiring GIS data of transmission lines and tower coordinate data, constructing a construction monitoring buffer zone for the power transmission corridor based on the centerline of the transmission line, and dynamically adjusting the construction monitoring buffer zone of the power transmission corridor through a digital elevation model combined with the terrain data of the transmission line to obtain the construction monitoring area of ​​the power transmission corridor.

[0013] The monitoring area division and path generation module is used to divide the construction monitoring area of ​​the power transmission corridor based on tower location coordinate data, resulting in multiple construction monitoring sub-areas. With UAV operating condition parameters as constraints, the module constructs UAV monitoring path optimization models for each sub-area and calculates the UAV monitoring paths for each sub-area with the minimum flight time and maximum monitoring coverage of the UAV as objectives.

[0014] The UAV swarm control and monitoring module is used to collect image data of each construction monitoring sub-area through the UAV swarm based on the UAV monitoring path of each sub-area and a distributed collaborative control algorithm, and to build a monitoring image dataset of the power transmission corridor area.

[0015] The image anomaly data comparison module uses a lightweight deep learning algorithm to perform semantic feature annotation and feature recognition based on the monitoring image dataset of the power transmission corridor area, and compares the features with historical image data of the power transmission corridor area to obtain an anomaly feature dataset of the power transmission corridor area.

[0016] The knowledge graph construction and construction anomaly monitoring module is used to acquire preset land parcel data and standard operating condition data of the power transmission corridor area, construct a construction knowledge graph of the power transmission corridor area, and perform feature comparison based on rule reasoning algorithm and anomaly feature dataset to obtain anomaly construction dataset of the power transmission corridor area, generate construction rectification report of the power transmission corridor area and push it to the construction unit for rectification.

[0017] The construction rectification monitoring module is used to acquire rectification frequency data, rectification progress data, and construction area sensitivity data of each construction unit, construct a construction violation risk value calculation model, calculate the construction violation risk level of each construction monitoring sub-area, and conduct drone swarm collaborative monitoring of each construction monitoring sub-area based on the construction violation risk level calculation until the construction rectification is completed.

[0018] The beneficial effects of this invention are as follows: It linearly segments the regulatory area based on the geographical orientation of the power transmission line corridor, divides the area into segments and generates sub-regions, and supports the collaborative scheduling of drone swarms; it utilizes lightweight computing units on drones to process high-resolution images in real time, creating annotations and performing differential comparisons with historical images, transmitting changed areas and recognition results; it performs pixel-level comparison and identification of changes, outputting a change layer containing location, type, and area information; it constructs a knowledge graph of "land parcel – use – approval – control conditions," locating pixel-level changed areas to specific land parcels, attaching image recognition tags, triggering a rule-based reasoning mechanism to determine whether there is a violation, automatically generating and pushing rectification notices, and completing a closed loop through follow-up patrols after rectification. Simultaneously, based on the density of violation identification and the progress of rectification, it dynamically adjusts the priority and frequency of flight missions, achieving differentiated high-frequency patrols and iterative supervision of high-risk areas. This method and system fully leverage the advantages of drone swarms and knowledge graph reasoning capabilities, realizing a closed-loop management of the entire process of illegal construction from discovery, notification, rectification to review, significantly improving the intelligence level and response efficiency of construction supervision in linear corridor areas. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of a construction supervision method for power transmission corridor areas based on unmanned aerial vehicle (UAV) swarm collaborative operation according to the present invention.

[0020] Figure 2 This is a schematic diagram of the construction supervision system for power transmission corridor areas based on the collaborative operation of unmanned aerial vehicle (UAV) swarms, according to the present invention.

[0021] Figure 3 This is a flowchart illustrating the process of determining the regulatory area according to an embodiment of the present invention.

[0022] Figure 4 This is a flowchart of a drone cluster collaborative patrol process according to an embodiment of the present invention.

[0023] Figure 5 This is a flowchart of real-time aerial survey data processing according to an embodiment of the present invention.

[0024] Figure 6 This is a flowchart of the violation identification and closed-loop management process according to an embodiment of the present invention.

[0025] Figure 7 This is a flowchart of the dynamic patrol strategy optimization according to an embodiment of the present invention. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention; that is, the described embodiments are only a part of the embodiments of the invention, and not all of them. The components of the embodiments of the invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0027] Therefore, all other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention. It should be noted that relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations.

[0028] Moreover, the term "comprising" or any other variation thereof is intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0029] Example 1:

[0030] This invention provides a method for construction supervision in power transmission corridor areas based on unmanned aerial vehicle (UAV) swarm collaborative operations, comprising the following steps:

[0031] S1. Obtain GIS data of transmission lines and tower coordinate data, construct a construction monitoring buffer zone for the transmission corridor based on the centerline of the transmission line, and dynamically adjust the construction monitoring buffer zone of the transmission corridor using a digital elevation model combined with the topographic data of the transmission line to obtain the construction monitoring area of ​​the transmission corridor.

[0032] S2. Based on tower location coordinate data, the construction monitoring area of ​​the power transmission corridor is divided into multiple construction monitoring sub-areas. With the operating parameters of the UAV as constraints, the UAV monitoring path optimization model of each sub-area is constructed. With the minimum flight time of the UAV and the maximum monitoring coverage of the UAV as objectives, the UAV monitoring path of each sub-area is calculated.

