Wind power line unmanned aerial vehicle flaming obstacle removing method, system and equipment and medium

Abstract

The invention discloses a method, a system, equipment and a medium for removing obstacles through flaming of a wind power line unmanned aerial vehicle, and relates to the technical field of power system operation and maintenance and intelligent inspection, and the method comprises the steps: synchronously collecting multi-source data, generating a power transmission line digital twinborn body through space-time calibration, carrying out the recognition and segmentation in the twinborn body, marking an inflammable grade, and constructing a risk base map. And according to the base map and the environmental parameters, calculating a minimum energy flaming instruction through a model, controlling the unmanned aerial vehicle to operate, monitoring safety in real time, rechecking an obstacle clearing effect and thermal damage by utilizing infrared after operation, storing data into a database, and analyzing and precipitating the data into a reusable strategy library. According to the method, the high-fidelity digital twin body is constructed, so that a whole-process closed loop from accurate risk modeling, minimum energy control to intelligent strategy precipitation is realized, and the safety, the efficiency and the intelligent level of power transmission line unmanned aerial vehicle flaming obstacle removal operation are remarkably improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power system operation and maintenance and intelligent inspection, and particularly relates to a wind power line unmanned aerial vehicle fire spraying and obstacle removing method, system, device and medium. BACKGROUND

[0002] The power transmission line of the wind power field is exposed to the environment of strong wind, seasonal dryness and complex terrain all year round, and the floating objects such as kites, dustproof nets, plastic films and dry branches are easily entangled near the conductor, ground wire or insulator; the traditional obstacle removing method has the problems of low efficiency, high risk and long power outage window, and the unmanned aerial vehicle carrying a heat source for "non-contact" obstacle removing has early exploration, but lacks high-fidelity three-dimensional environment, component-level risk layering and control strategy training capability for the power scene.

[0003] The existing unmanned aerial vehicle fire spraying and obstacle removing scheme can reduce the climbing operation risk under the condition of not stopping power supply, but a series of restricting factors are exposed in actual application: first, under the interference of firelight, the types, scales and relative spatial relationships of obstacles with conductors, ground wires, insulators and grading rings are difficult to be stably identified and positioned, resulting in insufficient decision basis; second, the spraying distance, angle and time are highly dependent on the experience of the pilot, and lack of quantitative constraints and interlocking mechanisms, which not only easily leads to "not burning" due to insufficient energy, but also may cause thermal shock or change the physical properties of the line body due to excessive energy, leaving long-term hidden dangers; at the same time, the safety boundary lacks visualization and rigid execution means, and it is difficult to strictly implement three-dimensional electronic fence, minimum approach distance, temperature rise red line and fire prohibition area in task planning and flight process; and moreover, the airborne fire spraying device is mostly a simple modification, and has problems of center of gravity deviation, vibration coupling and insufficient electromagnetic compatibility, and the spraying vector is difficult to accurately control, resulting in obvious fluctuation of obstacle removing effect; in addition, external disturbances such as crosswind and gust make the flame deflect, and the rain wet condition makes the ignition and combustion process unstable, and the real-time prediction and early warning of the flame influence cone and risk margin are lacking on site; the data between inspection, approval, execution and review are not connected, it is difficult to deposit reusable "scene-strategy library", and it is also difficult to realize multi-machine cooperation and large-scale promotion. SUMMARY

[0004] In view of the above existing problems, the present application provides a wind power line unmanned aerial vehicle fire spraying and obstacle removing method, system, device and medium to solve the problems of insufficient decision basis, center of gravity deviation, vibration coupling and insufficient electromagnetic compatibility, lack of real-time prediction and early warning of the flame influence cone and risk margin, and inability to realize multi-machine cooperation and large-scale promotion in the prior art.

[0005] To solve the above technical problems, a wind power line unmanned aerial vehicle fire spraying and obstacle removing method is provided, which comprises,

[0006] Synchronous acquisition of multi-source data of the power transmission line corridor, time synchronization and sensor external parameter calibration of multi-source data, obtaining the original data set of unified space-time reference, and generating the digital twin of the power transmission line and the surrounding environment by using the three-dimensional scene reconstruction technology; in the digital twin, using identification and classification methods for instance segmentation and attribute labeling, assigning flammable grade and allowable temperature rise threshold, constructing risk base map, and in real-time video stream, the safety boundary elements of the calculable risk base map and the operation parameter suggestions are fused and displayed to form a visual interactive interface; according to the risk base map and the real-time environmental parameters, the minimum energy control instruction sequence of the flame spraying operation is calculated through the flame propagation behavior model, the unmanned aerial vehicle is controlled to execute the flame spraying operation according to the minimum energy control instruction sequence, the safety constraint conditions are monitored in real time, and the failure protection mechanism is automatically triggered when the preset threshold is violated; after the operation is completed, the target area is checked for obstacle removal effect and thermal damage evaluation by infrared thermal imaging, the actual cooling curve is recorded and abnormal thermal spots are removed, the whole process data is written into the obstacle removal database, and the feature extraction and cluster analysis are used to deposit the reusable strategy library, and the scene feature vector and the recommended strategy are mapped.

[0007] As a preferred scheme of the wind power line unmanned aerial vehicle flame spraying obstacle removal method, the multi-source data of the power transmission line corridor includes acquiring multi-view image sequences through a visual acquisition device, acquiring three-dimensional point cloud data using a spatial detection device, and measuring wind speed, wind direction, air temperature and humidity using an environmental monitoring device, and time stamping and external parameter calibration of sensor data to obtain synchronized multi-source data set.

[0008] As a preferred scheme of the wind power line unmanned aerial vehicle flame spraying obstacle removal method, the three-dimensional scene reconstruction technology includes scene scale anchoring and coordinate unification based on multi-source data input, initializing a three-dimensional Gaussian primitive set, and using a micro-rendering-driven coarse-to-fine optimization strategy to refine the model structure.

