Unmanned aerial vehicle defect hidden danger autonomous identification closed-loop management method and device, and storage medium

Abstract

The invention provides an unmanned aerial vehicle defect hidden danger autonomous identification closed-loop management method and device, and a storage medium, and the method comprises the steps of unmanned aerial vehicle on-site patrol operation, manual or intelligent defect hidden danger registration, unmanned aerial vehicle patrol result returning, defect hidden danger manual confirmation and verification, defect work order distribution and defect elimination processing, and defect processing state updating. Whole-process online circulation of distribution network defects and hidden dangers is realized, whole-process management such as automatic identification, recording, classification, grading, processing and updating of the defects and hidden dangers is realized, inspection efficiency and quality are improved, and safe and stable operation of a power grid is guaranteed; the scheme comprises the following steps: step 1, compiling a patrol task, and distributing a patrol work order; step 2, unmanned aerial vehicle on-site patrol operation; step 3, manually or intelligently registering defect hidden dangers; 4, the unmanned aerial vehicle inspection result is transmitted back to the unmanned aerial vehicle inspection platform; step 5, manually confirming defects and hidden dangers, and auditing inspection results; step 6, pushing the defect work order to the PMS 3.0 system; and step 7, updating a defect processing state.

Description

Technical Field

[0001] The present invention relates to the field of drone recording technology, and in particular to a closed-loop management method for autonomously identifying defects and hidden dangers of drones. Background Art

[0002] Traditional methods of finding hidden dangers in distribution lines mainly rely on manual inspections, electrical testing equipment detection, and regular maintenance. Although these methods can detect and deal with defects and hidden dangers to a certain extent, they have many limitations. For example, manual inspections are time-consuming and labor-intensive, and are limited by the work ability and experience of operation and maintenance personnel, resulting in uneven quality and efficiency in the discovery of defects and hidden dangers. At the same time, although traditional electrical testing equipment can detect some electrical problems, it is often difficult to accurately judge complex defects and hidden dangers, such as equipment aging, overheating, and other phenomena.

[0003] Grassroots personnel need to promptly discover defects and hidden dangers in the distribution network equipment itself and the surrounding environment that affect the safe, economic and high-quality operation of the distribution network, and make unified records (including the location, position, discovery time, defect description, manufacturer and model of defective equipment, level, planned processing time, maintenance time, processing status, acceptance opinions, etc.), and promptly determine the maintenance strategy according to the classification and grading requirements, carry out defect elimination work, and conduct closed-loop management to ensure the safe and stable operation of the power grid. At present, there are three problems in the identification and management of defects and hidden dangers:

[0004] (I) Lack of defect and hidden danger identification strategy. At present, the monitoring and recording of defects and hidden dangers still rely on manual discovery during on-site operations or inspections. Intelligent operation and inspection equipment (such as distribution automation terminals such as FTU, DTU, TTU and other distribution secondary equipment, patrol robots, and online monitoring equipment) are not effectively utilized. There is a lack of intelligent means to use the collected power grid equipment monitoring data to automatically analyze and generate defects and hidden dangers. In addition, the classification and judgment of defects and hidden dangers depends on the work ability and work experience of the operation and maintenance personnel, resulting in uneven quality and efficiency of defect and hidden danger discovery, and a lack of intelligent analysis and decision-making means.

[0005] (II) There is a lack of intelligent auxiliary tools for defect elimination plans, and there is no convenient means to implement the classification of defect elimination methods. Simple defects do not require the preparation of defect elimination plans. For example, the defect of missing pole number plates can be solved by the defect discoverer on his own without the need to prepare plans or schedules. For complex defects, grassroots managers generally formulate feasible construction plans by referring to various information such as on-site defect hidden danger information, defect hidden danger information of the same line or nearby equipment, line topology relationship (single-line diagram, system diagram), material storage status, on-site working conditions, etc., and no effective intelligent auxiliary means are provided.

[0006] (III) Defect elimination acceptance is generally entered manually, and the quality of defect elimination cannot be verified online. At the same time, the value of the data such as equipment quality problems, supplier problems, and operation and maintenance quality problems reflected in the defect and hidden danger record data has not been effectively mined.

