Power inspection system and method based on unmanned aerial vehicle

Abstract

The invention discloses an electric power inspection system and method based on an unmanned aerial vehicle, which are used for solving the problem of how to realize autonomous path planning of the unmanned aerial vehicle in a complex three-dimensional environment so as to reduce redundant flight time and improve inspection efficiency. The electric power inspection system comprises a path planning unit, an obstacle avoidance positioning unit and a dynamic obstacle avoidance unit which are connected in sequence. Three-dimensional map data and real-time environment information of a to-be-inspected power line are obtained through a path planning unit, real-time path planning is carried out based on the three-dimensional map data and the real-time environment information, and an optimal flight path of the unmanned aerial vehicle is dynamically generated; obstacle avoidance positioning is carried out on the power equipment on the optimal flight path and the surrounding environment thereof through an obstacle avoidance positioning unit, and real-time positioning information of an obstacle is generated; and a dynamic obstacle avoidance unit constructs a dynamic obstacle map based on the real-time positioning information, and automatically plans an obstacle avoidance strategy according to the dynamic obstacle map, so that the unmanned aerial vehicle realizes dynamic obstacle avoidance based on the obstacle avoidance strategy.

Description

Technical Field

[0001] The present invention relates to the technical field of smart grid maintenance, and in particular to an unmanned aerial vehicle (UAV)-based power inspection system, an unmanned aerial vehicle (UAV)-based power inspection method, an electronic device and a storage medium. Background Art

[0002] In the field of smart grid maintenance, power inspection is an important part of ensuring the safe and stable operation of the power grid. Traditional power inspection methods mainly rely on manual work. This method has problems such as long inspection cycle, large workload, low efficiency, and high risk. Especially in the face of complex terrain and harsh weather conditions, the difficulty and danger of manual inspection will increase significantly.

[0003] With the rapid development of drone technology, its application in various fields is becoming more and more extensive. With its advantages of high flexibility, wide field of view and strong adaptability, drones have gradually become an important tool in the field of power inspection. However, the existing drone power inspection system still faces some challenges. How to realize autonomous path planning of drones in complex three-dimensional environments to reduce redundant flight time and improve inspection efficiency is an urgent problem to be solved in current power inspection. Summary of the invention

[0004] The present invention provides an electric power inspection system based on a drone, an electric power inspection method based on a drone, an electronic device and a storage medium, which are used to solve or partially solve the technical problem of how to realize autonomous path planning of a drone in a complex three-dimensional environment to reduce redundant flight time and improve inspection efficiency.

[0005] The present invention provides a power inspection system based on a drone, the power inspection system comprising a path planning unit, an obstacle avoidance and positioning unit, and a dynamic obstacle avoidance unit connected in sequence; wherein,

[0006] The path planning unit is used to obtain three-dimensional map data and real-time environmental information of the power line to be inspected, and perform real-time path planning based on the three-dimensional map data and the real-time environmental information to dynamically generate an optimal flight path for the UAV;

[0007] The obstacle avoidance and positioning unit is used to perform obstacle avoidance and positioning of the power equipment and its surrounding environment on the optimal flight path and generate real-time positioning information of obstacles during the inspection of the UAV based on the optimal flight path;

[0008] The dynamic obstacle avoidance unit is used to construct a dynamic obstacle map based on the real-time positioning information, and automatically plan an obstacle avoidance strategy according to the dynamic obstacle map, so that the UAV can achieve dynamic obstacle avoidance based on the obstacle avoidance strategy.

[0009] Optionally, the path planning unit is specifically used to:

[0010] Performing data preprocessing on the three-dimensional map data and the real-time environmental information to obtain processed environmental data;

[0011] Performing a comprehensive real-time path planning analysis on the processed environmental data in combination with a pre-trained path planning model, and dynamically generating multiple candidate flight paths in a three-dimensional map;

[0012] A flight path with the lowest comprehensive cost is selected from the multiple candidate flight paths as the optimal flight path of the UAV. The optimal flight path is used to formulate a detailed flight plan of the UAV, and the detailed flight plan includes a take-off point, a landing point, a flight altitude, a flight speed, and a turning point.

[0013] Optionally, the path planning unit is specifically used to:

[0014] Inputting the processed environmental data into a pre-trained path planning model;

[0015] Performing environmental path identification on the processed environmental data through the pre-trained path planning model to comprehensively analyze the current environment of the power line to be inspected and obtain environmental analysis results, wherein the environmental analysis results include potential risk areas identified through model analysis;

[0016] In combination with the potential risk areas, a path planning algorithm is adopted to comprehensively consider the actual flight cost and the estimated flight cost, and a plurality of candidate flight paths are dynamically generated in the three-dimensional map.

[0017] Optionally, the obstacle avoidance and positioning unit is specifically used for:

[0018] Collecting three-dimensional spatial information of the power equipment and its surrounding environment on the optimal flight path in real time;

[0019] According to the three-dimensional space information, combined with an obstacle avoidance algorithm, real-time positioning information of obstacles on the optimal flight path is generated, so that the UAV automatically identifies and avoids the obstacles based on the real-time positioning information.

[0020] Optionally, the obstacle avoidance and positioning unit is further used for:

[0021] Collecting image data and temperature distribution diagram of the power equipment and its surrounding environment on the optimal flight path in real time;

[0022] Preprocessing the image data to obtain processed image data, and importing the processed image data into a pre-trained image classification model to perform abnormality recognition and classification on the collected images and generate an abnormality detection report;

[0023] According to the temperature distribution diagram, combined with the ambient temperature and equipment load factors, the overheating of the power equipment is judged to determine the potential safety hazards on the optimal flight path;

[0024] Based on the abnormality detection report and the potential safety hazard, a comprehensive monitoring report for the power line to be inspected is generated.

[0025] Optionally, the dynamic obstacle avoidance unit is specifically used for:

[0026] Collecting image data of the power equipment and its surrounding environment on the optimal flight path in real time;

[0027] Convert the real-time positioning information into spatial coordinate information to determine the location of the obstacle and the distance between the obstacle and the drone;

[0028] Performing feature recognition on the image data based on a machine vision algorithm, extracting object features, and classifying and labeling the obstacles based on the object features to obtain labeled obstacles;

[0029] Constructing a dynamic obstacle map according to the spatial coordinate information and the marked obstacles;

[0030] Real-time analysis of the risk level of obstacles on the dynamic obstacle map to the flight of the UAV;

[0031] When it is detected that the risk level of the obstacle closest to the drone is greater than or equal to a preset risk level threshold, the obstacle avoidance strategy planning is automatically started, and the optimal obstacle avoidance path is replanned to re-search the safe flight path in the dynamic obstacle map, so that the drone changes the current flight path based on the safe flight path to achieve dynamic obstacle avoidance.

[0032] Optionally, the power inspection system further includes a data analysis unit, and the data analysis unit is used to:

[0033] Receive the inspection data sent back by the drone in real time;

[0034] The inspection data is first verified, then data preprocessed, to obtain processed inspection data, and the processed inspection data is stored;

[0035] Based on the anomaly detection algorithm, the processed inspection data is deeply mined and analyzed to automatically identify abnormal events in the inspection process and the inspection analysis results, and generate an inspection analysis report.