[0009] The continuity of conductor suspension and the consistency of insulator array are introduced as a priori, local errors are handled by adaptive densification and abnormal primitive removal, and after the model converges, quality check is performed to generate a Gaussian scene file.

[0010] As a preferred scheme of the wind power line unmanned aerial vehicle flame spraying obstacle removal method, the construction of the risk base map includes instance segmentation of line components and obstacles by image recognition technology, identification of type and spatial distribution.

[0011] The minimum spatial distance between components and obstacles, the volume feature and the included angle with the wind direction are calculated by spatial analysis method; and according to the material properties and surface state, the flammable grade and the allowable temperature rise threshold are assigned to each instance to construct the calculable risk base map.

[0012] As a preferred embodiment of the wind power line drone fire-spraying obstacle clearing method described in this invention, the formation of the visual interactive interface includes superimposing the three-dimensional electronic fence, minimum approach distance bubble, fire-prohibited mask, flame impact cone and suggested fire-spraying duration countdown from the computable risk base map onto the drone's real-time video stream through a unified mapping between the camera coordinate system and the world coordinate system.

[0013] Among them, the three-dimensional electronic fence marks the space areas where drones are prohibited from entering; the minimum approach distance bubble dynamically displays the closest distance between the drone and the charged body centered on the charged body; the fire-prohibited mask marks the buffer zone and safe area of ​​the wire accessories that are not allowed to be touched by flames; and the flame impact cone predicts the flame range based on real-time wind conditions. The operator can view the safety boundaries and parameter suggestions in real time through the current interface, and adjust the drone's posture according to the prompts until it meets the operation requirements.

[0014] The formula for calculating the closest distance between a drone and a charged object is as follows:

[0015]

[0016] in, For the minimum distance, Let be the spatial coordinates of the i-th key point on the obstacle. Two points and The Euclidean distance between two points is the straight-line distance in three-dimensional space. Let be the spatial coordinates of the j-th key point on the line component.

[0017] As a preferred embodiment of the wind power line drone fire-spraying obstacle clearing method described in this invention, the minimum energy control command sequence for calculating the fire-spraying operation includes taking nozzle diameter, spray angle, fuel flow rate, relative position of the drone and the target, and wind field elements at the operating height as input parameters, describing the weakening characteristics of heat flux with distance through a distance attenuation function, and introducing a crosswind deflection correction factor to compensate for the influence of the wind field on the flame direction.

[0018] The equivalent heat flux of the target surface is calculated and compared with the ignition threshold and minimum exposure time of the obstacle material to determine whether the ignition conditions are met. The minimum energy control command sequence, including the spray distance, spray angle and operation time, is solved by combining the fire-prevention mask and the minimum approach distance constraint.

[0019] The ignition condition is expressed as:

[0020]

[0021] in, This represents the total amount of heat energy accumulated on the target surface within the time interval [0, t]. is the instantaneous heat flux, is the ignition threshold of the obstacle material, i.e. the minimum heat flux required to ignite the material, is the minimum exposure time, i.e. the minimum duration required for the material to be ignited under the heat flux, is the time index;

[0022] The real-time monitoring safety constraints include calculating the minimum approach distance of the UAV to the live body, predicting the intersection state of the flame envelope and the fire-prohibited mask, and estimating whether the temperature rise touches the red line;

[0023] Wherein, the trigger logic of the failure protection mechanism includes that when it is monitored that the minimum approach distance is less than 3 meters, the flame envelope intersects with the fire-prohibited mask or the temperature rise exceeds the threshold, the system automatically performs the reduced fire spraying, reduces the flame influence range, or immediately stops the fire spraying and triggers the sequence of extinguishing, climbing and returning along the safety corridor to perform the UAV safety evacuation, and the monitoring and triggering process is based on the real-time returned UAV attitude, link quality and thermal image data.

[0024] As a preferred scheme of the wind power line UAV fire spraying obstacle removing method, the infrared thermal image review includes collecting the infrared thermal image data of the target area after the operation, comparing with the temperature baseline before the operation, and calculating the temperature distribution difference value.

[0025] The temperature distribution difference value formula is expressed as:

[0026]

[0027] Wherein, is the effective temperature rise value at the image coordinates (u, v) position, is the original temperature value measured at the coordinates (u, v) on the infrared thermal image after the operation, is the environmental temperature reference value measured in the reference area after the operation, is the original temperature value measured at the coordinates (u, v) on the infrared thermal image before the operation, is the environmental temperature reference value measured in the reference area before the operation, is the pixel coordinates of the infrared thermal image;

[0028] According to the allowable temperature rise threshold of the component, the abnormally high temperature data caused by reflection, background interference and transient noise points are removed, the removal standard is that the average temperature rise of the high temperature area is less than 10% of the allowable threshold or the short-time hot spot area is less than 1% of the total image area, the actual cooling curve is recorded, and the thermal damage is evaluated.

[0029] The feature extraction and cluster analysis comprises: standardizing wind conditions, terrain, obstacle types and execution parameters in the obstacle clearing database, eliminating dimensional differences, extracting main characteristic components by using principal component analysis method, generating scene characteristic vectors, and performing cluster analysis based on characteristic similarity, so that data is classified into similar scene categories, the execution parameter combination with the highest obstacle clearing success rate and the lowest energy consumption in each scene category is extracted, and is stored as a strategy library entry to realize mapping of scene characteristics and recommended strategies.

[0030] The beneficial effects of the preferred technical scheme are: through the flame propagation model and the energy criterion, precise energy delivery based on material characteristics is realized, the problems of insufficient or excessive energy caused by experience spraying are solved, the heat impact is minimized while ensuring the obstacle clearing effect, and through multi-dimensional safety monitoring and failure protection, a complete protection chain from prediction and early warning to emergency disposal is established, the unreliability problem of relying on human vigilance is solved, and active defense and automatic disposal of safety risks are realized.

