Unmanned aerial vehicle adaptive inspection path planning method, system and device

Abstract

The application discloses a kind of unmanned aerial vehicle adaptive inspection path planning method, system and equipment, it is related to the technical field of path planning, method includes: based on the current structure health score of each transmission tower is obtained to determine corresponding risk level by pre-constructed comprehensive health score system;According to risk level, the preset differentiated inspection accuracy mode is matched for each transmission tower;With risk weighted coverage as target, in combination with the endurance of unmanned aerial vehicle and flight safety constraint, initial adaptive inspection path is generated using multi-objective optimization algorithm;In the process of inspection, the transmission tower defect is identified in real time using the pre-trained ship detection model, when the transmission tower defect that satisfies the preset condition is identified, the initial adaptive inspection path is dynamically adjusted, the inspection mode of the current transmission tower is switched and the remaining path is re-planned;After inspection is completed, the detection result is fed back to the comprehensive health score system to update the risk level.The application realizes the differentiated intelligent allocation of transmission tower inspection resources.

Description

Technical Field

[0001] This application relates to the technical field of path planning, and in particular to a method, system and device for adaptive inspection path planning of unmanned aerial vehicles (UAVs). Background Technology

[0002] Transmission towers are crucial supporting structures for power transmission systems, and their structural safety directly impacts the reliability of power grid operation and public safety. With the continuous expansion of transmission line scale and the increase in service life, transmission towers face various structural degradation problems such as tilting, corrosion, loose bolts, and foundation settlement, requiring regular inspections to promptly identify potential hazards.

[0003] Currently, drone inspection has gradually replaced traditional manual inspection as the main means of power transmission line operation and maintenance. Existing drone inspection path planning methods typically employ preset fixed routes or geographical area divisions, applying the same inspection strategy and density to all towers within the transmission corridor. For example, existing technologies have proposed structural inspection path optimization methods based on genetic algorithms and greedy algorithms to minimize path length under coverage constraints; other technologies have proposed drone inspection strategy optimization frameworks based on physical models, utilizing bidirectional information exchange between structural diagnosis and prediction to optimize inspection intervals and distance parameters.

[0004] However, the above methods have the following shortcomings: First, a comprehensive multi-index structural health evaluation system for transmission towers has not been established, making it impossible to quantify the actual structural risk differences among towers; second, all towers are subject to the same inspection density and precision, resulting in insufficient inspection of high-risk towers and waste of resources on low-risk towers; third, there is a lack of a preventive inspection scheduling mechanism based on structural degradation trend prediction; fourth, meteorological conditions and structural status are not coupled for analysis to dynamically adjust inspection strategies; and fifth, there is a lack of real-time detection feedback-driven dynamic path adjustment capability during the inspection process.

[0005] The above content is only used to help understand the technical solution of this application and does not represent an admission that the above content is prior art. Summary of the Invention

[0006] The main purpose of this application is to provide an adaptive inspection path planning method, system and equipment for unmanned aerial vehicles (UAVs), which aims to solve the technical problems that existing technologies cannot perform differentiated inspections based on the actual structural risk differences of transmission towers, nor can they dynamically adjust the path based on real-time detection results during the inspection process.

[0007] To achieve the above objectives, this application proposes an adaptive inspection path planning method for unmanned aerial vehicles (UAVs), the method comprising: The current structural health score of each transmission tower is obtained based on a pre-built comprehensive health scoring system to determine the corresponding risk level; Based on the aforementioned risk level, a preset differentiated inspection accuracy mode is matched for each transmission tower; With risk-weighted coverage as the objective, and taking into account the constraints of UAV endurance and flight safety, a multi-objective optimization algorithm is used to generate an initial adaptive inspection path; During the inspection process, a pre-trained ship-based detection model is used to identify transmission tower defects in real time. When a transmission tower defect that meets the preset conditions is identified, the initial adaptive inspection path is dynamically adjusted, the inspection mode for the current transmission tower is switched, and the remaining path is replanned. After the inspection is completed, the test results will be fed back to the comprehensive health scoring system to update the risk level.

[0008] In one embodiment, the step of obtaining the current structural health score of each transmission tower based on a pre-built comprehensive health scoring system to determine the corresponding risk level includes: Collect structural status index data for each transmission tower. The structural status index data includes at least one or more of the following: tower tilt, component corrosion degree, bolt loosening rate, foundation settlement, and insulator deterioration degree. The structural state index data is normalized to obtain normalized index data; Based on the normalized index data, the subjective weights of each structural state index data are determined by the analytic hierarchy process (AHP), and the comprehensive weights of each structural comprehensive state index data are obtained based on the subjective weights and the objective weights determined by the entropy increase method. The comprehensive health score of each transmission tower is obtained by weighted calculation based on the normalized index data and the comprehensive weight. Based on the preset data range in which the comprehensive health score falls, each transmission tower is divided into three risk levels: high risk, medium risk, and low risk.

[0009] In one embodiment, the step of obtaining the current structural health score and corresponding risk level of each transmission tower based on a pre-built comprehensive health scoring system includes: The historical comprehensive health score sequence of each transmission tower is obtained, and the preset time series prediction model is input to predict the health score change curve within the target set time window. The time series prediction model is a long short-term memory network. The degradation rate of the health score change curve is calculated. When the degradation rate exceeds the preset threshold of the statistical distribution of degradation rates of the same type of transmission tower group, the current risk level of the corresponding transmission tower is increased, and the risk level after preventive correction is output.

[0010] In one embodiment, the step of matching each transmission tower with a preset differentiated inspection accuracy mode according to the risk level includes: Obtain the risk level after preventative correction; Based on the mapping relationship between risk level and inspection mode, high-risk towers are matched with fine inspection mode, medium-risk towers are matched with standard inspection mode, and low-risk towers are matched with rapid inspection mode.

[0011] In one embodiment, the step of generating an initial adaptive inspection path using a multi-objective optimization algorithm, with risk-weighted coverage as the objective and in combination with UAV endurance and flight safety constraints, includes: Based on the differentiated inspection accuracy pattern obtained by matching, the minimum inspection accuracy constraint corresponding to each transmission tower is determined. Obtain the current battery level and preset flight safety parameters of the drone, and construct a set of constraint conditions in conjunction with the minimum inspection accuracy constraint; Transmission towers of high-risk, medium-risk, and low-risk levels are assigned a first weight, a second weight, and a third weight, respectively, wherein the first weight is greater than the second weight, the second weight is greater than the third weight, and a path optimization model is constructed based on the first weight, the second weight, and the third weight, with the objective function being to maximize the risk-weighted coverage. A multi-objective optimization algorithm is used to calculate and generate the initial adaptive inspection path based on the objective function and the set of constraints.

[0012] In one embodiment, the constraint set further includes meteorological-structural coupled risk constraints, specifically: Obtain meteorological data to obtain ambient wind speed; When the ambient wind speed exceeds a first preset threshold, the inspection priority of the transmission towers located in wind-sensitive areas is increased, and the flight safety parameters of the UAV are adjusted simultaneously.

[0013] In one embodiment, the step of generating an initial adaptive inspection path using a multi-objective optimization algorithm further includes: When multiple drones are performing inspection tasks, the remaining power of each drone is obtained, and the risk spatial distribution is determined by combining the risk level and the spatial location of the transmission tower. An improved clustering-based task allocation algorithm is used to dynamically divide the area to be inspected into several sub-regions according to the remaining power and the risk spatial distribution. The inspection tasks in each of the sub-regions are assigned to the corresponding UAVs, and each UAV is controlled to independently execute the initial adaptive inspection path in the assigned sub-region.

[0014] In one embodiment, the step of using a pre-trained ship-based detection model to identify transmission tower defects in real time during the inspection process, and dynamically adjusting the initial adaptive inspection path, switching the inspection mode for the current transmission tower, and replanning the remaining path when a transmission tower defect that meets preset conditions is identified includes: A pre-trained ship-machine detection model is used to perform real-time defect detection on the transmission tower and obtain the defect detection results. The ship-machine detection model is a lightweight target detection network. Determine whether there are any serious defects in the defect detection results where the crack length exceeds the second preset threshold or the rust area ratio exceeds the third preset threshold; If the aforementioned serious defects exist, the inspection mode of the current transmission tower will be switched from the current mode to the fine inspection mode to take supplementary photos, and the inspection priority of towers of the same model adjacent to the current transmission tower will be increased. Based on the improved inspection priority, the path of the remaining uninspected transmission towers in the initial adaptive inspection path is replanned to generate an updated inspection path.

