Low-cost autonomous inspection method for unmanned aerial vehicle inspection of power distribution network

Abstract

The invention, which relates to the technical field of autonomous inspection, discloses a low-cost autonomous inspection method for unmanned aerial vehicle inspection of a power distribution network, comprising the following steps: establishing an inspection waypoint and generating an initial inspection path to obtain an inspection path; performing environment modeling based on the inspection path to obtain environment data; performing flight assistance based on the environment data to obtain flight stability data; performing inspection shooting according to the flight stability data to obtain inspection image data; performing defect detection according to the inspection image data to obtain defect information; performing infrared image supplementary verification according to the defect information to obtain verification data; performing low-altitude reinspection according to the verification data to obtain reinspection image data; and performing fault fine modeling according to the recheck image data. According to the method, unnecessary calculation is reduced, calculation resource consumption is reduced, the processing speed is increased, the performance is improved, the detection precision and the sensitivity to equipment defects are improved, and the real-time detection requirement of low-cost autonomous inspection is met.

Description

Technical Field

[0001] This invention relates to the field of autonomous inspection technology, specifically a low-cost autonomous inspection method for unmanned aerial vehicle (UAV) inspection of power distribution networks. Background Technology

[0002] With the continuous expansion of power distribution networks, the stable operation of power equipment is crucial to power supply security. Traditional manual inspection methods suffer from high labor costs, long inspection cycles, and low security, failing to meet the efficient operation and maintenance needs of modern power distribution networks. Unmanned aerial vehicle (UAV) inspection technology, with its advantages of high efficiency, flexibility, and wide coverage, has become an important means of power distribution network inspection.

[0003] Existing technologies have shortcomings in environmental modeling: traditional modeling has limitations, such as visual feature points relying too much on specific textures, which leads to the failure of mapping low-texture areas (such as substations and power distribution towers). UAVs with limited computing resources cannot support global high-density point cloud modeling, resulting in missing details. It is difficult to balance mapping accuracy and computational complexity, and drift accumulation is likely to occur in complex environments.

[0004] Existing technologies for defect detection have shortcomings: traditional defect detection methods typically require high computing resources, have slow processing speeds, and lack sensitivity to key defect features, affecting the reliability and accuracy of detection. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a low-cost autonomous inspection method for unmanned aerial vehicle (UAV) inspection of power distribution networks, thereby solving the problems mentioned in the background section.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] In a first aspect, embodiments of the present invention provide a low-cost autonomous inspection method for unmanned aerial vehicle (UAV) inspection of power distribution networks, comprising the following steps:

[0008] S1. Establish inspection waypoints and generate an initial inspection path to obtain the inspection path;

[0009] S2. Based on the inspection path, perform environmental modeling to obtain environmental data;

[0010] S3. Flight assistance based on environmental data to obtain flight stability data;

[0011] S4. Conduct inspection and photography based on flight stability data to obtain inspection image data;

[0012] S5. Perform defect detection based on the inspection image data to obtain defect information;

[0013] S6. Perform supplementary infrared image verification based on defect information to obtain verification data;

[0014] S7. Based on the verification data, perform a low-altitude re-inspection to obtain the re-inspection image data;

[0015] S8. Perform detailed fault modeling based on the re-inspection image data.

[0016] To further optimize this technical solution, the environmental modeling in S2 includes:

[0017] Based on the high-resolution images of the inspection path obtained by the camera when the UAV performs the inspection mission, an adaptive multi-scale modeling method is used to model the environment and construct a three-dimensional environment model of the inspection area.

[0018] To further optimize this technical solution, the adaptive multi-scale modeling includes:

[0019] ;

[0020] in:

[0021] Point cloud in pixels Density at that location;

[0022] Density adjustment coefficient;

[0023] : Adaptive weights for the nth frame image;

[0024] Gaussian blur kernel.

[0025] To further optimize this technical solution, the adaptive weights include:

[0026] ;

[0027] in:

[0028] : Multi-scale information entropy of the nth frame image;

[0029] The total number of frames in the image;

[0030] : Sum of the information entropy of all frames.

[0031] To further optimize this technical solution, the multi-scale information entropy includes:

[0032] ;

[0033] in:

[0034] : The nth frame image;

[0035] The total number of pixels in the image;

[0036] : The normalized probability of the gradient intensity of the i-th pixel in the image.

[0037] To further optimize this technical solution, the defect detection in S5 includes:

[0038] Based on the inspection image data, feature maps are extracted using the lightweight neural network MobileNetV3, and defect detection is performed using a defect detection model.