[0031] As a preferred scheme of the unmanned aerial vehicle flame spraying obstacle clearing system for a wind power transmission line, the system comprises a data acquisition and three-dimensional reconstruction module, a risk identification and base map construction module, a visual interaction and intelligent control module, and a safe execution and strategy learning module.

[0032] The data acquisition and three-dimensional reconstruction module is used for a multi-source sensor and a meteorological module arranged on the unmanned aerial vehicle, and is used for synchronously collecting multi-view visual data, three-dimensional space data and environmental meteorological data of the transmission line corridor, and performing time synchronization and external parameter calibration on all data to obtain an original data set with a unified space-time reference, and generating a digital twin based on a three-dimensional Gaussian splashing reconstruction technology.

[0033] The risk identification and base map construction module is used for realizing automatic instance segmentation and identification of conductors, insulators, fittings and temporary floating objects based on semantic segmentation and multi-view consistent projection technology based on deep learning on the basis of the reconstructed digital twin, and is used for combining a material dictionary and surface state estimation to assign a non-flammable, difficult-to-flammable or flammable flammable grade and an allowable temperature rise threshold to each instance, and is used for integrating a spatial analysis method to calculate a minimum distance, an obstacle volume and a wind direction angle to construct a calculable risk base map.

[0034] The visual interaction and intelligent control module is used for mapping safety boundary elements and operation parameter suggestions in the calculable risk base map through a camera-world coordinate system to form a real-time interactive interface by superimposing the safety boundary elements and the operation parameter suggestions into the unmanned aerial vehicle video stream, and is used for generating a minimum energy control instruction sequence based on a flame propagation behavior model combining distance attenuation and crosswind deflection effect, and combining an obstacle ignition criterion and safety constraints.

[0035] The safe execution and policy learning module is used for monitoring the safety constraints of the minimum approach distance, the fire prohibition mask intersection and the temperature rise red line in real time during the fire spraying execution process, and triggering a failure protection mechanism immediately when the safety constraints are exceeded, executing the logic of reducing the amount of fire spraying, stopping the work or automatically returning, reviewing through infrared thermal imaging after the work, eliminating abnormal heat spots and recording the cooling curve, evaluating the obstacle clearing effect and the thermal damage situation, and writing the whole process data into an obstacle clearing database, and precipitating into a policy library of scene-policy mapping through feature extraction and cluster analysis.

[0036] A computer device comprises a memory and a processor, the memory stores a computer program, and the processor implements the steps of the method for unmanned aerial vehicle fire spraying obstacle clearing of wind power lines when executing the computer program.

[0037] A computer readable storage medium stores a computer program, and the computer program implements the steps of the method for unmanned aerial vehicle fire spraying obstacle clearing of wind power lines when executed by a processor.

[0038] The present application has the following advantages: the present application establishes a real and reliable operation environment digital twin through multi-source data synchronous acquisition and three-dimensional reconstruction technology, solves the problem of insufficient decision basis caused by the complexity of the environment in the traditional method, and converts the abstract spatial relationship into intuitive visual prompts by means of risk base map construction and augmented reality visualization interaction, thereby significantly reducing the operation cognitive load and the failure probability; based on the flame propagation behavior model and the multi-level safety interlocking mechanism, the change from experience operation to quantitative control of the fire spraying operation is realized, and the rigid execution of the safety boundary is ensured; through the infrared thermal image review and the policy learning closed loop, the transformation channel from single operation to experience accumulation is established, and the continuous evolution of the obstacle clearing ability is realized. BRIEF DESCRIPTION OF DRAWINGS

[0039] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0040] Figure 1 A general flowchart of a wind power line unmanned aerial vehicle fire spraying obstacle clearing method provided by an embodiment of the present application.

[0041] Figure 2 A single flight technical flowchart of a wind power line unmanned aerial vehicle fire spraying obstacle clearing method provided by an embodiment of the present application.

[0042] Figure 3 A work operation schematic diagram of a wind power line unmanned aerial vehicle fire spraying obstacle clearing method provided by an embodiment of the present application.

[0043] Figure 4 A timing diagram of the operation process of a wind power line unmanned aerial vehicle fire spraying and obstacle clearing method provided for an embodiment of the present application

[0044] Figure 5 A system scheme flowchart of a wind power line unmanned aerial vehicle fire spraying and obstacle clearing system provided for an embodiment of the present application DETAILED DESCRIPTION

[0045] In order to make the above objectives, characteristics and advantages of the present application more obvious and easy to understand, the specific embodiments of the present application will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor should fall within the scope of protection of the present application.

[0046] Embodiment 1, refer to Figure 1 For an embodiment of the present application, a wind power line unmanned aerial vehicle fire spraying and obstacle clearing method is provided, comprising:

[0047] S100: synchronously collecting multi-source data of the power transmission line corridor, time synchronizing and sensor external parameter calibrating the multi-source data, obtaining an original data set with unified space-time reference, and generating a digital twin of the power transmission line and the surrounding environment by using three-dimensional scene reconstruction technology;

[0048] S200: in the digital twin, using recognition and classification methods for instance segmentation and attribute labeling, assigning flammable grade and allowable temperature rise threshold, constructing risk base map, and in real-time video stream, fusing and displaying safety boundary elements of the calculable risk base map and operation parameter suggestions to form a visual interactive interface;

[0049] S300: according to the risk base map and real-time environmental parameters, calculating the minimum energy control instruction sequence of the fire spraying operation through the flame propagation behavior model, controlling the unmanned aerial vehicle to execute the fire spraying operation according to the minimum energy control instruction sequence, monitoring the safety constraint conditions in real time, and automatically triggering the failure protection mechanism when the preset threshold is violated;

[0050] S400: after the operation is completed, the target area is checked for obstacle clearing effect and thermal damage evaluation by infrared thermal image, the actual cooling curve is recorded and the abnormal thermal spot is removed, the whole process data is written into the obstacle clearing database, and the feature extraction and cluster analysis are used to deposit the reusable strategy library, and the scene feature vector and the recommended strategy are mapped.