[0036] The present invention also provides a drone-based power inspection method, which is used in the drone-based power inspection system as described in any one of the above items, and the method comprises:

[0037] Acquire three-dimensional map data and real-time environmental information of the power line to be inspected, and perform real-time path planning based on the three-dimensional map data and the real-time environmental information to dynamically generate an optimal flight path for the UAV;

[0038] During the inspection process of the UAV based on the optimal flight path, obstacle avoidance positioning is performed on the power equipment and its surrounding environment on the optimal flight path to generate real-time positioning information of obstacles;

[0039] Based on the real-time positioning information, a dynamic obstacle map is constructed, and according to the dynamic obstacle map, an obstacle avoidance strategy is automatically planned, so that the UAV can achieve dynamic obstacle avoidance based on the obstacle avoidance strategy.

[0040] Optionally, performing real-time path planning based on the three-dimensional map data and the real-time environmental information to dynamically generate an optimal flight path for the UAV includes:

[0041] Performing data preprocessing on the three-dimensional map data and the real-time environmental information to obtain processed environmental data;

[0042] Performing a comprehensive real-time path planning analysis on the processed environmental data in combination with a pre-trained path planning model, and dynamically generating multiple candidate flight paths in a three-dimensional map;

[0043] A flight path with the lowest comprehensive cost is selected from the multiple candidate flight paths as the optimal flight path of the UAV. The optimal flight path is used to formulate a detailed flight plan of the UAV, and the detailed flight plan includes a take-off point, a landing point, a flight altitude, a flight speed, and a turning point.

[0044] Optionally, the pre-trained path planning model is combined to perform a real-time path planning comprehensive analysis on the processed environmental data, and a plurality of candidate flight paths are dynamically generated in the three-dimensional map, including:

[0045] Inputting the processed environmental data into a pre-trained path planning model;

[0046] Performing environmental path identification on the processed environmental data through the pre-trained path planning model to comprehensively analyze the current environment of the power line to be inspected and obtain environmental analysis results, wherein the environmental analysis results include potential risk areas identified through model analysis;

[0047] In combination with the potential risk areas, a path planning algorithm is adopted to comprehensively consider the actual flight cost and the estimated flight cost, and a plurality of candidate flight paths are dynamically generated in the three-dimensional map.

[0048] Optionally, performing obstacle avoidance positioning on the power equipment and its surrounding environment on the optimal flight path to generate real-time positioning information of obstacles includes:

[0049] Collecting three-dimensional spatial information of the power equipment and its surrounding environment on the optimal flight path in real time;

[0050] According to the three-dimensional space information, combined with an obstacle avoidance algorithm, real-time positioning information of obstacles on the optimal flight path is generated, so that the UAV automatically identifies and avoids the obstacles based on the real-time positioning information.

[0051] Optionally, the method further comprises:

[0052] Collecting image data and temperature distribution diagram of the power equipment and its surrounding environment on the optimal flight path in real time;

[0053] Preprocessing the image data to obtain processed image data, and importing the processed image data into a pre-trained image classification model to perform abnormality recognition and classification on the collected images and generate an abnormality detection report;

[0054] According to the temperature distribution diagram, combined with the ambient temperature and equipment load factors, the overheating of the power equipment is judged to determine the potential safety hazards on the optimal flight path;

[0055] Based on the abnormality detection report and the potential safety hazard, a comprehensive monitoring report for the power line to be inspected is generated.

[0056] Optionally, constructing a dynamic obstacle map based on the real-time positioning information, and automatically planning an obstacle avoidance strategy according to the dynamic obstacle map, includes:

[0057] Collecting image data of the power equipment and its surrounding environment on the optimal flight path in real time;

[0058] Convert the real-time positioning information into spatial coordinate information to determine the location of the obstacle and the distance between the obstacle and the drone;

[0059] Performing feature recognition on the image data based on a machine vision algorithm, extracting object features, and classifying and labeling the obstacles based on the object features to obtain labeled obstacles;

[0060] Constructing a dynamic obstacle map according to the spatial coordinate information and the marked obstacles;

[0061] Real-time analysis of the risk level of obstacles on the dynamic obstacle map to the flight of the UAV;

[0062] When it is detected that the risk level of the obstacle closest to the drone is greater than or equal to a preset risk level threshold, the obstacle avoidance strategy planning is automatically started, and the optimal obstacle avoidance path is replanned to re-search the safe flight path in the dynamic obstacle map, so that the drone changes the current flight path based on the safe flight path to achieve dynamic obstacle avoidance.

[0063] Optionally, the method further comprises:

[0064] Receive the inspection data sent back by the drone in real time;

[0065] The inspection data is first verified, then data preprocessed, to obtain processed inspection data, and the processed inspection data is stored;

[0066] Based on the anomaly detection algorithm, the processed inspection data is deeply mined and analyzed to automatically identify abnormal events in the inspection process and the inspection analysis results, and generate an inspection analysis report.

[0067] The present invention also provides an electronic device, the device comprising a processor and a memory:

[0068] The memory is used to store program code and transmit the program code to the processor;

[0069] The processor is used to execute the UAV-based power inspection method as described in any of the above items according to the instructions in the program code.

[0070] The present invention also provides a computer-readable storage medium, which is used to store program code, and the program code is used to execute the drone-based power inspection method as described in any of the above items.

[0071] It can be seen from the above technical solutions that the present invention has the following advantages:

[0072] Provided is a power inspection system based on a drone and a power inspection method thereof. The power inspection system comprises a path planning unit, an obstacle avoidance and positioning unit, and a dynamic obstacle avoidance unit connected in sequence; the path planning unit obtains three-dimensional map data and real-time environmental information of the power line to be inspected, and performs real-time path planning based on the three-dimensional map data and the real-time environmental information, and dynamically generates the optimal flight path of the drone; the obstacle avoidance and positioning unit performs obstacle avoidance and positioning on the power equipment and its surrounding environment on the optimal flight path, and generates real-time positioning information of obstacles; the dynamic obstacle avoidance unit constructs a dynamic obstacle map based on the real-time positioning information, and automatically plans an obstacle avoidance strategy based on the dynamic obstacle map, so that the drone can achieve dynamic obstacle avoidance based on the obstacle avoidance strategy. Thus, through intelligent autonomous path planning, obstacle avoidance and positioning, dynamic obstacle avoidance, and data analysis, the optimal flight path can be dynamically generated, and the flight plan can be adjusted in real time to deal with potential risk areas. Thus, during the inspection process, the redundant flight time of the drone can be significantly reduced, and the inspection efficiency can be greatly improved. Compared with the traditional manual inspection method, it can not only shorten the drone inspection cycle, but also reduce the workload of operation and maintenance personnel, and effectively improve the automation and intelligence level of power inspection. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0074] Figure 1 This is a structural diagram of a UAV-based power inspection system;

[0075] Figure 2 It is a flowchart of the steps of a power inspection method based on drone;

[0076] Figure 3 The figure is a schematic diagram of the overall process of a UAV-based power inspection method. DETAILED DESCRIPTION

[0077] Embodiments of the present invention provide a UAV-based power inspection system, a UAV-based power inspection method, an electronic device and a storage medium, which are used to solve or partially solve the technical problem of how to realize autonomous path planning of UAVs in complex three-dimensional environments to reduce redundant flight time and improve inspection efficiency.

[0078] In order to make the purpose, features and advantages of the present invention more obvious and easy to understand, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described below are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0079] As an example, with the rapid development of drone technology, its application in various fields is becoming more and more extensive. With its advantages such as high flexibility, wide field of view and strong adaptability, drones have gradually become an important tool in the field of power inspection. However, the existing drone power inspection system still faces some challenges. How to realize autonomous path planning of drones in complex three-dimensional environments to reduce redundant flight time and improve inspection efficiency is an urgent problem to be solved in current power inspection.

