Unmanned aerial vehicle electricity testing platform carrying electric field sensor and electricity testing method

Abstract

The invention discloses an electric field sensor-carrying unmanned aerial vehicle electricity testing platform and an electricity testing method, and relates to the technical field related to electricity testing, and the platform comprises an initialization unit which is used for calling electricity testing scene data and weather environment data according to an electricity testing task instruction to carry out the flight path planning of a multi-rotor unmanned aerial vehicle; the linkage acquisition unit is used for executing electric field signal acquisition; the alignment return unit is used for returning the mapped position attitude data and the electric field signal set to the edge calculation layer; a suppression unit for performing environmental interference suppression; and the discrimination unit is used for executing electricity verification correction discrimination of the electric field signal set and the mapped position attitude data and generating an electricity verification correction discrimination result. The technical problems that in the prior art, flight path planning is not scientific, electric field signal collection is prone to being interfered, data processing and judging efficiency is low, accuracy is poor, and electricity testing work efficiency and safety are poor are solved, and the technical effects of improving electric field signal processing accuracy, electricity testing work efficiency and safety are achieved.

Description

Technical Field

[0001] The present application relates to the technical field related to electrical testing, and specifically to an unmanned aerial vehicle electrical testing platform and an electrical testing method equipped with an electric field sensor. Background Art

[0002] Detecting the energized status of transmission lines and power equipment is a key link in ensuring the stable operation of the power grid. However, traditional drone electrical testing cannot fully consider complex testing scenarios and changeable weather conditions, resulting in suboptimal drone flight paths, low efficiency in electrical testing tasks, and even possible flight accidents due to unreasonable path planning. During the electric field signal acquisition and processing process, drones are subject to electromagnetic interference, environmental electric field fluctuations, etc., which seriously affect the accuracy and reliability of electric field signal acquisition. If the interference signal cannot be effectively removed and the electric field signal cannot be corrected, the electrical test results will be biased. In addition, the existing electrical testing methods lack efficient and intelligent analysis methods in the data processing and judgment links, making it difficult to quickly and accurately process and analyze the collected data, resulting in insufficient timeliness and accuracy of the electrical test results, thereby affecting operational safety and efficiency.

[0003] Therefore, in the current relevant technologies, there are technical problems such as unscientific flight path planning, susceptibility of electric field signal acquisition to interference, low efficiency and poor accuracy of data processing and judgment, which lead to poor efficiency and safety of electrical testing operations. Summary of the Invention

[0004] This application solves the technical problems in the prior art such as unscientific flight path planning, susceptibility to interference in electric field signal acquisition, low efficiency and poor accuracy in data processing and discrimination, which lead to poor efficiency and safety of electric testing operations by providing an unmanned aerial vehicle electrical testing platform and electrical testing method equipped with electric field sensors. It achieves the technical effect of improving the accuracy of electric field signal processing, efficiency and safety of electric testing operations.

[0005] The present application provides an unmanned aerial vehicle electrical testing platform equipped with an electric field sensor, the platform comprising: an initialization unit for receiving an electrical testing task instruction, calling electrical testing scene data and weather environment data according to the electrical testing task instruction to perform multi-rotor unmanned aerial vehicle flight path planning, and establishing a three-dimensional flight path; a linkage acquisition unit for starting the multi-rotor unmanned aerial vehicle and performing operation tasks along the three-dimensional flight path. When the position information of the multi-rotor unmanned aerial vehicle meets the starting interval, the electric field sensing array integrated in the multi-rotor unmanned aerial vehicle is activated to perform electric field signal acquisition; an alignment return unit for aligning the multi-rotor unmanned aerial vehicle with the electric field sensing array in time sequence, and then returning the mapped position and posture data and the electric field signal set to the edge computing layer; a suppression unit for denoising the electric field signal set using differential denoising, modeling the background electric field, and performing environmental interference suppression; a discrimination unit for activating the deep learning channel of the edge computing layer, performing electrical test correction discrimination on the electric field signal set after environmental interference suppression and the mapped position and posture data, and generating an electrical test correction discrimination result.

[0006] In a possible implementation, the drone electrical testing platform equipped with an electric field sensor also performs the following processing: parses and obtains the electrical testing target area and operation requirements based on the electrical testing task instructions; extracts the three-dimensional point cloud of the power facilities from a preset scene database based on the electrical testing target area, integrates the three-dimensional point cloud of the power facilities with the map information, and establishes electrical testing scene data; activates the remote weather interface to read weather forecast data and establishes weather environment data; and plans the flight path of the multi-rotor drone based on the electrical testing scene data and weather environment data.

[0007] In a possible implementation, the UAV electrical testing platform equipped with an electric field sensor also performs the following processing: establishing a safety distance constraint based on the electrical testing scene data; configuring the motion posture matching constraint according to the operation requirements; using the electrical testing scene data to perform electromagnetic interference fitting and establish electromagnetic interference avoidance constraints; after initializing the trajectory planning channel using the safety distance constraint, motion posture matching constraint, and electromagnetic interference avoidance constraint, perform balanced path planning of the path and acquisition effect to establish a three-dimensional flight path.

[0008] In a possible implementation, the UAV electrical inspection platform equipped with an electric field sensor also performs the following processing: a preprocessing module is used to perform time window alignment on the electric field signal set and the mapped position and posture data after environmental interference suppression, standardize them into a unified tensor structure, and input them into the feature extraction channel; an extraction module is used to activate the feature extraction channel, use the position embedding mechanism to encode the position and posture data sequence into vector features, introduce cross-attention to perform posture fusion of the electric field time series and vector features, and establish a fused feature tensor; an execution module is used to use the fused feature tensor as input data, execute the electrical inspection correction judgment of the deep learning channel, and generate the electrical inspection correction judgment result.

[0009] In a possible implementation, the UAV electrical testing platform equipped with an electric field sensor also performs the following processing: a real-time perception module, which is used to call the lidar to perform real-time path scene perception and establish real-time perception results when the multi-rotor UAV performs an operating task along the three-dimensional flight path; an obstacle avoidance reconstruction module, which is used to perform path hazard breakthrough analysis based on the real-time perception results. If the path hazard breakthrough analysis result meets the breakthrough threshold, a reconstructed path is established, and the multi-rotor UAV flight management is performed according to the reconstructed path.

[0010] In a possible implementation, the drone electrical testing platform equipped with an electric field sensor also performs the following processing: after the multi-rotor drone reaches the starting interval, a non-target area period is established as a background sampling area, and the sampling window is automatically calibrated; in the sampling window, the sampling points are divided into three-dimensional grid areas according to spatial coordinates, spatial domain modeling is performed, and an initial background electric field distribution map is generated; continuous time series sliding window processing is performed on the initial background electric field distribution map, and background electric field modeling is completed according to the sliding window processing results.