[0051] It should be noted that by constructing a high-fidelity digital twin, a data foundation is laid for accurate decision-making, abstract safety procedures are converted into a computable risk bottom map and are fused with visualization, the operation difficulty and failure rate are reduced, and through physical model solving, minimum energy control instructions are obtained, accurate, safe and economic obstacle removal operations are realized, and the safety, efficiency and economy of the power transmission line obstacle removal operation are significantly improved.

[0052] Embodiment 2, with reference to Figures 1-4 As a second embodiment of the present application, the embodiment provides a wind power line unmanned aerial vehicle fire spraying obstacle removal method, comprising:

[0053] In step S100, the collection of multi-source data of the power transmission line corridor includes steps S101-S103:

[0054] S101: collecting multi-view visual data and three-dimensional space data through multi-source sensors deployed on the unmanned aerial vehicle;

[0055] The multi-view visual data includes synchronous or near-synchronous image sequences obtained from the same target corridor at different spatial orientations and camera attitudes, specifically including overhead view, front view and rear view (head and tail directions), left and right oblique view (left and right oblique angles in the along-line and across-line directions), close-up component view (local magnification of insulator strings, fittings, jumpers, and grading rings), along-line view (following the direction of the conductor, for curvature and gap determination), across-line view (perpendicular to the direction of the conductor, for minimum clearance determination), point cloud and attitude / positioning data (sparse and dense three-dimensional point clouds generated by onboard laser radar and millimeter wave radar), each point containing X / Y / Z coordinates, echo intensity / reflectivity, echo order, timestamp, optional normal vector and semantic label.

[0056] S102: obtaining environmental meteorological data of wind speed, wind direction, gust probability and precipitation probability through a meteorological module, specifically including work height wind speed and wind direction, gust peak value and gust probability, crosswind / downwind component, vertical wind shear, turbulence intensity (TI), gust coefficient, air temperature, relative humidity, precipitation intensity and probability, visibility / haze dust concentration.

[0057] S103: time synchronization and sensor extrinsic calibration are performed on the data collected by the multi-source sensors and the meteorological module to form a synchronous multi-source data set with a unified space-time reference, providing a data foundation for subsequent three-dimensional scene reconstruction.

[0058] Further, in the embodiment of the present application, in step S101, the multi-source sensors include steps A1-A4:

[0059] A1: acquiring image sequences of overhead view, front view, rear view and oblique view through a visible light camera.

[0060] A2: Generate point cloud data with coordinates and reflectivity by laser radar and millimeter wave radar.

[0061] A3: Capture thermal distribution images by infrared thermal imager.

[0062] A4: After time stamp alignment and external parameter calibration of all sensor data, a multi-source data set in a unified coordinate system is formed for subsequent three-dimensional reconstruction and risk analysis.

[0063] In an optional embodiment, in step S101, the multi-source sensor further includes: acquiring visible light and near-infrared band images by a multispectral camera, identifying obstacles with different material properties (plastic film or dry branches), and generating distance and shape information of the obstacles by emitting sound waves and receiving echoes through a sonar sensor, especially supplementing visual data in smoke or low visibility conditions. After synchronous fusion of the data, a multi-modal data set is output.

[0064] In another optional embodiment, in step S101, the multi-source sensor can further include: capturing depth images from different angles by a stereo vision camera, and calculating three-dimensional structures using parallax, providing high-precision position and attitude data through GPS / IMU to assist spatial positioning, and integrating all data into an embedded system in real time for rapid preprocessing and coordinate unification.

[0065] Further, in step S100, the generating a digital twin includes steps S111-S112:

[0066] S111: Based on the synchronous multi-source data set, integrate visual data, spatial data, and environmental data into a unified coordinate system through multi-source data fusion technology to form a comprehensive three-dimensional data representation.

[0067] S112: Process the comprehensive three-dimensional data representation using three-dimensional modeling technology to generate a digital twin containing the geometric structure, spatial position, and surrounding terrain features of the power transmission line components.

[0068] Among them, the digital twin supports spatial query and visual interaction, multi-source data fusion ensures data consistency, and three-dimensional modeling technology constructs accurate three-dimensional models.

[0069] It should be noted that in the embodiments of the present application, in step S112, the three-dimensional modeling technology includes steps B1-B5:

[0070] B1: Complete scene scale anchoring and coordinate definition according to known tower positions, span distances, and GNSS reference.

[0071] B2: Initialize a three-dimensional Gaussian primitive set based on multi-view geometric solution results and extracted conductor centerlines and tower position connecting lines.

[0072] B3: A coarse-to-fine optimization strategy driven by micro-rendering is adopted to achieve fast convergence and progressive refinement, and wire sag continuity and insulator array consistency are introduced as structural priors to improve geometric stability.

[0073] B4: During optimization, adaptive densification from sparse to dense, abnormal primitive rejection, and re-optimization of local error areas are implemented.

[0074] B5: After the model converges, quality checks are carried out, including tower span end point position error, span and adjacent conductor spacing consistency, key component multi-view reprojection error evaluation, generation of Gaussian scene files for interlocking visualization and spatial query, and export of engineering exchange format and lightweight level of detail.

[0075] In an optional embodiment, in step S112, the three-dimensional modeling technique further includes calculating depth information by projecting a structured light pattern onto the surface of the power transmission line and capturing the deformed pattern using a camera, covering different viewing angles by multiple scans, and fusing data by point cloud registration, constructing a three-dimensional model using a mesh generation algorithm, and adding texture mapping to enhance realism.

[0076] In another optional embodiment, in step S112, the three-dimensional modeling technique can further include extracting feature points from multi-view images and generating a sparse point cloud using a Structure from Motion (SfM) algorithm, densifying with laser radar point cloud, and registering different view data by Iterative Closest Point (ICP) algorithm to generate a three-dimensional model with texture.

