An unmanned aerial vehicle intelligent inspection system for power distribution network

Abstract

This invention discloses an intelligent unmanned aerial vehicle (UAV) inspection system for power distribution networks, comprising: an automatic UAV nesting device for docking, charging, and dispatching UAVs; a UAV equipped with image acquisition equipment to perform inspection tasks; an edge computing device integrated into the UAV nesting device, with a built-in AI edge computing module and a multi-UAV multi-task collaborative scheduling module; and a distribution intelligent inspection application module deployed on a power grid management platform. The edge computing device receives inspection images transmitted in real-time from the UAVs, calls a pre-trained distribution network defect and problem identification algorithm model for local real-time analysis, generates defect identification results, and transmits them back to the application module, achieving closed-loop management with power operation and maintenance work order data. This invention enables rapid and accurate identification of distribution network defects and construction-related external damage hazards, intelligent dispatching of multiple UAVs, and full-process data security control, significantly improving inspection efficiency and automation levels, reducing operation and maintenance costs, and ensuring the safe and stable operation of the distribution network.

Description

Technical Field

[0001] This invention relates to the technical field of power distribution network maintenance, specifically to an unmanned aerial vehicle (UAV) intelligent inspection system for power distribution networks. Background Technology

[0002] The distribution network is the final link in the power system. It receives electrical energy from the transmission network (or regional power plants), steps it down through distribution transformers, and then distributes it to industrial, commercial, and residential users via distribution lines and switching equipment. It is a crucial link connecting the generation, transmission, and consumption sides, directly determining the reliability, stability, and power quality of users' electricity supply.

[0003] With the rapid development of my country's social economy and the advancement of urbanization, the existing conventional testing and inspection methods for the daily inspection and maintenance of power distribution networks cannot meet the requirements of speed and efficiency. Furthermore, due to the presence of numerous construction sites within the line protection zone, which pose potential safety hazards to the lines, the existing inspection methods suffer from low inspection efficiency. Summary of the Invention

[0004] The present invention aims to solve the above-mentioned technical problems and provide an intelligent unmanned aerial vehicle (UAV) inspection system for power distribution networks, which can significantly improve inspection efficiency and automation level, reduce operation and maintenance costs, and ensure the safe and stable operation of the power distribution network.

[0005] The technical solution provided by this invention is, as one aspect of this invention, a drone-based intelligent inspection system for power distribution networks, comprising: Automated drone nesting equipment for docking, charging, and scheduling at least one drone; Unmanned aerial vehicles (UAVs), equipped with image acquisition devices, are used to perform power distribution network inspection tasks in response to dispatch instructions; An edge computing device is used to perform edge computing and generate unified scheduling instructions for the UAV according to the needs of regional power grid inspection tasks, and send them to the automated drone nesting device. The distribution intelligent inspection application module is deployed on the power grid management platform to acquire power operation and maintenance work order data, generate the inspection tasks, and receive and process the inspection results. The edge computing device is further configured to include: The AI ​​edge computing module is used to receive the inspection image data transmitted back in real time by the drone, and call the pre-trained distribution network defect and problem identification algorithm model to perform local real-time analysis on the inspection image data to generate defect identification results; and transmit the defect identification results back to the distribution intelligent inspection application module through the automatic drone nest device to form a closed-loop management by associating it with the power operation and maintenance work order data. The multi-machine, multi-task collaborative scheduling module is used to generate unified scheduling instructions for the UAV based on the regional power grid inspection task instructions, and send them to the automated drone nesting device.

[0006] Preferably, the AI ​​edge computing module adopts an embedded platform with an integrated AI acceleration chip and integrates a 5G communication module; the AI ​​edge computing module supports iterative updates of the recognition algorithm model remotely.

[0007] Preferably, the pre-trained power distribution network defect and problem identification algorithm model is trained through multi-source data fusion and enhancement technology; the multi-source data includes historical data of UAV inspection and video surveillance data of key locations, and the data enhancement technology includes at least one or more of the following: rotation, translation, scaling, random occlusion and MOSAIC stitching enhancement of inspection images, to improve the robustness of identifying small target defects.

[0008] Preferably, the distribution network defect and problem identification algorithm model includes a defect identification sub-model and an external damage identification sub-model; the defect identification sub-model is used to identify at least tower corrosion, insulator damage, conductor and ground wire breakage, hardware deformation, foreign object suspension, and vibration damper displacement; the external damage identification sub-model is used to identify at least construction machinery intrusion, illegal construction, and tree obstruction hazards.