[0007] In the power industry, the identification, handling and closed-loop management of defects and hidden dangers are crucial. Traditional closed-loop management mainly relies on manual recording and processing, which is not only inefficient but also prone to errors. With the development of intelligent technology, it is particularly important to develop a method that can automatically identify defects and hidden dangers and conduct closed-loop management. This method can realize the full process management of automatic identification, recording, classification, grading, handling and acceptance of defects and hidden dangers, ensuring the safe and stable operation of the power grid.

[0008] Based on the above background and challenges, it is of great significance to develop a method for autonomous identification and closed-loop management of defects and hidden dangers of low-voltage distribution networks using drones. This method can make full use of the advantages of drone technology to realize automatic identification and processing of defects and hidden dangers, and improve inspection efficiency and quality; at the same time, through a closed-loop management mechanism, it can ensure that defects and hidden dangers are handled and accepted in a timely manner, ensuring the safe and stable operation of the power grid. Summary of the invention

[0009] The present invention aims to address the technical defects of the prior art and provide a closed-loop management method, equipment, and storage medium for autonomous identification of defects and hidden dangers by drones. Through the steps of on-site patrol operations of drones, manual or intelligent registration of defects and hidden dangers, feedback of drone patrol results, manual confirmation and review of defects and hidden dangers, issuance and elimination of defect work orders, and update of defect processing status, the entire process of distribution network defects and hidden dangers can be transferred online, and the entire process management of automatic identification, recording, classification, grading, processing, and updating of defects and hidden dangers can be realized, thereby improving the inspection efficiency and quality and ensuring the safe and stable operation of the power grid.

[0010] The closed-loop management method for autonomously identifying hidden dangers of drone defects of the present invention comprises the following steps:

[0011] Step 1: Prepare inspection tasks and issue inspection work orders;

[0012] Step 2: UAV on-site inspection operation;

[0013] Step 3: Manual or intelligent registration of defects and hidden dangers;

[0014] Step 4: The drone inspection results are transmitted back to the drone inspection platform;

[0015] Step 5: Manual confirmation of defects and hidden dangers and review of inspection results;

[0016] Step 6: Push the defect work order to the PMS3.0 system. PMS3.0 generates the corresponding defect work order, distributes the work order, closes the work order after the defect is eliminated, and returns the information to the drone inspection platform;

[0017] Step 7: The drone inspection platform updates the defect processing status.

[0018] Preferably, the preparation of the inspection task in step 1 includes the preparation of line details, and the line details include voltage level and line name. When the inspection is completed, the collection creation time and collection information statistics are automatically generated.

[0019] Preferably, the drone on-site patrol operation in step 2 autonomously completes the flight and takes photos according to the route mission preset for the current patrol mission.

[0020] Preferably, the drone uses Beidou positioning and RTK modules to obtain the coordinates of the drone's photography position and complete visible light photography of different parts and angles of the device;

[0021] After the coordinates are acquired, the coordinates of the photo and the actual tower are matched and associated;

[0022] The coordinate matching association adopts the following steps: after the drone takes a photo, the drone system is set to start matching the pole towers within a 10-meter radius of the drone coordinates. If only one pole tower is matched, the photo is associated with the pole tower. If multiple pole towers are matched, all the matched pole towers are listed and sorted from near to far by coordinate distance. The patrol pole tower associated with the photo is manually selected. If no pole tower is matched, all the pole towers under the current task are displayed, with the unpatrolled ones displayed in front and the patrolled ones displayed in the back. The ones with the closest coordinates are manually selected for association.

[0023] Preferably, in step 4, the drone inspection platform uses a defect recognition model to perform defect recognition on the drone inspection results;

[0024] The defect recognition model includes a data preprocessing module: preprocessing the images collected by the drone, including denoising, contrast enhancement, distortion correction, and cropping and scaling the images to meet the input requirements of the model;

[0025] Feature extraction module: Convolutional neural network (CNN) is used to extract features from preprocessed images;

[0026] Classifier module: Based on the extracted features, support vector machine (SVM) is used to classify images.

[0027] The classifier classifies the image into normal or defective categories;

[0028] Post-processing module: post-processes the output of the classifier, marks the defect location, and returns the defect type code to the drone inspection platform.