[0033] S3. Based on the drone monitoring paths of each sub-region, and using the drone swarm collaborative patrol mechanism, image data are collected from each construction monitoring sub-region by drone swarms to construct a monitoring image dataset for the power transmission corridor area.

[0034] S4. Based on the monitoring image dataset of the power transmission corridor area, semantic feature annotation and feature recognition are performed using a lightweight deep learning algorithm, and the features are compared with historical image data of the power transmission corridor area to obtain an abnormal feature dataset of the power transmission corridor area.

[0035] S5. Obtain the preset land parcel data and standard operating condition data of the power transmission corridor area, construct the construction knowledge graph of the power transmission corridor area, and perform feature comparison based on the rule reasoning algorithm and the abnormal feature dataset to obtain the abnormal construction dataset of the power transmission corridor area, generate the construction rectification report of the power transmission corridor area and push it to the construction unit for rectification;

[0036] S6. Obtain rectification frequency data, rectification progress data, and construction area sensitivity data from each construction unit, construct a construction risk value calculation model, calculate the construction risk level for each construction monitoring sub-area, and conduct drone swarm collaborative monitoring of each construction monitoring sub-area based on the construction risk level calculation until the construction rectification is completed.

[0037] Furthermore, step S1 includes the following sub-steps:

[0038] S101. Obtain GIS data of transmission lines and tower coordinate data, and perform three-dimensional spatial registration in the GIS platform to construct a three-dimensional spatial scene that includes the geometric features of transmission lines, the terrain features of transmission lines, and the features of ground objects along transmission lines;

[0039] S102. Based on the geometric feature data of the transmission line, taking the center line of the transmission line as a reference, and based on the voltage level data of the transmission line, calculate the basic buffer width of the monitoring area to obtain the boundary of the first monitoring corridor;

[0040] S103. Based on the topographic feature data of the transmission line and the pre-set topographic safety data of the transmission line, the boundary of the first monitoring corridor is dynamically adjusted through the digital elevation model to obtain the boundary of the second monitoring corridor.

[0041] S104. Based on the ground feature data of the transmission line, and based on the pre-set ground feature safety distance data of the transmission line, the boundary of the second monitoring corridor is corrected for safety distance avoidance in combination with the ground feature outline. The corrected corridor boundary range is then merged and spliced ​​with the boundary of the first monitoring corridor using a spatial Boolean algorithm to obtain the construction monitoring area of ​​the transmission corridor.

[0042] Specifically, the implementation principle and flow of each sub-step in the above embodiments are as follows:

[0043] like Figure 3 First, input GIS data of the transmission line, including transmission line coordinates, tower locations, voltage levels, and design safety distance parameters. Combined with DEM terrain data and feature vector data, spatial registration of multi-source data is completed to construct a three-dimensional spatial scene including line geometry, terrain undulations, and surrounding features. Then, using the line centerline as a reference, an initial buffer width is set according to the voltage level to generate a basic buffer zone. Based on the slope and aspect calculated by the DEM, complex terrain is corrected. When the slope is greater than 25°, the buffer zone is extended by 5-10 meters on the uphill side and reduced by 2-5 meters on the downhill side. When crossing rivers or canyons, the buffer zone is extended by 1.5 times the tower height on both sides along the crossing point. Next, features such as buildings and trees that are less than the safety distance from the line are avoided by extending the safety distance outward, and Boolean operations are performed with the basic buffer zone to eliminate monitoring blind spots. Finally, monitoring corridor vector polygon data with attributes such as line section ID, voltage level, and monitoring accuracy requirements are output to provide accurate spatial range basis for subsequent inspection tasks.

[0044] Specifically, a Geographic Information System (GIS) is a crucial spatial information system. Supported by computer hardware and software, it's a technological system for collecting, storing, managing, processing, analyzing, displaying, and describing geographic distribution data across the entire or part of the Earth's surface (including the atmosphere). A Digital Elevation Model (DEM) is a digital simulation of ground topography (i.e., a digital representation of the terrain's surface morphology) achieved through limited terrain elevation data. It's a physical ground model that represents ground elevation using an ordered array of numerical values.

[0045] Furthermore, step S2 includes the following sub-steps:

[0046] S201. Based on tower location coordinate data, the construction monitoring area of ​​the power transmission corridor is divided into multiple construction monitoring sub-areas;

[0047] S202. Acquire drone endurance time data, flight speed data, camera resolution data, and field of view data, and establish a drone performance parameter matrix;

[0048] S203. Using the boundary of the power transmission corridor construction monitoring area as a benchmark, establish the UAV obstacle avoidance parameter matrix based on the UAV performance parameter matrix;

[0049] S204. Based on the graph model construction algorithm, each construction monitoring sub-region is discretized into a node network, and the UAV performance parameter matrix is ​​used as the edge weight and the UAV obstacle avoidance parameter matrix is ​​used as the node value to construct the UAV monitoring path optimization model corresponding to the node network of each construction monitoring sub-region.

[0050] S205. Based on the UAV monitoring path optimization model corresponding to each node network, with the minimum UAV flight time and the maximum UAV monitoring coverage as objectives, the Dijkstra algorithm or A... * The algorithm calculates the drone monitoring paths for each construction monitoring sub-area.