[0077] In the embodiments of the present application, in step S200, the construction risk map includes steps S201-S203:

[0078] S201: In the digital twin, line components and obstacles are instance segmented by image recognition technology to identify types and spatial distribution.

[0079] S202: The minimum spatial distance between components and obstacles, volume characteristics, and the angle with the wind direction are calculated by spatial analysis method;

[0080] The minimum spatial distance formula is represented as:

[0081]

[0082] wherein, is the minimum distance, is the spatial coordinates of the i-th key point on the obstacle, is the Euclidean distance between two points and is the straight-line distance in three-dimensional space, is the spatial coordinate of the jth key point on the line assembly.

[0083] S203: According to the material properties and surface state, assign a flammable grade and an allowable temperature rise threshold to each instance, build a computable risk base map containing risk information, and provide a basis for visual interaction and fire control; wherein image recognition technology is used to classify components, and spatial analysis methods are used to quantify spatial relationships;

[0084] The risk level quantification formula is expressed as:

[0085]

[0086] wherein, is the comprehensive risk value, , and is the weight coefficient, is the volume of the obstacle, is the angle between the obstacle and the dominant wind direction;

[0087] The allowable temperature rise threshold determination formula is expressed as:

[0088]

[0089] wherein, is the allowable temperature rise threshold, is the maximum tolerance temperature of the component material, is the environmental temperature reference value, is the safety margin.

[0090] In an optional embodiment, in step S200, the building of the risk base map further includes extracting the conductor and obstacle contours using edge detection and morphological operations, classifying component types by template matching and historical data comparison, and assigning a flammable grade in combination with a rule engine (such as a threshold based on color and texture).

[0091] In another optional embodiment, in step S200, the building of the risk base map can further include segmenting connected regions in the image by region growing algorithm and extracting features (such as shape and texture), classifying features using support vector machine (SVM) or decision tree, and manually defining a material rule library to assign a flammable grade.

[0092] Further, in the embodiments of the present application, in step S200, the formation of the visual interaction interface includes steps S211-S213:

[0093] S211: Through the unified mapping of the camera coordinate system and the world coordinate system, the three-dimensional electronic fence, the minimum approach distance bubble, the fireproof mask, the flame influence cone, and the recommended fire duration countdown in the computable risk base map are superimposed into the real-time video stream of the UAV.

[0094] S212: The three-dimensional electronic fence marks the space area where the UAV is prohibited to enter, the minimum approach distance bubble dynamically displays the closest distance between the UAV and the live wire based on the live wire as the center and ensures that it is not less than 3 meters, the fireproof mask marks the buffer zone and the safety area where the flame is not allowed to reach the wire accessories, and the flame influence cone predicts the flame range according to the real-time wind condition;

[0095] When , it is displayed in green; when , it is displayed in yellow; and when , it is displayed in red.

[0096] The flame influence cone prediction formula is represented as:

[0097]

[0098]

[0099] wherein, is the flame influence range function considering the wind speed and angle, is the reference flame length under windless conditions, is the wind speed influence coefficient, is the nozzle pointing angle, is the flame influence cone boundary point set, is the nozzle position coordinates, is the coordinates of any point in the three-dimensional space, is the wind field speed.

[0100] S213: The operator views the safety boundary and parameter suggestions in real time through the current interface and adjusts the UAV pose according to the prompt, calculates the safety margin in real time and warns of the out-of-bound behavior;

[0101] wherein, the superimposed display adopts image plane and space dual-domain indication to ensure consistent safety boundary prompt under any viewing angle;

[0102] The calculation formula of the safety margin is represented as:

[0103]

[0104] wherein, is the current safety margin, is the safety distance threshold set to 3 meters, is the real-time dynamic distance between the UAV and the live wire; when , there is no warning; when Level 1 warning when Level 2 warning when

[0105] In an alternative embodiment, in step S200, the forming the visual interactive interface further comprises superimposing the risk base map elements into the operator's field of view through AR glasses, displaying the electronic fence and the flame cone, and adjusting the display parameters in combination with gesture recognition or voice commands, synchronizing the drone data in real time, and providing three-dimensional space navigation.

[0106] In another alternative embodiment, in step S200, the forming the visual interactive interface can further comprise loading the risk base map and real-time video stream through the browser, rendering the safety elements using HTML5 and WebGL, supporting multi-user collaborative viewing and operation log recording, and synchronizing the data through the cloud server to realize remote control.

[0107] In step S300, the calculating the minimum energy control instruction sequence of the flame spraying operation comprises steps S301-S303:

[0108] S301: Taking the nozzle diameter, the jet angle, the fuel flow, the relative position between the drone and the target, and the operation height wind field elements as input parameters, describing the attenuation characteristics of the heat flux with distance through a distance attenuation function, and introducing a side wind deflection correction factor to compensate for the influence of the wind field on the flame direction;

[0109] The equivalent heat flux of the flame in the direction of the nozzle centerline is represented as:

[0110]

[0111] wherein, is the equivalent heat flux of the target surface at a distance x from the nozzle, is the reference flux in the near field of the nozzle, and is an empirical parameter describing the attenuation characteristics of the heat flux with distance, is the straight-line distance from the nozzle to the target surface, is the equivalent diameter of the nozzle, is the deflection correction factor, is the pointing angle of the nozzle, is the wind field velocity.

[0112] S302: Calculating the equivalent heat flux of the target surface, and comparing it with the ignition threshold and the shortest exposure time of the obstacle material to determine whether the ignition condition is met;

[0113] The ignition condition is represented as:

[0114]

[0115] wherein, is the total amount of heat energy accumulated by the target surface in the time interval [0, t], is the instantaneous heat flux, is the ignition threshold of the barrier material, i.e., the minimum heat flux required to ignite the material, is the minimum exposure time, i.e., the minimum duration required for the material to be ignited under the action of the heat flux, is the time index.