[0011] In a possible implementation, the UAV electrical testing platform equipped with an electric field sensor also performs the following processing: an early warning unit is used to trigger analysis of the warning level according to the electrical test correction judgment result and establish a trigger warning instruction; a reporting unit is used to establish a scene early warning output mark in the electrical test scene data according to the trigger warning instruction.

[0012] The present application also provides a method for testing the electrical status of a drone equipped with an electric field sensor, the method comprising: receiving an electrical testing task instruction, calling electrical testing scene data and weather environment data according to the electrical testing task instruction to plan the flight path of a multi-rotor drone, and establishing a three-dimensional flight path; starting the multi-rotor drone, performing the operation task along the three-dimensional flight path, and when the position information of the multi-rotor drone meets the starting interval, activating the electric field sensing array integrated in the multi-rotor drone to perform electric field signal acquisition; after time-series alignment of the multi-rotor drone and the electric field sensing array, transmitting the mapped position and attitude data and the electric field signal set back to the edge computing layer; after denoising the electric field signal set using differential denoising, modeling the background electric field, and performing environmental interference suppression; activating the deep learning channel of the edge computing layer, performing electrical testing correction and discrimination of the electric field signal set after environmental interference suppression and the mapped position and attitude data, and generating an electrical testing correction and discrimination result.

[0013] The electric field sensor-equipped UAV electric test platform and electric test method proposed in this application include an initialization unit for calling the electric field test scene data and weather environment data according to the electric field test task instruction to plan the flight path of the multi-rotor UAV; a linkage acquisition unit for performing electric field signal acquisition; an alignment return unit for returning the mapped position and attitude data and the electric field signal set to the edge computing layer; a suppression unit for performing environmental interference suppression; and a discrimination unit for performing electric field correction discrimination between the electric field signal set and the mapped position and attitude data to generate an electric field correction discrimination result. This solves the technical problems existing in the prior art, such as unscientific flight path planning, susceptibility to interference in electric field signal acquisition, low efficiency and poor accuracy in data processing and discrimination, which lead to poor efficiency and safety in electric field test operations, and achieves the technical effect of improving the accuracy of electric field signal processing, the efficiency and safety of electric field test operations. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] To more clearly illustrate the technical solutions of the embodiments of the present disclosure, the accompanying drawings of the embodiments of the present disclosure are briefly introduced below. Flowcharts are used in this application to illustrate the operations performed by the systems according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed in precise order. Instead, various steps may be processed in reverse order or simultaneously as needed. Furthermore, other operations may be added to these processes, or one or more operations may be removed from these processes.

[0015] Figure 1 Schematic diagram of the structure of the UAV electrical testing platform equipped with electric field sensors provided in an embodiment of the present application.

[0016] Figure 2 A flow chart of a method for testing the electrical status of a drone equipped with an electric field sensor provided in an embodiment of the present application.

[0017] Description of the accompanying drawings: initialization unit 10, linkage collection unit 20, alignment return unit 30, suppression unit 40, and judgment unit 50. DETAILED DESCRIPTION

[0018] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below.

[0019] In order to make the purpose, technical solutions and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings. The described embodiments should not be regarded as limiting this application. All other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0020] In the following description, reference is made to “some embodiments”, which describes a subset of all possible embodiments, but it will be understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict, and the terms “first\second” involved are merely used to distinguish similar objects and do not represent a specific ordering of the objects. The terms “including” and “having” and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or modules that are not clearly listed or that are inherent to these processes, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this application belongs. The terms used herein are for the purpose of describing the embodiments of this application only.

[0021] The present application embodiment provides a UAV electrical inspection platform equipped with an electric field sensor, such as Figure 1 As shown, the platform includes: The initialization unit 10 is used to receive an electrical test task instruction, call the electrical test scene data and weather environment data according to the electrical test task instruction to perform multi-rotor UAV flight path planning, and establish a three-dimensional flight path.

[0022] Preferably, the power test task instruction includes key information such as the power test target (such as a certain section of the transmission line to be tested, a certain equipment in the substation), the power test operation type (live detection / power outage maintenance power test), the safety level (such as the minimum safe distance from the live body), the detection accuracy requirements, the operation time requirements, etc., which are used to clarify the operation object of the drone (such as the tower number and cable position to be tested), define the safety constraint boundary (such as the drone must stay in a safe area 5 meters away from the high-voltage line) and determine the task priority (such as emergency fault power testing must prioritize planning the shortest path); the drone receives the power test information from the power test platform After the mission instruction, the electrical test scene data and weather environment data are called according to the electrical test mission instruction to plan the flight path of the multi-rotor UAV. Among them, the electrical test scene data is used to construct a three-dimensional scene grid containing spatial obstacle constraints (such as avoiding buildings) and electrical test points (such as key nodes of equipment); the multi-rotor UAV is composed of multiple (commonly 3, 4, 6, 8) rotor motors and propellers, and flight attitude control is achieved through the motor speed difference. For example, all rotors accelerate / decelerate synchronously, changing the total lift to achieve vertical ascent or descent, and horizontal movement is achieved by generating torque by adjusting the speed difference of different rotors.

[0023] Preferably, the electrical inspection scene data includes a three-dimensional model of the power facility (such as tower location, conductor direction), spatial coordinates of the power facility (tower height, line direction), topographic data (elevation map, obstacle distribution), obstacle distribution (trees, buildings, towers) and surrounding electromagnetic environment distribution (substation, communication base station location); as well as structural parameters of the electrical equipment to be inspected (such as conductor radius, insulator distribution), and historical data of the energized state (such as recent fault records), which are used to determine the key locations that the drone needs to approach for inspection (such as the discharge-prone area at the insulator connection); and then integrate weather environment data, where the weather environment data includes real-time meteorological parameters (wind speed, wind direction, temperature, humidity), weather forecast data (rainfall probability, lightning warning), historical climate characteristics, etc., which are used to constrain the drone's mission execution. For example, when the wind speed exceeds the drone's maximum wind resistance level (such as level 6 wind), the path is automatically adjusted to avoid high wind speed areas or suspend operations; when planning the path, avoid flying for a long time in rainy days or high humidity environments to avoid affecting the accuracy of the electric field sensor.