[0051] Specifically, the implementation principle and flow of each sub-step in the above embodiments are as follows:

[0052] like Figure 4 First, the power line is divided into N inter-tower sections based on the location of the transmission towers, and the coordinate sequence of the centerline of each section is extracted as the basic path skeleton. Then, a path planning constraint model is established, combining performance parameters such as the maximum endurance of the UAV, the maximum flight speed of the UAV, and the resolution and maximum field of view of the UAV camera. Based on the minimum shooting resolution GSD ≤ 10cm, the optimal flight altitude is calculated using the formula H = focal length × ground distance / sensor size to meet the monitoring accuracy requirements, and the altitude is reduced by 10% within the monitoring corridor boundary. A safe flight path is formed, and obstacles such as trees and buildings are transformed into polygonal no-fly zones. Next, the inter-tower section is discretized into a node network with coordinate and flight altitude constraints. A path optimization model is constructed using flight time or energy consumption that takes into account the impact of wind speed as edge weights. The Dijkstra algorithm is used to generate the optimal inspection path, which includes reciprocating flight paths and hovering shooting points, with the goal of "full coverage monitoring + shortest flight time". Finally, the three-dimensional spatial path coordinate sequence (X,Y,Z) of each inter-tower section and the corresponding control parameters such as flight speed and shooting trigger point are output to provide precise execution instructions for UAV flight control.

[0053] Furthermore, the drone swarm mentioned in step S3 includes multiple drone sub-swarms, each consisting of a master drone and slave drones, and each corresponding to a construction monitoring sub-area. The master drone and slave drones transmit drone operating condition parameters in real time through a Mesh self-organizing network.

[0054] Furthermore, the UAV swarm collaborative patrol mechanism includes a multi-UAV swarm scheduling sub-mechanism and an adaptive patrol and obstacle avoidance sub-mechanism. The multi-UAV swarm scheduling sub-mechanism divides each construction monitoring sub-area into multiple task groups based on the UAV performance parameter matrix, using a block parallel and dynamic load balancing algorithm. With the master UAV as the central node and the slave UAVs as neighboring nodes, the operating parameters of the slave UAVs are transmitted to the master UAV in real time through a Mesh self-organizing network.

[0055] Furthermore, the adaptive patrol and obstacle avoidance sub-mechanism, based on the UAV obstacle avoidance parameter matrix and digital elevation model, uses a PID algorithm to control the flight attitude of the master UAV and master-slave UAV, uses a Venn diagram algorithm to generate local obstacle avoidance paths to control ground objects for the master UAV and master-slave UAV, uses GPS and ultra-wideband positioning technology to monitor the relative positions of the master UAV and slave UAV in real time, and the master UAV publishes a unified timestamp, which the slave UAV uses to perform obstacle avoidance control.

[0056] Specifically, the sub-steps and implementation principles of step S3 in the above embodiments are as follows:

[0057] like Figure 4First, task units are divided according to the monitoring corridor sections. Considering the number of drones and their endurance, a "blocked parallelism + dynamic load balancing" strategy is adopted to allocate N inter-tower sections to M task groups (each group has 5-10 sections, with a time difference ≤15%). The master drone is responsible for global scheduling, and slave drones synchronize status data in real time through a mesh self-organizing network to achieve cluster collaboration. For straight sections exceeding 2km in length, virtual nodes are set at the midpoint of the towers to facilitate side-by-side shooting. Then, during flight, parameters are adaptively adjusted based on weather conditions and terrain features. When the wind speed is greater than 10m / s, the flight speed is reduced to Vmax×80% and the altitude is increased by 5%. In foggy weather, the shooting interval is shortened to improve image overlap. DEM data and PID are also utilized. The algorithm climbs ahead of time when the altitude rises by more than 10 meters in a short distance, maintaining a relative height difference of ≤2 meters. Then, it uses a 3D grid map and Voronoi diagram algorithm to generate a local obstacle avoidance path that maintains a safe distance of ≥2 meters from obstacles. In a multi-machine environment, it achieves collaborative obstacle avoidance through obstacle coordinate broadcasting and priority sorting. Finally, it ensures safety by setting the horizontal distance between multiple machines to ≥50 meters and the vertical height difference to ≥10 meters, and by combining GPS and ultra-wideband positioning technology to monitor and trigger the speed compensation algorithm in real time. At the same time, in cross-segment tasks, it adopts a "main path synchronization + sub-path independence" mode, and ensures consistent image acquisition perspective by calibrating the position with a unified timestamp and synchronization point, thereby improving the accuracy of multi-machine collaboration.

[0058] Specifically, Mesh self-organizing networks are a special type of mesh networking that emphasizes the autonomous cooperation and self-organization capabilities among nodes. In a Mesh self-organizing network, nodes communicate with each other wirelessly and automatically adjust their routing paths based on the network topology and communication needs. This self-organizing characteristic makes Mesh self-organizing networks more adaptable and robust in complex environments. A Voronoi diagram, also known as a Thiessen polygon, is a set of continuous polygons composed of the perpendicular bisectors of line segments connecting two adjacent points.

[0059] Furthermore, step S4 includes the following sub-steps:

[0060] S401. Based on the monitoring image dataset of the power transmission corridor area, the construction target features are identified using a power construction-specific model improved based on YOLOv8n, and the construction target feature identification results are obtained.

[0061] S402. Based on the results of construction target feature identification, obtain historical image data of the power transmission corridor area, establish a pyramid hierarchical index, and obtain anomaly feature dataset of the power transmission corridor area through SIFT feature point matching.

