Photovoltaic intelligent unmanned aerial vehicle inspection system

Abstract

The invention relates to the related technical field of inspection, in particular to a photovoltaic intelligent unmanned aerial vehicle inspection system. The system comprises an SaaS platform for providing annotation platform management, merchant platform management, platform setting function and content management. The AI training system comprises a training system and an identification system. The training system is used for model training and continuously optimizing model performance through annotation data; the recognition system is used for recognizing defects, such as spot spots, multiple spots, battens, missing, no-load and shielding, in photovoltaic pictures shot by the unmanned aerial vehicle. And the labeling platform provides project management, task management, employee management, role management and picture labeling functions and is used for data labeling and preprocessing work. And the merchant platform performs photovoltaic project management, surveying and mapping management, map management, unmanned aerial vehicle management, task management, an infrared picture AI identification function and a photovoltaic defect report generation function.

Description

Technical Field

[0001] The present application relates to the field of inspection-related technologies, and specifically to a photovoltaic intelligent drone inspection system. Background Art

[0002] In the process of the vigorous development of the photovoltaic energy industry, the scale of photovoltaic power stations continues to expand and the layout becomes increasingly complex, and their operation and maintenance management faces severe challenges. Under the traditional manual inspection mode, operation and maintenance personnel need to walk between the vast photovoltaic arrays and check the photovoltaic components one by one with handheld detection tools. This method is extremely labor-intensive, time-consuming, and inefficient, and it is difficult to meet the high-frequency and refined inspection needs of large-scale power stations. According to industry statistics, a round of manual inspection of large photovoltaic power stations often takes several weeks, during which power generation failures are difficult to detect and repair in time, resulting in serious cumulative power generation losses. Summary of the invention

[0003] In view of this, the embodiments of the present application are directed to providing a photovoltaic intelligent drone inspection system.

[0004] This application provides a photovoltaic intelligent drone inspection system, including:

[0005] The merchant platform includes photovoltaic project management, surveying and mapping management, map management, drone management, task management, infrared image AI recognition function and photovoltaic defect report generation function. Among them, photovoltaic project management controls the basic information and status of photovoltaic power stations, surveying and mapping management supports drone surveying and mapping to generate photovoltaic area maps, map management realizes map generation and display, drone management provides map drawing area, waypoint setting and route setting functions to plan drone flight paths, task management coordinates drone inspection tasks, infrared image AI recognition function identifies photovoltaic defects in infrared images taken by drones based on AI algorithms, and photovoltaic defect report generation function automatically generates reports based on recognition results for reference by operation and maintenance personnel.

[0006] The AI ​​training system includes a training system and a recognition system. The training system optimizes the model performance by using the labeled data. The recognition system is used to identify defects in the photovoltaic images taken by the drone based on the model. The defects include spots, multiple spots, strips, missing, empty and occlusion, and is used to provide support for the AI ​​recognition function of infrared images;

[0007] A labeling platform, which has project management, task management, employee management, role management and image labeling functions to carry out data labeling and preprocessing work, and provide support for the merchant platform and the AI ​​training system;

[0008] The cloud service platform is used to provide preset application software for users to access the merchant platform, the AI ​​training system and the annotation platform based on the application software to manage the annotation platform, the merchant platform, platform setting functions and content management.

[0009] In some embodiments, the training system is used to periodically retrain the model.

[0010] In some embodiments, the model is a deep learning model.

[0011] In some embodiments, the drone management includes:

[0012] Control the drone to perform inspections.

[0013] In some embodiments, the preset application software is run on a user's smart device;

[0014] The smart device has a wireless communication function, which is used to communicate with the cloud service platform and perform drone inspections through the merchant platform, the labeling platform and the AI ​​training system.