[0039] To further optimize this technical solution, the defect detection model includes:

[0040] ;

[0041] in:

[0042] : The predicted probability of equipment defects;

[0043] Activation function;

[0044] The weights of the model;

[0045] : Adaptive weighted feature map;

[0046] : The bias term of the model.

[0047] To further optimize this technical solution, the adaptive weighting includes:

[0048] ;

[0049] in:

[0050] The number of channels in the feature map;

[0051] The weight coefficient of the i-th channel;

[0052] Feature map of the nth frame image The eigenvalues ​​in the i-th channel.

[0053] To further optimize this technical solution, the low-altitude re-inspection in S7 includes:

[0054] Based on the verification data, the defective equipment was identified. Using the inspection height adjustment model, combined with the defect type and environmental factors, a low-altitude re-inspection was conducted to obtain re-inspection image data.

[0055] To further optimize this technical solution, the inspection height adjustment model includes:

[0056] ;

[0057] in:

[0058] Optimal inspection height;

[0059] Minimum permitted flight altitude;

[0060] The size of the defect;

[0061] : The measured value of the i-th environmental factor;

[0062] : The weight coefficient of the i-th environmental factor;

[0063] The current wind speed;

[0064] , , : Adjust parameters.

[0065] In a second aspect, embodiments of the present invention provide a computer device, including a memory and a processor, wherein the memory stores a computer program, and the computer program instructions, when executed by the processor, implement the steps of a low-cost autonomous inspection method for distribution network drone inspection as described in the first aspect of the present invention.

[0066] Thirdly, embodiments of the present invention provide a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program instructions are executed by a processor, they implement the steps of a low-cost autonomous inspection method for distribution network unmanned aerial vehicle (UAV) inspection as described in the first aspect of the present invention.

[0067] Compared with existing technologies, this invention provides a low-cost autonomous inspection method for power distribution network unmanned aerial vehicle (UAV) inspections, which has the following advantages:

[0068] This low-cost autonomous inspection method for power distribution network drone inspections uses adaptive multi-scale modeling to dynamically adjust point cloud density based on the importance of the target area. It leverages the multi-scale characteristics of visual information to perform high-precision modeling of key components (such as insulators, connection points, and switching devices), while using low-precision modeling for background areas. This reduces redundant data, lowers unnecessary computation, improves processing speed, and enhances performance. It can still build high-precision maps in environments with poor lighting and no textured backgrounds.

[0069] By using a defect detection model, the consumption of computing resources is significantly reduced, enabling it to run efficiently on low-power devices, meeting the real-time detection requirements of low-cost autonomous inspection, and improving detection accuracy and sensitivity to equipment defects. Attached Figure Description

[0070] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0071] Figure 1 This is a flowchart illustrating a low-cost autonomous inspection method for power distribution network unmanned aerial vehicle (UAV) inspection proposed in this invention.

[0072] Figure 2 This is a flowchart illustrating the adaptive multi-scale modeling process of a low-cost autonomous inspection method for distribution network drone inspection proposed in this invention.

[0073] Figure 3 This is a flowchart illustrating the defect detection model of a low-cost autonomous inspection method for distribution network drone inspection proposed in this invention.

[0074] Figure 4 This is a flowchart illustrating the inspection height adjustment model of a low-cost autonomous inspection method for power distribution network drone inspection proposed in this invention. Detailed Implementation

[0075] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0076] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0077] Secondly, the term "an embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places throughout this specification does not necessarily refer to the same embodiment, nor is it a single embodiment or an embodiment selectively excluded from other embodiments.

[0078] Example 1:

[0079] Reference Figures 1-4 This is the first embodiment of the present invention, which provides a low-cost autonomous inspection method for power distribution network unmanned aerial vehicle (UAV) inspection, including the following steps:

[0080] S1. Establish inspection waypoints and generate an initial inspection path to obtain the inspection path.

[0081] In this embodiment, establishing waypoints and generating an initial flight path includes:

[0082] Before conducting a drone inspection mission, key target points for the power distribution network inspection are identified, including transmission lines, transformers, and switchgear. A high-precision RTK-GNSS system is used to mark waypoints in the target area and generate an initial inspection path. Factors considered in waypoint selection include equipment distribution (ensuring the drone's inspection range covers all key targets), drone endurance (reasonably dividing the inspection area to avoid excessively large single flight ranges), and obstacle avoidance requirements (avoiding tall buildings, trees, and other obstacles to ensure flight safety). RTK-GNSS provides centimeter-level positioning accuracy, ensuring the drone can cruise precisely and avoid deviations or target offsets.