[0080] Therefore, one of the core inventions of the embodiment of the present invention is to provide a power inspection system based on a drone and a power inspection method thereof. The power inspection system includes a path planning unit, an obstacle avoidance and positioning unit, and a dynamic obstacle avoidance unit connected in sequence, and a data analysis unit connected to these units respectively.

[0081] Among them, the three-dimensional map data and real-time environmental information of the power line to be inspected are obtained through the path planning unit, and the real-time path planning is performed based on the three-dimensional map data and the real-time environmental information to dynamically generate the optimal flight path of the UAV. During the inspection process of the UAV based on the optimal flight path, the obstacle avoidance and positioning unit performs obstacle avoidance positioning on the power equipment and its surrounding environment on the optimal flight path to generate real-time positioning information of obstacles. Based on the real-time positioning information, the dynamic obstacle avoidance unit constructs a dynamic obstacle map, and according to the dynamic obstacle map, the obstacle avoidance strategy is automatically planned so that the UAV can achieve dynamic obstacle avoidance based on the obstacle avoidance strategy. The inspection data returned by the UAV is received in real time through the data analysis unit, the inspection data is first verified, and then the data is pre-processed to obtain the processed inspection data, and the processed inspection data is stored. The processed inspection data is deeply mined and analyzed based on the anomaly detection algorithm to automatically identify abnormal events and inspection analysis results during the inspection process, and generate an inspection analysis report.

[0082] Through intelligent autonomous path planning, obstacle avoidance positioning, dynamic obstacle avoidance and data analysis, the optimal flight path can be dynamically generated, and the flight plan can be adjusted in real time to deal with potential risk areas. Therefore, during the inspection process, the redundant flight time of the drone can be significantly reduced, greatly improving the inspection efficiency. Compared with the traditional manual inspection method, it can not only shorten the drone inspection cycle, but also reduce the workload of operation and maintenance personnel, effectively improving the automation and intelligence level of power inspection.

[0083] In addition, through the use of drones for automated inspections, the risk of personnel directly participating in operations at high altitudes, in complex terrain or in harsh environments can be reduced. At the same time, due to the improvement in inspection efficiency, it can also indirectly reduce manpower and time costs, bringing significant economic benefits to power companies. Furthermore, based on the establishment of a data collection and analysis platform, operation and maintenance personnel can grasp the inspection progress in real time, adjust the inspection strategy in a timely manner, and further enhance the flexibility of the system and its ability to respond to emergencies. Therefore, through the technical solution of the present invention, it is possible to achieve automation, intelligence and efficiency of drone-based power inspections, providing a strong guarantee for the safe and stable operation of smart grids.

[0084] Reference Figure 1 , showing a schematic structural diagram of a UAV-based power inspection system provided in an embodiment of the present invention.

[0085] Combination Figure 1 The power inspection system 100 can mainly include a path planning unit 101, an obstacle avoidance and positioning unit 102, and a dynamic obstacle avoidance unit 103 connected in sequence. Among them, the path planning unit 101 can realize intelligent autonomous path planning for drone inspection flight. The obstacle avoidance and positioning unit 102 integrates machine vision and high-precision sensors to realize obstacle avoidance and positioning during drone flight. The dynamic obstacle avoidance system can realize autonomous planning of obstacle avoidance strategies and realize dynamic obstacle avoidance.

[0086] Furthermore, the power inspection system 100 may further include a data analysis unit 104. The data analysis unit 104 may be regarded as a data collection and analysis platform, which is mainly used for the collection and analysis of relevant data of power inspection.

[0087] In actual applications, the path planning unit 101 mainly uses algorithms to analyze the three-dimensional map data of the power lines. Combined with real-time environmental information, including but not limited to tree distribution, building location, and weather conditions, the optimal flight path is dynamically generated. Through the path planning unit 101, the system can identify and predict potential risk areas, automatically adjust the flight plan, reduce redundant flight time, and improve inspection efficiency.

[0088] In a specific implementation, the path planning unit 101 is mainly used to obtain three-dimensional map data and real-time environmental information of the power line to be inspected, and perform real-time path planning based on the three-dimensional map data and real-time environmental information to dynamically generate the optimal flight path of the drone.

[0089] More specifically, the path planning unit 101 can obtain accurate three-dimensional map data of the power line and its surrounding environment from a geographic information system and other data sources, including but not limited to the location of power towers, line directions, terrain heights, and vegetation coverage information.

[0090] The sensors carried by the drone collect the data of the current environment in real time (i.e. real-time environmental information). Then, the collected 3D map data and real-time environmental information are pre-processed by cleaning, denoising, formatting, etc. to obtain the processed environmental data.

[0091] Before the actual application of path planning, a path planning model needs to be pre-trained. Specifically, the path planning algorithm model is selected according to the task requirements. Among them, the path planning algorithm model is a type of algorithm used to find the optimal or feasible path in a complex environment. Path planning problems usually involve reaching the end point from the starting point while avoiding obstacles and satisfying certain constraints. Such as minimizing distance, time or cost. The current path planning algorithms mainly include A* (A-Star) algorithm, Dijkstra algorithm, D* Lite algorithm, RRT (Rapidly-exploring Random Trees) probabilistic path planning method, etc. For the path planning algorithm adopted, those skilled in the art can choose according to actual needs. It is understandable that the present invention is not limited to this.

[0092] Then, the constructed path algorithm model can be trained based on the collected historical inspection data and simulation data. First, the collected historical inspection data and simulation data are preprocessed, and key features are extracted from the preprocessed data, namely spatial features Spatial_Features=f CNN (3D_Map_Data), as the input of the algorithm model. Among them, f CNN Represents the forward propagation function, 3D_Map_Data represents the three-dimensional map data in the training data, which affects the search speed and path quality.

[0093] During the model training process, the goal is to minimize the loss function and continuously optimize the model parameters to improve the prediction accuracy. The loss function used in the training process is:

[0094]

[0095] in, represents the model parameters; represents the number of samples; Indicates The trajectory length of samples; Represents the discount factor. First, use the random strategy to initialize , for each sample , calculate the features and decision actions Then execute the action , observe the new state and rewards , calculate the cumulative reward , used to control the importance of future rewards. Update using gradient descent To minimize .

[0096] After obtaining the trained path planning model, you can combine it with the processed environmental data (corresponding to the actual collected real-time environmental information and three-dimensional map data) to use the model to conduct a comprehensive analysis of the current environment for real-time path planning. Specifically, the following formula is used for comprehensive analysis:

[0097]

[0098] in, Indicates the result of comprehensive analysis of the current environment, by combining real-time environmental information and 3D map data to calculate; Indicates the number of environmental factors considered; For each factor The weight of , which indicates the importance of the factor to the overall risk score; For each factor A specific evaluation function that combines real-time environmental information and 3D map data Calculate the risk value of the factor and identify potential risk areas.

[0099] Based on the environmental analysis results, a path planning algorithm is used to generate multiple candidate flight paths in a 3D map. By evaluating the function To select the next node. Represents the distance from the starting point to the current node The actual cost (price); Indicates that from the current node The estimated cost to reach the destination; represents the heuristic weight.

[0100] In the embodiment of the present invention, heuristic refers to the process of making decisions or optimizing based on empirical rules or heuristic methods. Heuristic weight refers to the weights assigned to different factors or variables based on a certain heuristic method in the process of solving a specific problem.

[0101] A heuristic method is an empirical rule or strategy used in the problem solving process. This method is not guaranteed to find the optimal solution, but it can usually find a good enough solution in a reasonable time. Heuristic methods are widely used in optimization problems, search problems, and machine learning.

