Unmanned aerial vehicle intelligent inspection operation management method and system

Abstract

The invention discloses an unmanned aerial vehicle intelligent inspection operation management method and system, and belongs to the technical field of inspection management. The system comprises a dynamic sensing module, an inspection analysis module, an operation management module and an instruction control module. The dynamic sensing module collects a GIS map and meteorological data of an inspection area and historical logs and operation data of each unmanned aerial vehicle; building a virtual scene and mapping the position of the unmanned aerial vehicle in real time; the inspection analysis module marks the position of each inspection point in the virtual scene, calculates the efficiency index between the unmanned aerial vehicle and the inspection point, and binds the inspection point for each unmanned aerial vehicle; the operation management module establishes an inspection route for each unmanned aerial vehicle, and plans nodes according to bound inspection points; continuing to calculate the efficiency index through the node position and binding other inspection points, and iteratively adjusting the inspection route until all the inspection points are covered; and the instruction control module issues the inspection route to the corresponding unmanned aerial vehicle and controls the unmanned aerial vehicle to execute inspection operation according to the inspection route.

Description

Technical Field

[0001] This invention relates to the field of inspection management technology, specifically to a method and system for managing intelligent inspection operations using unmanned aerial vehicles (UAVs). Background Technology

[0002] With the continuous development of technology and the improvement of industrial automation, the application scope of drones in various industries is expanding. Leveraging their flexibility, efficiency, and ability to access sparsely populated areas, drones are gradually becoming a key tool for inspection operations. Drone inspections effectively improve operational efficiency and safety, reduce labor costs, and are a core technological means to achieve intelligent inspection.

[0003] Currently, drone inspections typically employ fixed zoning or polling mechanisms, where each drone is pre-set to a fixed or nearest inspection point. These methods have significant limitations. Traditional fixed zoning strategies fail to adequately consider the impact of dynamic environmental changes on energy consumption prediction. In windy or low-temperature environments, drones in different locations consume differently in different flight directions: windward flight typically consumes more energy than leeward flight; in low-temperature environments, the temperature gradient between the windward and leeward sides of the drone is significant, and battery performance is highly sensitive to temperature changes, resulting in significant differences in battery discharge rates when drones fly at different speeds and in different directions. Traditional polling mechanisms lack dynamic performance assessment and conflict resolution capabilities, failing to consider dynamic environmental changes or differences in battery degradation rates or wind resistance performance between individual drones, still requiring manual intervention in unforeseen circumstances. Therefore, there is an urgent need to develop a more intelligent and efficient drone inspection operation management technology solution to address these issues. Summary of the Invention

[0004] The purpose of this invention is to provide a method and system for managing intelligent inspection operations using unmanned aerial vehicles (UAVs) to solve the problems mentioned in the background art.

[0005] To address the aforementioned technical problems, this invention provides a method for managing intelligent inspection operations using unmanned aerial vehicles (UAVs), comprising:

[0006] The S100 system collects GIS maps and meteorological data of the inspection area, as well as historical logs and operational data of each drone. A virtual scene is built using the GIS map and meteorological data, and the drone's location is mapped in real time within the virtual scene.

[0007] Meteorological data includes the distribution of meteorological elements within the inspection area, including wind direction, wind speed, temperature, and humidity.

[0008] The historical logs include the operation records of each drone flight, and each operation record includes the changes in the drone's operation data within the flight time.

[0009] The operational data includes a 3D model of the drone, as well as real-time GPS, battery, attitude, and temperature parameters. GPS parameters include spatial location, flight direction, and flight speed. Attitude parameters are acquired in real-time by inertial sensors, specifically including acceleration, angular velocity, and angle. Temperature parameters include the surface temperature field distribution of the drone.

[0010] A virtual scene is built based on a GIS map. The 3D models of each UAV are loaded into the virtual scene. The spatial position of each 3D model in the virtual scene, as well as the pitch, roll and yaw angles, are mapped in real time based on GPS parameters and attitude parameters.

[0011] By using acceleration and angular velocity data collected by inertial sensors and combining them with GPS spatial location, the Euler angles of the UAV's 3D model are calculated in real time in a virtual scene to ensure that the attitude simulation is consistent with the physical state.

[0012] By integrating GIS maps, meteorological data, and real-time drone operation data, a high-precision virtual scene is constructed and the three-dimensional attitude of the drone is dynamically mapped to achieve a global simulation of the environment and equipment status.

[0013] S200: Mark the locations of each inspection point in the virtual scene, and calculate the efficiency index between the drone and the inspection point by combining operational data and meteorological data, thereby assigning each drone to an inspection point. Specifically, this includes:

[0014] S201. Mark the spatial locations of inspection points and drones in the virtual scene, and plan the flight path between each drone and each inspection point as a reference path.

[0015] S202. Obtain the historical logs of each drone, analyze the relationship between surface temperature field changes and electrical charge changes, and construct a training set. Establish relational expressions for each drone based on its training set. Specifically, this includes:

[0016] S2021. Obtain the historical logs of the UAV and analyze the surface temperature field changes of the fuselage in each operation record. Set r sampling points on the surface of the UAV's 3D model and map the surface temperature field. Specific steps include:

[0017] First, locate the area where the battery is located in the 3D model of the UAV, analyze the position of the center point of the area and use it as the source point.