[0083] S2. Environmental modeling is performed based on the inspection path to obtain environmental data.

[0084] In this embodiment, environment modeling includes:

[0085] Based on the high-resolution images of the inspection path obtained by the camera when the UAV performs the inspection mission, an adaptive multi-scale modeling method is used to model the environment and construct a three-dimensional environment model of the inspection area.

[0086] Furthermore, the adaptive multi-scale modeling includes:

[0087] ;

[0088] in:

[0089] Point cloud in pixels Density at that location;

[0090] Density adjustment coefficient;

[0091] : Adaptive weights for the nth frame image;

[0092] Gaussian blur kernel, used to ensure smooth transitions in dense areas.

[0093] when At high altitudes, point cloud density Increase, achieve dense modeling, when At low point cloud density Reduce and optimize computational efficiency.

[0094] Furthermore, the adaptive weights include:

[0095] ;

[0096] in:

[0097] : Multi-scale information entropy of the nth frame image, used to measure the importance of pixels;

[0098] The total number of frames in the image;

[0099] The sum of the information entropy of all frames is used to normalize the weights.

[0100] Furthermore, the multi-scale information entropy includes:

[0101] ;

[0102] in:

[0103] : The nth frame image;

[0104] The total number of pixels in the image;

[0105] : The normalized probability of the gradient intensity of the i-th pixel in the image.

[0106] Environmental reconstruction is carried out by utilizing the multi-scale characteristics of visual information, using dense mapping in key areas and sparse mapping in irrelevant areas.

[0107] Compared to traditional models, this model effectively reduces unnecessary computation, improves inspection efficiency, and can still build high-precision maps in environments with poor lighting and no texture background. It adaptively adjusts point cloud density to achieve high-precision modeling of key parts and low-precision modeling of background areas, thereby improving performance.

[0108] The steps for using the above model include:

[0109] Information entropy calculation: based on the acquired high-resolution image data Calculate multi-scale information entropy To determine the level of detail in a region;

[0110] Adaptive weight calculation: based on the calculated multi-scale information entropy Calculate the adaptive weights for each frame This is used to measure the importance of pixels in order to determine the accuracy of map construction.

[0111] Map output: Combined with calculated adaptive weights Calculate point cloud density This achieves sparse-dense transformation and outputs optimized 3D environment modeling data based on the calculated point cloud density, which is then used by the UAV assisted flight method in step S3.

[0112] S3. Flight assistance based on environmental data to obtain flight stability data.

[0113] In this embodiment, flight assistance includes:

[0114] Based on environmental data, to prevent GPS signal interference and obstruction during inspection missions and to ensure stable flight of the UAV in complex environments, an inertial navigation IMU and optical flow sensor are used to assist flight, improving the UAV's inspection capabilities in complex environments, ensuring stable flight paths, and avoiding target deviation caused by GPS drift. The IMU is used to calculate the UAV's current attitude, while the optical flow sensor is used to calculate the UAV's relative displacement by utilizing ground texture changes when flying at low altitudes and entering areas with weak GNSS signals, improving positioning stability. Combined with environmental data, the IMU and optical flow data are corrected in real time to prevent drift and improve flight stability.

[0115] S4. Conduct inspection and photography based on flight stability data to obtain inspection image data.

[0116] In this embodiment, the inspection and photography includes:

[0117] Based on flight stability data, after flight stabilization, the inspection data acquisition phase begins. High-resolution RGB cameras are used for inspection image capture to obtain inspection image data. When capturing high-definition images of transmission lines and distribution equipment, it is ensured that the shooting area is fixed and clear, that image details are sharp, and that the images are unaffected by strong light or shadows, ensuring no omissions during the inspection process, thus providing a foundation for subsequent inspection image data analysis.

[0118] S5. Perform defect detection based on the inspection image data to obtain defect information.

[0119] In this embodiment, defect detection includes:

[0120] Based on the inspection image data, feature maps are extracted using the lightweight neural network MobileNetV3, and defect detection models are used for defect detection. This enables rapid detection and identification of defects in power distribution network equipment at the local end, avoiding latency and network bandwidth bottlenecks.

[0121] Furthermore, the defect detection model includes:

[0122] ;

[0123] in:

[0124] : The predicted probability of equipment defects;

[0125] : Activation function, using the Sigmoid activation function, to output probability values ​​between 0 and 1;

[0126] The weights of the model;

[0127] : Adaptive weighted feature map;

[0128] : The bias term of the model.