[0015] A photovoltaic intelligent drone inspection system provided in the present application includes: a merchant platform, including photovoltaic project management, surveying and mapping management, map management, drone management, task management, infrared image AI recognition function and photovoltaic defect report generation function, among which photovoltaic project management controls the basic information and status of photovoltaic power stations, surveying and mapping management supports drone surveying and mapping to generate photovoltaic area maps, map management realizes map generation and display, drone management provides map drawing area, waypoint setting and route setting functions to plan drone flight paths, task management coordinates drone inspection tasks, infrared image AI recognition function identifies photovoltaic defects in infrared images taken by drones based on AI algorithms, and photovoltaic defect report generation function automatically generates reports based on recognition results for reference by operation and maintenance personnel. The AI ​​training system includes a training system and a recognition system. The training system optimizes model performance by using labeled data. The recognition system is used to identify defects in photovoltaic images taken by drones based on the model. The defects include spots, multiple spots, strips, missing parts, no loads and occlusions, and is used to provide support for the AI ​​recognition function of infrared images; the labeling platform has project management, task management, employee management, role management and image labeling functions to carry out data labeling and preprocessing, and provide support for the merchant platform and the AI ​​training system; the cloud service platform is used to provide preset application software for users to access the merchant platform, the AI ​​training system and the labeling platform based on the application software, so as to manage the labeling platform, the merchant platform, the platform setting function and the content management. The synergistic beneficial effects of each module under the overall system architecture: Efficient operation and maintenance guarantee: The merchant platform integrates multiple management functions, photovoltaic project management ensures that the basic and dynamic data of the power station are under control, surveying and mapping and map management build an accurate geographic information framework, drone management plans scientific inspection routes, task management is orderly scheduled, infrared recognition and report generation functions quickly lock and feedback defects, and each link is seamlessly connected, turning the operation and maintenance of photovoltaic power stations from passive repairs to active prevention, greatly reducing the fault response time, effectively improving the continuity of power generation, and reducing the labor cost of operation and maintenance, laying a solid foundation for the stable and efficient operation of the power station. Advanced intelligent diagnosis: The AI ​​training system uses the massive and accurate annotation data of the annotation platform as nutrients to continuously improve the accuracy of the model. Based on this, the recognition system has a deep insight into various subtle defects in photovoltaic images, injects a strong intelligent core into the infrared recognition function, realizes intelligent identification of defects, and graded warning, helps operation and maintenance personnel accurately locate key hidden dangers, and accurately deploys maintenance resources, greatly improving the level of photovoltaic asset health management. Data-driven optimization: The labeling platform relies on all-round management of projects, tasks, personnel, roles and images to create high-quality data soil for the AI ​​training system, ensure unified data labeling standards and efficient and smooth processing, drive the acceleration of model evolution and iteration, inject continuous momentum into the system's intelligent decision-making capabilities, and enable the quality of the entire photovoltaic operation and maintenance process to leap forward with data intelligence.As the central nervous system of the system, the cloud service platform uses preset software to break through the interaction barriers between modules. Users can navigate through various functional domains through a unified entrance. Labeling management, merchant coordination, platform configuration and content operation are all done in one go. Multi-user cross-departmental collaboration is smooth, information is shared in real time, and instructions are conveyed quickly. The collaborative efficiency grows exponentially, the preparation period for new merchants to access the system is shortened, and the efficiency of system function expansion and upgrade deployment is improved, which activates the vitality of the system ecology in all aspects and accelerates the innovation and transformation of photovoltaic operation and maintenance business. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0017] Figure 1 It is a structural schematic diagram of a photovoltaic intelligent drone inspection system provided by an embodiment of the present application.

[0018] Figure 2 It is a schematic diagram of a photovoltaic intelligent drone inspection system provided in another embodiment of the present application. DETAILED DESCRIPTION

[0019] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0020] In the process of the vigorous development of the photovoltaic energy industry, the scale of photovoltaic power stations continues to expand and the layout becomes increasingly complex, and their operation and maintenance management faces severe challenges. Under the traditional manual inspection mode, operation and maintenance personnel need to walk between the vast photovoltaic arrays and check the photovoltaic components one by one with handheld detection tools. This method is extremely labor-intensive, time-consuming, and inefficient, and it is difficult to meet the high-frequency and refined inspection needs of large-scale power stations. According to industry statistics, a round of manual inspection of large photovoltaic power stations often takes several weeks, during which power generation failures are difficult to detect and repair in time, resulting in serious cumulative power generation losses.

[0021] With the rise of drone technology, it has made its first appearance in the field of photovoltaic inspection. Some inspection systems use drones to collect image data and initially improve inspection efficiency. However, most existing drone inspection systems are stuck in the dilemma of single function. Focusing only on image acquisition, they lack the ability to analyze and process in-depth data and are unable to accurately identify the complex and diverse defect types of photovoltaic components, resulting in a large number of potential fault hazards that are difficult to detect. For example, in the face of early fault manifestations such as microscopic cracks and hidden hot spots on photovoltaic panels, misjudgments and missed judgments often occur, resulting in a sharp increase in the risk of fault deterioration.