[0018] Secondly, starting from the source point, establish r three-dimensional vectors with different directions uniformly on the sphere. The angle between adjacent vectors in any plane transversely tangent to the source point is the same.

[0019] Finally, the intersection point of each vector with the surface of the UAV's 3D model is taken as a sampling point. The UAV has a total of the same number of sampling points as the number of vectors.

[0020] The surface temperature field distribution collected by the infrared sensor is mapped onto the three-dimensional model to provide input for the subsequent thermal management model.

[0021] S2022. Divide each operation record into time periods. Within each time period, use the temperature at each sampling point as the independent variable and the power consumption rate as the dependent variable, and package them into samples. Combine these samples to construct a training set for the UAV. Specifically, this includes:

[0022] S2022-1. Set the duration to time. Divide the flight duration of each run record by time to obtain the number of samples Q. Divide the run record into Q non-overlapping time periods with a duration of time.

[0023] S2022-2. Within each time period, set e time points evenly, analyze the flight direction and speed of the UAV at different time points, as well as the temperature at each sampling point, and substitute them into the formula to calculate the anomaly index (ANI) for each time period:

[0024]

[0025] In the formula, α, β, and γ are constants. V represents the temperature at the j-th sampling point at the i-th time point. i Let be the flight speed at the i-th time point.

[0026] ANG i,(i-1) Let be the angle between the flight direction vector at time point i and the flight direction vector at time point i-1.

[0027] By quantifying the coupling effect between temperature fluctuations and sudden motion changes, data from time periods with stable flight conditions are selected.

[0028] The numerator calculates the mean of the temperature change rate at all sampling points, while the denominator combines abrupt changes in heading and the rate of change in velocity to effectively eliminate noisy data under abnormal flight conditions. This ensures that the training set for the subsequent temperature-electricity relationship model only contains normal flight data that conforms to physical laws, thereby improving the reliability of model predictions.

[0029] S2022-3. Select time periods where the anomaly index is less than the threshold, and use the temperature of r sampling points at each time point in these time periods as independent variables and the power consumption rate as dependent variables, and package them into samples.

[0030] S2022-4. Combine these w×e samples to construct a training set for unmanned aerial vehicles (UAVs). Here, w represents the number of time periods where the anomaly index is less than the threshold.

[0031] By filtering out time periods of stable flight conditions using the anomaly index, the training set for temperature and power consumption rate constructed based on this data becomes more valuable for reference.

[0032] S2023. Establish a linear regression model, inputting all independent variables from each sample in the training set, and using the difference between the output value Y and the dependent variable as the difference coefficient for the corresponding sample. The expression for the linear regression model is:

[0033] Y = d + f1X1 + ... + f r X r

[0034] In the formula, d is the intercept, and f r Let X be the regression coefficient for the r-th sampling point. r Let be the temperature of the r-th sampling point.

[0035] A quantitative relationship between fuselage surface temperature distribution and power consumption rate was established based on historical operating data. The temperature of r sampling points was used as the independent variable, and the regression coefficient f was... r The weighting reflects the impact of temperature at different locations on power consumption, and the intercept d represents the base power consumption level.

[0036] After training, the rate of power consumption can be predicted by inputting any temperature distribution, providing energy consumption data support for the calculation of the efficiency index and realizing accurate modeling of the impact of the fuselage thermal field on battery degradation.

[0037] S2024. Adjust the intercept and regression coefficients until the sum of the difference coefficients of all samples is minimized, thus obtaining the relational expression of the trained UAV.

[0038] S2025, and so on, build a training set and a linear regression model for each drone, and train to obtain the relational expression.

[0039] S203. Each training set is input into the physical model for training. The trained physical model predicts the surface temperature field changes of the UAV when flying each reference path based on the current temperature parameters, meteorological data and reference path of the UAV.

[0040] A physical model is created for each drone, and each training set is put into the corresponding physical model for training.

[0041] The trained physical model simulates and analyzes the changes in the surface temperature field of the UAV as it flies along the reference path, based on the current surface temperature field distribution of the UAV and the distribution of meteorological elements in the vicinity of the reference path within the inspection area.

[0042] Surface temperature field change refers to the temperature distribution at different locations on the surface of a drone over a period of time and how this temperature changes over time. The duration of this change is specifically taken as the length of the reference path divided by the drone's normal flight speed.

[0043] S204. By analyzing the changes in surface temperature field and their relational expressions, the change in electrical charge is obtained. Based on the change in electrical charge and the reference path, the efficiency index between the drone and the inspection point is calculated, thereby assigning an inspection point to each drone. Specifically, this includes:

[0044] S2041. Based on the duration T of the surface temperature field change, K time points are uniformly set. Substitute the temperature of all sampling points at each time point into the relational expression to analyze the power consumption rate ER at each time point. c .

[0045] S2042. Analyze the UAV and inspection point INP corresponding to the reference path, and obtain the maximum duration T of the surface temperature field change between inspection point INP and other UAVs. max .