[0024] Preferably, the data is initialized, including converting the geographic coordinates of the power test scene (such as WGS84) into a local coordinate system available for UAV navigation (such as the Northeast Sky coordinate system) to ensure seamless connection between the position data and the flight control instructions; eliminating outliers (such as coordinate jump points in GIS data), and supplementing missing environmental data (such as wind speed estimation in areas without meteorological stations) through interpolation algorithms; then using the Dijkstra algorithm to search for the optimal path from the starting point (take-off point) to the end point (task end point) based on the scene grid model, while avoiding obstacles. Specifically, the flight constraint condition is set to a flight altitude of at least 10 meters above ground obstacles (such as trees) and within 5 meters below the height of the transmission line (for easy detection). The path needs to reserve a sufficient turning radius to prevent the UAV from losing control of its attitude due to sharp turns; dynamically adjust the path in combination with real-time meteorological data, for example, in the event of an emergency During strong winds, temporary hovering points are added to the path to stabilize flight. At the same time, a zigzag or spiral local path is generated for electrical inspection points (such as insulator strings) to ensure that the drone maintains a stable posture above each inspection point, facilitating multi-angle signal collection by the electric field sensor. The final three-dimensional flight path is output, including a timed sequence of path points with spatial coordinates (x, y, z), flight speed, and attitude angles (pitch / roll / yaw). The path flight is simulated through a drone simulation platform (such as Gazebo) to verify whether there is a collision risk, whether the sensor field of view covers all inspection points, and whether the flight energy consumption is within the drone's endurance range. This ensures that the drone approaches the inspection target at the optimal angle and distance. For example, flying in a horizontal posture directly above the transmission line allows the electric field sensor to collect signals vertically downward, thereby ensuring the drone's balanced safety, efficiency, and detection accuracy in complex environments.

[0025] Furthermore, the specific configuration of the initialization unit 10 also includes: obtaining the target area for electrical testing and the operation requirements according to the analysis of the electrical testing task instructions; extracting the three-dimensional point cloud of the power facilities from a preset scene database based on the electrical testing target area, fusing the three-dimensional point cloud of the power facilities with the map information, and establishing electrical testing scene data; activating the remote weather interface to read the weather forecast data and establish weather environment data; and planning the flight path of the multi-rotor UAV according to the electrical testing scene data and weather environment data.

[0026] Preferably, the electrical inspection task instructions are parsed to obtain the electrical inspection target area and operation requirements, wherein the electrical inspection target area includes the geographical coordinate range of the power station or equipment to be inspected (such as a rectangular area framed by longitude and latitude), and the operation requirements include safety distance, detection accuracy (such as electric field sensor sampling interval), priority and other constraints; based on the electrical inspection target area, the three-dimensional point cloud of the power facilities is extracted from the preset scene database, wherein the preset scene database stores the power facility point cloud generated by laser radar (LiDAR) scanning (such as pole tower, conductor, insulator three-dimensional model), map information includes GIS terrain data and electronic map (including the distribution of obstacles such as buildings and trees), and then the three-dimensional point cloud of the power facility is fused with the map information, that is, the point cloud data and the map data are aligned through the ICP (iterative closest point) algorithm, and then semantic labels (such as "conductor", "insulator", "obstacle") are added to the point cloud to identify the key areas of electrical inspection, and finally the three-dimensional space is discretized into a voxel grid to mark the flyable area and the no-fly zone, such as the safety buffer zone around the energized equipment.

[0027] Preferably, a remote weather interface is activated to read weather forecast data and establish weather environment data. Specifically, the OpenWeatherMap remote weather API is activated to obtain real-time data (such as wind speed, wind direction, temperature, humidity, and precipitation probability) for the target area and weather forecast data for the next 24 hours (such as typhoon paths and severe convective weather warnings). Weather parameters are then mapped into three-dimensional space. For example, a "wind field model" is established to represent the wind speed vector distribution at different altitudes, and "risk areas" are defined, such as areas with wind speeds greater than level 6 being marked as no-fly zones. Multi-rotor UAV flight path planning is performed based on the electrical testing scenario data and weather environment data. Specifically, a three-dimensional spatial model is constructed that incorporates scenario constraints and dynamic risks. A hierarchical planning strategy is used to plan the path, including using RRT (rapidly expanding randomized trees) to rapidly explore feasible regions in high-dimensional space and generate an initial global path. The initial global path is smoothed based on the scenario and weather environment data to obtain a flight path. Sudden weather changes (such as a sudden increase in wind speed) or the detection of new obstacles trigger local path replanning, ultimately generating a flight path that meets safety, efficiency, and quality requirements for the UAV electrical testing operation.

[0028] Furthermore, the specific configuration of the initialization unit 10 also includes establishing a safety distance constraint based on the electrical test scene data; configuring an action posture matching constraint according to the operation requirements; using the electrical test scene data to perform electromagnetic interference fitting and establish an electromagnetic interference avoidance constraint; after initializing the trajectory planning channel using the safety distance constraint, action posture matching constraint, and electromagnetic interference avoidance constraint, executing a balanced path planning of the path and acquisition effect to establish a three-dimensional flight path.

[0029] Preferably, the three-dimensional point cloud of power facilities (such as the spatial coordinates and dimensions of transmission lines and substation equipment) is extracted from the power test scene data, and combined with map information (terrain altitude, building distribution, etc.) to construct a three-dimensional spatial model including the spatial location of the power facilities and the surrounding environment. Then, safety distance constraints are imposed, including setting the minimum safety distance between the drone and the live body for power equipment of different voltage levels (such as 10kV, 110kV, 500kV, etc.) (for example, 10kV lines need to be maintained at ≥0.7 meters); at the same time, considering the drone's own size (such as rotor wheelbase, fuselage length) and flight shake error, a "no-fly zone" (such as a sphere with a radius of 2 meters centered on the equipment) is delineated around the power facilities to avoid drone collisions.

[0030] Preferably, according to the specific requirements in the electrical test task instructions (such as "fixed-point electrical test", "inspection along the line", "multi-angle scanning", etc.), the action posture that the drone needs to perform (such as hovering, pitching, side flying, circling, etc.) is determined as the operation requirement, and then the action posture matching constraints are configured, including ensuring that the flight posture can make the sensor face the target electrical test position (such as below the line, to the side of the equipment), avoiding signal collection blind spots due to posture deviation, and flight stability requirements. For example, when hovering for electrical testing, the drone's translation speed (such as ≤0.5m / s) and roll / pitch angle (such as ≤15°) are limited to ensure the accuracy of electric field signal collection; when circling, a uniform circular trajectory is planned to avoid centrifugal force causing attitude loss of control.