[0015] Furthermore, to achieve the above objectives, this application also proposes an adaptive inspection path planning system for unmanned aerial vehicles (UAVs), the UAV adaptive inspection path planning system comprising: The structural health assessment module is used to obtain the current structural health score of each transmission tower based on a pre-built comprehensive health scoring system to determine the corresponding risk level. The inspection accuracy matching module is used to match a preset differentiated inspection accuracy mode for each transmission tower according to the risk level. The path optimization module is used to generate an initial adaptive inspection path with risk-weighted coverage as the objective, combined with the drone's endurance and flight safety constraints, using a multi-objective optimization algorithm. The dynamic adjustment module is used to identify transmission tower defects in real time during the inspection process using a pre-trained ship detection model. When a transmission tower defect that meets the preset conditions is identified, the initial adaptive inspection path is dynamically adjusted, the inspection mode for the current transmission tower is switched, and the remaining path is replanned. The data update module is used to feed back the test results to the comprehensive health scoring system after the inspection is completed in order to update the risk level.

[0016] In addition, to achieve the above objectives, this application also proposes an adaptive inspection path planning device for unmanned aerial vehicles (UAVs), the device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the UAV adaptive inspection path planning method described above.

[0017] This application proposes an adaptive inspection path planning method for unmanned aerial vehicles (UAVs). The method includes: obtaining the current structural health score of each transmission tower based on a pre-built comprehensive health scoring system to determine the corresponding risk level; matching each transmission tower with a preset differentiated inspection accuracy mode according to the risk level; generating an initial adaptive inspection path using a multi-objective optimization algorithm with risk-weighted coverage as the objective, combined with UAV endurance and flight safety constraints; during the inspection process, using a pre-trained UAV-based detection model to identify transmission tower defects in real time; when a transmission tower defect meeting preset conditions is identified, dynamically adjusting the initial adaptive inspection path, switching the inspection mode for the current transmission tower, and replanning the remaining path; after the inspection is completed, feeding back the detection results to the comprehensive health scoring system to update the risk level. This application realizes differentiated intelligent allocation of transmission tower inspection resources, improving the defect detection rate of high-risk towers and the overall inspection efficiency. Attached Figure Description

[0018] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0019] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a flowchart illustrating an embodiment of the UAV adaptive inspection path planning method of this application. Figure 2 This is a flowchart illustrating Embodiment 2 of the UAV adaptive inspection path planning method of this application; Figure 3 This is a flowchart illustrating Embodiment 3 of the UAV adaptive inspection path planning method of this application; Figure 4 This is a flowchart illustrating Embodiment 4 of the UAV adaptive inspection path planning method of this application; Figure 5 This is a schematic diagram of the module structure of the UAV adaptive inspection path planning system according to an embodiment of this application; Figure 6 This is a schematic diagram of the equipment structure of the hardware operating environment involved in the UAV adaptive inspection path planning method in the embodiments of this application.

[0021] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0022] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0023] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0024] Because existing technology cannot conduct differentiated inspections based on the actual structural risks of transmission towers, nor can it dynamically adjust the path based on real-time detection results during the inspection process.

[0025] This application provides a solution that uses a pre-built comprehensive health scoring system to obtain the current structural health score of each transmission tower to determine its corresponding risk level; matches each transmission tower with a preset differentiated inspection accuracy mode according to the risk level; uses a risk-weighted coverage rate as the objective, combined with UAV endurance and flight safety constraints, and employs a multi-objective optimization algorithm to generate an initial adaptive inspection path; during the inspection process, a pre-trained UAV detection model is used to identify transmission tower defects in real time; when a transmission tower defect meeting preset conditions is identified, the initial adaptive inspection path is dynamically adjusted, the inspection mode for the current transmission tower is switched, and the remaining path is replanned; after the inspection is completed, the detection results are fed back to the comprehensive health scoring system to update the risk level. This application realizes differentiated intelligent allocation of transmission tower inspection resources, improving the defect detection rate of high-risk towers and the overall inspection efficiency.

[0026] Based on this, embodiments of this application provide an adaptive inspection path planning method for unmanned aerial vehicles (UAVs), referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the UAV adaptive inspection path planning method of this application.

[0027] In this embodiment, the UAV adaptive inspection path planning method includes steps S10 to S50: Step S10: Obtain the current structural health score of each transmission tower based on the pre-built comprehensive health scoring system to determine the corresponding risk level.

[0028] It should be noted that, in this embodiment, the comprehensive health scoring system refers to the algorithm and data framework used to quantitatively assess the structural safety status of transmission towers. The current structural health score refers to the quantitative value calculated by comprehensively analyzing various indicators of the transmission tower at the current moment, and the risk level refers to the level characterizing the degree of danger of the transmission tower structure based on the health score. Constructing the comprehensive health scoring system involves structural status indicator data such as tower tilt, component corrosion degree, bolt loosening rate, foundation settlement, and insulator deterioration degree. Structural status indicator data refers to the basic data reflecting the physical and electrical status of the transmission tower. Tower tilt refers to the angle or displacement of the tower body from the vertical reference plane, usually obtained through tilt sensors or laser ranging equipment. Component corrosion degree refers to the area or depth ratio of oxidation and corrosion on the surface of metal components, which can be detected through high-definition image analysis or coating thickness gauges. Bolt loosening rate refers to the proportion of loose bolts at connection nodes to the total number of bolts, usually detected by a torque wrench or identified based on vibration characteristics. Foundation settlement refers to the vertical displacement of the tower base, which can be monitored by a hydrostatic level or differential GPS. The degree of insulator deterioration refers to the extent to which the insulation performance of an insulator string decreases due to factors such as pollution, aging, and damage. It can be assessed through leakage current, ultraviolet imaging, or manual visual inspection.

[0029] This embodiment aims to transform the discrete and heterogeneous structural status monitoring data of various transmission towers distributed in different geographical locations into comparable and quantifiable comprehensive health scores through a unified standardized process, and based on this, classify the transmission towers into different risk levels, providing an objective basis for the formulation of differentiated inspection strategies.

[0030] In one possible implementation, the preset data range of the comprehensive health scoring system is divided as follows: when the comprehensive health score is less than a first threshold, the corresponding transmission tower is classified as high-risk; when the comprehensive health score is greater than or equal to the first threshold and less than a second threshold, the corresponding transmission tower is classified as medium-risk; and when the comprehensive health score is greater than or equal to the second threshold, the corresponding transmission tower is classified as low-risk. The first threshold can be 0.4, and the second threshold can be 0.7.

[0031] Additionally, it should be noted that structural condition index data can be automatically acquired through various sensor networks deployed on transmission towers, or periodically collected using detection equipment mounted on drones. Furthermore, structural condition index data is not limited to the five types mentioned above; it can also include supplementary indicators such as grounding resistance values, tower tilt azimuth angles, and bolt preload attenuation values ​​to enrich the dimensions of health assessment.

[0032] Step S20: Match a preset differentiated inspection accuracy mode to each transmission tower according to the risk level.

[0033] It should be noted that, in this embodiment, the differentiated inspection accuracy mode refers to inspection operation strategies with different levels of detail and resource consumption configured for transmission towers of different risk levels. The differentiated inspection accuracy mode defines a fine inspection mode for high-risk towers, a standard inspection mode for medium-risk towers, and a rapid inspection mode for low-risk towers. The fine inspection mode refers to a high-precision, detailed inspection method using low flight speed, close inspection distance, and multiple shooting angles; the standard inspection mode refers to a medium-precision inspection method using medium flight speed, medium inspection distance, and shooting of key parts; and the rapid inspection mode refers to a coarse inspection method using high flight speed, long inspection distance, and overall appearance shooting.

[0034] This embodiment aims to achieve precise matching between inspection resources and structural risks, breaking the traditional "one-size-fits-all" inspection mode. It enables high-risk towers to receive sufficient inspection attention to improve the defect detection rate, while allowing low-risk towers to pass through quickly to significantly save flight time and battery consumption, thereby maximizing overall inspection efficiency and resource utilization.

[0035] In one possible implementation, the fine inspection mode has a flight speed of 1m / s to 3m / s, an inspection distance of 5m to 8m, and a shooting strategy requiring the inspection and photography of tower joints, insulator strings, foundations, and other components from at least six different angles. The standard inspection mode has a flight speed of 3m / s to 5m / s, an inspection distance of 10m to 15m, and a shooting strategy requiring the photography of key parts such as the tower head, main tower structure, and insulator mounting points from at least three perspectives. The rapid inspection mode has a flight speed of 5m / s to 8m / s, an inspection distance of 15m to 20m, and a shooting strategy requiring only the capture of a photograph of the overall frontal appearance of the tower.

[0036] In addition, it should be noted that the differentiated inspection accuracy mode, in addition to including flight speed, inspection distance and shooting strategy, can also include strategies for equipping different sensors, such as activating the lidar scanning mode for high-risk towers.

[0037] Step S30: Taking the risk-weighted coverage rate as the objective, and combining the constraints of UAV endurance and flight safety, a multi-objective optimization algorithm is used to generate an initial adaptive inspection path.