[0062] Specifically, the implementation principle and flow of each sub-step in the above embodiments are as follows:

[0063] like Figure 5First, the Zhang Zhengyou calibration method is used to correct the distortion of the fisheye lens, ensuring that the residual distortion is <0.5%. This is combined with nonlocal mean filtering to remove Gaussian noise, improving the signal-to-noise ratio by ≥15dB and enhancing image quality. Visual SLAM positioning uses the ORB-SLAM3 algorithm to construct a sparse point cloud map in real time, achieving high-precision positioning with an accuracy of ≤0.3m in plains and ≤0.5m in mountainous areas. Multi-source data fusion uses extended Kalman filtering (EKF) to fuse GPS / IMU data, correcting positioning errors and outputting a geographic reference image with UTM coordinates (resolution 0.1m / pixel), providing a geographic coordinate benchmark for subsequent analysis.A lightweight model (parameters ≤ 10MB) specifically for power construction, based on an improved YOLOv8n architecture, is used to detect nine types of targets, including construction machinery and temporary buildings, in real-time from monitoring images of power transmission corridors. A five-tuple (category, confidence score, geographic coordinates, pixel coordinates, size) is generated for each detected target. Protocol Buffers compression is used to improve transmission efficiency by 40% and reduce data backhaul pressure. A historical image library containing RGB / multispectral images of the same area from the previous three months (interval 1 day / image) is constructed with a pyramid-level index to support fast multi-scale matching, providing historical data support for change detection. Subpixel-level alignment is achieved through SIFT (Scale-invariant feature transform) feature point matching (combined with RANSAC (Random sample consensus) to eliminate mismatches, achieving ≥20 matching pairs). Change detection is performed using SSIM (Structure Similarity Index Measure) (threshold < 0.8) and MSE (Mean Square Error) (threshold > 15), and NDVI (Normalized Difference Vegetation) is overlaid. The index excludes seasonal vegetation changes, uploading only ROIs with a compression ratio ≥10:1 and accompanied by identification tags and precise timestamps. Next, thin plate spline (TPS) interpolation is used for geometric correction of the images, and ground targets are used to periodically calibrate control points to ensure an error <0.2 pixels. Simultaneously, histogram matching (R / G / B band correlation coefficient >0.95) is performed on images from different time periods to achieve radiometric normalization and ensure image registration accuracy. Subsequently, multi-level change detection is conducted. First, regions with abrupt brightness changes (ΔL >20 and area >5m²) are extracted from the difference image for coarse detection. Then, LBP local binary mode is used to perform texture analysis on the coarsely detected regions to identify artificial structure features (straight edge density >10 lines / m²). Natural changes are masked using land use classification data, retaining only the change areas of built-up land and unused land. Finally, a GeoTIFF format change layer and attribute table are generated, recording the change type, geographic coordinates, pixel coordinates, area, change confidence level, and associated historical image ID. Information such as suggested verification levels is provided to ensure that the area detection error in plain areas is ≤5% and the position deviation is ≤0.5m, while the position deviation in mountainous areas is ≤1m, in order to meet the accuracy requirements for power transmission corridor change detection.

[0064] Furthermore, step S5 includes the following sub-steps:

[0065] S501. Obtain the preset land parcel data and construction condition data of the power transmission corridor area, and establish a multi-entity association network of land parcel data entities and construction condition data entities to construct a construction knowledge graph of the power transmission corridor area.

[0066] S502. Based on the SWRL (Semantic Web Rule Language) rule reasoning algorithm, according to the construction knowledge graph of the power transmission corridor area, the abnormal feature dataset is compared for compliance. Abnormal feature data that does not meet the requirements of the construction knowledge graph of the power transmission corridor area are regarded as abnormal construction data, and an abnormal construction dataset is obtained.

[0067] S503. Based on the spatial overlay analysis algorithm, the abnormal construction dataset is compared with the boundary of each construction monitoring sub-region. When the intersection area is greater than or equal to 50%, the current abnormal construction data is associated with the current construction monitoring sub-region, and a construction rectification report for the power transmission corridor area is generated.

[0068] S504. The obtained construction rectification report for the power transmission corridor area shall be sent to the corresponding construction unit for rectification.

[0069] Specifically, the implementation principle and flow of each sub-step in the above embodiments are as follows:

[0070] like Figure 6 First, the preset land parcel data and construction status data of the power transmission corridor area are obtained. The land parcel data of the power transmission corridor area includes land parcel number data, land parcel boundary coordinate data and land parcel use category data. The construction status data includes construction permit data, basic farmland protection line data, ecological protection line data, construction type data and construction duration data.

[0071] Define the following data as entity types in the graph structure: land parcel number data, land parcel boundary coordinate data, land parcel use category data, construction permit data, basic farmland protection line data, ecological protection line data, construction type data, and construction duration data. Define the following relationship types in the graph structure network: land parcel-use relationship, land parcel-construction permit relationship, land parcel-protected area relationship, land parcel-construction type relationship, and land parcel-construction duration relationship. Establish a multi-entity association network as a construction knowledge graph for the power transmission corridor area.

[0072] Then, based on the SWRL (Semantic Web Rule Language) rule reasoning algorithm as the reasoning language, rules are defined. According to the construction knowledge graph of the power transmission corridor area, feature compliance comparison is performed on the abnormal feature dataset. Abnormal feature data that does not meet the requirements of the construction knowledge graph of the power transmission corridor area is identified as abnormal construction data, resulting in an abnormal construction dataset. For example, if the plot entity is basic farmland and the construction type entity is concrete hardening, then the current plot is defined as violating the land use relationship; if the construction permit entity is an expired construction permit and the plot boundary coordinate entity overlaps with the basic farmland protection line entity or the ecological protection line entity, then the current plot is defined as violating the land protection zone relationship. Specifically, during reasoning, the matching time for a single rule reasoning is less than 100 milliseconds, and at least 1000 abnormal feature image data can be processed in a single rule reasoning iteration, meeting the requirements of real-time reasoning.