[0009] Preferably, the multi-machine multi-task collaborative scheduling module is used to plan cruise paths for multiple UAVs based on road network data and airspace grid management strategies, and dynamically adjust task allocation and avoid flight conflicts during flight. It also includes a satellite positioning anomaly processing unit, which is used to fuse inertial navigation sensor data and preset landmark identification information to assist the UAVs in accurate positioning and navigation when GPS signals are abnormal.

[0010] Preferably, the multi-machine multi-task collaborative scheduling module supports simultaneous task scheduling and flight management for 10 or more UAVs, and triggers an emergency response mechanism to interrupt the task or assign backup equipment to take over when a UAV or nest equipment failure is detected.

[0011] Preferably, the system further includes an automatic data labeling module, deployed in the edge computing device or power grid management platform; The automatic data labeling module is used to acquire historical data of UAV inspections and video surveillance data of key locations, construct a training sample set using multi-source data augmentation technology, and automatically label the distribution network defects and problems in the training sample set to generate labeled data. The labeled data is used to train and iteratively optimize the power distribution network defect and problem identification algorithm model.

[0012] Preferably, the automated nesting device further includes: The helipad module is used to enable precise docking and physical fixation of the UAV; Automatic charging module, used to automatically replace batteries or wirelessly charge docked drones; The environmental monitoring and communication module is used to monitor the environment around the drone nest and to achieve bidirectional data communication with the drone and the power grid management platform through TCP / IP protocol and 5G network.

[0013] Preferably, the power distribution intelligent inspection application module is further configured to: obtain external damage prevention visas and defect management work orders from the power grid management platform, and automatically generate drone inspection tasks based on the work orders; receive defect identification results returned by the edge computing device, generate alarm information and trigger new handling work orders, thereby realizing business collaboration between inspection and handling.

[0014] Preferably, the power distribution intelligent inspection application module also has a reserved data interface for data interaction with the monitoring terminal, so as to realize the real-time feedback of on-site verification and handling results of inspection alarms.

[0015] Implementing the embodiments of the present invention has the following beneficial effects: This invention provides an intelligent unmanned aerial vehicle (UAV) inspection system for power distribution networks. By employing multi-source data augmentation (rotation, translation, scaling, random occlusion, MOSAIC stitching) and a dynamic switching training strategy based on the Loss threshold, it significantly improves the robustness and accuracy of identifying small target defects (such as insulator damage and vibration damper displacement). In this invention, by deploying an edge computing device integrating a computing chip and a 5G communication module at the end of the automatic machine nest, the AI ​​recognition algorithm is brought down to the side closer to the data source, realizing local real-time analysis of inspection images and second-level defect output, getting rid of dependence on cloud computing power, greatly shortening the response time, and improving the real-time performance and autonomous decision-making ability of distribution network inspection. In this invention, a multi-machine, multi-task collaborative scheduling module based on a genetic algorithm, combined with airspace grid partitioning and spatiotemporal isolation mechanisms, achieves unified scheduling and dynamic task allocation for UAVs within a region, with rapid scheduling calculations. Simultaneously, it integrates satellite positioning anomaly handling, visual navigation assistance, and health status monitoring, automatically triggering emergency responses in the event of GPS signal loss or equipment failure, ensuring the continuity and safety of inspection tasks. In this invention, a secure data transmission link is established through a dedicated mobile SIM card and an encrypted intranet communication protocol. Domestic cryptographic technology is used to encrypt the entire process of raw inspection data, identification results, and control commands. Furthermore, network boundary access control, intrusion detection and early warning, and a national cryptographic algorithm authentication mechanism are added to ensure the security of system data transmission and interaction. The automated drone nest device is compatible with existing electric drones and supports both battery replacement and wireless charging modes, reducing equipment modification costs. In this invention, the distribution intelligent inspection application module is seamlessly integrated with the defect management and external damage prevention visa management modules of the power grid management platform. By establishing the association between work orders and patrol tasks, it realizes the closed-loop management of the entire process of "data collection, defect identification, work order generation, and handling tracking", breaks down information silos, promotes data flow and business collaboration, and comprehensively improves the intelligence and precision of distribution network operation and maintenance. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, obtaining other drawings based on these drawings without creative effort still falls within the scope of the present invention. Figure 1 This is a schematic diagram of the structure of an embodiment of an unmanned aerial vehicle (UAV) intelligent inspection system for power distribution networks according to the present invention. Figure 2 for Figure 1 A schematic diagram of the structure of the automatic nesting equipment in the diagram; Figure 3 for Figure 1 A schematic diagram of the structure of a mid-edge computing device; Figure 4 This is a schematic diagram illustrating the steps of building and running the system provided by the present invention in one embodiment. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings.