[0046] S2043. Calculate the efficiency index between the UAV and the inspection point INP, and assign each UAV to the inspection point with the lowest efficiency index. The efficiency index calculation formula is as follows:

[0047]

[0048] In the formula, N is a constant greater than 1, and EFI is the efficiency index.

[0049] The efficiency index combines time cost and energy cost to quantify the overall efficiency of a drone in performing a certain inspection point task.

[0050] Time cost: Amplify the difference in flight duration (T is the current path duration, T max (The longest time to bind all drones to this point) is selected, prioritizing the path with the shortest execution time.

[0051] Energy consumption cost: Cumulative path segmented power consumption (ER) c ), punish high-power-consuming paths.

[0052] The final output EFI value serves as the basis for binding decisions, achieving optimal inspection point binding with low time consumption and low energy consumption.

[0053] Based on historical operational data, a unique temperature-power coupling model is trained for each drone to predict temperature field changes and power consumption along the flight path. The performance index (EFI) is used to quantify and match the best and worst performance, achieving optimal binding between the drone and the inspection point and avoiding the risk of rapid power decay caused by local high or low temperatures on the drone surface during flight.

[0054] The S300 establishes inspection routes for each drone and plans nodes based on the bound inspection points. It continues to calculate and bind the efficiency index between nodes and other inspection points, iteratively adjusting the inspection routes until all inspection points are covered. Specifically, this includes:

[0055] S301. Establish an inspection route for each drone, using the bound inspection point as the first node. Analyze again the efficiency index between each drone and other inspection points when it is at the first node.

[0056] S302. Each drone continues to be bound to the inspection point with the lowest energy efficiency index as the second node. When the inspection point is bound repeatedly by different drones, the drone with the lowest energy efficiency index is automatically bound, and the other drones are unbound.

[0057] When an inspection point is bound to multiple drones, only the one with the lowest EFI is retained, and the remaining drones trigger an unbinding-rebinding cycle.

[0058] If drones A and B are both bound to inspection points, when EFI_A = 1.2 and EFI_B = 1.5, drone B is unbound and rebound to another inspection point with the lowest EFI.

[0059] S303. After the unbound drones are removed from the unbound inspection points, they are rebound to other inspection points with the lowest energy efficiency index. The energy efficiency index of each drone when it is at the second node position is analyzed again compared to other inspection points.

[0060] S304. By continuously iterating and adjusting the inspection points bound to the drones, the nodes of the inspection route are planned sequentially. The inspection route planning for each drone is completed when all inspection points are bound to the drones and there are no duplicate bindings.

[0061] By cascading and binding nodes and resolving conflicts, the inspection route is dynamically expanded to achieve full coverage through multi-machine collaboration and avoid task conflicts.

[0062] Once the S400 inspection route is planned, it is sent to the corresponding drone, which then controls the drone to perform inspection operations according to the inspection route.

[0063] The planned inspection routes are compiled into encrypted waypoint sequences and distributed to the corresponding drones.

[0064] The drone verifies the feasibility of the inspection points in real time and performs inspection operations in sequence according to the order of each node in the inspection route. Simultaneously, it dynamically adjusts its flight attitude based on environmental parameters such as wind speed to ensure the accurate execution of the inspection mission.

[0065] By issuing encrypted waypoint sequences and performing dynamic attitude corrections during flight, mission safety and execution accuracy are ensured, and environmental changes such as wind speed disturbances are adapted.

[0066] The present invention also provides an intelligent inspection operation management system for unmanned aerial vehicles (UAVs), including a dynamic perception module, an inspection analysis module, an operation management module, and a command control module.

[0067] The dynamic sensing module collects GIS maps and meteorological data of the inspection area, as well as historical logs and operational data of each drone. It then builds a virtual scene and maps the drone's location in real time.

[0068] Collect GIS maps and meteorological data of the inspection area, as well as historical logs and real-time operational data of each drone.

[0069] A virtual scene is built based on a GIS map, a 3D model of the drone is loaded, and GPS parameters and attitude parameters are used to map the spatial position and attitude state of the drone in the virtual scene in real time.

[0070] Through real-time data acquisition and scenario simulation, it provides high-precision visualization of the environment and UAV status, laying a data foundation for subsequent path planning and enhancing the system's ability to perceive dynamic environments.

[0071] The inspection analysis module marks the location of each inspection point in the virtual scene, calculates the efficiency index between the drone and the inspection point, and thus binds each drone to an inspection point.

[0072] The locations of inspection points are marked in the virtual scene to serve as reference paths between each drone and the inspection points. Historical logs are analyzed to build a training set, and a physical model is trained to predict the changes in the surface temperature field of the drones along the reference paths. This leads to the calculation of power changes and efficiency index, and finally, each drone is assigned the inspection point with the lowest efficiency index.

[0073] By optimizing the binding process through quantitative efficiency index, intelligent matching between drones and inspection points can be achieved, reducing energy consumption risks, improving inspection efficiency and reliability, and avoiding flight interruptions caused by abnormal temperature or power levels.

[0074] The operation management module establishes inspection routes for each drone and plans nodes based on the bound inspection points. It then calculates the efficiency index based on the node locations and binds other inspection points, iteratively adjusting the inspection routes until all inspection points are covered.