[0116] S303: Solve the minimum energy control instruction sequence, including the fire spraying distance, the spraying angle, and the operation time, in combination with the fire ban mask and the minimum approach distance constraint. The model parameters are fitted through ground bench and outdoor test, and are corrected online with short pulses in the field;

[0117] The minimum energy control instruction sequence calculation formula is represented as:

[0118]

[0119] wherein, is the total energy, is the total operation time, is the instantaneous heat flux, is the effective heating area of the target barrier, is the time index.

[0120] In step S300, the real-time monitoring safety constraint condition includes calculating the minimum approach distance of the unmanned aerial vehicle and the live body, predicting the intersection state of the flame envelope and the fire ban mask, and estimating whether the temperature rise reaches the red line;

[0121] The trigger logic of the failure protection mechanism includes that when it is monitored that the minimum approach distance is less than 3 meters, the flame envelope and the fire ban mask have an intersection, or the temperature rise exceeds the threshold, the system automatically performs the reduced fire spraying to reduce the flame influence range, or immediately stops the fire spraying and triggers the sequence of extinguishing, climbing, and returning along the safety corridor for the safe evacuation of the unmanned aerial vehicle. The monitoring and triggering process is based on the real-time returned unmanned aerial vehicle attitude, link quality, and thermal image data.

[0122] Further, in step S400, the infrared thermal image review includes steps S401-S403:

[0123] S401: Collect the infrared thermal image data of the target area after the operation, compare with the temperature baseline before the operation, and calculate the temperature distribution difference value;

[0124] The calculation formula of the temperature distribution difference value is represented as:

[0125]

[0126] wherein, is an effective temperature rise value at the image coordinate (u, v) position, is a raw temperature value measured at the coordinate (u, v) on the infrared thermal image after the operation, is an ambient temperature reference value measured in the reference area after the operation, is a raw temperature value measured at the coordinate (u, v) on the infrared thermal image before the operation, is an ambient temperature reference value measured in the reference area before the operation, is a pixel coordinate of the infrared thermal image.

[0127] S402: According to the component allowable temperature rise threshold value, abnormal high temperature data caused by reflection, background interference or transient noise points are removed, and the removal standard is that the average temperature rise of the high temperature area is lower than 10% of the allowable threshold value or the short time hot spot area is less than 1% of the total image area.

[0128] S403: Record the actual cooling curve and evaluate the thermal damage.

[0129] Further, in step S400, the infrared thermal image review includes steps S411-S414:

[0130] S411: Standardize the wind conditions, terrain, obstacle types and execution parameters in the obstacle removal database, and eliminate the dimensional differences.

[0131] S412: Extract the main characteristic components by using the principal component analysis method to generate the scene characteristic vector.

[0132] S413: Perform clustering analysis based on the feature similarity, and classify the data into similar scene categories.

[0133] S414: Extract the execution parameter combination with the highest success rate and the lowest energy consumption in each type of scene, store it as a strategy library entry, and realize the mapping of scene characteristics and recommended strategies.

[0134] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and are not limiting. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or replaced by equivalents without departing from the spirit and scope of the present application, and they should be covered in the scope of the claims of the present application.

[0135] Embodiment 3, refer to Figure 5 , is a third embodiment of the present application, which provides a wind power line unmanned aerial vehicle fire spraying obstacle removal system, including a data acquisition and three-dimensional reconstruction module, a risk identification and base map construction module, a visual interaction and intelligent control module, and a safe execution and strategy learning module.

[0136] The data acquisition and three-dimensional reconstruction module is used for the multi-source sensor and the weather module deployed on the unmanned aerial vehicle to synchronously collect multi-view visual data, three-dimensional space data and environmental weather data of the power transmission line corridor, and to perform time synchronization and external parameter calibration on all the data to obtain an original data set with unified space-time reference, and to generate a digital twin based on a three-dimensional Gaussian splashing reconstruction technology.

[0137] The risk identification and base map construction module is used for realizing automatic instance segmentation and identification of the conductor, the insulator, the hardware and the temporary floating object based on semantic segmentation and multi-view consistent projection technology based on deep learning on the basis of the reconstructed digital twin, and for assigning a non-flammable, difficult-to-flammable or flammable flammable grade and an allowable temperature rise threshold to each instance in combination with a material dictionary and surface state estimation, and for constructing a calculable risk base map by integrating a spatial analysis method to calculate a minimum distance, an obstacle volume and a wind direction angle.

[0138] The visual interaction and intelligent control module is used for mapping the safety boundary elements in the calculable risk base map and the operation parameter suggestion through a camera-world coordinate system to form a real-time interactive interface by superimposing the safety boundary elements and the operation parameter suggestion into the unmanned aerial vehicle video stream, and for generating a minimum energy control instruction sequence based on a flame propagation behavior model integrating distance attenuation and crosswind deflection effect in combination with an obstacle ignition criterion and safety constraints.

[0139] The safe execution and strategy learning module is used for monitoring the safety constraints of the minimum approach distance, the no-fire mask intersection and the temperature rise red line in real time during the fire spraying execution, and for triggering a failure protection mechanism immediately when the safety constraints are exceeded to execute a reduced fire spraying, a work suspension or an automatic return logic, and for evaluating the obstacle clearing effect and the thermal damage condition by eliminating abnormal heat spots and recording a cooling curve after the work through infrared thermal image review, and for writing the whole process data into an obstacle clearing database and precipitating a strategy library with scene-strategy mapping through feature extraction and cluster analysis.

[0140] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present application but not to limit the present application. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or replaced equivalently without departing from the spirit and scope of the technical solutions of the present application, and all of them should be covered in the scope of the claims of the present application.

[0141] Embodiment 4, the fourth embodiment of the present application, is different from the first three embodiments in that:

[0142] If the functions are implemented in the form of software function units and sold or used as independent products, they can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application or the part of the technical solutions that essentially contribute to the prior art or the part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in the various embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes.