[0129] Furthermore, the adaptive weighting includes:

[0130] ;

[0131] in:

[0132] The number of channels in the feature map;

[0133] The weight coefficient of the i-th channel is adaptively calculated based on the intensity of the defect feature;

[0134] Feature map of the nth frame image The eigenvalues ​​in the i-th channel, Obtained through the lightweight neural network MobileNetV3.

[0135] By adaptively weighting the feature maps, the predicted probability of equipment defects is calculated, and a threshold is used to determine whether the equipment has defects.

[0136] The lightweight architecture of MobileNetV3 reduces computational resource consumption, making it suitable for real-time detection on low-power devices. Adaptive weighted feature maps improve detection accuracy and sensitivity to device defects.

[0137] The steps for using this model include:

[0138] Data Acquisition: Image data is acquired from step S4, and feature maps are extracted using the lightweight neural network MobileNetV3. ;

[0139] Feature map weighting: Calculate the weight coefficient for each channel based on the intensity of the defect features. The feature map is adaptively weighted to obtain the adaptively weighted feature map. ;

[0140] Defect detection: based on the adaptively weighted feature map The calculation is performed, and the predicted probability of device defects is obtained after applying the Sigmoid activation function. The predicted probability is compared with a threshold to determine whether the equipment has a defect, for example. If the value is greater than 0.5, the equipment is considered defective; otherwise, the equipment is considered normal.

[0141] S6. Perform supplementary verification using infrared images based on the defect information to obtain verification data.

[0142] In this embodiment, infrared image supplementary verification includes:

[0143] When equipment malfunctions are detected based on defect information, an infrared thermal imaging camera is used to detect the temperature of the malfunctioning equipment, verifying the authenticity of the fault and improving the accuracy of inspections. By measuring the surface temperature of the equipment and comparing it with the normal temperature range, it is determined whether there is a temperature anomaly. The infrared image is then overlaid with an RGB image to improve the accuracy of defect identification.

[0144] S7. Perform a low-altitude re-inspection based on the verification data to obtain re-inspection image data.

[0145] In this embodiment, low-altitude re-inspection includes:

[0146] Based on verification data, defective equipment is identified. Using an inspection altitude adjustment model, combined with the defect type and environmental factors, the optimal inspection altitude is determined. Low-altitude re-inspection is then performed on the detected defective equipment, yielding re-inspection image data. By dynamically adjusting the low-altitude re-inspection, the drone's flight altitude is automatically adjusted, improving the accuracy and efficiency of the re-inspection.

[0147] Furthermore, the inspection height adjustment model includes:

[0148] ;

[0149] in:

[0150] Optimal inspection height;

[0151] Minimum permitted flight altitude to ensure the drone avoids obstacles;

[0152] The size of the defect indicates its severity; larger defects may require a lower inspection height.

[0153] : The measured value of the i-th environmental factor, including humidity and temperature;

[0154] : The weighting coefficient of the i-th environmental factor, used to reflect the degree of influence of this factor on flight altitude;

[0155] The current wind speed is high; when the wind speed is high, the inspection altitude needs to be increased to ensure flight stability.

[0156] , , Adjusting parameters controls the impact of defect size, environmental factors, and wind speed on inspection height.

[0157] By combining defect types and environmental factors, adaptive inspection height adjustment can be achieved.

[0158] Traditional re-inspection methods typically involve fixed flight altitudes and reliance on manual intervention to adjust the altitude, which reduces inspection efficiency and accuracy. By using adaptive inspection altitude adjustment, combined with defect type and environmental factors, the optimal inspection altitude under the current environment can be calculated, enabling dynamic adjustment of low-altitude re-inspection and improving the accuracy and efficiency of re-inspection.

[0159] The use of the model includes:

[0160] Influencing Factors: Determining the Size of the Defect Obtaining environmental factors and wind speed Used for calculating inspection height;

[0161] Inspection height calculation: Calculate the corresponding inspection height based on the factors that affect the inspection height.

[0162] Inspection altitude determination: Based on the calculated inspection altitude, compare it with the minimum permissible flight altitude. Compare the two values ​​and take the maximum value to determine the optimal inspection height. .

[0163] S8. Perform detailed fault modeling based on the re-inspection image data.