[0073] Next, based on the spatial overlay analysis algorithm, the abnormal construction dataset is spatially overlaid with each construction monitoring sub-region. When the abnormal image data in the abnormal construction dataset is greater than or equal to 50% of the image data in each construction monitoring sub-region, the current abnormal construction data is associated with the current construction monitoring sub-region, located as an illegal plot, and a construction rectification report for the power transmission corridor area is generated. The report includes the location of the construction violation, before and after comparison images, the type of construction violation relationship, the construction rectification period, and the responsible construction unit, so as to achieve accurate push of violation information.

[0074] According to the construction rectification report, the construction unit rectified the illegal construction behavior in accordance with the rectification report. After the responsible unit for the review task uploaded the rectification certificate, the unit dispatched drones to conduct key inspections of the illegal area (the flight altitude was reduced to 50m and the resolution was increased to 0.05m / pixel) to obtain high-definition review images.

[0075] The effectiveness of the rectification was verified using the DeepLabv3+ model for image semantic segmentation. The criteria for judgment were that the area of ​​artificial features was reduced by ≥90% and there were no new construction equipment. If the verification was successful, the loop was marked as closed; otherwise, it was automatically upgraded to manual verification to ensure that the rectification was implemented in place.

[0076] Furthermore, step S6 includes the following sub-steps:

[0077] S601. Obtain rectification frequency data and rectification progress data from each construction unit, and construct a construction risk value calculation model, wherein the construction risk value calculation model is expressed as follows:

[0078] ;

[0079] Where R represents the construction risk value, D represents the current rectification frequency, L represents the current rectification progress, and S represents the sensitivity of the construction area, which are weight parameters respectively;

[0080] S602. Based on the construction risk value calculation model, the construction risk level of each construction monitoring sub-area is calculated. When the construction risk value is greater than 0.8, it is a high-risk construction area; when the construction risk value is greater than 0.5 and less than or equal to 0.8, it is a medium-risk construction area; when the construction risk value is less than or equal to 0.5, it is a low-risk construction area.

[0081] S603. Based on the construction risk level calculation results, and based on the UAV monitoring paths of each sub-region, repeat monitoring is performed on each construction monitoring sub-region, and the construction knowledge graph of the power transmission corridor area is updated based on the repeat monitoring results.

[0082] S604. Based on the repeated monitoring results, through reinforcement learning algorithms, for construction areas where the same abnormal construction feature data appears at least twice, construct a drone swarm monitoring path based on the Archimedes spiral and conduct drone swarm monitoring until the construction rectification is completed.

[0083] Specifically, the implementation principle and flow of each sub-step in the above embodiments are as follows:

[0084] like Figure 7 ,

[0085] First, construct a risk value calculation model. Where R represents the construction risk value, D represents the current rectification frequency, L represents the current rectification progress, and S represents the sensitivity of the construction area. Taking into account the current violation density, rectification progress delay rate, and the number of nearby sensitive targets, a risk assessment model is established to quantify the regulatory risks of different areas. When the construction risk value is greater than 0.8, it is a high-risk construction area; when the construction risk value is greater than 0.5 and less than or equal to 0.8, it is a medium-risk construction area; and when the construction risk value is less than or equal to 0.5, it is a low-risk construction area.

[0086] Then, differentiated scheduling strategies are formulated based on the risk level classification. High-risk areas are patrolled 3 times a day with the flight altitude reduced to 80m and the camera frame rate increased to 30fps. Medium-risk areas are patrolled twice a day with multispectral detection enabled. Low-risk areas are maintained with one routine patrol per day and resources are dynamically allocated to newly connected high-risk areas to achieve intelligent patrol scheduling.

[0087] Next, the violation identification results are regularly fed back to the target detection model. The training dataset is updated quarterly and typical violation samples are added to improve the model's ability to identify hidden construction such as underground foundation excavation, thereby achieving iterative optimization of the identification model.

[0088] Subsequently, a spiral-shaped encrypted drone swarm monitoring path based on the Archimedes spiral is generated for areas with repeated violations using a reinforcement learning algorithm (Proximal Policy Optimization, PPO), and the flight path is dynamically optimized to improve patrol coverage and accuracy.

[0089] Finally, regularly collect statistics on indicators such as the recurrence rate of violations and the timeliness of rectification in each region, input the results into the intelligent scheduling system, and adjust the parameters of the risk assessment model through dynamic weighting to ensure that the inspection strategy continuously meets the actual regulatory needs until the construction and rectification are completed.

[0090] Example 2

[0091] like Figure 2 As shown, Embodiment 2 of the present invention also provides a construction supervision system for power transmission corridor areas based on UAV swarm collaborative operation. This system is based on a construction supervision method for power transmission corridor areas based on UAV swarm collaborative operation in Embodiment 1. It includes a multi-source data acquisition and fusion module for acquiring GIS data of transmission lines and tower coordinate data, constructing a construction monitoring buffer zone for the power transmission corridor based on the centerline of the transmission line, and dynamically adjusting the construction monitoring buffer zone for the power transmission corridor through a digital elevation model combined with the terrain data of the transmission line to obtain the construction monitoring area of ​​the power transmission corridor.