[0143] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered a list of executable instructions for implementing logic functions, and can be specifically embodied in any computer-readable medium for use by an instruction execution system, apparatus, or device, such as a computer-based system, a system including a processor, or other system that can fetch the instructions from the instruction execution system, apparatus, or device and execute the instructions, or in conjunction with these instructions execution systems, apparatuses, or devices. For the purpose of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport programs for use by an instruction execution system, apparatus, or device, or in conjunction with these instruction execution systems, apparatuses, or devices.

[0144] More specific examples (a non-exhaustive list) of the computer-readable medium include the following: an electrical connection having one or more wires (electrical devices), a portable computer diskette (magnetic devices), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CD ROM). In addition, the computer readable medium can even be paper or other suitable medium on which the program can be printed, as the program can be electronically obtained, for example, by optical scanning of the paper or other medium, followed by editing, interpreting, or otherwise processing, if necessary, in other suitable ways to be electronically obtained, and then stored in the computer memory.

[0145] It should be understood that aspects of the application can be implemented in hardware, software, firmware or combinations thereof. In the embodiments described above, various steps or methods can be implemented, for example, by software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, and in another embodiment, any of the following technology, known in the art, or combinations thereof, can be used: discrete logic circuitry having logic gates for implementing logic functions upon an application of data signals, application specific integrated circuits having appropriate combinational logic gates, programmable gate arrays (PGA), field programmable gate arrays (FPGA), and the like.

Claims

1. A wind power line unmanned aerial vehicle fire spraying obstacle removing method, characterized in that: The method comprises the following steps: Synchronization acquisition of multi-source data of the power transmission line corridor, time synchronization and sensor external parameter calibration of the multi-source data, obtaining an original data set with unified space-time reference, and generating a digital twin by using a three-dimensional scene reconstruction technology; In the digital twin, instance segmentation and attribute labeling are performed by using identification and classification methods, flammable grade and allowable temperature rise threshold are assigned, a risk base map is constructed, and in the real-time video stream, the safety boundary elements of the calculable risk base map and the operation parameter suggestions are fused and displayed to form a visual interactive interface; According to the risk base map and the real-time environmental parameters, the minimum energy control instruction sequence of the flame spraying operation is calculated through a flame propagation behavior model, the unmanned aerial vehicle is controlled to execute the flame spraying operation according to the minimum energy control instruction sequence, the safety constraint conditions are monitored in real time, and the failure protection mechanism is automatically triggered when the preset threshold is violated; After the operation is completed, the target area is checked for obstacle removal effect and thermal damage evaluation by infrared thermal imaging, the actual cooling curve is recorded and abnormal thermal spots are removed, the whole process data is written into the obstacle removal database, and the feature extraction and cluster analysis are used to deposit the reusable strategy library, and the scene feature vector and the recommended strategy are mapped.

2. The unmanned aerial vehicle fire spraying and obstacle removing method for wind power lines according to claim 1, characterized in that: The multi-source data acquisition of the power transmission line corridor comprises the following steps: acquiring multi-view image sequences by a visual acquisition device, acquiring three-dimensional point cloud data by a space detection device, and measuring wind speed, wind direction, air temperature and humidity by an environmental monitoring device; the sensor data is time stamped and aligned and the external parameter calibration is performed to obtain a synchronized multi-source data set.

3. The unmanned aerial vehicle fire spraying and obstacle removing method for wind power lines according to claim 2, characterized in that: The three-dimensional scene reconstruction technology comprises the following steps: based on multi-source data input, scene scale anchoring and coordinate unification are performed, a three-dimensional Gaussian primitive set is initialized, and a coarse-to-fine optimization strategy driven by micro-rendering is used to refine the model structure; The continuity of the conductor suspension and the consistency of the insulator array are introduced as a priori, local errors are handled by adaptive densification and abnormal primitive removal, and after the model converges, quality checking is performed to generate a Gaussian scene file.

4. The unmanned aerial vehicle fire spraying and obstacle removing method for wind power lines according to claim 3, characterized in that: The construction of the risk base map comprises the following steps: instance segmentation of line components and obstacles is performed by image recognition technology, and the type and spatial distribution are identified; The minimum spatial distance between components and obstacles, the volume feature and the included angle with the wind direction are calculated by a spatial analysis method; and according to the material properties and surface state, the flammable grade and the allowable temperature rise threshold are assigned to each instance to construct a calculable risk base map.

5. The unmanned aerial vehicle fire spraying and obstacle removing method for wind power lines according to claim 4, characterized in that: The formation of the visual interactive interface comprises the following steps: through the unified mapping of the camera coordinate system and the world coordinate system, the three-dimensional electronic fence, the minimum approach distance bubble, the fireproof mask, the flame influence cone and the recommended flame spraying time countdown in the calculable risk base map are superimposed into the real-time video stream of the unmanned aerial vehicle; The three-dimensional electronic fence identifies the space area where the unmanned aerial vehicle is prohibited to enter, the minimum approach distance bubble dynamically displays the closest distance between the unmanned aerial vehicle and the live body with the live body as the center, the fireproof mask marks the buffer zone and the safety area of the conductor accessories which are not allowed to be touched by the flame, and the flame influence cone predicts the flame range according to the real-time wind condition; the operator can view the safety boundary and the parameter suggestions in real time through the current interface, and adjust the unmanned aerial vehicle pose according to the prompt until the operation requirements are met; The closest distance between the unmanned aerial vehicle and the live body is calculated as follows: ， in, For the minimum distance, Let be the spatial coordinates of the i-th key point on the obstacle. Two points and The Euclidean distance between two points is the straight-line distance in three-dimensional space. Let be the spatial coordinates of the j-th key point on the line component.