[0164] In this embodiment, detailed fault modeling includes:

[0165] Based on the re-inspection image data obtained from the low-altitude re-inspection, a 3D model of the fault location is generated using structured light scanning and structured motion (SfM) modeling technology, providing more intuitive data support for maintenance decisions.

[0166] Example 2:

[0167] This embodiment also provides a computer device applicable to a low-cost autonomous inspection method for distribution network drone inspection, including a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to realize the low-cost autonomous inspection method for distribution network drone inspection proposed in the above embodiment.

[0168] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, it implements a low-cost autonomous inspection method for power distribution network drone inspection as proposed in the above embodiment.

[0169] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0170] If a function is implemented as 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 this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0171] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-including system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0172] More specific examples (a non-exhaustive list) of computer-readable media include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which programs can be printed, because programs can be obtained electronically, for example, by optically scanning the paper or other media, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0173] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0174] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A low-cost autonomous inspection method for unmanned aerial vehicle (UAV) inspection of power distribution networks, characterized in that, Includes the following steps: S1. Establish inspection waypoints and generate an initial inspection path to obtain the inspection path; S2. Based on the inspection path, perform environmental modeling to obtain environmental data; S3. Flight assistance based on environmental data to obtain flight stability data; S4. Conduct inspection and photography based on flight stability data to obtain inspection image data; S5. Perform defect detection based on the inspection image data to obtain defect information; S6. Perform supplementary infrared image verification based on defect information to obtain verification data; S7. Based on the verification data, perform a low-altitude re-inspection to obtain the re-inspection image data; S8. Perform detailed fault modeling based on the re-inspection image data.

2. The low-cost autonomous inspection method for unmanned aerial vehicle (UAV) inspection of power distribution networks according to claim 1, characterized in that, The environment modeling in S2 includes: Based on the high-resolution images of the inspection path obtained by the camera when the UAV performs the inspection mission, an adaptive multi-scale modeling method is used to model the environment and construct a three-dimensional environment model of the inspection area.

3. A low-cost autonomous inspection method for power distribution network unmanned aerial vehicle (UAV) inspection according to claim 2, characterized in that, The adaptive multi-scale modeling includes: ； in: Point cloud in pixels Density at that location; Density adjustment coefficient; : Adaptive weights for the nth frame image; Gaussian blur kernel.

4. A low-cost autonomous inspection method for unmanned aerial vehicle (UAV) inspection of power distribution networks according to claim 3, characterized in that, The adaptive weights include: ； in: : Multi-scale information entropy of the nth frame image; The total number of frames in the image; : Sum of the information entropy of all frames.

5. A low-cost autonomous inspection method for unmanned aerial vehicle (UAV) inspection of power distribution networks according to claim 4, characterized in that, The multi-scale information entropy includes: ； in: : The nth frame image; The total number of pixels in the image; : The normalized probability of the gradient intensity of the i-th pixel in the image.

6. A low-cost autonomous inspection method for unmanned aerial vehicle (UAV) inspection of power distribution networks according to claim 1, characterized in that, The defect detection in S5 includes: Based on the inspection image data, feature maps are extracted using the lightweight neural network MobileNetV3, and defect detection is performed using a defect detection model.

7. A low-cost autonomous inspection method for power distribution network unmanned aerial vehicle (UAV) inspection according to claim 6, characterized in that, The defect detection model includes: ； in: : The predicted probability of equipment defects; Activation function; The weights of the model; : Adaptive weighted feature map; : The bias term of the model.

8. A low-cost autonomous inspection method for unmanned aerial vehicle (UAV) inspection of power distribution networks according to claim 7, characterized in that, The adaptive weighting includes: ； in: The number of channels in the feature map; The weight coefficient of the i-th channel; Feature map of the nth frame image The eigenvalues ​​in the i-th channel.

9. A low-cost autonomous inspection method for unmanned aerial vehicle (UAV) inspection of power distribution networks according to claim 1, characterized in that, The low-altitude re-inspection in S7 includes: Based on the verification data, the defective equipment was identified. Using the inspection height adjustment model, combined with the defect type and environmental factors, a low-altitude re-inspection was conducted to obtain re-inspection image data.

10. A low-cost autonomous inspection method for unmanned aerial vehicle (UAV) inspection of power distribution networks according to claim 9, characterized in that, The inspection height adjustment model includes: ； in: Optimal inspection height; Minimum permitted flight altitude; The size of the defect; : The measured value of the i-th environmental factor; : The weight coefficient of the i-th environmental factor; The current wind speed; , , : Adjust parameters.

Images