[0018] like Figure 1 The diagram shown illustrates a structural schematic of an embodiment of an intelligent unmanned aerial vehicle (UAV) inspection system for power distribution networks provided by the present invention; in conjunction with... Figure 2 and Figure 3 As shown, in this embodiment, the unmanned aerial vehicle (UAV) intelligent inspection system for the power distribution network includes at least: Automated drone nesting equipment 1, used for docking, charging and scheduling at least one drone 2; Drone 2, equipped with image acquisition equipment, is used to perform power distribution network inspection tasks in response to dispatch instructions; Edge computing device 3 is used to perform edge computing and generate unified scheduling instructions for the UAV according to the needs of regional power grid inspection tasks, and send them to the automatic drone nest device 1; The distribution intelligent inspection application module 40 is deployed on the power grid management platform 4. It is used to acquire power operation and maintenance work order data, generate the inspection task, and receive and process the inspection results. The power grid management platform 4 is connected to a monitoring terminal 5, which can be such as an industrial control computer, a smart mobile device, or a smart tablet.

[0019] It is understood that a multi-agent collaborative mechanism is introduced in the system architecture of this invention. Multiple agents collaboratively process power inspection task instructions within the regional grid area and distribute them to the corresponding drones at the airport, thereby achieving intelligent task allocation and efficient execution.

[0020] like Figure 3 As shown, in a specific example, the edge computing device 3 is further configured to include: The automatic data annotation module 30 is used to acquire historical data of UAV inspections and video surveillance data of key locations, construct a training sample set using multi-source data augmentation technology, and automatically annotate the distribution network defects and problems in the training sample set to generate annotated data; the annotated data is used to train and iteratively optimize the distribution network defect and problem identification algorithm model. AI edge computing module 31 is used to receive the inspection image data transmitted back in real time by the UAV, and call the pre-trained distribution network defect and problem identification algorithm model to perform local real-time analysis on the inspection image data to generate defect identification results; and transmit the defect identification results back to the distribution intelligent inspection application module through the automatic drone nest device to form a closed-loop management by associating the power operation and maintenance work order data. The multi-machine multi-task collaborative scheduling module 32 is used to generate a unified scheduling instruction for the UAV based on the regional power grid inspection task instruction, and send it to the automatic drone nest device.

[0021] It is understood that, in other examples, the automatic data labeling module 30 may also be deployed in the power grid management platform 4.

[0022] Furthermore, in a specific example, the AI ​​edge computing module 31 adopts an embedded platform with an integrated AI acceleration chip and integrates a 5G communication module; the AI ​​edge computing module supports iterative updates of the recognition algorithm model remotely.

[0023] The pre-trained power distribution network defect and problem identification algorithm model is trained through multi-source data fusion and enhancement techniques. The multi-source data includes historical data from UAV inspections and video surveillance data from key locations. The data enhancement techniques include at least one or more of the following: rotation, translation, scaling, random occlusion, and MOSAIC stitching enhancement of the inspection images, used to improve the robustness of identifying small target defects.

[0024] Preferably, the distribution network defect and problem identification algorithm model may further include a defect identification sub-model and an external damage identification sub-model; the defect identification sub-model is used at least to identify tower corrosion, insulator damage, conductor and ground wire breakage, hardware deformation, foreign object suspension, and vibration damper displacement; the external damage identification sub-model is used at least to identify construction machinery intrusion, illegal construction, and tree obstruction hazards.

[0025] Furthermore, the multi-machine multi-task collaborative scheduling module 32 is used to plan cruise paths for multiple UAVs based on road network data and airspace grid management strategies, and dynamically adjust task allocation and avoid flight conflicts during flight. It also includes a satellite positioning anomaly processing unit, which is used to fuse inertial navigation sensor data and preset landmark identification information to assist the UAVs in accurate positioning and navigation when GPS signals are abnormal.