[0102] Based on dynamic path planning, the path can be adjusted dynamically according to real-time environmental changes, and a detailed flight plan can be formulated based on the optimal path, including but not limited to take-off point, landing point, flight altitude, speed, and turning point. The drone performs inspection tasks according to the flight plan while continuously monitoring real-time environmental information.

[0103] During the flight, when encountering emergencies, the system can respond quickly, reassess the environment and adjust the flight plan. During and after the inspection, the system can collect data sent back by the drone in real time, and conduct in-depth analysis of the collected data based on the data analysis unit 104 to evaluate the inspection effect and identify potential problems. Based on the data analysis results, the model and path planning algorithm can also be continuously optimized to improve the accuracy and efficiency of the system power inspection.

[0104] In combination with the above-mentioned contents, in a specific implementation, the path planning unit 101 may be used for:

[0105] Perform data preprocessing on three-dimensional map data and real-time environmental information to obtain processed environmental data;

[0106] Combined with the pre-trained path planning model, the processed environmental data is comprehensively analyzed for real-time path planning, and multiple candidate flight paths are dynamically generated in the three-dimensional map;

[0107] The flight path with the lowest comprehensive cost is selected from multiple candidate flight paths as the optimal flight path for the UAV. The optimal flight path is used to formulate a detailed flight plan for the UAV. The detailed flight plan includes the take-off point, landing point, flight altitude, flight speed and turning point.

[0108] Furthermore, for the comprehensive analysis of real-time path planning, the path planning unit 101 is specifically used for:

[0109] Input the processed environment data into the pre-trained path planning model;

[0110] The pre-trained path planning model is used to identify the environmental path of the processed environmental data, so as to comprehensively analyze the current environment of the power line to be inspected and obtain the environmental analysis results, which include the potential risk areas identified through model analysis;

[0111] Combined with potential risk areas, a path planning algorithm is used to dynamically generate multiple candidate flight paths in the three-dimensional map by comprehensively considering the actual flight cost and the estimated flight cost.

[0112] The obstacle avoidance and positioning unit 102 is based on the fusion of machine vision and high-precision sensors. It integrates high-definition cameras, infrared sensors and high-precision LiDAR (Light Detection and Ranging) to achieve all-round and high-precision monitoring of power towers, lines and surrounding environments.

[0113] Among them, machine vision technology is used to identify abnormal conditions such as line damage, dirty insulators, hanging foreign objects, etc. Cameras with high resolution, night vision and anti-shake functions can be selected to capture clear images of power lines and towers.

[0114] Infrared sensors are used to detect potential safety hazards caused by heating of circuits. Infrared thermal imagers with high sensitivity and wide temperature measurement range can be used to detect temperature changes of electrical equipment in real time.

[0115] Equipped with LiDAR sensors with high-precision ranging and point cloud generation capabilities to obtain three-dimensional spatial information of power facilities and their surroundings. Through the information collection of LiDAR sensors, drones can be accurately positioned in complex environments, build environmental models, and provide data support for dynamic obstacle avoidance.

[0116] At the same time, a drone with long endurance and appropriate load capacity is selected as the carrier for sensor integration. The high-definition camera, infrared sensor and LiDAR are wirelessly connected to the drone control system. The drone control system develops and integrates a data processing unit to process the data from each sensor in real time and output the monitoring results.

[0117] In actual applications, the obstacle avoidance and positioning unit 102 uses a high-definition camera to regularly or continuously shoot power lines, insulators and the surrounding environment to obtain clear image data. Then, the collected images are pre-processed. Then, a machine vision algorithm is applied to identify and classify abnormal conditions such as line damage, dirty insulators, and hanging foreign objects in the image using a pre-trained image classification model, so as to accurately mark the abnormal location and generate an abnormal detection report.

[0118] The obstacle avoidance and positioning unit 102 can also use an infrared thermal imager to perform a comprehensive scan of the power lines and equipment to obtain a temperature distribution map. Then, based on the temperature distribution of the infrared image, combined with the ambient temperature and equipment load factors, overheating is determined. Specifically, the relative temperature difference method is used to quantitatively evaluate potential safety hazards. The relative temperature difference calculation formula is as follows:

[0119]

[0120] in, Indicates the maximum temperature of the device under test (°C); Indicates the temperature of the ambient reference body (°C), usually the temperature of a similar device in normal operation or the ambient temperature; Indicates the ambient temperature (°C), which is the temperature of the air surrounding the device.

[0121] The obstacle avoidance and positioning unit 102 can also use the point cloud data obtained by the LiDAR sensor to construct a three-dimensional model (three-dimensional spatial information) of the power facility and the surrounding environment through automatic processing and fine classification. Then, the model is used to perform terrain analysis and spatial distance measurement to provide basic data for subsequent monitoring and obstacle avoidance.

[0122] During the flight of the drone, the three-dimensional spatial information generated in real time by LiDAR is used to achieve accurate positioning. Combined with the obstacle avoidance algorithm, the real-time positioning information of obstacles on the optimal flight path is generated, so that the drone can automatically identify and avoid obstacles based on the real-time positioning information. The calculation formula involved in this process is as follows:

[0123]

[0124] in, Indicates that the drone and The shortest distance between obstacles; min means the minimum value solution; Indicates at a point in time The flight path position of Indicates The location of obstacles; Represents the distance metric between two points.

[0125] Thus, the data support provided by LiDAR can be used to enable the drone to automatically identify and avoid obstacles. In addition, the data from high-definition cameras, infrared sensors and LiDAR are fused and processed to form a comprehensive monitoring report. The fused data is analyzed to extract valuable information and patterns. Based on the monitoring results and data analysis, decision support can be provided for the subsequent maintenance, inspection and safety management of power facilities. At the same time, early warning signals can be issued in time for the problems found, and corresponding treatment suggestions can be put forward. Based on the analysis results, the sensors can be calibrated and maintained regularly, and the drones and their control systems can be inspected and maintained. Furthermore, based on the feedback of actual application effects, the machine vision and obstacle avoidance algorithms can be continuously optimized to improve the accuracy and efficiency of monitoring.

[0126] Based on the above introduction, in a specific implementation, the obstacle avoidance and positioning unit 102 can be mainly used to perform obstacle avoidance and positioning of power equipment and its surrounding environment on the optimal flight path during the inspection process of the drone based on the optimal flight path, and generate real-time positioning information of obstacles.

[0127] Furthermore, the obstacle avoidance and positioning unit 102 can be specifically used for:

[0128] Collect three-dimensional spatial information of power equipment and its surrounding environment on the optimal flight path in real time;

[0129] According to the three-dimensional spatial information and combined with the obstacle avoidance algorithm, the real-time positioning information of obstacles on the optimal flight path is generated, so that the UAV can automatically identify and avoid obstacles based on the real-time positioning information.

[0130] The obstacle avoidance and positioning unit 102 may also be used for:

[0131] Real-time collection of image data and temperature distribution diagrams of power equipment and its surrounding environment on the optimal flight path;

[0132] Preprocess the image data to obtain processed image data, and import the processed image data into a pre-trained image classification model to perform abnormality recognition and classification on the collected images and generate an abnormality detection report;

[0133] According to the temperature distribution diagram, combined with the ambient temperature and equipment load factors, the overheating of the power equipment is judged, and the potential safety hazards on the optimal flight path are determined;

[0134] Based on the abnormal detection report and potential safety hazards, a comprehensive monitoring report for the power lines to be inspected is generated.