[0038] It should be noted that, in this embodiment, the risk-weighted coverage rate refers to a quantitative indicator used to measure the coverage of transmission towers of different risk levels by the inspection path. Its calculation method is as follows: First, second, and third weights are pre-assigned to high-risk, medium-risk, and low-risk transmission towers, respectively, with the first weight greater than the second weight, the second weight greater than the third weight, and the third weight greater than or equal to 0. The number of towers inspected at each risk level along the path is multiplied by their corresponding weights, summed, and then divided by the sum of all tower weights. The resulting ratio is the risk-weighted coverage rate. Flight safety constraints refer to the flight boundary conditions set to ensure the safety of the UAV and its surrounding environment, including but not limited to maximum flight altitude limits, minimum safe distance from charged objects, and no-fly zone avoidance. Multi-objective optimization algorithms refer to optimization algorithms capable of simultaneously handling multiple conflicting objective functions and finding a set of non-dominated solutions. In this embodiment, it is mainly used to maximize the risk-weighted coverage rate while also considering implicit objectives such as the shortest total path length or the lowest total energy consumption.

[0039] This embodiment aims to find an optimal route that maximizes safety benefits under complex constraints such as limited power, variable environment, and heavy workload, while taking into account safety avoidance under extreme weather conditions and efficient scheduling of multi-aircraft collaborative operations.

[0040] In one possible implementation, the multi-objective optimization algorithm can be the NSGA-II genetic algorithm based on non-dominated sorting, or it can be an improved ant colony algorithm or particle swarm optimization algorithm.

[0041] Step S40: During the inspection process, a pre-trained ship-machine detection model is used to identify transmission tower defects in real time. When a transmission tower defect that meets the preset conditions is identified, the initial adaptive inspection path is dynamically adjusted, the inspection mode for the current transmission tower is switched, and the remaining path is replanned.

[0042] It should be noted that, in this embodiment, the airborne detection model refers to a lightweight target detection network deployed on the edge computing unit of an unmanned aerial vehicle (UAV), used for online inference analysis of video streams or image frames collected in real time during the inspection process to identify various visible light defects in key parts of the transmission tower. The lightweight target detection network refers to a deep neural network model with a small number of parameters, low computational complexity, and suitable for real-time operation on embedded devices with limited computing power, such as YOLOv5-nano, YOLOv8-nano, or MobileNet-SSD. Dynamic path adjustment refers to the online modification of the currently executing inspection path during the inspection flight based on the defect events detected in real time by the airborne detection model. This includes, but is not limited to, changing the inspection mode of the current tower, rearranging the access order of the remaining towers, adding or skipping certain tower positions, etc.

[0043] This embodiment utilizes a lightweight model for high-frequency real-time inference. Once a preset critical defect condition is hit, the current routine process is immediately interrupted, execution mode switching and local priority reconstruction are performed, and finally, a replanning algorithm is triggered to generate an updated inspection path.

[0044] In one possible implementation, the defects in the transmission tower under the preset conditions may include, in addition to cracks and corrosion, defects such as severely damaged insulators or detached hardware.

[0045] In one specific implementation, during the flight of the UAV along the initial adaptive inspection path, the dynamic adjustment module uses a preset lightweight target detection network as the ship-to-aircraft detection model to perform real-time defect detection on the current transmission tower and obtain defect detection results. It then determines whether there are serious defects in the defect detection results, such as crack length exceeding a second preset threshold (e.g., 5cm) or corrosion area exceeding a third preset threshold (e.g., 10%). If such a serious defect exists, the inspection mode of the current transmission tower is immediately switched from the standard inspection mode to the fine inspection mode to perform supplementary shooting from no less than 6 perspectives. At the same time, since towers of the same batch often have the same weak points, the system increases the inspection priority of towers of the same model adjacent to the current transmission tower. Based on the increased inspection priority, the path of the remaining uninspected transmission towers in the initial adaptive inspection path is replanned, an updated inspection path is generated, and the path is sent to the UAV for execution.

[0046] Step S50: After the inspection is completed, the test results are fed back to the comprehensive health scoring system to update the risk level.

[0047] It should be noted that, in this embodiment, the detection results refer to all inspection data recorded by the onboard storage device or transmitted back in real time via the data link after the UAV completes a single inspection mission. This includes, but is not limited to, high-definition visible light photographs, infrared thermal images, defect identification and annotation files, defect type and severity classification records, UAV flight logs, and actual inspection mode execution records for each tower. Feedback to the comprehensive health scoring system refers to importing the above detection results into the database of the comprehensive health scoring system according to a preset data interface format, using them as new evidence data to participate in the recalculation of the structural status index values ​​of each transmission tower. Updating the risk level refers to correcting the comprehensive health score of each transmission tower based on the new detection results and re-determining the risk level to form the input basis for planning a new round of inspection missions.

[0048] This embodiment aims to achieve a closed-loop iteration between inspection data and structural health assessment, so that the results of each inspection task can be reflected in the health record of the transmission tower in real time, ensuring that the formulation of inspection plans is always based on the latest and most accurate structural status information and approaches the optimal inspection resource allocation scheme.

[0049] This application quantitatively assesses the risk level of transmission towers to match differentiated inspection strategies, integrates multiple constraints to solve the initial adaptive path, and realizes dynamic adjustment of the inspection path and system closed-loop iteration based on real-time defect detection and feedback data. It achieves precise tilting and differentiated intelligent allocation of inspection resources to high-risk towers, upgrading traditional passive inspection to proactive preventive inspection based on trend prediction and real-time feedback. Under the premise of ensuring safety in complex meteorological environments and multi-machine collaborative operations, it significantly improves the overall inspection efficiency and the detection rate of serious defects, and forms a data-driven self-iterative optimization closed loop.

[0050] Furthermore, referring to Figure 2 The second embodiment of the UAV adaptive inspection path planning method of this application provides a flowchart, based on the above. Figure 2 The embodiment shown further refines the step S10, "obtaining the current structural health score of each transmission tower based on a pre-built comprehensive health scoring system to determine the corresponding risk level," including steps A201 to A205: Step A201: Collect structural status index data for each transmission tower. The structural status index data includes at least one or more of the following: tower tilt, component corrosion degree, bolt loosening rate, foundation settlement, and insulator deterioration degree. Step A202: Normalize the structural state index data to obtain normalized index data; Step A203: Based on the normalized index data, the subjective weights of each structural state index data are determined by the analytic hierarchy process (AHP), and the comprehensive weights of each structural comprehensive state index data are obtained based on the subjective weights and the objective weights determined by the entropy increase method. Step A204: Based on the normalized index data and the comprehensive weight, a weighted calculation is performed to obtain the comprehensive health score of each transmission tower; Step A205: Based on the preset data range in which the comprehensive health score is located, each transmission tower is divided into three risk levels: high risk, medium risk, and low risk.

[0051] It should be noted that, in this embodiment, normalization refers to a data preprocessing operation that maps original structural state indicator data with different dimensions and value ranges to a unified numerical range through specific mathematical transformations; the analytic hierarchy process (AHP) is a subjective weighting method that decomposes complex multi-criteria decision-making problems into target, criterion, and alternative layers, and determines the relative importance of each evaluation indicator by constructing pairwise comparison judgment matrices and calculating the maximum eigenvalue and corresponding eigenvector of the judgment matrices; subjective weight refers to the importance weight coefficient of each structural state indicator calculated using the AHP; entropy increase method is an objective weighting method based on information entropy theory, which determines the indicator weight according to the dispersion of the observed data of each indicator. The smaller the information entropy of the indicator data, the greater the degree of data variation and the more information it provides, and the higher the objective weight assigned to the indicator; comprehensive weight refers to the final weight vector obtained by fusing subjective weight and objective weight through certain combination rules. The preset data range refers to the pre-set numerical boundary range used to distinguish different risk levels.

[0052] This embodiment comprehensively characterizes the tower's condition through multi-indicator data collection, then eliminates computational interference caused by different physical dimensions through normalization. Subsequently, it incorporates prior knowledge from power operation and maintenance experts using the analytic hierarchy process (AHP) and mines the discreteness information of the data itself using the entropy increase method. The combination of these two approaches forms a precise comprehensive weight. Finally, a weighted calculation is performed based on this weight to obtain a single-dimensional comprehensive health score, which is then transformed into intuitive high, medium, and low risk levels according to a preset data range. This achieves the transformation from complex and heterogeneous data to standardized risk classification. The aim is to accurately quantify the true structural condition of each transmission tower through multi-dimensional data fusion and a weight allocation that combines subjective and objective factors, thereby overcoming the shortcomings of existing technologies that cannot quantify the differences in actual structural risks of each tower.