[0031] Preferably, the distribution and operating status of power facilities in the electrical test scenario data (such as current size and magnetic field strength) are used to fit the electromagnetic field distribution around the power facilities through electromagnetic simulation algorithms (such as finite element method and moment method), identify strong electromagnetic interference areas (such as near insulators and busbar connections), and then configure electromagnetic interference avoidance constraints, including interference source isolation, that is, setting an "avoidance buffer zone" (such as a radius of 1 meter) around the strong electromagnetic interference area to prevent the drone from entering the area and causing distortion of the electric field sensor signal (such as the noise amplitude exceeds 30% of the effective value of the signal), and when planning the drone's flight path, give priority to paths with weaker electromagnetic interference, or adjust the flight altitude / angle to "pass sideways" rather than "pass directly through" the interference area to reduce the interference duration.

[0032] Preferably, the trajectory planning channel is initialized using safety distance constraints, motion posture matching constraints, and electromagnetic interference avoidance constraints, including converting safety distance constraints, motion posture matching constraints, and electromagnetic interference avoidance constraints into boundary conditions that can be recognized by the algorithm, and inputting them into the path planning algorithm. The three-dimensional space (X / Y / Z axis) of the target area of ​​the electrical inspection is then divided into grids or nodes, each node containing position coordinates, safety status (whether it touches the restricted area), electromagnetic interference level, reachable posture set, etc.; then, balanced path planning of the path and acquisition effect is performed. Specifically, first, ensure that all planned paths meet the safety distance and electromagnetic interference avoidance constraints, and eliminate paths that cross restricted areas or high interference areas; in the safe path, select the path that can make the electric field sensor The system plans a path with the best viewing angle of the device (such as perpendicular to the direction of the wire), the most stable flight attitude, and the longest acquisition time window. For example, a zigzag or spiral trajectory is planned to enable the drone to sample the same device multiple times at different heights and angles. Through path pruning (deleting redundant waypoints) and curvature optimization (smoothing trajectory inflection points), the flight distance is shortened and energy consumption is reduced. For example, the paths of adjacent power test points are merged to avoid repeated flights. At the same time, dynamic adjustments are made based on weather and environmental data. For example, when weather and environmental data (such as wind speed > 5m / s, rainfall probability > 70%) triggers the preset threshold, the path height is automatically adjusted (such as raising it to more than 10 meters) or switched to "emergency inspection mode" to reduce flight risks and ultimately obtain the three-dimensional flight path of the multi-rotor drone.

[0033] The linkage collection unit 20 is used to start the multi-rotor drone and perform the operation task along the three-dimensional flight path. When the position information of the multi-rotor drone meets the starting interval, the electric field sensing array integrated in the multi-rotor drone is activated to perform electric field signal collection.

[0034] Preferably, after receiving the command, the multi-rotor drone is started, and the drone flight control module loads the pre-planned three-dimensional flight path. At this time, the drone's power module (motor, propeller) starts to operate, and the attitude control module (such as gyroscope, accelerometer) is initialized to ensure that the drone is in a stable hovering state; then the operation task is performed along the three-dimensional flight path, that is, the drone gradually adjusts its own state according to the position, speed and attitude information in the path point sequence. For example, according to the path planning requirements, the drone accelerates in certain sections to shorten the operation time, and slows down and adjusts its attitude when approaching the target area to ensure a smooth transition to the detection point; and during the flight, the flight control module receives sensor data such as GPS and inertial measurement unit (IMU) in real time, corrects flight deviations, and ensures that the drone always moves along the preset path.

[0035] Preferably, the starting interval is a pre-set spatial range based on the position of the electrical test target, the safe distance between the drone and the target, and the effective working range of the sensor, which is used to define the reasonable time for electric field signal collection. If the position information of the multi-rotor drone meets the starting interval, for example, when the drone flies to a rectangular area 5 meters below and 10 meters in front and behind the power transmission line to be tested, it is determined that the position information meets the starting interval requirements, and then the electric field sensing array integrated in the multi-rotor drone is self-checked (such as checking whether the sensor power supply and sensitivity are normal) and activated. Specifically, an activation command is sent to the electric field sensing array, wherein the electric field sensing array is composed of multiple high-precision electric field sensors, which can detect the electric field strength and distribution around the power equipment. Then, the electric field sensors are used to collect electric field signals at a set sampling frequency (such as 100 times per second) and record the corresponding timestamp, drone position and attitude data. Multiple sensors in the electric field sensing array need to be started synchronously to ensure that the collected electric field signals have temporal and spatial consistency and can truly reflect the charged condition of the power equipment.

[0036] Furthermore, the specific configuration of the linkage acquisition unit 20 also includes a real-time perception module, which is used to call the laser radar to perform real-time path scene perception and establish real-time perception results when the multi-rotor drone performs an operating task along the three-dimensional flight path; an obstacle avoidance reconstruction module, which is used to perform path hazard breakthrough analysis based on the real-time perception results. If the path hazard breakthrough analysis result meets the breakthrough threshold, a reconstructed path is established, and the multi-rotor drone flight management is performed according to the reconstructed path.

[0037] Preferably, the real-time perception module relies on the high-frequency scanning capability of the laser radar (LiDAR). By actively emitting laser beams and receiving reflected signals, it constructs three-dimensional scene information in the flight path, thereby enabling the drone to "see" the surrounding environment clearly during flight and avoid collisions with obstacles (such as trees, buildings, other power facilities, etc.). Specifically, the laser radar on the drone emits laser pulses in all directions at a fixed frequency (such as 100,000 times per second), measures the distance from each laser point to the obstacle, and forms a dense "point cloud" data to reflect the spatial structure of the drone's surrounding environment (such as object position, shape, distance, etc.) in real time; then, noise is removed through filtering algorithms (such as outlier removal and voxel filtering), and then through coordinate system conversion (converting the laser radar coordinate system to the drone's global coordinate system), it is integrated with the preset electrical inspection scene data (such as the three-dimensional point cloud of power facilities) to form a real-time environment map, that is, the real-time perception result; the processed point cloud data is compared with the preset safety distance threshold to mark the dangerous areas in the path (such as obstacles less than 2 meters from the drone) and output in the form of visual or structured data (such as obstacle position coordinates and threat level) for call by the obstacle avoidance reconstruction module.