[0092] The monitoring area division and path generation module is used to divide the construction monitoring area of ​​the power transmission corridor based on tower location coordinate data, resulting in multiple construction monitoring sub-areas. With UAV operating condition parameters as constraints, the module constructs UAV monitoring path optimization models for each sub-area and calculates the UAV monitoring paths for each sub-area with the minimum flight time and maximum monitoring coverage of the UAV as objectives.

[0093] The UAV swarm control and monitoring module is used to collect image data of each construction monitoring sub-area through the UAV swarm based on the UAV monitoring path of each sub-area and a distributed collaborative control algorithm, and to build a monitoring image dataset of the power transmission corridor area.

[0094] The image anomaly data comparison module uses a lightweight deep learning algorithm to perform semantic feature annotation and feature recognition based on the monitoring image dataset of the power transmission corridor area, and compares the features with historical image data of the power transmission corridor area to obtain an anomaly feature dataset of the power transmission corridor area.

[0095] The knowledge graph construction and construction anomaly monitoring module is used to acquire preset land parcel data and standard operating condition data of the power transmission corridor area, construct a construction knowledge graph of the power transmission corridor area, and perform feature comparison based on rule reasoning algorithm and anomaly feature dataset to obtain anomaly construction dataset of the power transmission corridor area, generate construction rectification report of the power transmission corridor area and push it to the construction unit for rectification.

[0096] The construction rectification monitoring module is used to acquire rectification frequency data, rectification progress data, and construction area sensitivity data of each construction unit, construct a construction risk value calculation model, calculate the construction risk level of each construction monitoring sub-area, and conduct drone swarm collaborative monitoring of each construction monitoring sub-area based on the construction risk level calculation until the construction rectification is completed.

[0097] Specifically, the workflow is as follows:

[0098] First, the system acquires GIS data of transmission lines and tower coordinate data through a multi-source data acquisition and fusion module, and constructs a construction monitoring buffer zone for the transmission corridor based on the centerline of the transmission line. Combining the digital elevation model (DEM) with the transmission line topographic data, the monitoring buffer zone is dynamically adjusted, ultimately generating a construction monitoring area for the transmission corridor that covers the actual terrain conditions.

[0099] Then, the monitoring area division and path generation module divides the power transmission corridor construction monitoring area into blocks, dividing it into multiple construction monitoring sub-areas according to the tower coordinate data. Combining the UAV operating condition parameters (such as maximum flight time, maximum monitoring range, flight speed, etc.), a UAV monitoring path optimization model is established, and the optimal UAV monitoring path for each construction monitoring sub-area is calculated with "minimum flight time + maximum monitoring coverage" as the objective function.

[0100] Next, the UAV swarm control and monitoring module, based on the optimal monitoring path for each sub-region, employs a distributed collaborative control algorithm to schedule the UAV swarm, enabling efficient inspection and image acquisition tasks. As the UAVs cover all sub-regions, they collect construction image data in real time and aggregate it to generate a monitoring image dataset for the power transmission corridor area.

[0101] Subsequently, the image anomaly data comparison module performs lightweight deep learning processing on the collected monitoring image dataset to achieve semantic feature annotation and target feature recognition. The extracted monitoring features are then compared with historical image data of the transmission corridor area to automatically generate an anomaly feature dataset for the transmission corridor area, providing a basis for subsequent construction compliance assessments.

[0102] Based on this, the knowledge graph construction and construction anomaly monitoring module acquires preset site data and standard operating condition data to establish a construction knowledge graph for the power transmission corridor. The system, based on a rule-based reasoning algorithm, compares the anomaly feature dataset obtained in the previous stage with the knowledge graph, identifies abnormal construction behaviors, generates a construction anomaly dataset for the power transmission corridor area, and automatically generates a construction rectification report, which is then pushed to the responsible construction unit to initiate the rectification process.

[0103] Finally, the construction rectification monitoring module dynamically tracks the rectification status of each construction unit, collects rectification frequency data, rectification progress data, and construction area sensitivity data, establishes a construction risk value calculation model, and conducts risk level assessments for each construction monitoring sub-area. Based on the assessment results, the monitoring frequency and flight path planning of the UAV swarm are dynamically adjusted, and intensified patrols are conducted in key risk areas until the rectification of construction violations is completed, thereby achieving closed-loop management throughout the entire process.

[0104] This invention is described from the perspectives of its intended use, effectiveness, progress, and novelty. Its practical and progressive features meet the functional enhancement and use requirements emphasized by the Patent Law. The above description and drawings are merely preferred embodiments of this application and are not intended to limit this application. Therefore, all structures, devices, features, etc., that are similar to or identical to those of this application, i.e., all equivalent substitutions or modifications made in accordance with the scope of this patent application, shall fall within the scope of protection of this patent application.