[0053] In one specific implementation, the tower tilt angle θ is measured by a tilt sensor, with a normal range of 0 to 0.5°, and exceeding 1° is considered severe; the degree of component corrosion R is evaluated through image recognition and manual recording, using a dimensionless score of 0 to 1, where 0 indicates no corrosion and 1 indicates severe corrosion; the bolt loosening rate B is determined by ultrasonic testing or manual inspection, with a value range of 0 to 1; the foundation settlement S is calculated through leveling or tilt sensor measurement, with a unit of millimeters (mm), and a normal range of 0 to 10 mm; and the insulator deterioration degree I is evaluated through infrared thermography and zero-value insulator testing, using a dimensionless score of 0 to 1.

[0054] In one possible implementation, index normalization can be achieved by using the Min-Max normalization method to map each index to the [0,1] interval. Combined weighting can be performed using AHP to determine the subjective weight vector W_s=[w_θ^s, w_R^s, w_B^s, w_S^s, w_I^s], and the objective weight vector W_o=[w_θ^o, w_R^o, w_B^o, w_S^o, w_I^o] can be determined using the entropy weight method. The comprehensive weight W=α·W_s+(1-α)·W_o, where α∈[0,1] is the balance coefficient between subjective and objective weights; in this embodiment, α=0.5.

[0055] In one possible implementation, the health score is calculated by inverting the normalized values ​​of each indicator (a higher health score indicates better health), and then weighted summing to obtain the comprehensive health score H_i = Σ(w_j·(1-x_ij)), where w_j is the comprehensive weight of the j-th indicator, and x_ij is the normalized value of the j-th indicator for the i-th tower. Furthermore, based on a preset data range in which the comprehensive health score falls, transmission towers with H < 0.4 are classified as high-risk, those with 0.4 ≤ H < 0.7 as medium-risk, and those with H ≥ 0.7 as low-risk, thus achieving a precise quantitative determination of the transmission tower's risk level.

[0056] Furthermore, the step of obtaining the current structural health score and corresponding risk level of each transmission tower based on the pre-built comprehensive health scoring system includes A301-A302: Step A301: Obtain the historical comprehensive health score sequence of each transmission tower, input the preset time series prediction model, and predict the health score change curve within the target set time window. The time series prediction model is a long short-term memory network.

[0057] It should be noted that, in the embodiments of this application, the historical comprehensive health score sequence refers to the set of health score data of each transmission tower recorded in chronological order, the preset time series prediction model refers to the mathematical model that mines the time series patterns based on historical time series data to predict future trends, the target setting time window refers to a specific future time period that needs to be predicted forward, the health score change curve refers to the continuous trajectory of the predicted comprehensive health score changing over time, and the long short-term memory network refers to a deep learning recurrent neural network that includes a gating mechanism to learn long-term dependencies.

[0058] This embodiment organizes past discrete scoring data into an ordered sequence and inputs it into an artificial intelligence model with long-term memory capabilities. Through the model's feature extraction and fitting capabilities, it outputs the continuous scoring evolution trajectory over a future period, achieving a leap from cross-sectional state assessment to time-series trend prediction.

[0059] In one possible implementation, the preset time series prediction model can employ not only a long short-term memory network, but also a grey prediction model or a Transformer-based time series prediction network to adapt to prediction needs with different data volumes and feature complexities.

[0060] Step A302: Calculate the degradation rate of the health score change curve. When the degradation rate exceeds the preset threshold of the statistical distribution of degradation rates of the same type of transmission tower group, the current risk level of the corresponding transmission tower is increased, and the risk level after preventive correction is output.

[0061] It should be noted that, in the embodiments of this application, the degradation rate refers to the speed at which the health score decreases per unit time; the statistical distribution of degradation rates of a group of transmission towers of the same type refers to the statistical distribution law of degradation speed of a group of transmission towers of the same design model; the preset threshold refers to the critical degradation rate value that triggers the risk level upgrade operation; the current risk level refers to the risk level divided based on the static score in the previous operation; and the risk level after preventive correction refers to the risk level that is adjusted upward in advance after degradation trend prediction analysis.

[0062] In this embodiment, the predicted health score change curve is mathematically differentiated or differentially calculated to quantify the degradation rate. Then, the individual rate is compared with the statistical characteristics of the population. Once an individual's degradation rate is found to be abnormally fast, the original static rating result is immediately broken and the rating is upgraded preventively to ensure that high-risk hazards are intercepted as early as possible.

[0063] In one possible implementation, the mapping rule between degradation rate and risk level increase in the preventive inspection scheduling is as follows: Calculate the predicted degradation rate r_i = ΔH_i / ΔT for each transmission tower within a set time window ΔT, where ΔH_i is the predicted decrease in health score for the i-th tower; using the historical average degradation rate μ_d and standard deviation σ_d of a group of transmission towers of the same type as a benchmark, set a degradation rate threshold Th_d = μ_d + k·σ_d, where k is a confidence coefficient; when r_i > Th_d, increase the risk level of the transmission tower by one level; when the predicted health score crosses the risk level boundary value at the end of the time window, directly adjust the risk level of the transmission tower to the risk level corresponding to the predicted value; the degradation rate threshold is adaptively updated with the accumulation of historical data. For example, a confidence coefficient k = 2 can be taken, corresponding to approximately 95% confidence level.

[0064] In one specific implementation, risk level mapping can be performed based on the following rules: (a) when the predicted degradation rate r_i of a tower exceeds Th_d, the risk level of the tower is increased by one level (low risk → medium risk, medium risk → high risk); (b) when the predicted health score crosses the risk level boundary value at the end of the time window (e.g., from H≥0.7 to H<0.7, or from H≥0.4 to H<0.4), the tower is directly adjusted to the risk level corresponding to the predicted value; (c) the degradation rate threshold Th_d is updated quarterly with the accumulation of historical data.

[0065] This application analyzes the future health degradation trend of transmission towers through a time series prediction model, and proactively raises the risk level of individual towers when their degradation rate abnormally exceeds the statistical threshold of the group, thus outputting preventive correction results. It breaks the lag of traditional static assessment and upgrades the inspection mode from a passive approach of "responding only when hidden dangers appear" to a proactive preventive approach of "predicting degradation trends in advance". It effectively intercepts "latent" hidden danger towers that are degrading rapidly but have not yet fallen below the static risk threshold, thereby reducing the safety risk of sudden structural failure and realizing the advanced and precise scheduling of inspection resources in the time dimension.

[0066] Furthermore, the step of matching a preset differentiated inspection accuracy mode to each transmission tower according to the risk level includes A401-A402: Step A401: Obtain the risk level after preventative correction; Step A402: Based on the mapping relationship between the risk level and the inspection mode, high-risk towers are matched with the fine inspection mode, medium-risk towers are matched with the standard inspection mode, and low-risk towers are matched with the rapid inspection mode.

[0067] It should be noted that, in the embodiments of this application, the risk level after preventive correction refers to the final risk level that is adjusted upward in advance after combining the structural degradation trend prediction analysis; the mapping relationship between risk level and inspection mode refers to the correspondence rules between different risk levels and specific inspection operation strategies established in advance.

[0068] This embodiment uses the risk level after preventive correction as an index to directly call the predefined differentiated inspection accuracy mode, so as to realize the direct mapping and linkage from risk assessment results to specific flight action commands, as well as the precise matching of inspection resources and structural risks.

[0069] In one possible implementation, the mapping relationship between risk level and inspection mode can be achieved not only through a preset three-level discrete lookup table mapping, but also by dynamically generating flight speed and inspection distance parameters using linear interpolation based on continuous values ​​of the comprehensive health score. Furthermore, the differentiated inspection accuracy mode, in addition to including flight speed, inspection distance, and shooting strategy, can further include a strategy of equipping different sensors. For example, when matching a fine inspection mode to a high-risk tower, a lidar scanning mode can be simultaneously activated to obtain three-dimensional point cloud data.

[0070] Furthermore, referring to Figure 3 The third embodiment of the UAV adaptive inspection path planning method of this application provides a flowchart, based on the above. Figure 3 The embodiment shown further refines the step S30, "using risk-weighted coverage as the objective, combined with UAV endurance and flight safety constraints, to generate an initial adaptive inspection path using a multi-objective optimization algorithm," including steps A501 to A504: Step A501: Based on the differentiated inspection accuracy mode obtained by matching, determine the minimum inspection accuracy constraint corresponding to each transmission tower.

[0071] It should be noted that, in the embodiments of this application, the minimum inspection accuracy constraint refers to the most basic parameter limitation condition that ensures the quality of the inspection image can meet the defect identification requirements of the corresponding risk level, so as to prevent the sacrifice of the detection quality of key parts due to excessive pursuit of the shortest path.

[0072] Step A502: Obtain the current battery level and preset flight safety parameters of the drone, and construct a set of constraint conditions in conjunction with the minimum inspection accuracy constraint.