[0029] Preferably, the defects and hidden dangers manually registered or intelligently identified during the drone inspection process are summarized and displayed on the inspection task page of the drone inspection platform based on a map of the power grid, including inspection progress, drone flight trajectory, inspection task information, and inspection equipment details. Defects and hidden dangers can be confirmed, modified, and reviewed manually.

[0030] Preferably, the route mission is generated by having a drone fly over the area of ​​the equipment to be inspected in advance along a preset path. During the flight, the drone will collect relevant geographic location, equipment layout, and possible obstacle information data. Based on the collected data, the system will generate and save a route file.

[0031] Preferably, the push of the defect work order in step 6 calls a standard interface provided by the business center.

[0032] A computer device comprises: a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the program, a closed-loop management method for autonomously identifying hidden dangers of defects in unmanned aerial vehicles is implemented.

[0033] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements a closed-loop management method for autonomously identifying defects and hidden dangers of unmanned aerial vehicles.

[0034] Compared with the prior art, the present invention has the following beneficial effects:

[0035] The present invention provides a closed-loop management method for autonomous identification of defects and hidden dangers by drones, which is an efficient and intelligent distribution network operation and maintenance management method, realizing the online circulation of the entire process of distribution network defects and hidden dangers, making full use of the advantages of drone technology, and realizing automatic identification and processing of defects and hidden dangers, mainly including the following steps:

[0036] 1. Prepare patrol tasks and issue patrol work orders. At this stage, the operation and maintenance personnel will prepare patrol tasks in the drone inspection platform according to the actual operation status and patrol cycle of the distribution network. After the preparation is completed, the patrol work order will be issued to the designated drone and operator to ensure that the drone can perform patrol operations according to the predetermined plan;

[0037] 2. Drone on-site inspection operations

[0038] After receiving the inspection work order, the drone will conduct on-site inspections according to the predetermined route tasks and actions. During the inspection process, the drone will use the high-definition camera, infrared thermal imager and other sensor equipment to conduct a full range of monitoring and recording of the distribution line and its surrounding environment. These monitoring data will be transmitted back to the drone inspection platform in real time for subsequent analysis and processing;

[0039] 3. Hidden dangers of defects in manual or intelligent registration

[0040] After the drone inspection operation is completed, the operation and maintenance personnel need to conduct a preliminary analysis of the monitoring data sent back by the drone. For obvious defects and hidden dangers, such as equipment damage and overheating, the operation and maintenance personnel can directly register them on the drone inspection platform. At the same time, the drone inspection platform can also use intelligent algorithms to automatically analyze the monitoring data, identify potential defects and hidden dangers, and automatically register them in the system;

[0041] 4. UAV inspection results are transmitted back to the UAV inspection platform

[0042] After the drone inspection operation is completed, all monitoring data and identified defects and hidden dangers will be transmitted back to the drone inspection platform;

[0043] 5. Manual confirmation of defects and hidden dangers, and review of inspection results

[0044] Operation and maintenance personnel need to manually confirm and review the defects and hidden dangers on the drone inspection platform. According to the actual situation on site and monitoring data, they will make a preliminary judgment on the level, type and treatment method of the defects and hidden dangers. At the same time, they will also review the inspection results of the drone to ensure the accuracy and completeness of the inspection data.

[0045] 6. Push defect work orders to the PMS3.0 system for work order distribution, defect elimination and closure

[0046] After manual confirmation and review, the drone inspection platform will push the defect hidden danger information to the PMS3.0 system. The PMS3.0 system will automatically generate the corresponding defect work order according to the level and type of the defect hidden danger, and distribute it to the relevant operation and maintenance personnel for processing. It is driven by the business work order and autonomously distributed to the corresponding management person in charge, and notified to the corresponding responsible person in the form of mobile phone text messages or i State Grid, realizing the "prompt-supervision-closed loop" and other similar defects to merge similar items. After receiving the work order, the operation and maintenance personnel will carry out defect elimination according to the predetermined plan and plan. After processing, the work order will be closed in the PMS3.0 system and the processing results will be returned to the drone inspection platform;