[0105] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for construction supervision in power transmission corridor areas based on unmanned aerial vehicle (UAV) swarm collaborative operations, characterized in that, Includes the following steps: S1. Obtain GIS data of transmission lines and tower coordinate data, construct a construction monitoring buffer zone for the transmission corridor based on the centerline of the transmission line, and dynamically adjust the construction monitoring buffer zone of the transmission corridor using a digital elevation model combined with the topographic data of the transmission line to obtain the construction monitoring area of ​​the transmission corridor. S2. Based on tower location coordinate data, the construction monitoring area of ​​the power transmission corridor is divided into multiple construction monitoring sub-areas. With the operating parameters of the UAV as constraints, the UAV monitoring path optimization model of each sub-area is constructed. With the minimum flight time of the UAV and the maximum monitoring coverage of the UAV as objectives, the UAV monitoring path of each sub-area is calculated. S3. Based on the drone monitoring paths of each sub-region, and using the drone swarm collaborative patrol mechanism, image data are collected from each construction monitoring sub-region by drone swarms to construct a monitoring image dataset for the power transmission corridor area. S4. Based on the monitoring image dataset of the power transmission corridor area, semantic feature annotation and feature recognition are performed using a lightweight deep learning algorithm, and the features are compared with historical image data of the power transmission corridor area to obtain an abnormal feature dataset of the power transmission corridor area. S5. Obtain the preset land parcel data and standard operating condition data of the power transmission corridor area, construct the construction knowledge graph of the power transmission corridor area, and perform feature comparison based on the rule reasoning algorithm and the abnormal feature dataset to obtain the abnormal construction dataset of the power transmission corridor area, generate the construction rectification report of the power transmission corridor area and push it to the construction unit for rectification. S6. Obtain rectification frequency data, rectification progress data, and construction area sensitivity data from each construction unit, construct a construction violation risk value calculation model, calculate the construction violation risk level for each construction monitoring sub-area, and conduct drone swarm collaborative monitoring of each construction monitoring sub-area based on the construction violation risk level calculation until the construction rectification is completed.

2. The method for construction supervision in power transmission corridor areas based on unmanned aerial vehicle (UAV) swarm collaborative operation according to claim 1, characterized in that, Step S1 includes the following steps: S101. Obtain GIS data of transmission lines and tower coordinate data, and perform three-dimensional spatial registration in the GIS platform to construct a three-dimensional spatial scene that includes the geometric features of transmission lines, the terrain features of transmission lines, and the features of ground objects along transmission lines; S102. Based on the geometric feature data of the transmission line, taking the center line of the transmission line as a reference, and based on the voltage level data of the transmission line, calculate the basic buffer width of the monitoring area to obtain the boundary of the first monitoring corridor; S103. Based on the topographic feature data of the transmission line and the pre-set topographic safety data of the transmission line, the boundary of the first monitoring corridor is dynamically adjusted through the digital elevation model to obtain the boundary of the second monitoring corridor. S104. Based on the ground feature data of the transmission line, and based on the pre-set ground feature safety distance data of the transmission line, the boundary of the second monitoring corridor is corrected for safety distance avoidance in combination with the ground feature outline. The corrected corridor boundary range is then merged and spliced ​​with the boundary of the first monitoring corridor using a spatial Boolean algorithm to obtain the construction monitoring area of ​​the transmission corridor.

3. The method for construction supervision in power transmission corridor areas based on unmanned aerial vehicle (UAV) swarm collaborative operation according to claim 1, characterized in that, Step S2 includes the following steps: S201. Based on tower location coordinate data, the construction monitoring area of ​​the power transmission corridor is divided into multiple construction monitoring sub-areas; S202. Acquire drone endurance time data, flight speed data, camera resolution data, and field of view data, and establish a drone performance parameter matrix; S203. Using the boundary of the power transmission corridor construction monitoring area as a benchmark, establish the UAV obstacle avoidance parameter matrix based on the UAV performance parameter matrix; S204. Based on the graph model construction algorithm, each construction monitoring sub-region is discretized into a node network, and the UAV performance parameter matrix is ​​used as the edge weight and the UAV obstacle avoidance parameter matrix is ​​used as the node value to construct the UAV monitoring path optimization model corresponding to the node network of each construction monitoring sub-region. S205. Based on the UAV monitoring path optimization model corresponding to each node network, with the minimum UAV flight time and the maximum UAV monitoring coverage as objectives, the UAV monitoring path for each construction monitoring sub-area is calculated using the Dijkstra algorithm.

4. A method for construction supervision in power transmission corridor areas based on unmanned aerial vehicle (UAV) swarm collaborative operations as described in claim 1, characterized in that, The drone swarm described in step S3 includes multiple drone sub-swarms; The drone subgroup includes master drones and slave drones, and corresponds one-to-one with each construction monitoring sub-area; The master drone and slave drones transmit drone operating parameters in real time through a Mesh self-organizing network.

5. A method for construction supervision in power transmission corridor areas based on unmanned aerial vehicle (UAV) swarm collaborative operations as described in claim 4, characterized in that, The UAV swarm collaborative patrol mechanism described in step S3 includes a multi-UAV swarm scheduling sub-mechanism and an adaptive patrol and obstacle avoidance sub-mechanism. The multi-UAV cluster scheduling sub-mechanism divides each construction monitoring sub-area into multiple task groups based on the UAV performance parameter matrix and through block parallelism and dynamic load balancing algorithms. With the master UAV as the central node and the slave UAVs as neighboring nodes, the operating status parameters of the slave UAVs are transmitted to the master UAV in real time through a Mesh self-organizing network. The adaptive patrol and obstacle avoidance sub-mechanism uses the UAV obstacle avoidance parameter matrix and a digital elevation model to control the flight attitude of the master UAV and master-slave UAVs through a PID algorithm, generates local obstacle avoidance paths through a Venn diagram algorithm to control ground object avoidance for the master UAV and master-slave UAVs, monitors the relative positions of the master UAV and slave UAVs in real time through GPS and ultra-wideband positioning technology, and publishes a unified timestamp by the master UAV, which then performs obstacle avoidance control based on the timestamp.