[0073] It should be noted that, in this embodiment, the current battery life refers to the actual flight time or mileage corresponding to the remaining available charge in the battery pack before the drone takes off; the preset flight safety parameters refer to the physical boundary values ​​set to ensure that the drone does not collide or crash during flight; and the constraint set refers to the set of mathematical inequalities consisting of all physical limitations, task requirements, and environmental constraints. This embodiment obtains the real-time battery status of the drone and the safety boundaries of the geographical environment, and logically integrates them with the minimum inspection accuracy constraint to form a multi-dimensional joint constraint space, thereby trunculating unrealistic solutions during the algorithm optimization process.

[0074] In one possible implementation, the preset flight safety parameters may include a fixed maximum permissible flight speed and a minimum safe distance, or they may include safety envelope surface parameters that are dynamically adjusted according to the density of obstacles.

[0075] Step A503: Assign a first weight, a second weight, and a third weight to transmission towers of high-risk, medium-risk, and low-risk levels, respectively, wherein the first weight is greater than the second weight, the second weight is greater than the third weight, and construct a path optimization model based on the first weight, the second weight, and the third weight, with the objective function being the maximization of risk-weighted coverage.

[0076] It should be noted that, in this embodiment, the risk-weighted coverage rate refers to the comprehensive quantitative evaluation index obtained by dividing the sum of the products of the number of inspected towers at each risk level and their corresponding weights by the sum of the weights of all towers after assigning weights to high-risk, medium-risk, and low-risk towers; the objective function refers to the mathematical expression that the optimization algorithm needs to maximize or minimize during the solution process; and the path optimization model refers to the complete solution framework formed by mathematically abstracting the objective function and the set of constraints. This embodiment transforms differentiated risks into a unified objective orientation by setting gradient weight coefficients, constructing a mathematical function that can accurately measure the safety benefits of each take-off and landing operation, and assembling it together with the set of constraints into a rigorous path optimization model.

[0077] In one possible implementation, the specific values ​​of the first weight, the second weight, and the third weight can be statically set according to the importance level of the power grid operation, or they can be dynamically calculated in real time according to the current load status and the expected power outage loss.

[0078] Step A504: Using a multi-objective optimization algorithm, the initial adaptive inspection path is generated based on the objective function and the set of constraints.

[0079] It should be noted that, in the embodiments of this application, the multi-objective optimization algorithm refers to an intelligent heuristic optimization algorithm that can find the Pareto optimal solution set among multiple conflicting objectives; the initial adaptive inspection path refers to the flight trajectory generated by the system based on the current state before takeoff, which includes a series of waypoint coordinates and corresponding operational parameters.

[0080] In this embodiment, the geographical coordinates of the transmission tower, the boundary values ​​of the constraints, and the scoring rules of the objective function are input into the algorithm engine. Through the iterative evolution of the population or the positive feedback mechanism of pheromones, inferior routes are continuously eliminated, and finally a globally optimal initial adaptive inspection path is converged and output for the UAV to execute.

[0081] In one possible implementation, the multi-objective optimization algorithm can employ the NSGA-II genetic algorithm based on non-dominated sorting to obtain a uniformly distributed Pareto front, or it can employ an improved ant colony algorithm to accelerate the convergence speed in complex transmission corridor terrain.

[0082] In one specific implementation, this embodiment employs the NSGA-II genetic algorithm based on non-dominated sorting as a multi-objective optimization algorithm. The coordinates of all transmission towers within the inspection area, the power consumption and safety limit parameters in the constraint set, and a function expression aiming to maximize risk-weighted coverage are input into the algorithm. NSGA-II initializes the waypoint population, performs crossover and mutation operations, and iterates through multiple generations based on non-dominated sorting and congestion distance to ultimately solve for the Pareto optimal path set. The path optimization module selects the path with the highest comprehensive score from this set, generating an initial adaptive inspection path that includes a precise flight trajectory, hovering photo points, and speed commands. Optionally, the population size is 100, the iterations are 500 generations, the crossover probability is 0.9, and the mutation probability is 0.1.

[0083] This application constructs a path optimization model by integrating multiple constraints such as inspection accuracy, power consumption, and safety. With the goal of maximizing risk-weighted coverage, a multi-objective optimization algorithm is used to generate an initial adaptive inspection path. It can be seen that the differentiated inspection accuracy requirements are transformed into hard mathematical constraints to ensure the bottom-line quality of defect detection. At the same time, by introducing a risk weight mechanism, the path planning is transformed from the traditional "blind full coverage" or "simply shortest distance" to "maximizing overall safety benefits". Under the premise of strictly ensuring the physical boundaries of UAV flight and power safety, the application achieves precise tilting and efficient scheduling of limited endurance resources towards high-risk towers.

[0084] Furthermore, the set of constraints also includes meteorological-structural coupled risk constraints, specifically A601-A602: Step A601: Obtain meteorological data to obtain the ambient wind speed; Step A602: When the ambient wind speed exceeds the first preset threshold, the inspection priority of the power transmission tower located in the wind-sensitive area is increased, and the flight safety parameters of the UAV are adjusted simultaneously.

[0085] It should be noted that in this embodiment, the meteorological-structural coupling risk constraint refers to a type of adaptive constraint condition in which the meteorological coupling submodule in the path optimization module dynamically adjusts the transmission tower inspection priority and the UAV flight safety boundary based on real-time acquired or forecasted meteorological data when constructing the path optimization constraint condition set. Its core feature is to correlate changes in the external meteorological environment with the structural stress sensitivity of the transmission tower itself. Meteorological data refers to observed or forecasted values ​​reflecting the atmospheric environment status of the inspection area obtained from the meteorological forecast service interface, micro-meteorological monitoring stations along the line, or UAV-borne meteorological sensors, including at least one or more of environmental wind speed, environmental temperature, and environmental icing thickness. Environmental wind speed refers to the relative velocity scalar value of airflow within the UAV inspection flight altitude layer, usually measured in meters per second. The first preset threshold is a pre-set environmental wind speed critical value used to trigger the meteorological-structural coupling risk constraint mechanism. When the measured or forecasted wind speed reaches or exceeds this critical value, it is considered that the current meteorological conditions have a significant impact on the structural safety of the transmission tower or the flight stability of the UAV. Wind-sensitive areas refer to geographical sections pre-delineated based on historical wind damage records, analysis of transmission tower structural dynamics, or design wind load verification results. Transmission towers located within these areas are more prone to vortex-induced vibration, component fatigue damage, or tower collapse under strong winds. Flight safety parameters refer to a set of adjustable control parameters used to limit the flight envelope boundaries of UAVs, including the maximum permissible flight speed and minimum safe distance. These values ​​can be dynamically adjusted according to external environmental conditions to ensure flight safety.

[0086] This embodiment deeply couples and analyzes meteorological environment, transmission tower structural properties and UAV flight status, and automatically and intelligently adjusts inspection strategy under severe weather conditions. This can both maximize the absolute safety of UAV flight and meet the needs of identifying potential structural damage to transmission towers in wind-sensitive areas under extreme weather conditions.

[0087] In one possible implementation, the triggering condition for the meteorological-structural coupled risk constraint is not limited to wind speed alone. Environmental icing thickness can also be introduced as an additional triggering condition. When the environmental icing thickness exceeds a preset icing threshold, it also triggers an increase in the inspection priority of transmission towers located in heavy icing zones and adjustments to flight safety parameters. Furthermore, the adjustment range of flight safety parameters can also adopt a piecewise linear adjustment strategy. That is, based on the degree to which the environmental wind speed exceeds the first preset threshold, the maximum permissible flight speed is adjusted proportionally at each level, and the minimum safe distance is increased by a multiplier, achieving more refined safety boundary control.

[0088] In one possible implementation, real-time or forecast meteorological data is acquired, including wind speed, temperature, and icing thickness; the meteorological-structural coupling risk index R_c = β1·R_w + β2·R_t + β3·R_i is calculated for each transmission tower, where β1, β2, and β3 are the weighting coefficients of each component and β1 + β2 + β3 = 1; the wind speed risk component R_w = (v / v_cr)² × (1 - H_i), where v is the current or forecast wind speed, v_cr is the critical wind speed determined by the tower type and height, and H_i is the comprehensive health score of the transmission tower; the temperature risk component R_t = |T - T_ref| / T_range × K_material, where... T represents the current temperature, T_ref represents the reference temperature, T_range represents the temperature range within which the material's properties are affected by temperature, and K_material represents the material's temperature sensitivity coefficient. The icing risk component R_i = (t_ice / t_ice_design) × K_span, where t_ice is the measured or predicted icing thickness, t_ice_design is the designed maximum icing thickness, and K_span is the icing sensitivity coefficient determined based on span and altitude. When R_c exceeds the preset coupling risk threshold, the inspection priority of the transmission tower is automatically increased, and the flight safety envelope parameters of the UAV are adjusted simultaneously. The flight safety envelope parameters include the maximum permissible flight speed and the minimum safe distance.