[0047] 7. Update defect handling status

[0048] After receiving the defect handling results returned by the PMS3.0 system, the drone inspection platform will update the handling status of the defect hidden dangers;

[0049] To summarize, the closed-loop management method for autonomous identification of defects and hidden dangers by drones achieves efficient and intelligent management and processing of defects and hidden dangers in the distribution network through steps such as on-site drone inspection operations, manual or intelligent registration of defects and hidden dangers, feedback of drone inspection results, manual confirmation and review of defects and hidden dangers, issuance and elimination of defect work orders, and update of defect processing status. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 is a flow chart of the present invention;

[0051] Figure 2 It is the inspection and patrol task interface for drones;

[0052] Figure 3 It is the device inspection interface;

[0053] Figure 4 This is the patrol task details interface;

[0054] Figure 5 It is the inspection equipment details interface;

[0055] Figure 6 Defect recognition model recognition interface. DETAILED DESCRIPTION

[0056] 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. All other embodiments obtained by ordinary technicians in the field without making creative work based on the embodiments of the present invention shall fall within the scope of protection of the present invention.

[0057] like Figures 1 to 6 As shown, it shows a specific embodiment of the present invention:

[0058] The closed-loop management method for autonomously identifying hidden dangers of drone defects in the technical solution of the present invention comprises the following steps:

[0059] Step 1: Prepare inspection tasks and issue inspection work orders;

[0060] Step 2: UAV on-site inspection operation;

[0061] Step 3: Manual or intelligent registration of defects and hidden dangers;

[0062] Step 4: The drone inspection results are transmitted back to the drone inspection platform;

[0063] Step 5: Manual confirmation of defects and hidden dangers and review of inspection results;

[0064] Step 6: Push the defect work order to the PMS3.0 system. PMS3.0 generates the corresponding defect work order, distributes the work order, closes the work order after the defect is eliminated, and returns the information to the drone inspection platform;

[0065] Step 7: The drone inspection platform updates the defect processing status.

[0066] Preferably, Figure 2 As shown, the preparation of the inspection task in step 1 includes the preparation of line details, and the line details include voltage level and line name. After the inspection is completed, the collection creation time and collection information statistics are automatically generated.

[0067] In the process of preparing inspection tasks, the preparation of line details is the basis and premise of the inspection tasks. Only when the key information such as the voltage level and line name of the line is clear, can the inspection plan and program be formulated. At the same time, the completion of the inspection task will in turn affect the preparation of line details. Through the execution of the inspection task and the statistics of the collected information, we can continuously improve and update the line details to make them more accurate, comprehensive and practical.

[0068] In summary, the compilation of line details and the automatic generation of collection creation time and collection information statistics after the inspection are important links in the inspection task. Together, they constitute the basis and support system of the inspection task, providing a strong guarantee for the safe operation and equipment management of the line.

[0069] Preferably, the drone on-site patrol operation in step 2 autonomously completes the flight and takes photos according to the route mission preset for the current patrol mission.

[0070] In step 2 of the closed-loop management method for autonomous identification of defects and hidden dangers of drones, the drone on-site inspection operation is to autonomously complete the key links of flight and photography according to the route mission preset by the current inspection mission. After receiving the take-off command, the drone will automatically take off according to the preset route mission. The drone will use the RTK module for navigation to ensure the accuracy and stability of the flight. During the flight, the drone will take photos and videos of the distribution line and its surrounding environment according to the preset parameters such as the photography interval and ground resolution. These photos and videos will be transmitted back to the drone inspection platform in real time for subsequent analysis and processing.

[0071] If an emergency occurs during flight (such as weather changes, signal loss, etc.), the drone can suspend the mission and wait for recovery. Once recovered, the drone can continue to fly and take photos according to the preset route mission, realizing breakpoint flight. At the same time, the inspection personnel can monitor the flight status and photo data of the drone in real time through the drone inspection platform to ensure the safety and effectiveness of the inspection operation.

[0072] In summary, the process of drone on-site inspection, which autonomously completes flight and photography according to the preset route mission of the current inspection mission, is a highly automated and intelligent process. It can not only improve the efficiency and accuracy of inspection operations, but also reduce the workload and risks of operation and maintenance personnel.