[0022] Traditional manual inspection methods for photovoltaic power stations are time-consuming, labor-intensive, and inefficient. With the rapid development of drone technology and artificial intelligence technology, it has become possible to use drones to inspect photovoltaic power stations. However, most existing drone inspection systems have single functions and lack complete SaaS platform support, making it impossible to achieve multi-merchant sharing and efficient management.

[0023] In order to solve the above problems, the present application provides a solution, which relates to the field of photovoltaic energy monitoring technology, specifically a photovoltaic smart drone inspection system and its operation method, which aims to realize fast and accurate inspection of photovoltaic areas through drone technology and artificial intelligence algorithms, and improve the operation and maintenance efficiency of photovoltaic power stations. Including: SaaS platform: provides annotation platform management, merchant platform management, platform setting function and content management. Among them, the annotation platform management is used for data annotation; the merchant platform management supports the use of multiple merchants to realize the management of photovoltaic projects, surveying and mapping, maps and drones; the platform setting function is used for the overall configuration of the system; content management is used for content publishing and review. AI training system: including training system and recognition system. The training system is used for model training, and the model performance is continuously optimized by annotating data; the recognition system is used to identify defects in photovoltaic pictures taken by drones, such as spots, multiple spots, strips, missing, empty and occlusion. Annotation platform: provides project management, task management, employee management, role management and picture annotation functions for data annotation and preprocessing. Merchant platform: Volt project management: used to manage the basic information and status of photovoltaic power stations. Surveying and mapping management: supports drone surveying and mapping functions to generate photovoltaic area maps. Map management: including map generation and map display, used to display the geographical location and layout of photovoltaic power stations. Drone management: provides map drawing area function, waypoint setting function and route setting function, used to plan the flight path of drones. Task management: used to manage drone inspection tasks. Infrared image AI recognition function: use AI algorithm to identify photovoltaic defects in infrared images taken by drones. Photovoltaic defect report generation function: according to the recognition results, photovoltaic defect reports are automatically generated for reference by operation and maintenance personnel. In this way, through the combination of drones and AI technology, fast and accurate inspection of photovoltaic power stations is achieved, greatly improving operation and maintenance efficiency. The design of the SaaS platform enables multiple merchants to share system resources and reduce operation and maintenance costs. The integration of the annotation platform and the AI ​​training system enables the system to continuously optimize model performance and improve recognition accuracy. The merchant platform provides a wealth of management functions, allowing operation and maintenance personnel to easily manage photovoltaic power stations and drone inspection tasks.

[0024] After introducing the basic principles of the present application, various non-limiting embodiments of the present application will be described in detail with reference to the accompanying drawings.

[0025] Reference Figure 1 and Figure 2 The photovoltaic intelligent drone inspection system provided in this application includes:

[0026] The merchant platform includes photovoltaic project management, surveying and mapping management, map management, drone management, task management, infrared image AI recognition function and photovoltaic defect report generation function. Among them, photovoltaic project management controls the basic information and status of photovoltaic power stations, surveying and mapping management supports drone surveying and mapping to generate photovoltaic area maps, map management realizes map generation and display, drone management provides map drawing area, waypoint setting and route setting functions to plan drone flight paths, task management coordinates drone inspection tasks, infrared image AI recognition function identifies photovoltaic defects in infrared images taken by drones based on AI algorithms, and photovoltaic defect report generation function automatically generates reports based on recognition results for reference by operation and maintenance personnel.

[0027] The AI ​​training system includes a training system and a recognition system. The training system optimizes the model performance by using the labeled data. The recognition system is used to identify defects in the photovoltaic images taken by the drone based on the model. The defects include spots, multiple spots, strips, missing, empty and occlusion, and is used to provide support for the AI ​​recognition function of infrared images;

[0028] A labeling platform, which has project management, task management, employee management, role management and image labeling functions to carry out data labeling and preprocessing work, and provide support for the merchant platform and the AI ​​training system;

[0029] The cloud service platform is used to provide preset application software for users to access the merchant platform, the AI ​​training system and the annotation platform based on the application software to manage the annotation platform, the merchant platform, platform setting functions and content management.

[0030] Specifically, the merchant platform is used to execute photovoltaic project management: comprehensively control the core elements of photovoltaic power stations, from basic parameters such as power station name, geographic coordinates, installed capacity, and grid-connected time, to dynamic indicators such as real-time power, power generation fluctuations, and equipment health status, all of which are accurately collected, updated in real time, and deeply analyzed, laying a solid data foundation for power station operation strategy formulation, benefit evaluation, and fault warning, ensuring that operation and maintenance decisions are targeted, accurate, and efficient.