[0089] Furthermore, the synergistic mechanism between risk-weighted coverage and multi-objective optimization is as follows: The first objective is to maximize the risk-weighted coverage C_w, where C_w = Σ(w_k·n_k) / Σ(w_k·N_k), where k ∈ {high, medium, low}, w_k is the weight corresponding to each risk level (w1 > w2 > w3), n_k is the number of inspected k-th level towers, and N_k is the total number of k-th level towers; the second objective is to minimize the total inspection task time; when the meteorological-structural coupled risk index of the i-th tower... When the number R_c exceeds the coupling risk threshold, the risk weight of the tower body in the coverage calculation is dynamically amplified to w_k×(1+R_c), so that the tower body with high coupling risk gets a higher weight in the coverage target, realizing the dynamic driving of meteorological coupling risk on the path optimization target; a Pareto optimal solution set is generated by using non-dominated sorting and congestion distance selection strategies, and the solution with risk-weighted coverage not lower than the preset minimum coverage threshold and the shortest task time is selected from the Pareto optimal solution set as the final inspection path.

[0090] Furthermore, referring to Figure 4 The fourth embodiment of the UAV adaptive inspection path planning method of this application provides a flowchart, based on the above. Figure 4 The embodiment shown further refines the step of "generating an initial adaptive inspection path using a multi-objective optimization algorithm" in step S30, including steps A701 to A703: Step A701: When multiple drones are performing inspection tasks, obtain the remaining power of each drone and determine the spatial distribution of risk based on the risk level and the spatial location of the transmission tower.

[0091] It should be noted that, in this embodiment, multiple drones performing inspection tasks refers to a work mode in which two or more drones are dispatched simultaneously to conduct parallel inspections of transmission towers in the same transmission line section or different sub-regions within the same inspection operation period. Multi-drone collaborative operation can significantly shorten the overall inspection cycle and improve emergency response efficiency. Remaining power refers to the amount of electrical energy stored in the batteries of each drone at the current moment, which can be used to continue performing flight and inspection tasks. It is usually expressed in watt-hours or as a percentage of remaining power and is a key constraint parameter determining the amount of inspection tasks each drone can undertake. Spatial location refers to the latitude and longitude coordinates and altitude of each transmission tower in a unified geographical coordinate system, which is the basic data describing the spatial distribution characteristics of transmission towers. Risk spatial distribution refers to the two-dimensional or three-dimensional spatial field description formed by fusing the risk level attributes of each transmission tower with its spatial location information, used to intuitively characterize the risk density and risk gradient changes at different locations within the inspection area.

[0092] In one possible implementation, determining the spatial distribution of risk can be achieved by using a kernel density estimation algorithm to generate a continuous risk heatmap to visually represent risk clusters.

[0093] Step A702: Using an improved clustering-based task allocation algorithm, the area to be inspected is dynamically divided into several sub-regions according to the remaining power and the risk spatial distribution.

[0094] It should be noted that, in this embodiment, the improved clustering task allocation algorithm refers to a multi-drone collaborative inspection task partitioning method that is improved based on the traditional K-means spatial clustering algorithm by introducing the risk level of the transmission tower as a weighting factor and considering the constraint of the difference in the remaining power of the drones. The core improvement of this algorithm is that: the initial cluster center is determined by the current position coordinates of each drone, rather than being randomly initialized; the distance metric in the clustering iteration process adopts risk-weighted Euclidean distance, which makes the transmission towers with high risk levels generate a stronger attraction effect in distance calculation, and thus tend to be assigned to the clusters corresponding to drones with more remaining power. The area to be inspected refers to the entire geographical range planned to be covered by this inspection task, which is usually a narrow strip-shaped area composed of one or more transmission line corridors. A sub-region refers to a continuous or quasi-continuous geographical block that is independently inspected by a drone after the spatial partitioning of the area to be inspected by the improved clustering task allocation algorithm. Each sub-region does not overlap and together covers all the towers to be inspected.

[0095] This embodiment aims to decompose the large-scale centralized inspection path planning problem into multiple smaller sub-problems, reduce the computational complexity of single UAV path optimization, and at the same time ensure the balance of workload and reasonable allocation of risk coverage among UAVs.

[0096] In one specific implementation, the algorithm first uses the current initial position coordinates of each UAV as K initial cluster centers, where K equals the number of UAVs participating in the collaboration. Then, the algorithm enters an iterative optimization phase. In each iteration, the distance from each transmission tower to each cluster center is calculated. This distance is defined as the risk-weighted Euclidean distance, which is the standard Euclidean distance multiplied by a penalty factor positively correlated with the risk level of the target transmission tower. This makes high-risk towers more sensitive to the location of the cluster center in the cluster assignment decision. During the iteration, the cluster centers are updated based on the weighted average position of the transmission towers within their respective clusters. After convergence, the algorithm outputs K clusters, each corresponding to a sub-region. The transmission towers contained within the cluster are the inspection targets assigned to the corresponding UAV. The improved clustering task allocation algorithm, through the introduction of risk-weighted distance, implicitly considers the uneven distribution of risk during spatial clustering, allowing UAVs with sufficient remaining power or closer to high-risk areas to naturally undertake more inspection tasks for high-risk towers.

[0097] Step A703: Assign the inspection tasks in each sub-region to the corresponding UAVs, and control each UAV to independently execute the initial adaptive inspection path in the assigned sub-region.

[0098] It should be noted that, in this embodiment, the inspection task refers to the set of inspection operation instructions for all transmission towers to be inspected within a sub-region, including the spatial location of each tower, the matched inspection accuracy mode, the minimum inspection accuracy constraint, and the corresponding shooting strategy requirements. The corresponding UAV refers to the UAV determined by the task allocation algorithm that is responsible for executing the inspection task within a specific sub-region. Independent execution means that after receiving the assigned sub-region inspection task, each UAV, under the guidance of its own task planning module or in collaboration with the ground station, autonomously generates and executes an initial adaptive inspection path only for the transmission towers within its assigned sub-region. The flight paths and operation sequences of each UAV are decoupled and do not interfere with each other.

[0099] In one possible implementation, to ensure coverage integrity at the boundaries of each sub-region, the multi-machine collaborative sub-module sets up a sub-region overlap buffer zone when allocating tasks. That is, a certain width of overlap zone is set near the boundary of adjacent sub-regions. The power transmission towers within the overlap zone can be assigned to either of the two drones, and the final assignment is determined by the algorithm based on the power balance principle.

[0100] In another possible implementation, each UAV periodically reports its current location and remaining battery power to the ground station while independently performing inspection tasks. The ground station can monitor the overall task progress in real time and dynamically adjust the task boundaries of adjacent UAVs to take over the unfinished parts when a UAV is unable to complete its sub-area task due to insufficient battery power or malfunction, thereby realizing online task redistribution for multi-UAV collaboration.

[0101] This application dynamically divides sub-regions by integrating the remaining battery power of drones with the risk spatial distribution and using an improved clustering algorithm. It also controls each drone to independently execute an initial adaptive inspection path within its assigned region. This breaks the traditional fixed area distribution mode and achieves precise matching between multi-drone task load and actual endurance. It avoids the risk of drone crashes caused by low-battery drones undertaking tasks in high-risk, densely populated areas. Furthermore, through the parallel and non-interfering independent operation mode of multiple drones, it significantly reduces the overall inspection time of large-scale power transmission lines.

[0102] Furthermore, the steps of using a pre-trained ship-based detection model to identify transmission tower defects in real time during the inspection process, and dynamically adjusting the initial adaptive inspection path, switching the inspection mode for the current transmission tower, and replanning the remaining path when a transmission tower defect meeting preset conditions is identified include A801-A804: Step A801: Use a pre-trained ship detection model to perform real-time defect detection on the transmission tower and obtain the defect detection results. The ship detection model is a lightweight target detection network.

[0103] It should be noted that, in this embodiment, the pre-trained airborne detection model refers to a deep learning inference model deployed on the edge computing unit of the UAV, which has been offline trained and solidified before the inspection task is executed. This model is used for online target detection and defect classification of video streams or image frames acquired in real time during UAV inspections. The airborne detection model is a lightweight target detection network, which refers to a type of deep neural network model with fewer network parameters, lower computational complexity, faster inference speed, and suitability for real-time operation on embedded devices with limited computing power, such as YOLOv5-nano, MobileNet-SSD, or EfficientDet-D0. Real-time defect detection refers to the process during UAV flight inspections where the airborne detection model analyzes and processes each frame or several frames of image data acquired by the camera in real time, and outputs the location bounding box and category label of the defect target within milliseconds. The defect detection result refers to the set of structured information output by the airborne detection model after analyzing the current detection frame. It includes at least a Boolean identifier indicating whether a defect exists, a defect category label, the coordinates of the defect bounding box, and a quantitative description value of the defect attributes. The defect categories include at least two types: cracks and rust. The quantitative description values ​​of the defect attributes include at least the crack length value and the rust area percentage value.