[0073] Preferably, the drone uses Beidou positioning and RTK modules to obtain the coordinates of the drone's photography position and complete visible light photography of different parts and angles of the device;

[0074] After the coordinates are acquired, the coordinates of the photo and the actual tower are matched and associated;

[0075] The coordinate matching association adopts the following steps: after the drone takes a photo, the drone system is set to start matching the pole towers within a 10-meter radius of the drone coordinates. If only one pole tower is matched, the photo is associated with the pole tower. If multiple pole towers are matched, all the matched pole towers are listed and sorted from near to far by coordinate distance. The patrol pole tower associated with the photo is manually selected. If no pole tower is matched, all the pole towers under the current task are displayed, with the unpatrolled ones displayed in front and the patrolled ones displayed in the back. The ones with the closest coordinates are manually selected for association.

[0076] The drone uses Beidou positioning and RTK modules to obtain the coordinates of the photo location, and completes visible light photography of different parts and angles of the equipment, as well as a detailed description of the subsequent matching and association process between the photos and the actual tower coordinates:

[0077] 1. Drone photography and coordinate acquisition

[0078] Beidou positioning and RTK module:

[0079] The drone is equipped with the Beidou satellite navigation system and RTK (real-time dynamic differential positioning) module. These high-precision positioning devices can ensure that the drone obtains its precise position coordinates in real time during flight. The Beidou system provides basic navigation and positioning services, while the RTK module further corrects the positioning error by receiving the differential signal sent by the ground base station, achieving centimeter-level positioning accuracy.

[0080] Visible light photography:

[0081] During the flight, the drone takes visible light photos of the distribution lines and their surroundings according to the preset route mission and photography parameters (such as photography interval, ground resolution, exposure time, etc.). When taking photos, the drone automatically adjusts the camera angle and focal length to ensure that images of different parts and angles of the equipment can be clearly captured.

[0082] 2. Matching and associating photos with actual tower coordinates

[0083] Coordinate acquisition and storage:

[0084] After each photo is taken, the drone will store the coordinate information (longitude, latitude, altitude) of its current location together with the photo to its internal storage system or transmit it to the ground control station.

[0085] Matching process:

[0086] When the drone completes taking photos and stores the coordinate information, the system begins the matching and association process between the photos and the actual tower coordinates.

[0087] The system first sets a matching radius (such as 10 meters), and then searches for known tower coordinates within the radius.

[0088] If only one tower is found within the matching radius, the system automatically associates the photo with that tower.

[0089] If multiple towers are found within the matching radius, the system will sort the towers by coordinate distance from near to far and list them on the interface. At this time, the inspector needs to manually select the inspected tower associated with the photo.

[0090] Manual selection and unmatched processing:

[0091] When multiple poles and towers need to be manually selected, the inspectors can select the most suitable pole and tower for association based on the actual scene and pole and tower features in the photo.

[0092] If no tower is found within the matching radius, the system will display a list of all towers under the current task. The uninspected towers will be displayed in front, and the inspected towers will be displayed in the back. At this time, the inspector needs to manually select the tower with the closest coordinates for association.

[0093] Confirmation and storage of association results:

[0094] Whether it is automatic matching or manual selection, once the association result is determined, the system will associate and store the photo with the corresponding tower information.

[0095] This related information will be used for subsequent defect and hidden danger identification, data analysis and other links.

[0096] During the matching and association process, the accuracy and stability of the drone positioning system should be ensured to avoid matching errors caused by positioning errors. For complex terrain or dense tower areas, it may be necessary to adjust the matching radius or use more advanced matching algorithms to improve matching accuracy.

[0097] Inspection personnel should have certain professional knowledge and experience so that they can accurately judge the correspondence between photos and towers when manually selecting.

[0098] Through the above steps and precautions, the drone can efficiently complete the tasks of taking pictures and obtaining coordinates, and accurately match and associate the photos with the actual towers, providing reliable data support for subsequent analysis and processing.

[0099] Preferably, Figure 6 As shown, in step 4, the UAV inspection platform uses a defect recognition model to perform defect recognition on the UAV inspection results;

[0100] The defect recognition model includes a data preprocessing module: preprocessing the images collected by the drone, including denoising, contrast enhancement, distortion correction, and cropping and scaling the images to meet the input requirements of the model;

[0101] Feature extraction module: Convolutional neural network (CNN) is used to extract features from preprocessed images;

[0102] Classifier module: Based on the extracted features, support vector machine (SVM) is used to classify images.