[0135] On the basis of the functions realized by the obstacle avoidance and positioning unit 102, the dynamic obstacle avoidance unit 103 mainly builds a dynamic obstacle map based on the real-time scanning data of LiDAR and the results of machine vision recognition. Based on the constructed dynamic obstacle map, the system can evaluate the obstacle risk on the flight path in real time, automatically plan the obstacle avoidance strategy, optimize the obstacle avoidance decision-making process through algorithms, and improve the accuracy and real-time performance of obstacle avoidance.

[0136] As can be seen from the above content, the LiDAR sensor can scan the surrounding environment in real time to obtain three-dimensional spatial information such as distance and angle data. The obstacle avoidance and positioning unit 102 can generate real-time positioning information of obstacles on the optimal flight path based on the three-dimensional spatial information and the obstacle avoidance algorithm. Then, in the dynamic obstacle avoidance unit 103, the power equipment and its surrounding environment can be identified by machine vision technology to obtain image information, including but not limited to the shape, color, and texture features of the object.

[0137] Integrate the real-time positioning information obtained after LiDAR scanning and processing and the results based on machine vision recognition to build a dynamic obstacle map. Specifically, the real-time positioning information is converted into spatial coordinate information to determine the location and distance of obstacles. At the same time, the object features recognized by machine vision are used to classify and annotate obstacles. Based on the spatial coordinate information and the annotated obstacles, a dynamic obstacle map is built.

[0138] Analyze obstacles on the flight path in real time, and assess the risk level of obstacles to flight based on the location, distance, and speed of the obstacles. You can pre-set different risk levels, such as mild, moderate, and severe. And set different risk ranges to correspond to different risk levels. When obstacles are detected on the flight path and the risk reaches a certain level, the system automatically starts obstacle avoidance strategy planning, considers multiple factors, formulates the optimal obstacle avoidance path, and re-searches for a safe flight path in the dynamic obstacle map.

[0139] In addition, the obstacle avoidance decision-making process can be continuously optimized to improve the accuracy and real-time performance of obstacle avoidance. The obstacle map and risk assessment results are updated in real time, and the performance of the algorithm is monitored and evaluated to adjust and improve it according to actual conditions.

[0140] Based on the above introduction, in a specific implementation, the dynamic obstacle avoidance unit 103 is mainly used to construct a dynamic obstacle map based on real-time positioning information, and automatically plan an obstacle avoidance strategy according to the dynamic obstacle map, so that the drone can achieve dynamic obstacle avoidance based on the obstacle avoidance strategy.

[0141] Furthermore, the dynamic obstacle avoidance unit 103 is specifically used for:

[0142] Collect image data of power equipment and its surrounding environment on the optimal flight path in real time;

[0143] Convert real-time positioning information into spatial coordinate information to determine the location of obstacles and their distance from the drone;

[0144] Based on the machine vision algorithm, the image data is identified and the object features are extracted. Then, obstacles are classified and labeled based on the object features to obtain labeled obstacles.

[0145] Construct a dynamic obstacle map based on spatial coordinate information and marked obstacles;

[0146] Real-time analysis of the risk level of obstacles to drone flight on the dynamic obstacle map;

[0147] When it is detected that the risk level of the obstacle closest to the drone is greater than or equal to the preset risk level threshold, the obstacle avoidance strategy planning is automatically started, and the optimal obstacle avoidance path is re-planned to re-search for a safe flight path in the dynamic obstacle map, so that the drone can change the current flight path based on the safe flight path to achieve dynamic obstacle avoidance.

[0148] Thus, in combination with the obstacle avoidance and positioning unit 102 and the dynamic obstacle avoidance unit 103, the three-dimensional point cloud data of the power equipment and its surrounding environment on the inspection path can be obtained in real time during the inspection mission of the drone. These data contain the position, shape and distance information of objects in the environment. The image data of the power equipment and its surrounding environment are collected by the camera and the camera combination. These data provide rich visual information and help identify obstacles and scene features.

[0149] The data processing speed and efficiency are improved by filtering, denoising and downsampling the LiDAR 3D point cloud data. At the same time, the height information and density information in the LiDAR 3D point cloud data are used to identify obstacles in the 3D space. The machine vision image data is processed by image enhancement, feature extraction and other processing operations to facilitate subsequent obstacle identification. The machine vision algorithm is used to detect obstacles on the preprocessed image data, which can identify obstacles in the image and determine their position and size. By fusing the obstacle information identified by machine vision and LiDAR, a dynamic obstacle map can be constructed. The map is updated in real time to reflect the location and status of obstacles in the current environment.

[0150] According to the dynamic obstacle map, the obstacle risk on the current flight path is evaluated. This process can be achieved by calculating the minimum distance between the flight path and the obstacle and predicting the movement trajectory of the obstacle. Then, according to the obstacle risk assessment results, the obstacle avoidance strategy is automatically planned to find the optimal or feasible obstacle avoidance path.

[0151] For obstacle avoidance strategy training, historical flight data and simulation data can be used to train the algorithm model. The model can learn obstacle avoidance strategies in different scenarios and continuously optimize its decision-making ability. During the actual flight process, the real-time collected LiDAR and machine vision data are pre-processed and input into the algorithm model. The model can quickly identify obstacles and predict their movement trajectories based on the input data, and then output the optimal obstacle avoidance strategy.

[0152] In summary, integrating LiDAR, machine vision, and algorithm modules into the flight control system of a drone can form a complete obstacle avoidance system. Testing the system in different scenarios can verify the accuracy and real-time performance of the obstacle avoidance system. Furthermore, the system can be optimized and improved based on the test results to improve its performance and reliability.

[0153] The data analysis unit 104 is mainly used to establish a cloud data collection and analysis platform. By receiving the inspection data sent back by the drone in real time, using big data analysis technology, in-depth mining of the inspection data, developing anomaly detection algorithms, automatically identifying abnormal events during the inspection process, generating data analysis reports, summarizing key findings and trends during the inspection process, and providing a scientific basis for power line maintenance.

[0154] In a specific implementation, the data analysis unit 104 can be mainly used for:

[0155] Receive inspection data sent back by drones in real time;

[0156] The inspection data is first verified, then pre-processed, the processed inspection data is obtained, and the processed inspection data is stored;

[0157] Based on the anomaly detection algorithm, the processed inspection data is deeply mined and analyzed to automatically identify abnormal events in the inspection process and the inspection analysis results, and generate an inspection analysis report.

[0158] In addition, the data analysis unit 104 may also be used to:

[0159] Determine the specific requirements of drone inspections, including but not limited to inspection areas, inspection frequency, data types, and real-time requirements. Define the remote monitoring and command functions and abnormal event recognition accuracy key performance indicators that the system needs to support.

[0160] Design the system architecture, including data collection layer, data processing layer, data storage layer, data analysis layer, and user interaction layer. At the same time, determine the corresponding technology selection and configure the drone.

[0161] Install necessary sensors to collect inspection data. Build a server cluster on the selected cloud computing platform and configure the necessary network environment and security policies.

[0162] Install and configure the big data analysis framework and database system. Develop the data interface between the drone and the cloud system, and design the data transmission protocol, including data format, transmission frequency, error handling, etc.

[0163] Deploy data receiving services in the cloud to receive data sent back by drones in real time, verify and clean the received data, remove invalid, duplicate or abnormal data, and store the processed data in the cloud database. Design efficient data indexing and query mechanisms to support subsequent data analysis and retrieval.

[0164] By designing a user-friendly interactive interface, operation and maintenance personnel are supported to view inspection progress, abnormal events and data analysis results in real time.