[0104] In one possible implementation, the lightweight target detection network is trained offline using a large-scale transmission tower inspection image dataset before deployment. The training samples cover labeled images with different lighting conditions, different shooting angles, different tower types, and different defect severity. Data augmentation strategies are used to improve the model's generalization ability, and the model is pruned and quantized to adapt to the computing power and power consumption limitations of the UAV-borne embedded computing platform.

[0105] In another possible implementation, the airborne detection model is not limited to visible light image detection, but can also integrate infrared thermal imaging data to identify insulator deterioration or contact heating defects through temperature anomaly areas in the infrared image, forming a multimodal complement to the visible light detection results.

[0106] Step A802: Determine whether there are any serious defects in the defect detection results where the crack length exceeds the second preset threshold or the rust area ratio exceeds the third preset threshold.

[0107] It should be noted that, in this embodiment, crack length refers to the physical extension dimension of a linear crack caused by fatigue, stress corrosion, or material defects on or inside the surface of a transmission tower metal component, typically measured in millimeters. The second preset threshold is a pre-set length threshold used to determine whether a crack defect constitutes a serious defect. When the detected crack length exceeds the second preset threshold, the crack is considered to pose a substantial threat to the load-bearing capacity or fatigue life of the component. The corrosion area ratio refers to the ratio of the area covered by the corrosion oxide layer on the surface of the transmission tower metal component to the total visible surface area of ​​the component under the shooting angle, typically expressed as a percentage. The third preset threshold is a pre-set corrosion area ratio threshold used to determine whether a corrosion defect constitutes a serious defect. When the detected corrosion area ratio exceeds the third preset threshold, the cross-sectional loss of the component is considered to potentially affect the overall structural safety. A serious defect refers to a defect type that simultaneously meets one of the following conditions: crack length exceeds the second preset threshold, or corrosion area ratio exceeds the third preset threshold. Once a serious defect is confirmed, a dynamic path adjustment mechanism is triggered.

[0108] In addition, it should be noted that the criteria for determining serious defects can be further expanded to include other defect types such as exceeding the limit for the number of loose bolts and exceeding the limit for the number of self-exploding insulator discs. By expanding the attributes of the defect detection results and the corresponding preset thresholds, a comprehensive judgment of multiple types of defects can be achieved.

[0109] Step A803: If the serious defect exists, the inspection mode of the current transmission tower is switched from the current mode to the fine inspection mode for supplementary shooting, and the inspection priority of the same type of tower adjacent to the current transmission tower is increased.

[0110] It should be noted that, in this embodiment, the current mode refers to the inspection accuracy mode currently used by the UAV when performing inspection of the current transmission tower. This mode may be any one of the fine inspection mode, standard inspection mode, or rapid inspection mode, determined by the initial inspection path planning stage or the preceding dynamic adjustment stage. Supplementary shooting refers to an additional set of more detailed close-range multi-angle shooting operations performed after the original inspection mode of the current transmission tower has been completed. This aims to provide a more comprehensive image record of the discovered serious defects and their surrounding related structures, providing sufficient data support for manual review or quantitative defect analysis. Adjacent towers of the same model refer to related transmission towers that are geographically close to the transmission tower currently found to have serious defects and belong to the same design model, the same production batch, or the same construction period. Inspection priority refers to the weight level or access priority of each transmission tower in the path planning. The higher the priority, the more likely it is to be accessed first during path replanning.

[0111] In one possible implementation, the supplementary shooting includes repeatedly shooting the component where the defect is located and its adjacent connection parts from no less than 8 preset perspectives, and using optical zoom to get a high-resolution close-up image of the defect.

[0112] In addition, it should be noted that while increasing the priority of inspection of associated towers, real-time alarm information containing defect images, defect location coordinates, and suggested maintenance levels is also sent to the ground operation and maintenance monitoring center so that ground personnel can simultaneously initiate emergency response preparations.

[0113] Step A804: Based on the improved inspection priority, replan the path of the remaining uninspected transmission towers in the initial adaptive inspection path to generate an updated inspection path.

[0114] It should be noted that, in this embodiment, the remaining uninspected transmission towers refer to the set of all transmission towers that have not yet been inspected at the current moment, including towers that were not originally planned to be visited and towers that have been newly included in the priority inspection scope due to their association. Replanning refers to the process of online secondary optimization of the access order of the remaining uninspected transmission towers, using the current location of the UAV as the new path starting point and the improved inspection priority as the new weight input, under the conditions of remaining power and other constraints. The updated inspection path refers to a new flight path generated after replanning, starting from the current location of the UAV, traversing all remaining uninspected transmission towers, and finally returning to the take-off and landing point or the designated recovery point. This path will replace the original remaining path and be executed by the UAV.

[0115] In one possible implementation, the lightweight path replanning algorithm has a preset solution time limit to ensure that the real-time performance of dynamic adjustments does not affect the continuous flight operation of the UAV.

[0116] In another possible implementation, the replanning can employ a global replanning algorithm to re-include all remaining uninspected transmission towers in the calculation, or a local replanning algorithm to quickly fine-tune the local routes of only a few adjacent waypoints affected by priority changes in order to save computation time.

[0117] This application identifies critical defects in real time using an airborne model, triggering a linkage adjustment mechanism that combines precise re-examination of the current tower with priority enhancement of adjacent towers, and replans the remaining inspection path accordingly. It achieves intelligent adaptive decision-making, moving from "following the map" to "adapting to changing circumstances," completely eliminating the lag of traditional offline analysis. Through a space contagion defense strategy of "single-point discovery and full-line early warning," it not only ensures precise evidence collection of current critical defects but also accurately directs limited remaining flight resources to adjacent towers with potential degradation risks, significantly improving the ability to respond to emergencies and the overall defect interception rate during inspection operations.

[0118] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the UAV adaptive inspection path planning method of this application. Any simple modifications based on this technical concept are within the protection scope of this application.

[0119] This application also provides an adaptive inspection path planning system for unmanned aerial vehicles (UAVs). Please refer to [link / reference]. Figure 5 The UAV adaptive inspection path planning system includes: The structural health assessment module 10 is used to obtain the current structural health score of each transmission tower based on a pre-built comprehensive health scoring system to determine the corresponding risk level. The inspection accuracy matching module 20 is used to match a preset differentiated inspection accuracy mode for each transmission tower according to the risk level. The path optimization module 30 is used to generate an initial adaptive inspection path with risk-weighted coverage as the objective, combined with the drone's endurance and flight safety constraints, using a multi-objective optimization algorithm. The dynamic adjustment module 40 is used to identify transmission tower defects in real time during the inspection process using a pre-trained ship detection model. When a transmission tower defect that meets the preset conditions is identified, the initial adaptive inspection path is dynamically adjusted, the inspection mode for the current transmission tower is switched, and the remaining path is replanned. The data update module 50 is used to feed back the test results to the comprehensive health scoring system after the inspection is completed in order to update the risk level.

[0120] The UAV adaptive inspection path planning system provided in this application, employing the UAV adaptive inspection path planning method in the above embodiments, can solve the technical problems of existing technologies that cannot perform differentiated inspections based on the actual structural risk differences of transmission towers, nor can they dynamically adjust the path based on real-time detection results during the inspection process. Compared with the prior art, the beneficial effects of the UAV adaptive inspection path planning system provided in this application are the same as those of the UAV adaptive inspection path planning method provided in the above embodiments, and other technical features of the UAV adaptive inspection path planning system are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0121] This application provides an adaptive inspection path planning device for unmanned aerial vehicles (UAVs). The UAV adaptive inspection path planning device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the UAV adaptive inspection path planning method in the above embodiment 1.

[0122] The following is for reference. Figure 6The diagram illustrates a structural schematic of a drone adaptive inspection path planning device suitable for implementing embodiments of this application. The drone adaptive inspection path planning device in this application embodiment may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Description), PMPs (Portable Media Players), vehicle terminals (e.g., vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 6 The drone adaptive inspection path planning device shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0123] like Figure 6 As shown, the UAV adaptive inspection path planning device may include a processing unit 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory 1002 or a program loaded from a storage device 1003 into a random access memory 1004. The random access memory 1004 also stores various programs and data required for the operation of the UAV adaptive inspection path planning device. The processing unit 1001, the read-only memory 1002, and the random access memory 1004 are interconnected via a bus 1005. An input / output interface 1006 is also connected to the bus. Typically, the following systems can be connected to the input / output interface 1006: input devices 1007 including, for example, a touchscreen, touchpad, keyboard, mouse, image sensor, microphone, accelerometer, gyroscope, etc.; output devices 1008 including, for example, a liquid crystal display (LCD), speaker, vibrator, etc.; storage devices 1003 including, for example, magnetic tape, hard disk, etc.; and communication devices 1009. Communication device 1009 allows the UAV adaptive inspection path planning device to communicate wirelessly or wiredly with other devices to exchange data. While the figure shows a UAV adaptive inspection path planning device with various systems, it should be understood that implementation or possession of all the systems shown is not required. More or fewer systems may be implemented alternatively.