[0103] The classifier classifies the image into normal or defective categories;

[0104] Post-processing module: post-processes the output of the classifier, marks the defect location, and returns the defect type code to the drone inspection platform.

[0105] The drone inspection platform uses a defect recognition model to identify defects in drone inspection results, including the specific functions and workflow of the data preprocessing module, feature extraction module, classifier module, and post-processing module:

[0106] 1. Data Preprocessing Module

[0107] Denoising: The images collected by drones may be interfered by various noises, such as Gaussian noise, salt and pepper noise, etc. The data preprocessing module will first perform denoising on these images to improve image quality. Common denoising methods include mean filtering, median filtering, Gaussian filtering, etc.

[0108] Enhance contrast:

[0109] In order to make the details in the image clearer, the data preprocessing module will perform contrast enhancement on the image. Contrast enhancement can be achieved through methods such as histogram equalization, adaptive histogram equalization, and Laplace operator.

[0110] Correct distortion:

[0111] Since the lens of the drone camera may be distorted, the image may be distorted. Therefore, the data preprocessing module will perform distortion correction on the image. Distortion correction is usually based on the camera's intrinsic parameters and distortion coefficients, which can be obtained through camera calibration.

[0112] Image cropping and scaling:

[0113] In order to meet the input requirements of the model, the data preprocessing module will crop and scale the image. Cropping can remove unnecessary parts of the image, such as the background and edges, while scaling can adjust the image to the size required by the model.

[0114] 2. Feature Extraction Module

[0115] Convolutional Neural Network CNN:

[0116] The feature extraction module uses convolutional neural network (CNN) to extract features from the preprocessed images.

[0117] CNN is a deep learning model that extracts discriminative features from images through structures such as convolutional layers, pooling layers, and fully connected layers. During the feature extraction process, CNN automatically learns information such as texture, shape, and color in the image and represents this information as a high-dimensional feature vector.

[0118] 3. Classifier Module

[0119] Support Vector Machine SVM:

[0120] The classifier module uses support vector machine (SVM) to classify the extracted features. SVM is a classification method based on statistical learning theory, which separates samples of different categories by finding an optimal hyperplane.

[0121] During the classification process, the SVM is trained using the extracted feature vectors and known class labels to obtain a classification model. The model can then classify new unknown images into normal or defective categories.

[0122] 4. Post-processing module

[0123] Mark defect locations: The post-processing module further processes the output of the classifier and marks the defect locations in the image. This is usually achieved through positioning algorithms, such as feature-based matching algorithms, deep learning-based target detection algorithms, etc. The marked defect locations can help inspectors quickly locate and handle defects.

[0124] Returns the defect type code:

[0125] The post-processing module will also return the identified defect type code to the drone inspection platform.

[0126] The defect type codes are shown in Table 1:

[0127] Table 1:

[0128]

[0129]

[0130] Through the processing of the above four modules, the drone inspection platform can realize automatic defect recognition of drone inspection results. This recognition method not only improves the accuracy and efficiency of recognition, but also reduces the cost and risk of manual intervention. With the continuous development of deep learning technology, these modules can be further optimized and improved in the future to improve the performance and generalization ability of defect recognition.

[0131] Preferably, Figures 4-5 As shown, the defects and hidden dangers manually registered or intelligently identified during the drone inspection process are summarized and displayed on the inspection task page of the drone inspection platform based on a map of the power grid, including inspection progress, drone flight trajectory, inspection task information, and inspection equipment details. Defects and hidden dangers can be confirmed, modified, and reviewed manually.

[0132] Defect hazards can be confirmed, modified and reviewed manually, mainly considering the problems of misidentification and under-identification of defects. Deletions and additions can be made manually for correction.

[0133] The inspection task page of the drone inspection platform is the core management interface of drone inspection operations. It integrates functions such as inspection progress monitoring, drone flight trajectory display, inspection task information management, inspection equipment details viewing, and defect hidden danger handling. Through this page, inspection personnel can fully understand the status of drone inspection operations and handle defects and hidden dangers in a timely and effective manner.