[0026] In some examples, the multi-machine multi-task collaborative scheduling module supports simultaneous task scheduling and flight management for 10 or more UAVs, and triggers an emergency response mechanism to interrupt the task or assign backup equipment to take over when a UAV or nest equipment failure is detected.

[0027] In some examples, such as Figure 2 As shown, the automated nesting device 1 further includes: The helipad module 10 is used to achieve precise docking and physical fixation of the UAV; Automatic charging module 11 is used to automatically replace batteries or wirelessly charge docked drones. The environmental monitoring and communication module 12 is used to monitor the environment around the drone nest and to achieve bidirectional data communication with the drone and the power grid management platform through TCP / IP protocol and 5G network.

[0028] In a specific example, the power distribution intelligent inspection application module 40 is further configured to: obtain external damage prevention visas and defect management work orders from the power grid management platform, and automatically generate drone inspection tasks based on the work orders; receive defect identification results returned by the edge computing device, generate alarm information and trigger new handling work orders, thereby realizing business collaboration between inspection and handling.

[0029] It is understandable that the power distribution intelligent inspection application module 40 also has a reserved data interface for data interaction with the monitoring terminal, so as to realize the real-time feedback of on-site verification and handling results of inspection alarms.

[0030] The invention provides an intelligent drone inspection system for power distribution networks. In practical applications, it acquires work order data from modules such as external damage prevention and defect management on the power grid management platform. The intelligent distribution inspection application module creates drone inspection tasks. The drone nesting device schedules and dispatches drones using a multi-drone, multi-task collaborative scheduling algorithm. The drone executes the task and inputs photo data into the edge computing module. After AI recognition, the results are transmitted back to the drone nesting device, which then transmits the data back to the intelligent distribution inspection application module. This application module reserves space for the power grid management module's basic module and for extracting application data from mobile devices. This system aims to build an efficient and intelligent distribution inspection system by integrating drone inspection, video surveillance of key locations, AI edge computing, and 5G communication technologies. The system can automatically identify and label distribution network defects and problems, achieve real-time video transmission and edge computing processing, significantly improve inspection efficiency and accuracy, reduce operation and maintenance costs, and ensure the safe and stable operation of the power grid. Through refined task decomposition and collaborative advancement, an intelligent distribution inspection system integrating drone inspection, AI edge computing, and 5G communication has been constructed. The various tasks are closely interconnected and mutually supportive, working together to achieve the overall project goals. In the future, with continuous technological upgrades, the system will further improve inspection efficiency and accuracy, providing strong support for the safe and stable operation of the power grid.

[0031] To further understand the system provided by this invention, the following specific example illustrates the specific construction and operation process of the system involved in this invention.

[0032] Please refer to Figure 4 As shown, in this embodiment, the construction and operation of the entire system involves the following key technical steps: Step S1: Deploy Automatic Labeling Technology and Algorithm Training: Research and deploy automatic labeling technology for distribution network defects and problems that combines UAV inspection data and video surveillance data from key locations, and train the recognition algorithm. Specifically, Step S1 aims to overcome the bottleneck of multi-source fusion of "UAV inspection data + video surveillance data from key locations," construct a high-precision and high-efficiency automatic labeling system for distribution network defects, train a defect recognition model adapted to complex scenarios, and achieve automatic detection, classification, and location of more than 10 typical defects, such as insulator spontaneous explosion, conductor strand breakage, hardware corrosion, and tree obstruction hazards.

[0033] During training, the training methods were optimized to improve the model's adaptability, and multi-source data augmentation methods were used to increase data complexity. Specifically, this included rotating (0°-360° adjustable), translating (±50 pixels), scaling (0.5-1.5 times), and randomly occluding (occlusion area ≤10%) the inspection images. Simultaneously, MOSAIC enhancement technology was used, where multiple images were randomly scaled, cropped, and stitched together for model training. When the training loss value from small target defects was insufficient, the stitched image was prioritized for training in the next iteration to enhance the robustness of small target recognition. The labeled data was used for subsequent training and iterative optimization of the recognition algorithm model.

[0034] Step S2: Deploy intelligent inspection technology based on edge computing at the distribution network: Step S2 involves researching and deploying four sets of distribution network-specific defect identification and external damage identification algorithm models based on the effective data provided in S1. Specifically, this involves constructing an edge computing hardware architecture and software system at the distribution network to achieve local real-time processing of inspection data, rapid defect identification, anomaly warning, and local decision-making.