[0075] Starting from the bound inspection point, an inspection route is established for each drone; the efficiency index of the drone at the node position and other inspection points is calculated iteratively, new nodes are bound and duplicate bindings are handled, and the route is adjusted until all inspection points are covered.

[0076] It enables collaborative path optimization among multiple drones, ensuring no inspection points are missed or conflicted through iterative adjustments, improving resource utilization and operational coverage, reducing redundant energy consumption, and supporting intelligent scheduling in large-scale complex areas.

[0077] The command and control module sends the inspection route to the corresponding drone and controls the drone to perform the inspection operation according to the inspection route.

[0078] The planned inspection routes are compiled into encrypted waypoint sequences and distributed to the corresponding drones. The drones verify the feasibility of the inspection points in real time and dynamically adjust their flight attitude according to environmental parameters to ensure that the inspection operations are performed in the order of the nodes.

[0079] Encrypted transmission and real-time verification ensure mission security and data integrity, while dynamic adjustments enhance flight stability and accuracy, ensuring efficient and reliable execution of inspection missions and adaptability to changing environmental conditions.

[0080] Compared with the prior art, the beneficial effects achieved by the present invention are:

[0081] Precision thermal management: A power consumption prediction model is established based on the temperature field distribution on the fuselage surface. Through regression analysis of sampling points, the impact of local high temperature on battery degradation is quantified, overcoming the shortcomings of traditional methods that ignore the thermal conduction of the fuselage and significantly reducing operational risks.

[0082] Dynamic collaborative optimization: Introducing the Efficiency Index (EFI) to integrate time and energy costs, combined with an iterative binding mechanism (such as unbinding inefficient drones), enables flexible adjustment of multi-drone task allocation, solving coverage blind spots and conflict issues caused by fixed partitions.

[0083] Enhanced anti-interference control: Ensuring command security through encrypted waypoint sequence transmission and dynamically correcting flight attitude based on wind speed vectors (such as yaw angle compensation) to improve positioning accuracy and execution stability under complex weather conditions.

[0084] Intelligent data cleaning: The Anomalous Movement Index (ANI) formula is used to filter stable flight data, eliminating the interference of sudden motion changes (sharp turns / acceleration / deceleration) on the temperature-electricity relationship model, thereby enhancing the reliability of predictions. Attached Figure Description

[0085] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0086] Figure 1 This is a flowchart illustrating an intelligent inspection operation management method for unmanned aerial vehicles (UAVs) according to the present invention.

[0087] Figure 2 This is a schematic diagram of the structure of an intelligent inspection and management system for unmanned aerial vehicles (UAVs) according to the present invention. Detailed Implementation

[0088] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0089] Please see Figure 1 This invention provides a method for managing intelligent inspection operations using unmanned aerial vehicles (UAVs), comprising:

[0090] The S100 system collects GIS maps and meteorological data of the inspection area, as well as historical logs and operational data of each drone. A virtual scene is built using the GIS map and meteorological data, and the drone's location is mapped in real time within the virtual scene.

[0091] Meteorological data includes the distribution of meteorological elements within the inspection area, including wind direction, wind speed, temperature, and humidity.

[0092] The historical logs include the operation records of each drone flight, and each operation record includes the changes in the drone's operation data within the flight time.

[0093] The operational data includes a 3D model of the drone, as well as real-time GPS, battery, attitude, and temperature parameters. GPS parameters include spatial location, flight direction, and flight speed. Attitude parameters are acquired in real-time by inertial sensors, specifically including acceleration, angular velocity, and angle. Temperature parameters include the surface temperature field distribution of the drone.

[0094] A virtual scene is built based on a GIS map. The 3D models of each UAV are loaded into the virtual scene. The spatial position of each 3D model in the virtual scene, as well as the pitch, roll and yaw angles, are mapped in real time based on GPS parameters and attitude parameters.

[0095] By using acceleration and angular velocity data collected by inertial sensors and combining them with GPS spatial position, the Euler angles (pitch / roll / yaw) of the UAV's 3D model are calculated in real time in a virtual scene to ensure that the attitude simulation is consistent with the physical state.

[0096] By integrating GIS maps, meteorological data, and real-time UAV operational data (GPS, attitude, and temperature parameters), a high-precision virtual scene is constructed and the UAV's three-dimensional attitude (pitch / roll / yaw) is dynamically mapped to achieve a global simulation of the environment and equipment status.

[0097] S200: Mark the locations of each inspection point in the virtual scene, and calculate the efficiency index between the drone and the inspection point by combining operational data and meteorological data, thereby assigning each drone to an inspection point. Specifically, this includes:

[0098] S201. Mark the spatial locations of inspection points and drones in the virtual scene, and plan the flight path between each drone and each inspection point as a reference path.

[0099] S202. Obtain the historical logs of each drone, analyze the relationship between surface temperature field changes and electrical charge changes, and construct a training set. Establish relational expressions for each drone based on its training set. Specifically, this includes:

[0100] S2021. Obtain the historical logs of the UAV and analyze the surface temperature field changes of the fuselage in each operation record. Set r sampling points on the surface of the UAV's 3D model and map the surface temperature field. Specific steps include:

[0101] First, locate the area where the battery is located in the 3D model of the UAV, analyze the position of the center point of the area and use it as the source point.