6. A method for construction supervision in power transmission corridor areas based on unmanned aerial vehicle (UAV) swarm collaborative operations as described in claim 1, characterized in that, Step S4 includes the following steps: S401. Based on the monitoring image dataset of the power transmission corridor area, the construction target features are identified using a power construction-specific model improved based on YOLOv8n, and the construction target feature identification results are obtained. S402. Based on the results of construction target feature identification, obtain historical image data of the power transmission corridor area, establish a pyramid hierarchical index, and obtain anomaly feature dataset of the power transmission corridor area through SIFT feature point matching.

7. A method for construction supervision in power transmission corridor areas based on unmanned aerial vehicle (UAV) swarm collaborative operations as described in claim 1, characterized in that, Step S5 includes the following steps: S501. Obtain the preset land parcel data and construction condition data of the power transmission corridor area, and establish a multi-entity association network of land parcel data entities and construction condition data entities to construct a construction knowledge graph of the power transmission corridor area. S502. Based on the SWRL rule reasoning algorithm, according to the construction knowledge graph of the transmission corridor area, the abnormal feature dataset is compared for compliance. Abnormal feature data that does not meet the requirements of the construction knowledge graph of the transmission corridor area are regarded as abnormal construction data, and the abnormal construction dataset is obtained. S503. Based on the spatial overlay analysis algorithm, the abnormal construction dataset is compared with the boundary of each construction monitoring sub-region. When the intersection area is greater than or equal to 50%, the current abnormal construction data is associated with the current construction monitoring sub-region, and a construction rectification report for the power transmission corridor area is generated. S504. The obtained construction rectification report for the power transmission corridor area shall be sent to the corresponding construction unit for rectification.

8. A method for construction supervision in power transmission corridor areas based on unmanned aerial vehicle (UAV) swarm collaborative operations as described in claim 1, characterized in that, Step S6 includes the following steps: S601. Obtain rectification frequency data and rectification progress data of each construction unit, and construct a construction violation risk value calculation model, wherein the construction violation risk value calculation model is expressed as: ； Where R represents the risk value of construction violations, D represents the current rectification frequency, L represents the current rectification progress, and S represents the sensitivity of the construction area; S602. Based on the construction violation risk value calculation model, calculate the construction violation risk level for each construction monitoring sub-area. When the construction violation risk value is greater than 0.8, it is a high-risk construction area; when the construction violation risk value is greater than 0.5 and less than or equal to 0.8, it is a medium-risk construction area; when the construction violation risk value is less than or equal to 0.5, it is a low-risk construction area. S603. Based on the calculation results of the construction violation risk level, and based on the drone monitoring path of each sub-region, repeat monitoring is performed on each construction monitoring sub-region, and the construction knowledge graph of the power transmission corridor area is updated based on the repeat monitoring results. S604. Based on the repeated monitoring results, through reinforcement learning algorithms, for construction areas where the same abnormal construction feature data appears at least twice, construct a drone swarm monitoring path based on the Archimedes spiral and conduct drone swarm monitoring until the construction rectification is completed.

9. A construction supervision system for power transmission corridor areas based on unmanned aerial vehicle (UAV) swarm collaborative operation, the system being implemented based on the construction supervision method for power transmission corridor areas based on unmanned aerial vehicle (UAV) swarm collaborative operation as described in any one of claims 1-8, characterized in that, include: The multi-source data acquisition and fusion module is used to acquire GIS data of transmission lines and tower coordinate data, construct a construction monitoring buffer zone of the transmission corridor based on the centerline of the transmission line, and dynamically adjust the construction monitoring buffer zone of the transmission corridor by combining the digital elevation model with the topographic data of the transmission line to obtain the construction monitoring area of ​​the transmission corridor. The monitoring area division and path generation module is used to divide the construction monitoring area of ​​the power transmission corridor based on tower location coordinate data, resulting in multiple construction monitoring sub-areas. With UAV operating condition parameters as constraints, the module constructs UAV monitoring path optimization models for each sub-area and calculates the UAV monitoring paths for each sub-area with the minimum flight time and maximum monitoring coverage of the UAV as objectives. The UAV swarm control and monitoring module is used to collect image data of each construction monitoring sub-area through the UAV swarm based on the UAV monitoring path of each sub-area and a distributed collaborative control algorithm, and to build a monitoring image dataset of the power transmission corridor area. The image anomaly data comparison module uses a lightweight deep learning algorithm to perform semantic feature annotation and feature recognition based on the monitoring image dataset of the power transmission corridor area, and compares the features with historical image data of the power transmission corridor area to obtain an anomaly feature dataset of the power transmission corridor area. The knowledge graph construction and construction anomaly monitoring module is used to acquire preset land parcel data and standard operating condition data of the power transmission corridor area, construct a construction knowledge graph of the power transmission corridor area, and perform feature comparison based on rule reasoning algorithm and anomaly feature dataset to obtain anomaly construction dataset of the power transmission corridor area, generate construction rectification report of the power transmission corridor area and push it to the construction unit for rectification. The construction rectification monitoring module is used to acquire rectification frequency data, rectification progress data, and construction area sensitivity data of each construction unit, construct a construction violation risk value calculation model, calculate the construction risk level of each construction monitoring sub-area, and conduct drone swarm collaborative monitoring of each construction monitoring sub-area based on the construction risk level calculation until the construction rectification is completed.

Images