[0035] To address this, an edge computing device was designed and deployed, importing drone inspection video and image information into the algorithm developed and deployed in step S1, and integrating the algorithm into the power distribution network edge computing device. Its research and deployment objectives can be divided into the integration of 5G networks and AI edge computing, the application of real-time video transmission technology, the hardware development and deployment of the edge computing box, and algorithm optimization and update mechanisms. The device utilizes the RK3588 chip and supporting products, integrating a 5G communication module and a dedicated power network card. It imports drone inspection video and image information into the algorithm developed and deployed in step S1 in real time, enabling on-site analysis and processing of inspection data through edge computing, reducing data backhaul volume. Based on the effective labeled data provided in step S1, four sets of distribution network-specific defect identification algorithm models and external damage identification algorithm models were researched and deployed to complete the system. The defect identification algorithm models target six common defects: tower corrosion, insulator damage, conductor / ground wire breakage, hardware deformation, foreign object suspension, and vibration damper displacement. The external damage identification algorithm models target three common external damage hazards: construction machinery intrusion, illegal construction, and excessive tree obstruction. The model recognition accuracy is ≥90%, and the output time for a single image recognition result is ≤1 second. This eliminates reliance on cloud computing power, shortening the defect identification response time to the second level and improving the real-time performance and autonomy of distribution network inspections. The edge computing device supports remote iterative updates of the algorithm, utilizing a multimodal large model to reduce the initial training load, and subsequently optimizing the model through updates to the recognition material library.

[0036] Step S3: Deploy unmanned automated drone nesting equipment to empower existing power grid drones: Research and deploy unmanned automated drone nesting equipment to empower existing power grid drones. Step S3 involves developing and deploying a fully automated, highly available automated drone nesting device. This device aims to be compatible with existing power grid equipment and integrates research and deployment results on hardware and software, including automatic labeling technology for distribution network defects and problems combining drone inspection data and video surveillance data from key locations, and intelligent inspection technology based on edge computing at the drone nesting end. Specifically, this involves overcoming the compatibility bottleneck between existing power grid drones and automated drone nesting systems. Through hardware modification, functional development, and software integration, it enables unmanned functions such as automatic take-off and landing, battery replacement, data transmission, and status monitoring for existing drones.

[0037] Step S3's research and deployment objectives include four parts: ① Nest design and technology verification: achieving low cost and easy maintenance by streamlining unnecessary functions and hardware, and verifying the nest's structural strength, stability, and environmental adaptability; ② Intelligent empowerment design for drones: completing the technology verification of drone docking and positioning, intelligent wake-up, and rapid recharging, ensuring stable docking and efficient recharging of drones within the nest; ③ Communication and data transmission between the nest and drones: using TCP / IP protocol and 5G network to achieve efficient communication between the two, ensuring stable real-time video transmission, inspection image feedback, and other functions; ④ Integration and application of edge computing devices: integrating edge computing devices at the nest end to achieve real-time analysis and processing of inspection data and feedback of problem information. The devices possess high-performance computing capabilities and low power consumption characteristics, improving the autonomy and continuity of drone inspections and reducing the cost of manual intervention.

[0038] Step S4: Deploy a multi-machine, multi-task collaborative scheduling algorithm: Research and deploy a multi-machine, multi-task collaborative scheduling algorithm to achieve unified scheduling of UAVs. Step S4 involves developing a multi-machine, multi-task collaborative scheduling algorithm to achieve unified scheduling and automatic task assignment for all power grid inspection UAVs within the region. Specifically, this involves developing a multi-machine, multi-task collaborative scheduling algorithm adapted to power grid inspection scenarios, constructing a unified UAV scheduling system, and realizing dynamic task allocation, path planning, conflict avoidance, and resource optimization for multiple UAVs.

[0039] Step S4's research and deployment objectives include five parts: ① Spatiotemporal isolation and emergency handling: integrating a spatiotemporal isolation mechanism to avoid interference from UAV missions and establishing a fault emergency response mechanism to ensure inspection continuity; ② Data fusion and analysis: fusing and processing UAV-transmitted data, extracting key information, and generating inspection reports; ③ Multi-aircraft, multi-task scheduling based on road networks: incorporating route data and surface data into the road network to achieve gridded airspace segmentation and highly refined management, and setting safe distances for UAVs; ④ Navigation decision algorithm: planning flight paths based on sensor data, environmental maps, and mission requirements, and addressing obstacles and flight restrictions; ⑤ Satellite positioning anomaly handling algorithm: combining inertial navigation sensors and landmark recognition technology to ensure accurate positioning and navigation of UAVs when GPS signals are abnormal, improving the efficiency and safety of multi-UAV collaborative inspections, and supporting simultaneous scheduling of ≤10 UAVs.