[0134] The inspection task page of the drone inspection platform provides a comprehensive, convenient and efficient inspection operation management platform for inspection personnel by integrating inspection progress display, drone flight trajectory display, inspection task information management, inspection equipment details viewing and defect hidden danger handling functions. This helps inspection personnel to timely grasp the progress and situation of inspection operations, quickly discover and handle defects and hidden dangers, and improve the efficiency and accuracy of inspection operations.

[0135] Preferably, the route mission is generated by having a drone fly over the area of ​​the equipment to be inspected in advance along a preset path. During the flight, the drone will collect relevant geographic location, equipment layout, and possible obstacle information data. Based on the collected data, the system will generate and save a route file.

[0136] This route mission involves using a drone to conduct a preliminary flight inspection of the equipment to be inspected. Specifically, the drone will fly over the area of ​​the equipment to be inspected in advance according to a preset or autonomously planned path. During the flight, the drone will collect data such as relevant geographical location, equipment layout and possible obstacle information. Based on this data, the system will generate a detailed route file and save it. This route file contains key parameters such as the precise path, altitude, speed, etc. of the drone during flight, as well as specific areas or equipment that may need attention.

[0137] Once generated and saved, this route file becomes an optional flight plan. In subsequent patrol missions, operators can choose to use this pre-generated route file to guide the flight of the drone according to actual needs. Doing so can not only improve the efficiency and accuracy of patrols, but also reduce the errors and risks caused by human operation.

[0138] Preferably, the push of the defect work order in step 6 calls a standard interface provided by the business center.

[0139] In the drone inspection process, when a defect is identified and a defect work order needs to be generated and pushed, this step is usually implemented by calling the standard interface provided by the business middle platform.

[0140] Standard interfaces are a standardized and standardized data exchange method provided by the business platform. These interfaces define the format, protocol and rules of data exchange, ensuring seamless data exchange and integration between different business systems. In drone inspection systems, standard interfaces are widely used for the push of defect work orders and feedback on defect processing.

[0141] The PMS3.0 system is a new generation of lean management system for equipment assets built by State Grid Corporation of China. With the power grid resource business middle platform as the core, PMS3.0 covers a new digital architecture of "three zones and four layers". It horizontally connects the production control area, management information area, and Internet area, and vertically covers the perception layer, network layer, platform layer, and application layer. Based on enterprise-level middle platforms such as the data middle platform and the technical middle platform, it follows a unified model, service, and organizational structure. In actual applications, the PMS3.0 system has realized the application of "one map for the power grid" and strengthened the integration of equipment management business. Through the mobile phone i State Grid mobile operation APP, the dual-flow of business flow and information flow of daily operations such as distribution inspection, maintenance, and testing is realized.

[0142] A computer device comprises: a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the program, a closed-loop management method for autonomously identifying hidden dangers of defects in unmanned aerial vehicles is implemented.

[0143] The embodiment of the present application also provides a computer program product including instructions, which, when executed by a computer, enables the computer to execute the method of the above method embodiment.

[0144] A computer-readable storage medium having a computer program stored thereon, characterized in that when the computer program is executed by a processor, a closed-loop management method for autonomously identifying defects and hidden dangers of unmanned aerial vehicles is implemented.

[0145] The computer-readable storage medium may be any available medium that can be accessed by a computer or a data storage device such as a server or a data center that includes one or more available media. The available medium may be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a digital video disc (DVD)), or a semiconductor medium (e.g., a solid state disk (SSD)).

[0146] The preferred embodiments of the present invention are described in detail above in conjunction with the accompanying drawings, but the present invention is not limited to the above embodiments. Various changes can be made within the knowledge scope of ordinary technicians in the field without departing from the purpose of the present invention. These changes involve related technologies well known to those skilled in the art, which all fall within the scope of protection of the patent of the present invention.

[0147] Many other changes and modifications may be made without departing from the concept and scope of the present invention.It should be understood that the present invention is not limited to the specific embodiments, and the scope of the present invention is defined by the appended claims.