[0165] Support remote monitoring and command functions, so that operation and maintenance personnel can grasp the inspection progress in real time and adjust the inspection strategy in time. Through the remote monitoring function, operation and maintenance personnel are allowed to view the real-time video and images of the drone inspection in real time. Through the command function, operation and maintenance personnel are allowed to send instructions to the drone through the cloud system to adjust the inspection strategy or perform specific tasks.

[0166] It can also implement comprehensive system testing, including functional testing, performance testing, and security testing. The system can be optimized and improved based on the test results and user feedback to continuously optimize the algorithm and data processing flow, thereby improving system performance and user experience.

[0167] In an embodiment of the present invention, a power inspection system based on a drone is provided. The power inspection system includes a path planning unit, an obstacle avoidance and positioning unit, and a dynamic obstacle avoidance unit connected in sequence, and a data analysis unit connected to these units respectively. Based on the power inspection system based on a drone, through intelligent autonomous path planning, obstacle avoidance and positioning, dynamic obstacle avoidance and data analysis, the optimal flight path can be dynamically generated, and the flight plan can be adjusted in real time to deal with potential risk areas. Therefore, during the inspection process, the redundant flight time of the drone can be significantly reduced, and the inspection efficiency can be greatly improved. Compared with the traditional manual inspection method, it can not only shorten the drone inspection cycle, but also reduce the workload of the operation and maintenance personnel, and effectively improve the automation and intelligence level of power inspection. In addition, through the automated inspection of drones, the risk of personnel directly participating in operations at high altitudes, complex terrain or harsh environments can be reduced. At the same time, due to the improvement of inspection efficiency, the manpower and time costs can also be indirectly reduced, bringing significant economic benefits to power companies. Further, based on the establishment of a data collection and analysis platform, the operation and maintenance personnel can grasp the inspection progress in real time, adjust the inspection strategy in time, and further enhance the flexibility of the system and the ability to respond to emergencies. Therefore, through the technical solution of the present invention, automation, intelligence and efficiency of UAV-based power inspection can be achieved, providing strong guarantee for the safe and stable operation of the smart grid.

[0168] Reference Figure 2 , shows a flowchart of a method for power inspection based on a drone provided in an embodiment of the present invention, the method is used in the power inspection system based on a drone as described in the previous embodiment, and the method may specifically include the following steps:

[0169] Step 201, obtaining three-dimensional map data and real-time environmental information of the power line to be inspected, and performing real-time path planning based on the three-dimensional map data and the real-time environmental information to dynamically generate an optimal flight path for the UAV;

[0170] Step 202, during the inspection process of the UAV based on the optimal flight path, obstacle avoidance positioning is performed on the power equipment and its surrounding environment on the optimal flight path to generate real-time positioning information of obstacles;

[0171] Step 203: construct a dynamic obstacle map based on the real-time positioning information, and automatically plan an obstacle avoidance strategy according to the dynamic obstacle map, so that the UAV can achieve dynamic obstacle avoidance based on the obstacle avoidance strategy.

[0172] In an optional embodiment, the real-time path planning based on the three-dimensional map data and the real-time environmental information to dynamically generate the optimal flight path of the UAV includes:

[0173] Performing data preprocessing on the three-dimensional map data and the real-time environmental information to obtain processed environmental data;

[0174] Performing a comprehensive real-time path planning analysis on the processed environmental data in combination with a pre-trained path planning model, and dynamically generating multiple candidate flight paths in a three-dimensional map;

[0175] A flight path with the lowest comprehensive cost is selected from the multiple candidate flight paths as the optimal flight path of the UAV. The optimal flight path is used to formulate a detailed flight plan of the UAV, and the detailed flight plan includes a take-off point, a landing point, a flight altitude, a flight speed, and a turning point.

[0176] In an optional embodiment, the pre-trained path planning model is combined to perform a real-time path planning comprehensive analysis on the processed environmental data, and a plurality of candidate flight paths are dynamically generated in a three-dimensional map, including:

[0177] Inputting the processed environmental data into a pre-trained path planning model;

[0178] Performing environmental path identification on the processed environmental data through the pre-trained path planning model to comprehensively analyze the current environment of the power line to be inspected to obtain an environmental analysis result, wherein the environmental analysis result includes a potential risk area identified through model analysis;

[0179] In combination with the potential risk areas, a path planning algorithm is adopted to comprehensively consider the actual flight cost and the estimated flight cost, and a plurality of candidate flight paths are dynamically generated in the three-dimensional map.

[0180] In an optional embodiment, performing obstacle avoidance positioning on the power equipment and its surrounding environment on the optimal flight path to generate real-time positioning information of obstacles includes:

[0181] Collecting three-dimensional spatial information of the power equipment and its surrounding environment on the optimal flight path in real time;

[0182] According to the three-dimensional space information, combined with an obstacle avoidance algorithm, real-time positioning information of obstacles on the optimal flight path is generated, so that the UAV automatically identifies and avoids the obstacles based on the real-time positioning information.

[0183] In an optional embodiment, the method further includes:

[0184] Collecting image data and temperature distribution diagram of the power equipment and its surrounding environment on the optimal flight path in real time;

[0185] Preprocessing the image data to obtain processed image data, and importing the processed image data into a pre-trained image classification model to perform abnormality recognition and classification on the collected images and generate an abnormality detection report;

[0186] According to the temperature distribution diagram, combined with the ambient temperature and equipment load factors, the overheating of the power equipment is judged to determine the potential safety hazards on the optimal flight path;

[0187] Based on the abnormality detection report and the potential safety hazard, a comprehensive monitoring report for the power line to be inspected is generated.

[0188] In an optional embodiment, the step of constructing a dynamic obstacle map based on the real-time positioning information and automatically planning an obstacle avoidance strategy according to the dynamic obstacle map includes:

[0189] Collecting image data of the power equipment and its surrounding environment on the optimal flight path in real time;

[0190] Convert the real-time positioning information into spatial coordinate information to determine the location of the obstacle and the distance between the obstacle and the drone;

[0191] Performing feature recognition on the image data based on a machine vision algorithm, extracting object features, and classifying and labeling the obstacles based on the object features to obtain labeled obstacles;

[0192] Constructing a dynamic obstacle map according to the spatial coordinate information and the marked obstacles;

[0193] Real-time analysis of the risk level of obstacles on the dynamic obstacle map to the flight of the UAV;

[0194] When it is detected that the risk level of the obstacle closest to the drone is greater than or equal to a preset risk level threshold, the obstacle avoidance strategy planning is automatically started, and the optimal obstacle avoidance path is replanned to re-search the safe flight path in the dynamic obstacle map, so that the drone changes the current flight path based on the safe flight path to achieve dynamic obstacle avoidance.

[0195] In an optional embodiment, the method further includes:

[0196] Receive the inspection data sent back by the drone in real time;

[0197] The inspection data is first verified, then data preprocessed, to obtain processed inspection data, and the processed inspection data is stored;

[0198] Based on the anomaly detection algorithm, the processed inspection data is deeply mined and analyzed to automatically identify abnormal events in the inspection process and the inspection analysis results, and generate an inspection analysis report.

[0199] As for the method embodiment, since it is basically similar to the system embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the aforementioned system embodiment.