[0031] Surveying and mapping management: UAVs are equipped with cutting-edge surveying and mapping instruments (lidar, high-resolution optical cameras, etc.) to cruise and collect massive geographic and photovoltaic equipment spatial data. After processing with professional algorithms, two-dimensional and three-dimensional digital maps with centimeter-level accuracy are constructed to accurately outline the topography, component layout, and electrical wiring architecture of power stations, providing intuitive and accurate spatial references for the entire life cycle management of power station planning, construction, operation and maintenance inspections, and disaster assessment.

[0032] Map management: Integrate advanced geographic information system (GIS) technology to efficiently generate electronic maps and achieve multi-source map data fusion (satellite images, terrain vector maps, etc.). Give users a smooth interactive experience, such as stepless zooming, arbitrary perspective switching, and layer overlay control, so that operation and maintenance personnel can immersively manage power station assets, plan inspection routes, analyze fault distribution patterns, and optimize operation and maintenance resource allocation in a virtual map environment.

[0033] Drone management: In the map interactive interface, you can easily draw the inspection area boundary, flexibly set waypoints and plan routes according to the power station layout and inspection needs, and accurately define flight speed, altitude, heading angle and other parameters. Integrate intelligent navigation algorithms to achieve autonomous obstacle avoidance and dynamic path re-planning in complex environments, adapt to changing lighting and weather conditions, and ensure drone flight safety, accurate and efficient mission execution, and complete and reliable data collection.

[0034] Task management: Based on the power station operation and maintenance procedures, equipment aging models, and meteorological forecast data, intelligently arrange inspection task plans, carefully set task priorities, execution periods, and execution frequencies, and accurately dispatch and adapt drone resources. Monitor task progress throughout the process, provide real-time feedback on task status (executing, paused, completed, abnormal interruption), seamlessly connect the fault reporting process, and provide central dispatch support for the automation and collaborative operation of the operation and maintenance process.

[0035] Infrared image AI recognition function: deeply integrates convolutional neural network (CNN) and deep learning algorithm, deeply mines infrared image spectrum and texture features, accurately identifies potential defects such as hot spots, hidden cracks, diode failures, and cell failures of photovoltaic modules, and outputs quantitative diagnostic information such as defect category, location, area, and temperature anomaly in real time, reducing the traditional manual visual inspection process of several hours to milliseconds, improving defect screening efficiency and accuracy, and reducing dependence on operation and maintenance manpower.

[0036] Generate photovoltaic defect report function: Based on AI recognition results and power station operation data, the standardized report template is automatically filled to generate a photovoltaic defect analysis report with rich pictures and texts and detailed data. The content covers the panoramic distribution heat map of defects, detailed parameters of defective components, fault evolution trend prediction, repair strategy suggestions and cost estimation, and pushes them to the operation and maintenance team's mobile terminal and management terminal with one click, providing key decision-making basis for fault repair, performance optimization, and asset management, and accelerating the operation and maintenance closed-loop response process.

[0037] Training system in AI training system: Based on the annotation platform that gathers a large number of photovoltaic defect samples, it adopts a deep neural network architecture (such as ResNet and VGG variants), follows the supervised learning paradigm, takes the labeled defect labels as the true value, and iteratively fine-tunes the network weight parameters through the back-propagation optimization algorithm to deeply fit the intrinsic pattern of the data. It uses data enhancement technology (rotation, flipping, noise injection) to expand sample diversity, alleviate overfitting, and continuously improve the model generalization and feature learning capabilities, laying a solid intelligent foundation for accurate defect identification.

[0038] Recognition system in AI training system: Relying on mature training models, real-time preprocessing (normalization, grayscale stretching, filtering and noise reduction) of photovoltaic images (visible light and infrared bands) collected by drones is performed, and the input model extracts abstract feature representations, and the output defect category confidence vector is determined by multiple classifiers, and non-maximum suppression screening is used to accurately locate the defect area. Its multi-category defect recognition has high accuracy and low misjudgment rate, providing reliable intelligent diagnosis support for power station operation and maintenance, and effectively curbing the risk of fault spread.

[0039] The annotation platform is used to: Execute project management: Create a project resource library, centrally manage the project documents of the entire life cycle of photovoltaic power stations (design drawings, construction records, acceptance reports, operation and maintenance logs), and associate corresponding inspection annotation tasks and team member role permissions. Flexible allocation of resources according to project progress nodes, accurate control of key indicators of project cost, progress, and quality, and achieve standardized, visual and fine management of the entire process from project planning and establishment to decommissioning and demolition.