Claims

1. A closed-loop management method for autonomous identification of defects and hidden dangers of drones, characterized by: Includes steps: Step 1: Prepare inspection tasks and issue inspection work orders; Step 2: UAV on-site inspection operation; Step 3: Manual or intelligent registration of defects and hidden dangers; Step 4: The drone inspection results are transmitted back to the drone inspection platform; Step 5: Manual confirmation of defects and hidden dangers and review of inspection results; Step 6: Push the defect work order to the PMS3.0 system. PMS3.0 generates the corresponding defect work order, distributes the work order, closes the work order after the defect is eliminated, and returns the information to the drone inspection platform; Step 7: The drone inspection platform updates the defect processing status.

2. The closed-loop management method for autonomous identification of defects and hidden dangers of unmanned aerial vehicles according to claim 1 is characterized in that: The preparation of the inspection task in step 1 includes the preparation of line details, and the line details include voltage level and line name. After the inspection is completed, the collection creation time and collection information statistics are automatically generated.

3. The closed-loop management method for autonomous identification of defects and hidden dangers of unmanned aerial vehicles according to claim 2 is characterized in that: In step 2, the drone on-site patrol operation autonomously completes the flight and takes photos according to the route mission preset in the current patrol mission.

4. The closed-loop management method for autonomous identification of defects and hidden dangers of unmanned aerial vehicles according to claim 3 is characterized in that: The drone uses Beidou positioning and RTK modules to obtain the coordinates of the drone's photo-taking position and complete visible light photography of different parts and angles of the device; After the coordinates are acquired, the coordinates of the photo and the actual tower are matched and associated; The coordinate matching association adopts the following steps: after the drone takes a photo, the drone system is set to start matching the pole towers within a 10-meter radius of the drone coordinates. If only one pole tower is matched, the photo is associated with the pole tower. If multiple pole towers are matched, all the matched pole towers are listed and sorted from near to far by coordinate distance. The patrol pole tower associated with the photo is manually selected. If no pole tower is matched, all the pole towers under the current task are displayed, with the unpatrolled ones displayed in front and the patrolled ones displayed in the back. The ones with the closest coordinates are manually selected for association.

5. The closed-loop management method for autonomous identification of defects and hidden dangers of unmanned aerial vehicles according to any one of claims 1 to 4, characterized in that: In step 4, the UAV inspection platform uses a defect recognition model to identify defects in the UAV inspection results; The defect recognition model includes a data preprocessing module: preprocessing the images collected by the drone, including denoising, contrast enhancement, distortion correction, and cropping and scaling the images to meet the input requirements of the model; Feature extraction module: Convolutional neural network (CNN) is used to extract features from preprocessed images; Classifier module: Based on the extracted features, support vector machine (SVM) is used to classify the image. The classifier divides the image into normal or defective categories. Post-processing module: post-processes the output of the classifier, marks the defect location, and returns the defect type code to the drone inspection platform.

6. The closed-loop management method for autonomous identification of defects and hidden dangers of unmanned aerial vehicles according to claim 5 is characterized in that: The defects and hidden dangers manually registered or intelligently identified during the drone inspection process are summarized and displayed on the inspection task page of the drone inspection platform based on a map of the power grid, including inspection progress, drone flight trajectory, inspection task information, and inspection equipment details. Defects and hidden dangers can be confirmed, modified, and reviewed manually.

7. The closed-loop management method for autonomous identification of defects and hidden dangers of unmanned aerial vehicles according to claim 6 is characterized in that: The generation of the route mission uses a drone to fly over the area of ​​the equipment to be inspected in advance according to a preset path. During the flight, the drone will collect relevant geographical location, equipment layout, and possible obstacle information data. Based on the collected data, the system will generate a route file and save it.

8. The closed-loop management method for autonomous identification of defects and hidden dangers of unmanned aerial vehicles according to claim 7 is characterized in that: The push of the defect work order in step 6 calls the standard interface provided by the business center.

9. A computer device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, a closed-loop management method for autonomous identification of defects and hidden dangers of unmanned aerial vehicles as described in any one of claims 1 to 8 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, the closed-loop management method for autonomous identification of defects and hidden dangers of drones as described in any one of claims 1 to 8 is implemented.

Images