[0102] Secondly, starting from the source point, establish r three-dimensional vectors with different directions uniformly on the sphere. The angle between adjacent vectors in any plane transversely tangent to the source point is the same.

[0103] Finally, the intersection point of each vector with the surface of the UAV's 3D model is taken as a sampling point. The UAV has a total of the same number of sampling points as the number of vectors.

[0104] The surface temperature field distribution collected by the infrared sensor is mapped onto the three-dimensional model to provide input for the subsequent thermal management model.

[0105] S2022. Divide each operation record into time periods. Within each time period, use the temperature at each sampling point as the independent variable and the power consumption rate as the dependent variable, and package them into samples. Combine these samples to construct a training set for the UAV. Specifically, this includes:

[0106] S2022-1. Set the duration to time. Divide the flight duration of each run record by time to obtain the number of samples Q. Divide the run record into Q non-overlapping time periods with a duration of time.

[0107] S2022-2. Within each time period, set e time points evenly, analyze the flight direction and speed of the UAV at different time points, as well as the temperature at each sampling point, and substitute them into the formula to calculate the anomaly index (ANI) for each time period:

[0108]

[0109] In the formula, α, β, and γ are constants. V represents the temperature at the j-th sampling point at the i-th time point. i Let be the flight speed at the i-th time point.

[0110] ANG i,(i-1) Let be the angle between the flight direction vector at time point i and the flight direction vector at time point i-1.

[0111] By quantifying the coupling effect between temperature fluctuations and sudden motion changes, data from time periods when the flight status is stable (without sharp turns or rapid accelerations) are selected.

[0112] The numerator calculates the mean of the temperature change rate at all sampling points, while the denominator combines abrupt changes in heading and the rate of change in velocity to effectively eliminate noisy data under abnormal flight conditions. This ensures that the training set for the subsequent temperature-electricity relationship model only contains normal flight data that conforms to physical laws, thereby improving the reliability of model predictions.

[0113] S2022-3. Select time periods where the anomaly index is less than the threshold, and use the temperature of r sampling points at each time point in these time periods as independent variables and the power consumption rate as dependent variables, and package them into samples.

[0114] S2022-4. Combine these w×e samples to construct a training set for unmanned aerial vehicles (UAVs). Here, w represents the number of time periods where the anomaly index is less than the threshold.

[0115] By filtering out time periods of stable flight conditions using the anomaly index, the training set for temperature and power consumption rate constructed based on this data becomes more valuable for reference.

[0116] S2023. Establish a linear regression model, inputting all independent variables from each sample in the training set, and using the difference between the output value Y and the dependent variable as the difference coefficient for the corresponding sample. The expression for the linear regression model is:

[0117] Y = d + f1X1 + ... + f r X r

[0118] In the formula, d is the intercept, and f r Let X be the regression coefficient for the r-th sampling point. r Let be the temperature of the r-th sampling point.

[0119] A quantitative relationship between fuselage surface temperature distribution and power consumption rate was established based on historical operating data. The temperature of r sampling points was used as the independent variable, and the regression coefficient f was... r The weighting reflects the impact of temperature at different locations on power consumption, and the intercept d represents the base power consumption level.

[0120] After training, the rate of power consumption can be predicted by inputting any temperature distribution, providing energy consumption data support for the calculation of the efficiency index and realizing accurate modeling of the impact of the fuselage thermal field on battery degradation.

[0121] S2024. Adjust the intercept and regression coefficients until the sum of the difference coefficients of all samples is minimized, thus obtaining the relational expression of the trained UAV.

[0122] S2025, and so on, build a training set and a linear regression model for each drone, and train to obtain the relational expression.

[0123] S203. Each training set is input into the physical model for training. The trained physical model predicts the surface temperature field changes of the UAV when flying each reference path based on the current temperature parameters, meteorological data and reference path of the UAV.

[0124] A physical model is created for each drone, and each training set is put into the corresponding physical model for training.

[0125] The trained physical model simulates and analyzes the changes in the surface temperature field of the UAV as it flies along the reference path, based on the current surface temperature field distribution of the UAV and the distribution of meteorological elements in the vicinity of the reference path within the inspection area.

[0126] Surface temperature field change refers to the temperature distribution at different locations on the surface of a drone over a period of time and how this temperature changes over time. The duration of this change is specifically taken as the length of the reference path divided by the drone's normal flight speed.

[0127] S204. By analyzing the changes in surface temperature field and their relational expressions, the change in electrical charge is obtained. Based on the change in electrical charge and the reference path, the efficiency index between the drone and the inspection point is calculated, thereby assigning an inspection point to each drone. Specifically, this includes:

[0128] S2041. Based on the duration T of the surface temperature field change, K time points are uniformly set. Substitute the temperature of all sampling points at each time point into the relational expression to analyze the power consumption rate ER at each time point. c .

[0129] S2042. Analyze the UAV and inspection point INP corresponding to the reference path, and obtain the maximum duration T of the surface temperature field change between inspection point INP and other UAVs.max .

[0130] S2043. Calculate the efficiency index between the UAV and the inspection point INP, and assign each UAV to the inspection point with the lowest efficiency index. The efficiency index calculation formula is as follows:

[0131]

[0132] In the formula, N is a constant greater than 1, and EFI is the efficiency index.