[0038] Preferably, the obstacle avoidance reconstruction module mainly determines whether the current path is safe based on the real-time environmental perception results, and generates a new flight path when necessary to ensure that the drone can complete the electrical testing task and avoid obstacles. Specifically, based on the real-time perception results, the distance, relative speed and other parameters between the obstacle and the current path of the drone are calculated to evaluate the collision risk. For example, if the obstacle is within 1 meter in front of the drone and the relative speed is greater than 0.5m / s, it is judged as "high-risk breakthrough"; a preset breakthrough threshold (such as a safe distance threshold, a time-to-collision threshold) is set. When the obstacle distance is less than the threshold or the expected collision time is less than the preset value, the path reconstruction is triggered, and the safety distance constraint (maintaining at least 3 meters away from power facilities), the action posture constraint (the maximum turning angle of the drone does not exceed 60°), and the electromagnetic interference avoidance constraint are met. Based on preset conditions such as beam (avoiding strong electromagnetic field areas), a path planning algorithm such as fast search random tree may be used. With the current UAV position as the starting point and the next safe point on the original path as the end point, a detour path is generated outside the obstacle avoidance area. For example, if there are trees blocking the front, the path may be offset by 5 meters to the left or right, and then return to the original path after bypassing the obstacle. Finally, the multi-rotor UAV flight management is performed according to the reconstructed path, including sending the reconstructed path to the UAV flight control module, controlling the UAV to fly along the new path by adjusting the motor speed, servo angle, etc., and during the detour, the real-time perception module continuously monitors the environmental changes of the new path. If an obstacle is detected again, the path reconstruction will continue until the UAV returns to the original path or completes the mission, thereby ensuring the efficiency of the UAV electrical inspection operation and the safety of the flight mission.

[0039] The alignment return unit 30 is used to align the multi-rotor drone with the electric field sensing array in time sequence, and then return the mapped position and posture data and electric field signal set to the edge computing layer.

[0040] Preferably, the multi-rotor drone and the electric field sensing array are time-synchronized to ensure the accuracy of data collection. If the electric field signals (such as voltage amplitude and phase) collected by the electric field sensing array do not completely correspond to the drone's position and attitude (such as longitude, latitude, altitude, and heading angle), fault location deviation may occur. Specifically, when the flight control module generates position and attitude data (such as longitude, latitude, altitude, roll angle, pitch angle, and yaw angle), it adds an accurate timestamp (such as UTC time, accurate to microseconds) to each data item. At the same time, when the electric field sensing array collects electric field signals (such as the electric field strength and frequency components of each channel), it synchronously marks the same timestamp and establishes a time axis mapping relationship. The electric field signal is then bound to the spatial state (position and attitude) of the drone at the time of collection. For example, the abnormal point of the electric field signal must correspond to the drone's current position (such as near the K32 tower of the transmission line) and attitude (such as the fuselage tilted 3° to the right, and the sensor facing the wire) in order to accurately locate the abnormal source. The position and attitude data include longitude, latitude, altitude, roll angle, pitch angle, yaw angle, as well as flight speed, acceleration, etc. The two types of data are encapsulated into data frames in chronological order. Each data frame contains a timestamp, position and attitude data, and the original value of the electric field signal. Finally, the encapsulated data frames are sent to the edge computing layer in real time via wireless transmission methods (such as 4G / 5G, Wi-Fi, and digital radio). It is usually deployed on a server or embedded device (such as an industrial-grade gateway) close to the work site. It has the characteristics of low latency and high bandwidth, and can quickly receive and process data, thereby ensuring data analysis accuracy and processing efficiency.

[0041] The suppression unit 40 is configured to perform denoising on the electric field signal set using differential denoising, model the background electric field, and perform environmental interference suppression.

[0042] Preferably, the continuity of the electric field signal in time or space is utilized to identify and suppress random noise by calculating the difference (difference) between adjacent data points. For example, the change of the real electric field signal is usually smooth (such as gradually attenuating with increasing distance), while random noise (such as sensor electronic noise, radio frequency interference) is irregular and will be significantly amplified after the difference calculation, so that it can be detected and eliminated. This may include differentiating the signals collected by the same sensor at consecutive moments (such as the signal value at the current moment minus the signal value at the previous moment). If the difference exceeds a preset threshold (such as the maximum possible value of normal electric field change), it is determined to be a noise point and replaced with the average of adjacent effective values ​​or interpolation method. In addition, the signals of adjacent sensors in the electric field sensing array are differentiated (such as the signal difference between sensor A and sensor B). If the difference exceeds the physical law (such as the electric field strength of adjacent sensors at the same position should be close to the same), it is determined that at least one of the sensors is interfered with by noise, and correction is performed through multi-sensor data fusion (such as voting method, weighted average).

[0043] Preferably, the background electric field refers to the non-target electric field that exists stably in the operating environment (such as the earth's electrostatic field, the power frequency electric field of the surrounding non-power-off equipment, the atmospheric electric field in the natural environment, etc.), which is not the target signal to be detected (such as the residual charge electric field of the power-off equipment), but will produce superimposed interference on the collected data. Then, the background electric field is modeled. Specifically, when the UAV is not close to the target equipment (such as in the initial flight phase outside a safe distance), the electric field signal is collected for a period of time. Through statistical analysis (such as mean, variance, and spectrum analysis), based on electromagnetic field theory (such as Maxwell's equations), combined with the geographical information of the operating scene (such as the distribution of surrounding transmission lines, the location of buildings) and meteorological conditions (such as humidity and temperature affecting air conductivity), the background electric field is modeled. Through simulation calculation, a normal model of the background electric field is established. The electric field signal of each collection point is then compared with the corresponding background electric field signal. That is, the background electric field value predicted by the model is subtracted from the electric field signal of each collection point to obtain the net electric field signal (that is, the superposition of the target electric field and the residual noise). The net electric field signal is then subjected to spectrum analysis to identify whether there is any unmodeled environmental interference (such as occasional radio frequency signals and electromagnetic pulses generated by vehicle engines), which usually manifests as spikes at specific frequencies (such as the 2.4GHz frequency band for mobile phone communications). Digital filters (such as low-pass filters and band-stop filters) are then used to suppress interference at specific frequencies, thereby enhancing the signal-to-noise ratio of the electric field signal and avoiding mistaking the background electric field for the target signal to improve the accuracy of electrical testing.

[0044] Furthermore, the specific configuration of the suppression unit 40 also includes, after the multi-rotor drone reaches the starting interval, establishing a non-target area period as a background sampling area and automatically calibrating the sampling window; dividing the sampling points into three-dimensional grid areas according to spatial coordinates in the sampling window, performing spatial domain modeling, and generating an initial background electric field distribution map; performing continuous time series sliding window processing on the initial background electric field distribution map, and completing background electric field modeling based on the sliding window processing results.