[0040] Step S5: Deploy the integrated application of the intelligent distribution network inspection module: Research and deploy the integrated application of the intelligent distribution network inspection module, integrate power maintenance work order data with UAV data, and complete the demonstration of the results. Step S5 is used to build and integrate the intelligent distribution network inspection module on the power grid management platform. Specifically, it integrates the technical achievements of S1-S4, develops an integrated application module for intelligent distribution network inspection, realizes deep integration of UAV inspection data and power maintenance work order data, builds a full-process inspection system of "data collection, defect identification, work order generation, handling tracking - closed-loop management", and completes the demonstration of the results in typical distribution network areas to improve the efficiency of distribution network inspection and maintenance.

[0041] Based on the above steps, the system comprises a proprietary intelligent identification algorithm model for distribution network equipment defects and common construction-related external damage issues, software application modules, automatic machine nesting equipment, and a multi-machine, multi-task collaborative scheduling algorithm. By establishing the relationship between work orders and patrol tasks, it enables practical applications such as patrol task issuance, patrol data management, defect and external damage area identification, and alarm information management. It also supports seamless information integration and sharing for applications such as power grid terminal applications, power grid management platform anti-external damage certification, and defect management, breaking down information silos, promoting data flow and application collaboration, enhancing comprehensive monitoring and emergency response capabilities of the distribution network status, and ultimately driving the intelligent upgrade and significant improvement of comprehensive efficiency in distribution network operation and maintenance management, thus completing integrated application.

[0042] Implementing the embodiments of the present invention has the following beneficial effects: This invention provides an intelligent unmanned aerial vehicle (UAV) inspection system for power distribution networks. By employing multi-source data augmentation (rotation, translation, scaling, random occlusion, MOSAIC stitching) and a dynamic switching training strategy based on the Loss threshold, it significantly improves the robustness and accuracy of identifying small target defects (such as insulator damage and vibration damper displacement). In this invention, by deploying an edge computing device integrating a computing chip and a 5G communication module at the end of the automatic machine nest, the AI ​​recognition algorithm is brought down to the side closer to the data source, realizing local real-time analysis of inspection images and second-level defect output, getting rid of dependence on cloud computing power, greatly shortening the response time, and improving the real-time performance and autonomous decision-making ability of distribution network inspection. In this invention, a multi-machine, multi-task collaborative scheduling module based on a genetic algorithm, combined with airspace grid partitioning and spatiotemporal isolation mechanisms, achieves unified scheduling and dynamic task allocation for UAVs within a region, with rapid scheduling calculations. Simultaneously, it integrates satellite positioning anomaly handling, visual navigation assistance, and health status monitoring, automatically triggering emergency responses in the event of GPS signal loss or equipment failure, ensuring the continuity and safety of inspection tasks. In this invention, a secure data transmission link is established through a dedicated mobile SIM card and an encrypted intranet communication protocol. Domestic cryptographic technology is used to encrypt the entire process of raw inspection data, identification results, and control commands. Furthermore, network boundary access control, intrusion detection and early warning, and a national cryptographic algorithm authentication mechanism are added to ensure the security of system data transmission and interaction. The automated drone nest device is compatible with existing electric drones and supports both battery replacement and wireless charging modes, reducing equipment modification costs. In this invention, the distribution intelligent inspection application module is seamlessly integrated with the defect management and external damage prevention visa management modules of the power grid management platform. By establishing the association between work orders and patrol tasks, it realizes the closed-loop management of the entire process of "data collection, defect identification, work order generation, and handling tracking", breaks down information silos, promotes data flow and business collaboration, and comprehensively improves the intelligence and precision of distribution network operation and maintenance.

[0043] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 Units that specify functions within one or more boxes.

[0044] The above description is merely a preferred embodiment of the present invention and should not be construed as limiting the scope of the invention. Therefore, any equivalent variations made in accordance with the claims of the present invention are still within the scope of the present invention.