[0133] The efficiency index combines time cost and energy cost to quantify the overall efficiency of a drone in performing a certain inspection point task.

[0134] Time cost: Amplify the difference in flight duration (T is the current path duration, T max (The longest time to bind all drones to this point) is selected, prioritizing the path with the shortest execution time.

[0135] Energy consumption cost: Cumulative path segmented power consumption (ER) c ), punish high-power-consuming paths.

[0136] The final output EFI value serves as the basis for binding decisions, achieving optimal inspection point binding with low time consumption and low energy consumption.

[0137] Based on historical operational data, a unique temperature-power coupling model is trained for each drone to predict temperature field changes and power consumption along the flight path. The performance index (EFI) is used to quantify and match the best and worst performance, achieving optimal binding between the drone and the inspection point and avoiding the risk of rapid power decay caused by local high or low temperatures on the drone surface during flight.

[0138] The S300 establishes inspection routes for each drone and plans nodes based on the bound inspection points. It continues to calculate and bind the efficiency index between nodes and other inspection points, iteratively adjusting the inspection routes until all inspection points are covered. Specifically, this includes:

[0139] S301. Establish an inspection route for each drone, using the bound inspection point as the first node. Analyze again the efficiency index between each drone and other inspection points when it is at the first node.

[0140] S302. Each drone continues to be bound to the inspection point with the lowest energy efficiency index as the second node. When the inspection point is bound repeatedly by different drones, the drone with the lowest energy efficiency index is automatically bound, and the other drones are unbound.

[0141] When an inspection point is bound to multiple drones, only the one with the lowest EFI is retained, and the remaining drones trigger an unbinding-rebinding cycle.

[0142] If drones A and B are both bound to inspection points, when EFI_A = 1.2 and EFI_B = 1.5, drone B is unbound and rebound to another inspection point with the lowest EFI.

[0143] S303. After the unbound drones are removed from the unbound inspection points, they are rebound to other inspection points with the lowest energy efficiency index. The energy efficiency index of each drone when it is at the second node position is analyzed again compared to other inspection points.

[0144] S304. By continuously iterating and adjusting the inspection points bound to the drones, the nodes of the inspection route are planned sequentially. The inspection route planning for each drone is completed when all inspection points are bound to the drones and there are no duplicate bindings.

[0145] By cascading and binding nodes and resolving conflicts (such as retaining the lowest EFI drone when binding repeatedly), the inspection route can be dynamically expanded to achieve full coverage of multi-drone collaboration and avoid task conflicts.

[0146] Once the S400 inspection route is planned, it is sent to the corresponding drone, which then controls the drone to perform inspection operations according to the inspection route.

[0147] The planned inspection routes are compiled into encrypted waypoint sequences and distributed to the corresponding drones.

[0148] The drone verifies the feasibility of the inspection points in real time and performs inspection operations in sequence according to the order of each node in the inspection route. Simultaneously, it dynamically adjusts its flight attitude based on environmental parameters such as wind speed to ensure the accurate execution of the inspection mission.

[0149] By issuing encrypted waypoint sequences and performing dynamic attitude corrections during flight, mission safety and execution accuracy are ensured, and environmental changes such as wind speed disturbances are adapted.

[0150] Please see Figure 2 The present invention also provides an intelligent inspection operation management system for unmanned aerial vehicles (UAVs), including a dynamic perception module, an inspection analysis module, an operation management module, and a command control module.

[0151] The dynamic sensing module collects GIS maps and meteorological data of the inspection area, as well as historical logs and operational data of each drone. It then builds a virtual scene and maps the drone's location in real time.

[0152] Collect GIS maps and meteorological data (including wind direction, wind speed, temperature and humidity) of the inspection area, as well as historical logs (recording changes in operational data during flight time) and real-time operational data (such as 3D models, GPS parameters, battery parameters, attitude parameters and temperature parameters) of each UAV.

[0153] A virtual scene is built based on a GIS map, a 3D model of the UAV is loaded, and GPS parameters (spatial position, flight direction, flight speed) and attitude parameters (acceleration, angular velocity, angle) are used to map the UAV's spatial position and attitude state (pitch, roll, yaw) in the virtual scene in real time.

[0154] Through real-time data acquisition and scenario simulation, it provides high-precision visualization of the environment and UAV status, laying a data foundation for subsequent path planning and enhancing the system's ability to perceive dynamic environments.

[0155] The inspection analysis module marks the location of each inspection point in the virtual scene, calculates the efficiency index between the drone and the inspection point, and thus binds each drone to an inspection point.

[0156] The locations of inspection points are marked in the virtual scene to serve as reference paths between each drone and the inspection points. Historical logs are analyzed to construct a training set (including a linear regression model of sampling point temperature and power consumption rate). The physical model is trained to predict the surface temperature field changes of the drones on the reference paths, thereby calculating power changes and efficiency indices. Finally, the inspection point with the lowest efficiency index is assigned to each drone.

[0157] By optimizing the binding process through quantitative efficiency index, intelligent matching between drones and inspection points can be achieved, reducing energy consumption risks, improving inspection efficiency and reliability, and avoiding flight interruptions caused by abnormal temperature or power levels.