[0124] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from read-only memory 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.

[0125] The UAV adaptive inspection path planning device provided in this application, employing the UAV adaptive inspection path planning method described in the above embodiments, can solve the technical problems of existing technologies that cannot perform differentiated inspections based on the actual structural risk differences of transmission towers, nor can they dynamically adjust the path based on real-time detection results during the inspection process. Compared with the prior art, the beneficial effects of the UAV adaptive inspection path planning device provided in this application are the same as those of the UAV adaptive inspection path planning method provided in the above embodiments, and other technical features of this UAV adaptive inspection path planning device are the same as those disclosed in the previous embodiment method, and will not be repeated here.

[0126] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0127] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0128] This application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, which are used to execute the UAV adaptive inspection path planning method in the above embodiments.

[0129] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems or devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0130] The aforementioned computer-readable storage medium may be included in the UAV adaptive inspection path planning device; or it may exist independently and not be assembled into the UAV adaptive inspection path planning device.

[0131] The aforementioned computer-readable storage medium carries one or more programs. When these programs are executed by the UAV adaptive inspection path planning device, the UAV adaptive inspection path planning device: obtains the current structural health score of each transmission tower based on a pre-built comprehensive health scoring system to determine the corresponding risk level; matches a preset differentiated inspection accuracy mode to each transmission tower according to the risk level; generates an initial adaptive inspection path using a multi-objective optimization algorithm with risk-weighted coverage as the objective, combined with UAV endurance and flight safety constraints; during the inspection process, it uses a preset aircraft-ship detection model to identify transmission tower defects in real time; when a transmission tower defect that meets preset conditions is identified, it dynamically adjusts the initial adaptive inspection path, switches the inspection mode for the current transmission tower, and replans the remaining path; after the inspection is completed, it feeds back the detection results to the comprehensive health scoring system to update the risk level.

[0132] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a Local Area Network (LAN) or a Wide Area Network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0133] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0134] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0135] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A method for adaptive inspection path planning of unmanned aerial vehicles (UAVs), characterized in that, The UAV adaptive inspection path planning method includes: The current structural health score of each transmission tower is obtained based on a pre-built comprehensive health scoring system to determine the corresponding risk level; Based on the aforementioned risk level, a preset differentiated inspection accuracy mode is matched for each transmission tower; With risk-weighted coverage as the objective, and taking into account the constraints of UAV endurance and flight safety, a multi-objective optimization algorithm is used to generate an initial adaptive inspection path; During the inspection process, a pre-trained ship-based detection model is used to identify transmission tower defects in real time. When a transmission tower defect that meets the preset conditions is identified, the initial adaptive inspection path is dynamically adjusted, the inspection mode for the current transmission tower is switched, and the remaining path is replanned. After the inspection is completed, the test results will be fed back to the comprehensive health scoring system to update the risk level.

2. The UAV adaptive inspection path planning method as described in claim 1, characterized in that, The steps for obtaining the current structural health score of each transmission tower based on a pre-built comprehensive health scoring system to determine the corresponding risk level include: Structural condition data for each transmission tower are collected. These data include at least one or more of the following: tower tilt, component corrosion degree, bolt loosening rate, foundation settlement, and insulator deterioration degree. The structural state index data is normalized to obtain normalized index data; Based on the normalized index data, the subjective weights of each structural state index data are determined by the analytic hierarchy process (AHP), and the comprehensive weights of each structural comprehensive state index data are obtained based on the subjective weights and the objective weights determined by the entropy increase method. The comprehensive health score of each transmission tower is obtained by weighted calculation based on the normalized index data and the comprehensive weight. Based on the preset data range in which the comprehensive health score falls, each transmission tower is divided into three risk levels: high risk, medium risk, and low risk.

3. The UAV adaptive inspection path planning method as described in claim 1, characterized in that, The step of obtaining the current structural health score and corresponding risk level of each transmission tower based on the pre-built comprehensive health scoring system includes: The historical comprehensive health score sequence of each transmission tower is obtained, and the preset time series prediction model is input to predict the health score change curve within the target set time window. The time series prediction model is a long short-term memory network. The degradation rate of the health score change curve is calculated. When the degradation rate exceeds the preset threshold of the statistical distribution of degradation rates of the same type of transmission tower group, the current risk level of the corresponding transmission tower is increased, and the risk level after preventive correction is output.

4. The UAV adaptive inspection path planning method as described in claim 3, characterized in that, The step of matching a preset differentiated inspection accuracy mode to each transmission tower according to the risk level includes: Obtain the risk level after preventative correction; Based on the mapping relationship between risk level and inspection mode, high-risk towers are matched with fine inspection mode, medium-risk towers are matched with standard inspection mode, and low-risk towers are matched with rapid inspection mode.

5. The UAV adaptive inspection path planning method as described in claim 4, characterized in that, The steps for generating an initial adaptive inspection path using a multi-objective optimization algorithm, with risk-weighted coverage as the objective and considering UAV endurance and flight safety constraints, include: Based on the differentiated inspection accuracy pattern obtained by matching, the minimum inspection accuracy constraint corresponding to each transmission tower is determined. Obtain the current battery level and preset flight safety parameters of the drone, and construct a set of constraint conditions in conjunction with the minimum inspection accuracy constraint; Transmission towers of high-risk, medium-risk, and low-risk levels are assigned a first weight, a second weight, and a third weight, respectively, wherein the first weight is greater than the second weight, the second weight is greater than the third weight, and a path optimization model is constructed based on the first weight, the second weight, and the third weight, with the objective function being to maximize the risk-weighted coverage. A multi-objective optimization algorithm is used to calculate and generate the initial adaptive inspection path based on the objective function and the set of constraints.

6. The UAV adaptive inspection path planning method as described in claim 5, characterized in that, The constraint set also includes meteorological-structural coupled risk constraints, specifically: Obtain meteorological data to obtain ambient wind speed; When the ambient wind speed exceeds a first preset threshold, the inspection priority of the transmission towers located in wind-sensitive areas is increased, and the flight safety parameters of the UAV are adjusted simultaneously.

7. The UAV adaptive inspection path planning method as described in claim 5, characterized in that, The step of generating the initial adaptive inspection path using a multi-objective optimization algorithm also includes: When multiple drones are performing inspection tasks, the remaining power of each drone is obtained, and the risk spatial distribution is determined by combining the risk level and the spatial location of the transmission tower. An improved clustering-based task allocation algorithm is used to dynamically divide the area to be inspected into several sub-regions according to the remaining power and the risk spatial distribution. The inspection tasks in each of the sub-regions are assigned to the corresponding UAVs, and each UAV is controlled to independently execute the initial adaptive inspection path in the assigned sub-region.

8. The UAV adaptive inspection path planning method as described in claim 5, characterized in that, The steps of using a pre-trained ship-based detection model to identify transmission tower defects in real time during the inspection process, and dynamically adjusting the initial adaptive inspection path, switching the inspection mode for the current transmission tower, and replanning the remaining path when a defect meeting preset conditions is identified include: A pre-trained ship-machine detection model is used to perform real-time defect detection on the transmission tower and obtain the defect detection results. The ship-machine detection model is a lightweight target detection network. Determine whether there are any serious defects in the defect detection results where the crack length exceeds the second preset threshold or the rust area ratio exceeds the third preset threshold; If the aforementioned serious defects exist, the inspection mode of the current transmission tower will be switched from the current mode to the fine inspection mode to take supplementary photos, and the inspection priority of towers of the same model adjacent to the current transmission tower will be increased. Based on the improved inspection priority, the path of the remaining uninspected transmission towers in the initial adaptive inspection path is replanned to generate an updated inspection path.

9. An adaptive inspection path planning system for unmanned aerial vehicles (UAVs), characterized in that, The UAV adaptive inspection path planning system includes: The structural health assessment module is used to obtain the current structural health score of each transmission tower based on a pre-built comprehensive health scoring system to determine the corresponding risk level. The inspection accuracy matching module is used to match a preset differentiated inspection accuracy mode for each transmission tower according to the risk level. The path optimization module is used to generate an initial adaptive inspection path with risk-weighted coverage as the objective, combined with the drone's endurance and flight safety constraints, using a multi-objective optimization algorithm. The dynamic adjustment module is used to identify transmission tower defects in real time during the inspection process using a pre-trained ship detection model. When a transmission tower defect that meets the preset conditions is identified, the initial adaptive inspection path is dynamically adjusted, the inspection mode for the current transmission tower is switched, and the remaining path is replanned. The data update module is used to feed back the test results to the comprehensive health scoring system after the inspection is completed in order to update the risk level.

10. An adaptive inspection path planning device for unmanned aerial vehicles (UAVs), characterized in that, The device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the UAV adaptive inspection path planning method as described in any one of claims 1 to 8.

Images