6. The unmanned aerial vehicle fire spraying and obstacle removing method for wind power lines according to claim 5, characterized in that: The minimum energy control instruction sequence of the fire spraying operation comprises, taking the nozzle diameter, the spraying angle, the fuel flow, the relative position of the unmanned aerial vehicle and the target, and the operation height wind field element as input parameters, describing the weakening characteristics of the heat flux with distance through a distance attenuation function, and introducing a side wind deflection correction factor to compensate for the influence of the wind field on the flame direction; The equivalent heat flux of the target surface is calculated, and the ignition threshold and the shortest exposure time of the obstacle material are compared to determine whether the ignition condition is met, and the minimum energy control instruction sequence is solved, including the spraying distance, the spraying angle, and the operation time, in combination with the fireproof mask and the minimum approach distance constraint; The ignition condition is expressed as: ， wherein, is the total amount of heat energy accumulated by the target surface over the time interval [0, t], is the instantaneous heat flux, is the ignition threshold of the barrier material, i.e. the minimum heat flux required to ignite the material, is the minimum exposure time, i.e. the minimum duration of time required for the material to ignite under the action of the heat flux, is the time index; The real-time monitoring of the safety constraint condition comprises calculating the minimum approach distance of the unmanned aerial vehicle and the live body, predicting the intersection state of the flame envelope and the fireproof mask, and estimating whether the temperature rise reaches the red line; Wherein, the trigger logic of the failure protection mechanism comprises: when it is monitored that the minimum approach distance is less than 3 meters, the flame envelope intersects with the fireproof mask, or the temperature rise exceeds the threshold, the system automatically performs reduced fire spraying to reduce the influence range of the flame, or immediately stops the fire spraying and triggers the sequence of extinguishing, climbing and returning along the safety corridor for safe evacuation of the unmanned aerial vehicle, and the monitoring and triggering process is based on the real-time returned unmanned aerial vehicle attitude, link quality and thermal image data.

7. The unmanned aerial vehicle fire spraying and obstacle removing method for wind power lines according to claim 6, characterized in that: The infrared thermal image review comprises collecting infrared thermal image data of the target area after the operation, comparing with the temperature baseline before the operation, and calculating the temperature distribution difference value; The formula for calculating the temperature distribution difference value is: ， wherein, is an effective temperature rise value at an image coordinate (u, v) position, is an original temperature value measured after the work at the coordinate (u, v) on the infrared thermograph, is an environmental temperature reference value measured after the work at the reference region, is an original temperature value measured before the work at the coordinate (u, v) on the infrared thermograph, is an environmental temperature reference value measured before the work at the reference region, is a pixel coordinate of the infrared thermograph; According to the allowable temperature rise threshold of the component, abnormal high temperature data caused by reflection, background interference and transient noise points are removed, the removal standard is that the average temperature rise of the high temperature area is less than 10% of the allowable threshold or the short time hot spot area is less than 1% of the total image area, the actual cooling curve is recorded, and the thermal damage is evaluated; The feature extraction and cluster analysis comprises standardizing the wind condition, terrain, obstacle type and execution parameter in the obstacle removal database to eliminate the dimension difference, extracting the main characteristic components by using the principal component analysis method, generating the scene feature vector, and performing cluster analysis based on the feature similarity to classify the data into similar scene categories, extracting the execution parameter combination with the highest success rate and the lowest energy consumption in each scene category as the strategy library entry for storage, and mapping the scene features and the recommended strategy.

8. A wind power line unmanned aerial vehicle fire spraying and obstacle clearing system, applying a wind power line unmanned aerial vehicle fire spraying and obstacle clearing method according to any one of claims 1-7, characterized in that, It comprises a data acquisition and three-dimensional reconstruction module, a risk identification and base map construction module, a visual interaction and intelligent control module, and a safe execution and strategy learning module; The data acquisition and three-dimensional reconstruction module is used for deploying a multi-source sensor and a meteorological module on an unmanned aerial vehicle, synchronously collecting multi-view visual data, three-dimensional space data and environmental meteorological data of a power transmission line corridor, time-synchronizing and external parameter calibrating all the data, obtaining an original data set with a unified time and space reference, and generating a digital twin based on a three-dimensional Gaussian splashing reconstruction technology; The risk identification and base map construction module is used for realizing automatic instance segmentation and identification of the conductor, insulator, hardware and temporary floating object on the basis of the reconstructed digital twin through semantic segmentation based on deep learning and multi-view consistent projection technology, and assigning the non-combustible, difficult-to-combust or combustible grade and allowable temperature rise threshold to each instance in combination with a material dictionary and surface state estimation, and constructing a calculable risk base map by integrating a spatial analysis method to calculate the minimum distance, obstacle volume and wind direction angle. The visual interaction and intelligent control module is used for mapping the safety boundary elements in the calculable risk base map and operation parameter suggestions to the real-time superposition of the unmanned aerial vehicle video stream through camera-world coordinate unification, forming a real-time interaction interface, and generating a minimum energy control instruction sequence based on a flame propagation behavior model integrating distance attenuation and crosswind deflection effect, in combination with an obstacle ignition criterion and safety constraints. The safe execution and strategy learning module is used for monitoring the safety constraints of the minimum approach distance, fire-prohibited mask intersection and temperature rise red line in real time during the fire spraying execution, triggering a failure protection mechanism immediately when the constraints are exceeded, and executing the logic of reduced fire spraying, work suspension or automatic return, and after the work, eliminating abnormal heat spots and recording cooling curves through infrared thermal image review, evaluating the obstacle clearing effect and thermal damage, and writing the whole process data into an obstacle clearing database, and depositing the data into a strategy library of scene-strategy mapping through feature extraction and cluster analysis. 9.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-8 when the computer program is executed by the processor. The processor executes the computer program to realize the steps of the wind power line unmanned aerial vehicle fire spraying obstacle clearing method in any one of claims 1 to 7.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to realize the steps of the wind power line unmanned aerial vehicle fire spraying obstacle clearing method in any one of claims 1 to 7.

Images