[0158] The operation management module establishes inspection routes for each drone and plans nodes based on the bound inspection points. It then calculates the efficiency index based on the node locations and binds other inspection points, iteratively adjusting the inspection routes until all inspection points are covered.

[0159] Starting from the bound inspection point, an inspection route is established for each drone; the efficiency index of the drone at the node position and other inspection points is calculated iteratively, new nodes are bound and duplicate bindings are handled (such as unbinding and rebinding the drone), and the route is adjusted until all inspection points are covered.

[0160] It enables collaborative path optimization among multiple drones, ensuring no inspection points are missed or conflicted through iterative adjustments, improving resource utilization and operational coverage, reducing redundant energy consumption, and supporting intelligent scheduling in large-scale complex areas.

[0161] The command and control module sends the inspection route to the corresponding drone and controls the drone to perform the inspection operation according to the inspection route.

[0162] The planned inspection routes are compiled into encrypted waypoint sequences and distributed to the corresponding drones. The drones verify the feasibility of the inspection points in real time and dynamically adjust their flight attitude (e.g., by correcting the angle through inertial sensor data) based on environmental parameters (such as wind speed) to ensure that the inspection operations are performed in the order of the nodes.

[0163] Ensure the security of tasks and the integrity of data through encrypted transmission and real-time verification, dynamically adjust to improve flight stability and accuracy, ensure the efficient and reliable execution of inspection tasks, and adapt to variable environmental conditions.

[0164] Example 1:

[0165] Suppose UAV1 and UAV2 are repeatedly bound to the same inspection point, and their power consumption speeds at each time point are as follows:

[0166] UAV1:

[0167] Time point 1: 60 w / h; Time point 2: 65 w / h; Time point 3: 75 w / h;

[0168] UAV2:

[0169] Time point 1: 60 w / h; Time point 2: 55 w / h; Time point 3: 45 w / h;

[0170] When the durations of the surface temperature field changes of UAV1 and UAV2 are 0.1 h and 0.12 h respectively, the maximum duration between the inspection point and other UAVs is 0.5 h, and N = 2, substitute into the formula to calculate the effectiveness indices of UAV1 and UAV2 respectively:

[0171]

[0172] Since EFI2 < EFI1, UAV2 is bound to the inspection point and UAV1 is unbound from the inspection point.

[0173] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device.

[0174] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for managing intelligent inspection operations using unmanned aerial vehicles (UAVs), characterized in that: The method includes: S100 collects GIS maps and meteorological data of the inspection area, as well as historical logs and operational data of each drone; it builds virtual scenes using GIS maps and meteorological data, and maps the drone positions in real time within the virtual scenes; S200: Mark the location of each inspection point in the virtual scene, and calculate the efficiency index between the drone and the inspection point by combining the operation data and meteorological data, so as to bind each drone to an inspection point. S300 establishes inspection routes for each drone and plans nodes based on the bound inspection points; it continues to calculate and bind the efficiency index between nodes and other inspection points, iteratively adjusting the inspection routes until all inspection points are covered; Once the S400 inspection route is planned, it is sent to the corresponding drone, which then controls the drone to perform inspection operations according to the inspection route.

2. The intelligent inspection operation management method for unmanned aerial vehicles (UAVs) according to claim 1, characterized in that: In S100, meteorological data includes the distribution of meteorological elements within the inspection area, including wind direction, wind speed, temperature, and humidity. The historical log includes the operation records of each drone flight, and each operation record includes the changes in the drone's operation data within the flight time. The operational data includes a 3D model of the drone, as well as real-time GPS parameters, battery parameters, attitude parameters, and temperature parameters; GPS parameters include spatial position, flight direction, and flight speed; attitude parameters are collected in real time by inertial sensors, specifically including acceleration, angular velocity, and angle; temperature parameters include the surface temperature field distribution of the drone. A virtual scene is built based on a GIS map. The 3D models of each UAV are loaded into the virtual scene. The spatial position of each 3D model in the virtual scene, as well as the pitch, roll and yaw angles, are mapped in real time based on GPS parameters and attitude parameters.

3. The intelligent inspection operation management method for unmanned aerial vehicles (UAVs) according to claim 2, characterized in that: S200 includes: S201. Mark the spatial locations of inspection points and drones in the virtual scene, and plan the flight path between each drone and each inspection point as a reference path. S202. Obtain the historical logs of each drone, analyze the relationship between surface temperature field changes and power changes, and construct a training set. Establish relational expressions based on the training set of each drone. S203. Each training set is input into the physical model for training. The trained physical model predicts the surface temperature field changes when the UAV flies each reference path based on the current temperature parameters, meteorological data and reference path of the UAV. S204. By analyzing the changes in surface temperature field and the relational expression, the change in power is obtained; based on the change in power and the reference path, the efficiency index between the UAV and the inspection point is calculated, thereby binding each UAV to an inspection point.