[0045] Preferably, the background sampling area refers to the "non-target area" (such as an open area or an area known to be free of strong electric field interference) at a safe distance before the drone reaches the target operating area (such as a power outage line to be inspected). After the multi-rotor drone enters the starting range (such as 50 meters away from the target device), a non-target area period is established as the background sampling area, and the sampling window is automatically calibrated, including automatically triggering a fixed-time sampling (such as 30 seconds) to avoid unstable flight posture (such as shaking during takeoff). At the same time, the drone is required to maintain a relatively stable flight state (such as hovering or uniform straight flight) during the sampling period; then all sampling points in the sampling window are pressed The spatial coordinates (x, y, z) are divided into several small cubic grids (e.g., each grid has a side length of 1 meter). Each grid contains multiple sampling points (the coordinate error of points within the same grid is less than half the grid side length). Spatial domain modeling is performed. That is, by statistically analyzing the electric field signal characteristics (such as mean and variance) of the sampling points within each grid, the distribution of the electric field intensity in three-dimensional space is established, and an initial background electric field distribution map is obtained (for example, showing that the electric field near the ground is stronger and the electric field at high altitude is weaker). The mean electric field value within the same grid represents the background electric field baseline value of the area, and the variance reflects the stability of the electric field in the area. A small variance indicates a uniform background electric field, while a large variance may indicate the presence of a local interference source.

[0046] Preferably, a continuous time series sliding window processing is performed on the initial background electric field distribution map, that is, the time series data of the sampling window is divided into multiple 5-second sub-windows (such as window 1: 0 to 5 seconds, window 2: 3 to 8 seconds, window 3: 6 to 11 seconds, and so on), each sub-window contains about 500 sampling points (assuming the sampling frequency is 100 Hz), and the dynamic characteristics of the background electric field over time (such as the slow fluctuation of the electrostatic field caused by weather changes, or the periodic change caused by power frequency interference) are captured by analyzing the changes in the electric field distribution in different sub-windows; then, the three-dimensional grid division and spatial domain construction are repeated for each sub-window according to the sliding window results. The background electric field distribution map of each sub-window is obtained by modeling, and the changes in the electric field strength values ​​of the same grid in different sub-windows are compared. If the difference is less than a preset threshold (such as ±10%), the background electric field is considered stable, and the mean of all sub-windows is taken as the final value. If the difference is large (such as exceeding ±30%), it is determined that there is time-varying interference in the area (such as the electromagnetic signal of a moving vehicle). Then, more sub-window data or outlier elimination corrections are made, and finally the analysis results of all sub-windows are integrated to establish a dynamic background electric field model that includes spatial distribution and temporal characteristics, so as to facilitate real-time elimination of environmental interference and accurate identification of target electric fields (such as residual charge of power-off equipment).

[0047] The discrimination unit 50 is used to activate the deep learning channel of the edge computing layer, perform electrical detection and correction discrimination on the electric field signal set after environmental interference suppression and the mapped position and posture data, and generate an electrical detection and correction discrimination result.

[0048] Preferably, the deep learning channel of the edge computing layer is activated, that is, a pre-trained neural network model (such as a convolutional neural network CNN or a recurrent neural network RNN) is used to automatically identify the features in the electric field signal, and determine whether the currently detected electric field belongs to the normal electric field of the target charged body (such as a transmission line), or whether there is an interference signal that needs to be corrected (such as electromagnetic noise from other equipment); then the electric field signal set after environmental interference suppression and the mapped position and posture data are subjected to electrical inspection and correction. Specifically, based on the input data, it is determined whether the currently detected electric field meets the electric field characteristics of the target charged body (such as the law that the electric field strength around the high-voltage transmission line decays with distance), and the output is whether a charged body is detected (yes / no), the electric field Whether the intensity exceeds the safety threshold and the signal credibility score (such as probability value); if there is an abnormality in the signal (such as a large deviation between the signal strength and the theoretical value, and the spatial distribution is not in line with expectations), correction and judgment will be performed, which may include adjusting the sensor calibration parameters of the electric field sensing array (such as sensitivity, offset), correcting the mapping relationship between the drone position and the electric field signal (such as coordinate system conversion error) and eliminating residual non-target interference (such as the model recognizes a certain type of specific noise pattern). Finally, the electrical test correction judgment result is output, including the judgment conclusion (such as "normal", "charged body detected", "signal abnormality requires correction") and correction suggestions (such as sensor calibration instructions, flight path fine-tuning prompts), so as to ensure the accuracy and safety of the drone electrical test results.

[0049] Furthermore, the specific configuration of the discrimination unit 50 also includes a preprocessing module, which is used to perform time window alignment on the electric field signal set and the mapped position and posture data after environmental interference suppression, standardize them into a unified tensor structure, and input them into a feature extraction channel; an extraction module, which is used to activate the feature extraction channel, use the position embedding mechanism to encode the position and posture data sequence into vector features, introduce cross-attention to perform posture fusion of the electric field time series and the vector features, and establish a fused feature tensor; an execution module, which is used to use the fused feature tensor as input data, execute the electrical test and correction discrimination of the deep learning channel, and generate the electrical test and correction discrimination result.

[0050] Preferably, the preprocessing module is used to perform time window alignment on the electric field signal set after environmental interference suppression and the mapped position and posture data, that is, based on the timestamp of the electric field signal, the position and posture data are divided into time windows (such as 1 second), and the two types of data in each time window are ensured to be strictly corresponding in the time dimension by filling or interpolation, so as to avoid "time and space dislocation" caused by sampling frequency differences or transmission delays. For example, the electric field signal is sampled at 100Hz and the position data is updated at 20Hz. The position data needs to be synchronized to 100Hz through an interpolation algorithm (such as linear interpolation). , so that each electric field signal point corresponds to an accurate position and posture; then the electric field signal is normalized (such as scaling to the range of -1 to 1), and the position and posture data (such as latitude and longitude, angle) are transformed (such as converting into relative coordinates with the target device as the origin). The processed data is organized into multi-dimensional tensors, such as batch dimension, time dimension (time series within each time window, such as 100 time points in 1 second) and feature dimension (number of electric field signal channels + position and posture parameters, such as 3D position + 3D posture = 6 dimensions), and finally input into the feature extraction channel.

[0051] Preferably, the extraction module is used to activate the feature extraction channel and use the position embedding mechanism to encode the position posture data sequence into a vector feature. That is, an embedding layer (such as a neural network layer) is used to map each position posture parameter (such as longitude and latitude) into a high-dimensional vector (such as 128 dimensions). Each element in the vector represents the potential feature of the parameter (such as "close to the target device" and "in a high-risk area"). Then, cross attention is introduced to fuse the electric field time series and vector features. Specifically, the time series features of the electric field signal (such as waveform trends and abnormal pulses) are extracted through a one-dimensional convolutional neural network (1D-CNN) or a recurrent neural network (RNN). , the position embedding vector is processed through a fully connected network or Transformer encoder to generate features containing spatial semantics; then the electric field time series feature is used as the "query", and the position feature is used as the "key" and "value", and the similarity (such as dot product) between each electric field time point and the position feature is calculated through the cross-attention mechanism to generate attention weights. Then, the position feature is fused into the electric field time series according to the weight to obtain a fused feature tensor, which also contains the time dependency of the electric field signal (such as fluctuations over time) and the spatial context of the position posture (such as the relative position of the drone and the target device).