[0200] In an embodiment of the present invention, a corresponding power inspection method is provided on the basis of the power inspection system based on drones introduced above. Through intelligent autonomous path planning, obstacle avoidance positioning, dynamic obstacle avoidance and data analysis, the optimal flight path can be dynamically generated, and the flight plan can be adjusted in real time to deal with potential risk areas. Therefore, during the inspection process, the redundant flight time of the drone can be significantly reduced, and the inspection efficiency can be greatly improved. Compared with the traditional manual inspection method, it can not only shorten the drone inspection cycle, but also reduce the workload of operation and maintenance personnel, and effectively improve the automation and intelligence level of power inspection. In addition, through the automated inspection of drones, the risk of personnel directly participating in operations at high altitudes, complex terrain or harsh environments can be reduced. At the same time, due to the improvement of inspection efficiency, the manpower and time costs can also be indirectly reduced, bringing significant economic benefits to power companies. Further, based on the establishment of a data collection and analysis platform, the operation and maintenance personnel can grasp the inspection progress in real time, adjust the inspection strategy in time, and further enhance the flexibility of the system and the ability to deal with emergencies. Therefore, through the technical solution of the present invention, the automation, intelligence and efficiency of drone-based power inspection can be realized, providing a strong guarantee for the safe and stable operation of the smart grid.

[0201] For better explanation, refer to Figure 3 , showing a schematic diagram of the overall process of a power inspection method based on a drone provided by an embodiment of the present invention. It should be pointed out that this embodiment only briefly describes the general process of power inspection based on a drone, and the specific implementation process of each step can be understood by referring to the relevant content in the aforementioned embodiment, which will not be repeated here. It can be understood that the present invention is not limited to this.

[0202] To further simplify the process, the data analysis unit is described in the embodiment of the present invention by taking the example of the drone transmitting data back to the data analysis unit for processing after the inspection. It can be known that during or after the inspection, the data collected by the drone can be transmitted back to the data analysis unit for analysis and processing.

[0203] 1. Execution process corresponding to the path planning unit

[0204] Collect three-dimensional map data and real-time environmental information of the power lines to be inspected;

[0205] Perform data preprocessing on three-dimensional map data and real-time environmental information to obtain processed environmental data;

[0206] The processed environmental data is input into the pre-trained path planning model for environmental path identification, so as to comprehensively analyze the current environment of the power line to be inspected and obtain environmental analysis results, which include potential risk areas identified through model analysis;

[0207] Combined with potential risk areas, a path planning algorithm is used to consider the actual flight cost and the estimated flight cost, and multiple candidate flight paths are dynamically generated in the three-dimensional map;

[0208] The flight path with the lowest comprehensive cost is selected from multiple candidate flight paths as the optimal flight path for the UAV, and a detailed flight plan for the UAV is formulated based on the optimal flight path.

[0209] (II) Execution process corresponding to the obstacle avoidance and positioning unit

[0210] When the drone is ready, it takes off according to the flight plan, starts to perform inspection tasks, and loads machine vision and high-precision sensors;

[0211] Real-time collection of three-dimensional spatial information, image data and temperature distribution diagrams of power equipment and its surrounding environment on the optimal flight path;

[0212] Based on the three-dimensional spatial information and combined with the obstacle avoidance algorithm, the real-time positioning information of obstacles on the optimal flight path is generated;

[0213] Preprocess the image data to obtain processed image data, and import the processed image data into a pre-trained image classification model to perform abnormality recognition and classification on the collected images and generate an abnormality detection report;

[0214] According to the temperature distribution diagram, combined with the ambient temperature and equipment load factors, the overheating of the power equipment is judged, and the potential safety hazards on the optimal flight path are determined;

[0215] Based on the abnormal detection report and potential safety hazards, a comprehensive monitoring report for the power lines to be inspected is generated.

[0216] (III) Execution process corresponding to the dynamic obstacle avoidance unit

[0217] Convert real-time positioning information into spatial coordinate information to determine the location of obstacles and their distance from the drone;

[0218] Based on the machine vision algorithm, the image data is identified and the object features are extracted. Then, obstacles are classified and labeled based on the object features to obtain labeled obstacles.

[0219] Construct a dynamic obstacle map based on spatial coordinate information and marked obstacles;

[0220] Real-time analysis of the risk level of obstacles to drone flight on the dynamic obstacle map;

[0221] When it is detected that the risk level of the obstacle closest to the drone is greater than or equal to the preset risk level threshold, the obstacle avoidance strategy planning is automatically started, and the optimal obstacle avoidance path is re-planned to re-search for a safe flight path in the dynamic obstacle map, so that the drone can change the current flight path based on the safe flight path to achieve dynamic obstacle avoidance.

[0222] (IV) Execution process corresponding to the data analysis unit

[0223] After the inspection is completed, receive the inspection data sent back by the drone;

[0224] The inspection data is first verified, and then pre-processed to obtain the processed inspection data;

[0225] Store and process inspection data;

[0226] Based on the anomaly detection algorithm, the processed inspection data is deeply mined and analyzed to automatically identify abnormal events in the inspection process and the inspection analysis results, and generate an inspection analysis report.

[0227] For ease of understanding, the technical solution of the present invention is briefly described below through two embodiments.

[0228] Embodiment 1

[0229] The various sensors carried by drones are used to collect real-time environmental data of power lines and their surroundings, including but not limited to the location of power towers, line directions, terrain heights, vegetation coverage information, tree distribution, building locations, and weather conditions. The collected raw data is preprocessed by cleaning, denoising, and formatting to eliminate redundancy, errors, and noise in the data and improve the accuracy and availability of the data. At the same time, the preprocessed data is integrated into the geographic information system to form an accurate three-dimensional map of the power lines and their surroundings.

[0230] Select the path planning algorithm model according to the task requirements, and use historical inspection data and simulation data to train the model. During the model training process, the goal is to minimize the loss function and continuously optimize the model parameters to improve the prediction accuracy. The loss function used in the training process is:

[0231]

[0232] in, represents the model parameters; represents the number of samples; Indicates The trajectory length of samples; Represents the discount factor. First, use the random strategy to initialize , for each sample , calculate the features and decision actions Then execute the action , observe the new state and rewards , calculate the cumulative reward , used to control the importance of future rewards. Update using gradient descent To minimize .

[0233] By continuously optimizing model parameters, the path planning model's ability to recognize complex environments and the path planning accuracy can be improved.

[0234] The pre-processed 3D map data and real-time environmental information are input into the path planning algorithm model, and the current environment is comprehensively analyzed based on the model to identify potential risk areas. Based on the environmental analysis results, the path planning algorithm is used to generate multiple candidate flight paths in the 3D map.

[0235] Afterwards, the candidate paths are evaluated, and the optimal path is selected by comprehensively considering factors such as flight time, safety, and energy consumption. A detailed flight plan is then developed based on the optimal path, including take-off point, landing point, flight altitude, speed, and turning point. The drone performs inspection tasks according to the flight plan and continuously monitors real-time environmental information during the flight.

[0236] During the flight, when encountering emergencies, the system can respond quickly, re-evaluate the environment and adjust the flight plan. By optimizing the obstacle avoidance strategy, the drone can safely avoid obstacles and continue to complete the mission. During and after the inspection, the system can collect the data sent back by the drone in real time, and conduct in-depth mining and analysis of the collected data based on the data analysis unit, identify abnormal events during the inspection, evaluate the inspection results, and identify potential problems. The model and path planning algorithm can also be continuously optimized based on the data analysis results. Thus, the accuracy and efficiency of the system power inspection can be improved through continuous iterative optimization.

[0237] Embodiment 2

[0238] The drone is equipped with high-precision sensors including GPS (Global Positioning System), inertial navigation system and lidar to comprehensively collect accurate three-dimensional map data of power lines and their surrounding environment. These data cover key elements such as the location of power towers, line direction, terrain height, and vegetation coverage information. During the flight, the drone collects current environmental data in real time to ensure the timeliness and accuracy of the data as much as possible. Then, the collected three-dimensional map data and real-time environmental information are pre-processed by cleaning, denoising, and formatting to improve the accuracy and availability of the data, providing a basis for subsequent analysis.