[0040] Task management: Disassemble task orders according to the requirements of the annotation project, set task priorities, delivery deadlines, and acceptance criteria, intelligently assign tasks to the task queues of annotation personnel, and track task progress in real time (not started, in progress, pending review, completed). Integrate the performance management module to quantitatively evaluate personnel performance based on task volume, annotation accuracy, and completion time, motivate the annotation team to collaborate efficiently, ensure timely and high-quality delivery of annotation tasks, and provide a stable and high-quality data supply for system model training.

[0041] Employee management: Build an employee information database covering basic information, professional qualifications, project experience, skills and expertise, etc., to provide all-round data support for personnel recruitment, training, performance evaluation, and career development planning. Dynamically adjust personnel division of labor and authority allocation according to project task load, accurately match human resources and business needs, build a high-performance labeling team, and improve overall operational efficiency and quality stability.

[0042] Role management: define the multiple role permission system for labelers, auditors, administrators, etc., and set up hierarchical permissions for data access, labeling operations, task allocation, quality review, system configuration, etc. Adapt differentiated operation interfaces and workflows according to personnel roles to ensure data access security and compliance, standardized and orderly labeling processes, efficient and coordinated task transfer, and build a data labeling ecosystem with clear responsibilities and strict control.

[0043] Image annotation function: integrated with professional image annotation tools (rectangular selection, polygon outline, point annotation, semantic segmentation tools), annotators can accurately annotate the defect area boundaries, attribute labels, and damage degree descriptions of photovoltaic images according to the predefined defect classification system. It supports multiple rounds of annotation quality review and annotation conflict arbitration mechanism, injects high-value information into AI model training with high-precision annotation data, drives model learning optimization, and improves the intelligent level of photovoltaic defect recognition.

[0044] The cloud service platform provides application software service integration services: integrating cutting-edge cloud computing infrastructure and microservice architecture, prefabricating general application software suites (system management portal, data visualization screen, mobile inspection APP, annotation work terminal), and opening up multi-platform user access channels with unified identity authentication and single sign-on technology. Each software module is loosely coupled and independently iterated and upgraded, and seamless interactive collaboration is achieved through the API gateway, providing users with a one-stop, full-scenario, cross-terminal system access experience, accelerating business innovation and digital transformation.

[0045] Collaborative support of platform functions: Users can access the full-function modules of merchant platform operation and maintenance management, AI training system intelligent diagnosis, and annotation platform data processing in one stop through cloud application software. Functional domains are set on the platform to flexibly configure system parameters (algorithm model hyperparameters, data storage strategies, network communication protocols), manage user authority roles, and monitor system operation indicators (resource utilization, task throughput, response latency). Relying on the content management module, system announcements, operation manuals, and industry information can be efficiently reviewed and released to ensure system knowledge inheritance, smooth user communication, and accurate and efficient information dissemination, so as to activate the collaborative innovation vitality of the system ecology and improve the overall digital and intelligent operation level of photovoltaic operation and maintenance management.

[0046] In some embodiments, the training system is used to periodically retrain the model. The model is a deep learning model.

[0047] In the photovoltaic smart drone inspection system, the training system is specifically used to regularly retrain the deep learning model to adapt to changing data and improve the accuracy of identifying photovoltaic power station defects. The following is a detailed introduction to this function:

[0048] 1. Characteristics of deep learning models

[0049] Deep learning models, especially convolutional neural networks (CNNs), have performed well in image recognition and classification tasks. They can automatically learn complex feature representations from data without manually designing feature extraction algorithms. This is particularly important for image recognition in photovoltaic power plant inspections, as the types of defects in photovoltaic panels are diverse and complex.

[0050] 2. Regular retraining of the training system

[0051] Data update: As time goes by, the number of images of photovoltaic power plants taken by drones will continue to increase, and this new data can be used to retrain the model to adapt to new situations and environmental changes.

[0052] Model optimization: Regular retraining allows the model to learn the latest data features, thereby improving its accuracy and robustness in identifying PV defects.

[0053] Technology iteration: The field of deep learning is developing rapidly. Regularly retraining models can integrate the latest algorithms and technological improvements to improve model performance.

[0054] 3. Key functions of the training system

[0055] Automated data preprocessing: Perform preprocessing steps such as standardization, normalization, and data augmentation on newly collected image data to improve the efficiency and effectiveness of model training.