[0052] Preferably, the execution module is used to use the fused feature tensor as input data to execute the electrical test and correction judgment of the deep learning channel, including electrical test judgment and correction judgment. Specifically, the electrical test judgment is used to determine whether the current electric field signal belongs to the normal electric field of the target charged body (such as whether there is residual charge in the line after a power outage), and the correction judgment is used to determine whether correction is required when the signal is abnormal (such as sensor deviation, position solution error), and locate the correction direction (such as adjusting sensor sensitivity, correcting flight attitude), and then output category labels (such as "normal", "charged", "interference"), correction instructions (such as "calibrate sensor channel 1", "adjust flight altitude +2 meters") or confidence scores (such as "charged probability 95%", "interference confidence 80%"), to achieve full process automation from data preprocessing to judgment result output, reduce manual intervention, meet the real-time requirements of drone electrical testing (such as millisecond-level response of the edge computing layer), thereby helping to evaluate the safety status of power equipment.

[0053] Furthermore, the UAV electrical testing platform equipped with an electric field sensor also includes an early warning unit for triggering and analyzing the warning level according to the electrical test correction judgment result, and establishing a trigger early warning instruction; and a reporting unit for establishing a scene early warning output mark in the electrical test scene data according to the trigger early warning instruction.

[0054] Preferably, the early warning unit is used to perform trigger analysis of the early warning level according to the electrical test and correction judgment result, that is, to map the electrical test and correction judgment result to different early warning levels, as shown in Table 1: Table 1 Warning level mapping table

[0055] The warning level is automatically matched through the rule engine (such as if-else conditional judgment). For example, if the judgment result is "powered" and the distance between the drone and the equipment is less than the safety threshold, a level one warning is triggered; if the judgment result is "interference" and the confidence level is greater than 70%, a level two warning is triggered. Then, trigger warning instructions are established, that is, operation instructions corresponding to each warning level are generated. For example, for a level one warning, "immediately hover and return home" and "send an emergency stop signal"; for a level two warning, "adjust the flight path to a safe distance and retest" and "extend the detection time to 60 seconds"; for a level three warning, "record sensor number CH2 abnormality and recommend calibration after operation."

[0056] Preferably, the reporting unit is used to establish a scene early warning warning label in the power inspection scene data according to the trigger warning instruction, that is, according to the drone position attitude data (longitude, latitude, altitude) when the warning is triggered, the specific location is marked in the power inspection scene data (such as GIS map, three-dimensional model of power facilities). For example, the warning point is marked with a red icon on the two-dimensional map to display "live abnormality", and the abnormal equipment (such as tower number K32) is selected with a highlighted color in the three-dimensional point cloud model; finally, the warning label is displayed graphically on the monitoring platform of the edge computing layer, for example, a red warning window pops up, accompanied by sound and light alarms, or an animation is used to simulate the risk area in the three-dimensional scene (such as a flashing red halo around live equipment), thereby improving the timeliness and accuracy of risk response and the efficiency of emergency response, thereby ensuring the intelligence level and safety of power equipment operation and maintenance.

[0057] In the above, refer to Figure 1 The UAV electrical inspection platform equipped with an electric field sensor according to an embodiment of the present invention is described in detail. Figure 2 A method for testing the electrical state of a drone equipped with an electric field sensor according to an embodiment of the present invention is described.

[0058] The method of testing the electricity of drones equipped with electric field sensors, such as Figure 2 As shown, the method includes: receiving an electrical test task instruction, calling the electrical test scene data and weather environment data according to the electrical test task instruction to plan the flight path of the multi-rotor UAV, and establishing a three-dimensional flight path; starting the multi-rotor UAV, performing the operation task along the three-dimensional flight path, and when the position information of the multi-rotor UAV meets the starting interval, activating the electric field sensing array integrated in the multi-rotor UAV to perform electric field signal acquisition; after the multi-rotor UAV and the electric field sensing array are time-series aligned, the mapped position and posture data and the electric field signal set are transmitted back to the edge computing layer; after the electric field signal set is denoised using differential denoising, the background electric field is modeled and environmental interference suppression is performed; the deep learning channel of the edge computing layer is activated, and electrical test correction judgment is performed on the electric field signal set after environmental interference suppression and the mapped position and posture data, and an electrical test correction judgment result is generated.

[0059] In one possible implementation, the method for testing electricity using a drone equipped with an electric field sensor also includes: obtaining the target area for testing electricity and the operating requirements based on the test task instructions; extracting a three-dimensional point cloud of power facilities from a preset scene database based on the test target area, fusing the three-dimensional point cloud of power facilities with map information, and establishing test scene data; activating a remote weather interface to read weather forecast data and establish weather environment data; and planning the flight path of a multi-rotor drone based on the test scene data and weather environment data.

[0060] In one possible implementation, the method for testing the electrical state of an unmanned aerial vehicle equipped with an electric field sensor further includes: establishing a safety distance constraint based on the electrical test scene data; configuring an action posture matching constraint according to operation requirements; performing electromagnetic interference fitting using the electrical test scene data to establish an electromagnetic interference avoidance constraint; after initializing the trajectory planning channel using the safety distance constraint, action posture matching constraint, and electromagnetic interference avoidance constraint, executing a balanced path planning of the path and acquisition effect to establish a three-dimensional flight path.

[0061] In one possible implementation, the method for testing the electrical status of a drone equipped with an electric field sensor also includes: a preprocessing module, which is used to perform time window alignment on the electric field signal set and the mapped position and posture data after environmental interference suppression, standardize them into a unified tensor structure, and input them into a feature extraction channel; an extraction module, which is used to activate the feature extraction channel, use the position embedding mechanism to encode the position and posture data sequence into vector features, introduce cross-attention to perform posture fusion of the electric field time series and the vector features, and establish a fused feature tensor; an execution module, which is used to use the fused feature tensor as input data, execute the electrical test correction judgment of the deep learning channel, and generate the electrical test correction judgment result.

[0062] In one possible implementation, the method for testing the electrical status of a drone equipped with an electric field sensor also includes: a real-time perception module, which is used to call a lidar to perform real-time path scene perception and establish real-time perception results when the multi-rotor drone performs an operating task along the three-dimensional flight path; an obstacle avoidance reconstruction module, which is used to perform path hazard breakthrough analysis based on the real-time perception results. If the path hazard breakthrough analysis result meets the breakthrough threshold, a reconstructed path is established, and the multi-rotor drone flight management is performed according to the reconstructed path.