Claims

1. A drone-based intelligent inspection system for power distribution networks, characterized in that, include: Automated drone nesting equipment for docking, charging, and scheduling at least one drone; Unmanned aerial vehicles (UAVs), equipped with image acquisition devices, are used to perform power distribution network inspection tasks in response to dispatch instructions; An edge computing device is used to perform edge computing and generate unified scheduling instructions for the UAV according to the needs of regional power grid inspection tasks, and send them to the automated drone nesting device. The distribution intelligent inspection application module is deployed on the power grid management platform to acquire power operation and maintenance work order data, generate the inspection tasks, and receive and process the inspection results. The edge computing device is further configured to include: The AI ​​edge computing module is used to receive the inspection image data transmitted back in real time by the drone, and call the pre-trained distribution network defect and problem identification algorithm model to perform local real-time analysis on the inspection image data to generate defect identification results; and transmit the defect identification results back to the distribution intelligent inspection application module through the automatic drone nest device to form a closed-loop management by associating it with the power operation and maintenance work order data. The multi-machine, multi-task collaborative scheduling module is used to generate unified scheduling instructions for the UAV based on the regional power grid inspection task instructions, and send them to the automated drone nesting device.

2. The system according to claim 1, characterized in that, The AI ​​edge computing module adopts an embedded platform with an integrated AI acceleration chip and a 5G communication module; the AI ​​edge computing module supports iterative updates of the recognition algorithm model remotely.

3. The system according to claim 2, characterized in that, The pre-trained power distribution network defect and problem identification algorithm model is trained through multi-source data fusion and enhancement technology; the multi-source data includes historical data of UAV inspection and video surveillance data of key locations, and the data enhancement technology includes at least one or more of the following: rotation, translation, scaling, random occlusion and MOSAIC stitching enhancement of inspection images.

4. The system according to claim 3, characterized in that, The power distribution network defect and problem identification algorithm model includes a defect identification sub-model and an external damage identification sub-model. The defect identification sub-model is used to identify at least tower corrosion, insulator damage, conductor and ground wire breakage, hardware deformation, foreign object suspension, and vibration damper displacement. The external damage identification sub-model is used to identify at least construction machinery intrusion, illegal construction, and tree obstruction hazards.

5. The system according to claim 4, characterized in that, The multi-machine, multi-task collaborative scheduling module is used to plan cruise paths for multiple UAVs based on road network data and airspace grid management strategies, and dynamically adjust task allocation and avoid flight conflicts during flight. It also includes a satellite positioning anomaly processing unit, which is used to fuse inertial navigation sensor data and preset landmark identification information to assist the UAVs in accurate positioning and navigation when GPS signals are abnormal.

6. The system according to claim 5, characterized in that, The multi-machine multi-task collaborative scheduling module supports simultaneous task scheduling and flight management for 10 or more UAVs, and triggers an emergency response mechanism to interrupt the task or assign backup equipment to take over when a UAV or nest equipment failure is detected.

7. The system according to any one of claims 1 to 6, characterized in that, The system also includes an automatic data labeling module, which is deployed on the edge computing device or power grid management platform; The automatic data labeling module is used to acquire historical data of UAV inspections and video surveillance data of key locations, construct a training sample set using multi-source data augmentation technology, and automatically label the distribution network defects and problems in the training sample set to generate labeled data. The labeled data is used to train and iteratively optimize the power distribution network defect and problem identification algorithm model.

8. The system according to claim 7, characterized in that, The automated nesting equipment also includes: The helipad module is used to enable precise docking and physical fixation of the UAV; Automatic charging module, used to automatically replace batteries or wirelessly charge docked drones; The environmental monitoring and communication module is used to monitor the environment around the drone nest and to achieve bidirectional data communication with the drone and the power grid management platform through TCP / IP protocol and 5G network.

9. The system according to claim 8, characterized in that, The power distribution intelligent inspection application module is further configured to: obtain external damage prevention certificates and defect management work orders from the power grid management platform, and automatically generate drone inspection tasks based on the work orders; The system receives the defect identification results returned by the edge computing device, generates alarm information, and triggers a new handling work order to achieve business collaboration between inspection and handling.

10. The system according to claim 9, characterized in that, The power distribution intelligent inspection application module also has a reserved data interface for data interaction with the monitoring terminal, so as to realize the real-time feedback of on-site verification and handling results of inspection alarms.

Images