4. The intelligent inspection operation management method for unmanned aerial vehicles according to claim 3, characterized in that: S202 includes: S2021. Obtain the historical logs of the UAV and analyze the changes in the surface temperature field of the fuselage in each operation record; set r sampling points on the surface of the UAV's 3D model and map the surface temperature field. S2022. Divide each operation record into time periods. The temperature of each sampling point within the time period is used as the independent variable and the power consumption rate is used as the dependent variable. Pack them into samples and combine the samples to build a training set for the UAV. S2023. Establish a linear regression model, inputting all independent variables from each sample in the training set, and using the difference between the output value Y and the dependent variable as the difference coefficient for the corresponding sample; the linear regression model expression is: Y=d+f1X1+...+f r X r In the formula, d is the intercept, and f r Let X be the regression coefficient for the r-th sampling point. r Let r be the temperature of the r-th sampling point; S2024. Adjust the intercept and regression coefficients until the sum of the difference coefficients of all samples is minimized, and obtain the relational expression of the trained UAV. S2025, and so on, build a training set and a linear regression model for each drone, and train to obtain the relational expression.

5. The intelligent inspection operation management method for unmanned aerial vehicles according to claim 4, characterized in that: S2022 includes: S2022-1. Set the duration to time. Divide the flight duration of each operation record by time to obtain the number of samples Q. Divide the operation record into Q non-overlapping time periods with a duration of time. S2022-2. Within each time period, set e time points evenly, analyze the flight direction and speed of the UAV at different time points, as well as the temperature at each sampling point, and substitute them into the formula to calculate the anomaly index (ANI) for each time period: In the formula, α, β, and γ are constants. V represents the temperature at the j-th sampling point at the i-th time point. i Let be the flight speed at the i-th time point; ANG i,(i-1) The angle between the flight direction vector at the i-th time point and the flight direction vector at the (i-1)-th time point; S2022-3. Select the time periods when the anomaly index is less than the threshold, and use the temperature of r sampling points at each time point in these time periods as independent variables and the power consumption rate as dependent variable, and package them into samples. S2022-4. Combine these w×e samples to construct a training set for UAVs; where w is the number of time periods when the anomaly index is less than the threshold.

6. The intelligent inspection operation management method for unmanned aerial vehicles according to claim 3, characterized in that: In S203, a physical model is built for each drone, and each training set is put into the corresponding physical model for training. The trained physical model simulates and analyzes the changes in the surface temperature field of the UAV as it flies along the reference path, based on the current surface temperature field distribution of the UAV and the distribution of meteorological elements in the vicinity of the reference path within the inspection area.

7. The intelligent inspection operation management method for unmanned aerial vehicles (UAVs) according to claim 4, characterized in that: S204 includes: S2041. Based on the duration T of the surface temperature field change, K time points are uniformly set. Substitute the temperature of all sampling points at each time point into the relational expression to analyze the power consumption rate ER at each time point. c ; S2042. Analyze the UAV and inspection point INP corresponding to the reference path, and obtain the maximum duration T of the surface temperature field change between inspection point INP and other UAVs. max ; S2043. Calculate the efficiency index between the UAV and the inspection point INP, and assign each UAV to the inspection point with the lowest efficiency index; the efficiency index calculation formula is as follows: In the formula, N is a constant greater than 1, and EFI is the efficiency index.

8. The intelligent inspection operation management method for unmanned aerial vehicles according to claim 3, characterized in that: The S300 includes: S301. Establish an inspection route for each drone and take the bound inspection point as the first node; analyze the efficiency index between each drone and other inspection points when it is at the first node. S302. Each drone continues to be bound to the inspection point with the lowest energy efficiency index as the second node. When the inspection point is bound repeatedly by different drones, the drone with the lowest energy efficiency index is automatically bound, and the other drones are unbound. S303. After the unbound drones are removed from the unbound inspection points, they are rebound to other inspection points with the lowest energy efficiency index; the energy efficiency index of each drone when it is in the second node position is analyzed again between it and other inspection points. S304. By continuously iterating and adjusting the inspection points bound to the drones, the nodes of the inspection route are planned sequentially; until all inspection points are bound to the drones and there are no duplicate bindings, the inspection route planning for each drone is completed.

9. The intelligent inspection operation management method for unmanned aerial vehicles according to claim 8, characterized in that: In the S400, the planned inspection route is compiled into an encrypted waypoint sequence and sent to the corresponding UAV. The UAV verifies the feasibility of the inspection points in real time and performs the inspection operation in sequence according to the order of each node in the inspection route. Simultaneously, the flight attitude is dynamically adjusted according to environmental parameters such as wind speed to ensure the accurate execution of the inspection task.

10. A drone intelligent inspection operation management system, characterized in that: The system includes a dynamic sensing module, an inspection and analysis module, an operation management module, and an instruction control module; The dynamic perception module collects GIS maps and meteorological data of the inspection area, as well as historical logs and operational data of each drone; it builds a virtual scene and maps the drone's position in real time. The inspection analysis module marks the location of each inspection point in the virtual scene, calculates the efficiency index between the drone and the inspection point, and thus binds each drone to an inspection point. The operation management module establishes inspection routes for each drone and plans nodes based on the bound inspection points; it continues to calculate the efficiency index based on the node locations and binds other inspection points, iteratively adjusting the inspection routes until all inspection points are covered; The command and control module sends the inspection route to the corresponding drone and controls the drone to perform the inspection operation according to the inspection route.

Images