[0063] In one possible implementation, the method for testing the electrical properties of a drone equipped with an electric field sensor further includes: after the multi-rotor drone reaches the starting interval, establishing a non-target area period as a background sampling area and automatically calibrating the sampling window; dividing the sampling points in the sampling window into three-dimensional grid areas according to spatial coordinates, performing spatial domain modeling, and generating an initial background electric field distribution map; performing continuous time series sliding window processing on the initial background electric field distribution map, and completing background electric field modeling based on the sliding window processing results.

[0064] In one possible implementation, the method for testing the electrical status of a drone equipped with an electric field sensor further includes: an early warning unit for performing a trigger analysis of the warning level based on the electrical test correction judgment result and establishing a trigger early warning instruction; and a reporting unit for establishing a scene early warning output mark in the electrical test scene data based on the trigger early warning instruction.

[0065] The UAV electrical testing platform equipped with an electric field sensor provided in the embodiment of the present invention can execute the UAV electrical testing method equipped with an electric field sensor provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0066] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules may be used and run on the user terminal and / or server, and the various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other and are not used to limit the scope of protection of the present invention.

[0067] The above specific embodiments do not constitute a limitation on the scope of protection of this application. Those skilled in the art should understand that various modifications, combinations, and substitutions may be made based on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application shall be included within the scope of protection of this application.

Claims

1. The UAV electrical inspection platform equipped with electric field sensors is characterized by: The electrical testing platform includes: An initialization unit is used to receive an electrical test task instruction, call the electrical test scene data and weather environment data according to the electrical test task instruction to perform multi-rotor UAV flight path planning, and establish a three-dimensional flight path; A linkage acquisition unit is used to start the multi-rotor drone and perform the operation task along the three-dimensional flight path. When the position information of the multi-rotor drone meets the start interval, the electric field sensing array integrated in the multi-rotor drone is activated to perform electric field signal acquisition; An alignment and return unit is used to align the multi-rotor drone with the electric field sensing array in time sequence and then return the mapped position and posture data and electric field signal set to the edge computing layer; A suppression unit, configured to perform background electric field modeling and environmental interference suppression after performing electric field signal set denoising using differential denoising; The discrimination unit is used to activate the deep learning channel of the edge computing layer, perform electrical inspection and correction discrimination on the electric field signal set after environmental interference suppression and the mapped position and posture data, and generate the electrical inspection and correction discrimination result.

2. The UAV electrical testing platform equipped with an electric field sensor as claimed in claim 1, characterized in that: The initialization unit is used for: Analyze and obtain the target area and operation requirements of the electrical inspection task according to the electrical inspection task instruction; Extracting a three-dimensional point cloud of power facilities from a preset scene database based on the target area for power inspection, fusing the three-dimensional point cloud of power facilities with map information, and establishing power inspection scene data; Activate the remote weather interface to read weather forecast data and establish weather environment data; The flight path of the multi-rotor UAV is planned based on the electrical test scene data and weather environment data.

3. The UAV electrical testing platform equipped with an electric field sensor as claimed in claim 2, characterized in that: In the initialization unit, establishing a three-dimensional flight path includes: Establishing a safety distance constraint based on the electrical inspection scenario data; Configure motion posture matching constraints according to job requirements; Using the electrical test scenario data to perform electromagnetic interference fitting and establish electromagnetic interference avoidance constraints; After initializing the trajectory planning channel using the safety distance constraint, motion posture matching constraint, and electromagnetic interference avoidance constraint, a balanced path planning of the path and acquisition effect is performed to establish a three-dimensional flight path.

4. The UAV electrical testing platform equipped with an electric field sensor as claimed in claim 1, characterized in that: The discrimination unit includes: The preprocessing module is used to perform time window alignment on the electric field signal set after environmental interference suppression and the mapped position and posture data, standardize them into a unified tensor structure, and input them into the feature extraction channel; The extraction module is used to activate the feature extraction channel, encode the position and posture data sequence into vector features using the position embedding mechanism, introduce cross attention to fuse the electric field time series and vector features, and establish a fused feature tensor; An execution module is used to use the fused feature tensor as input data, execute the electrical test and correction judgment of the deep learning channel, and generate an electrical test and correction judgment result.

5. The UAV electrical testing platform equipped with an electric field sensor as claimed in claim 1, characterized in that: The linkage acquisition unit includes: A real-time perception module is used to call the laser radar to perform real-time path scene perception and establish real-time perception results when the multi-rotor drone performs an operation task along the three-dimensional flight path; The obstacle avoidance reconstruction module is used to perform path hazard violation analysis based on the real-time perception results. If the path hazard violation analysis results meet the violation threshold, a reconstructed path is established and multi-rotor UAV flight management is performed according to the reconstructed path.

6. The UAV electrical testing platform equipped with an electric field sensor as claimed in claim 1, characterized in that: In the suppression unit, modeling the background electric field and performing environmental interference suppression includes: After the multi-rotor drone reaches the starting interval, a non-target area period is established as the background sampling area, and the sampling window is automatically calibrated; Dividing the sampling points in the sampling window into three-dimensional grid areas according to spatial coordinates, performing spatial domain modeling, and generating an initial background electric field distribution map; Continuous time-series sliding window processing is performed on the initial background electric field distribution map, and background electric field modeling is completed according to the sliding window processing result.

7. The UAV electrical testing platform equipped with an electric field sensor as claimed in claim 1, characterized in that: The electrical inspection platform also includes: An early warning unit, configured to perform trigger analysis of an early warning level based on the electrical test and correction judgment result, and establish a trigger early warning instruction; The reporting unit is used to establish a scene pre-alarm output mark in the electrical inspection scene data according to the trigger warning instruction.

8. A method for testing the electrical properties of a drone equipped with an electric field sensor, characterized in that: The method is applied to the UAV electrical testing platform equipped with an electric field sensor according to any one of claims 1 to 7, and the method comprises: Receive an electrical inspection task instruction, call the electrical inspection scene data and weather environment data according to the electrical inspection task instruction to perform multi-rotor UAV flight path planning, and establish a three-dimensional flight path; Starting the multi-rotor drone to perform an operation task along the three-dimensional flight path, and activating an electric field sensing array integrated in the multi-rotor drone to perform electric field signal acquisition when the position information of the multi-rotor drone meets the starting interval; After the multi-rotor drone is time-sequentially aligned with the electric field sensing array, the mapped position and posture data and electric field signal set are transmitted back to the edge computing layer; After performing denoising on the electric field signal set using differential denoising, the background electric field is modeled and environmental interference suppression is performed; Activate the deep learning channel of the edge computing layer, perform electrical inspection and correction judgment on the electric field signal set after environmental interference suppression and the mapped position and posture data, and generate the electrical inspection and correction judgment result.

Images