[0239] Combined with the pre-trained path planning model, the pre-processed data is comprehensively analyzed to obtain environmental analysis results and identify potential risk areas. Based on the environmental analysis results, an advanced path planning algorithm is used to dynamically generate multiple candidate flight paths in the three-dimensional map. These paths take into account the layout of power lines, terrain characteristics, and potential risk areas. The candidate paths are evaluated, and the optimal path is selected by comprehensively considering flight time, safety, and energy consumption factors. At the same time, a detailed flight plan is formulated so that the drone can complete the inspection task efficiently and safely.

[0240] During the flight, the drone continuously monitors real-time environmental information to respond promptly to emergencies such as sudden weather changes and new obstacles. Once an emergency occurs, the system can quickly assess the current environment, automatically adjust the flight plan, and re-plan the flight path so that the drone can safely avoid obstacles and continue to complete the mission. During and after the inspection, the data sent back by the drone is collected and analyzed, and the collected data is deeply analyzed to continuously optimize the model and path planning algorithm to improve the accuracy and efficiency of the system. Through continuous learning and optimization, the drone can respond to various complex environments more intelligently and further improve the inspection effect.

[0241] An embodiment of the present invention further provides an electronic device, the device comprising a processor and a memory:

[0242] The memory is used to store the program code and transmit the program code to the processor;

[0243] The processor is used to execute the power inspection method based on a drone of any embodiment of the present invention according to the instructions in the program code.

[0244] An embodiment of the present invention further provides a computer-readable storage medium, which is used to store program code, and the program code is used to execute the drone-based power inspection method of any embodiment of the present invention.

[0245] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0246] In the several embodiments provided by the present invention, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0247] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0248] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0249] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc., various media that can store program codes.

[0250] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features thereof may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A power inspection system based on drones, characterized in that: The power inspection system includes a path planning unit, an obstacle avoidance and positioning unit, and a dynamic obstacle avoidance unit connected in sequence; wherein, The path planning unit is used to obtain three-dimensional map data and real-time environmental information of the power line to be inspected, and perform real-time path planning based on the three-dimensional map data and the real-time environmental information to dynamically generate an optimal flight path for the UAV; The obstacle avoidance and positioning unit is used to perform obstacle avoidance and positioning of the power equipment and its surrounding environment on the optimal flight path and generate real-time positioning information of obstacles during the inspection of the UAV based on the optimal flight path; The dynamic obstacle avoidance unit is used to construct a dynamic obstacle map based on the real-time positioning information, and automatically plan an obstacle avoidance strategy according to the dynamic obstacle map, so that the UAV can achieve dynamic obstacle avoidance based on the obstacle avoidance strategy.

2. The power inspection system according to claim 1, characterized in that: The path planning unit is specifically used for: Performing data preprocessing on the three-dimensional map data and the real-time environmental information to obtain processed environmental data; Performing a comprehensive real-time path planning analysis on the processed environmental data in combination with a pre-trained path planning model, and dynamically generating multiple candidate flight paths in a three-dimensional map; A flight path with the lowest comprehensive cost is selected from the multiple candidate flight paths as the optimal flight path of the UAV. The optimal flight path is used to formulate a detailed flight plan of the UAV, and the detailed flight plan includes a take-off point, a landing point, a flight altitude, a flight speed, and a turning point.

3. The power inspection system according to claim 2, characterized in that: The path planning unit is specifically used for: Inputting the processed environmental data into a pre-trained path planning model; Performing environmental path identification on the processed environmental data through the pre-trained path planning model to comprehensively analyze the current environment of the power line to be inspected to obtain an environmental analysis result, wherein the environmental analysis result includes a potential risk area identified through model analysis; In combination with the potential risk areas, a path planning algorithm is adopted to comprehensively consider the actual flight cost and the estimated flight cost, and a plurality of candidate flight paths are dynamically generated in the three-dimensional map.

4. The power inspection system according to claim 1, characterized in that: The obstacle avoidance and positioning unit is specifically used for: Collecting three-dimensional spatial information of the power equipment and its surrounding environment on the optimal flight path in real time; According to the three-dimensional space information, combined with an obstacle avoidance algorithm, real-time positioning information of obstacles on the optimal flight path is generated, so that the UAV automatically identifies and avoids the obstacles based on the real-time positioning information.

5. The power inspection system according to claim 4, characterized in that: The obstacle avoidance and positioning unit is also used for: Collecting image data and temperature distribution diagram of the power equipment and its surrounding environment on the optimal flight path in real time; Preprocessing the image data to obtain processed image data, and importing the processed image data into a pre-trained image classification model to perform abnormality recognition and classification on the collected images and generate an abnormality detection report; According to the temperature distribution diagram, combined with the ambient temperature and equipment load factors, the overheating of the power equipment is judged to determine the potential safety hazards on the optimal flight path; Based on the abnormality detection report and the potential safety hazard, a comprehensive monitoring report for the power line to be inspected is generated.

6. The power inspection system according to claim 4, characterized in that: The dynamic obstacle avoidance unit is specifically used for: Collecting image data of the power equipment and its surrounding environment on the optimal flight path in real time; Convert the real-time positioning information into spatial coordinate information to determine the location of the obstacle and the distance between the obstacle and the drone; Performing feature recognition on the image data based on a machine vision algorithm, extracting object features, and classifying and labeling the obstacles based on the object features to obtain labeled obstacles; Constructing a dynamic obstacle map according to the spatial coordinate information and the marked obstacles; Real-time analysis of the risk level of obstacles on the dynamic obstacle map to the flight of the UAV; When it is detected that the risk level of the obstacle closest to the drone is greater than or equal to a preset risk level threshold, the obstacle avoidance strategy planning is automatically started, and the optimal obstacle avoidance path is re-planned to re-search the safe flight path in the dynamic obstacle map, so that the drone changes the current flight path based on the safe flight path to achieve dynamic obstacle avoidance.

7. The power inspection system according to any one of claims 1 to 6, characterized in that: The power inspection system further includes a data analysis unit, which is used to: Receive the inspection data sent back by the drone in real time; The inspection data is first verified, then data preprocessed, to obtain processed inspection data, and the processed inspection data is stored; Based on the anomaly detection algorithm, the processed inspection data is deeply mined and analyzed to automatically identify abnormal events in the inspection process and the inspection analysis results, and generate an inspection analysis report.

8. A power inspection method based on drone, characterized in that: The power inspection method is used in the power inspection system based on a drone according to any one of claims 1 to 7, and the method comprises: Acquire three-dimensional map data and real-time environmental information of the power line to be inspected, and perform real-time path planning based on the three-dimensional map data and the real-time environmental information to dynamically generate an optimal flight path for the UAV; During the inspection process of the UAV based on the optimal flight path, obstacle avoidance positioning is performed on the power equipment and its surrounding environment on the optimal flight path to generate real-time positioning information of obstacles; Based on the real-time positioning information, a dynamic obstacle map is constructed, and according to the dynamic obstacle map, an obstacle avoidance strategy is automatically planned, so that the UAV can achieve dynamic obstacle avoidance based on the obstacle avoidance strategy.

9. An electronic device, characterized in that: The device comprises a processor and a memory: The memory is used to store program code and transmit the program code to the processor; The processor is used to execute the UAV-based power inspection method of claim 8 according to the instructions in the program code.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium is used to store program code, and the program code is used to execute the drone-based power inspection method described in claim 8.

Images