[0056] Model selection and tuning: Select the appropriate deep learning model architecture based on task requirements and adjust hyperparameters to optimize performance.

[0057] Distributed training: Leverage cloud computing resources for distributed training to accelerate the model training process, especially when dealing with large-scale data sets.

[0058] Performance monitoring: Monitor the performance of the model during training, including indicators such as accuracy, recall, and F1 score, to ensure the effectiveness of model training.

[0059] Model version control: Record the model version and performance of each training to facilitate comparison and rollback to a model version with better performance.

[0060] 4. The coordination between the training system and other parts of the system

[0061] Collaboration with annotation platforms: The training system can directly obtain high-quality annotated data from the annotation platform, which is crucial for training deep learning models.

[0062] Collaboration with AI recognition system: Models regularly updated by the training system can be directly deployed into the recognition system to improve its performance.

[0063] Collaboration with merchant platforms: Merchant platforms can provide feedback to help the training system understand the performance of the model in actual applications and guide further optimization of the model.

[0064] 5. Technical Implementation

[0065] Use deep learning frameworks: such as TensorFlow, Keras, PyTorch, etc. These frameworks provide tools and libraries for building and training deep learning models.

[0066] Use cloud computing resources: Use the elasticity and scalability of cloud computing to provide the necessary computing resources for large-scale data training.

[0067] Through such a design, the training system can ensure the continuous optimization and updating of the deep learning model, and work closely with other parts of the system to jointly improve the overall performance and efficiency of the photovoltaic smart drone inspection system.

[0068] The drone management includes: controlling the drone to perform inspections. In the photovoltaic smart drone inspection system, drone management is a key component, which involves the scheduling, control and monitoring of drones to perform inspection tasks. The following is a detailed introduction to the drone management function:

[0069] 1. Inspection task planning

[0070] Automatic flight: Drones can automatically perform inspection tasks according to preset routes, reducing human intervention.

[0071] Route setting: Drone management allows users to set the flight route of the drone according to the layout and inspection requirements of the PV power station.

[0072] Task Scheduling: The system can intelligently schedule inspection tasks based on task priority and drone status.

[0073] 2. Data Collection

[0074] Image acquisition: The camera equipped on the drone can take images of the photovoltaic panels at predetermined intervals for subsequent AI recognition and analysis.

[0075] Infrared thermal imaging: Some drones may be equipped with infrared thermal imaging equipment to detect thermal anomalies in photovoltaic panels, which is another effective way to identify defects.

[0076] 3. Safety and emergency response

[0077] Obstacle avoidance system: The drone management integrates an obstacle avoidance system to ensure that the drone can automatically avoid obstacles during flight.

[0078] Emergency Return: In case of emergency or low battery, the drone can automatically return to the preset landing point.

[0079] 4. Drone maintenance

[0080] Maintenance reminder: The system will remind the operation and maintenance personnel to perform necessary maintenance work based on the usage and maintenance cycle of the drone.

[0081] Health Monitoring: Drone management may also include monitoring the health of each drone component to prevent potential failures.

[0082] 5. Collaboration with other parts of the system

[0083] Collaboration with task management: Drone management works closely with task management to ensure that drone inspection tasks can be executed as planned.

[0084] Collaboration with AI recognition system: Image data captured by the drone will be sent to the AI ​​recognition system for analysis to identify defects in photovoltaic panels.

[0085] Through such drone management functions, the photovoltaic intelligent drone inspection system can realize efficient and safe drone inspection operations, and improve the operation and maintenance efficiency and quality of photovoltaic power stations.

[0086] The preset application software is used to run on the user's smart device; the smart device has a wireless communication function, which is used to communicate with the cloud service platform and perform drone inspections through the merchant platform, the labeling platform and the AI ​​training system.

[0087] In the photovoltaic smart drone inspection system, the preset application software is a bridge connecting the user's smart device and the cloud service platform, allowing users to remotely and efficiently manage and monitor drone inspection activities. The following is a detailed introduction to this function:

[0088] 1. Features of the default application software

[0089] Cross-platform compatibility: The application software should be compatible with multiple operating systems, such as iOS, Android, Windows, Linux, etc., so that it can run on different smart devices.

[0090] User-friendly interface: Provides an intuitive, easy-to-use user interface that allows users to easily navigate and operate drone inspection tasks.

[0091] Real-time data synchronization: Ensure real-time synchronization of data between the application on the smart device and the cloud service platform so that users can obtain the latest inspection information.

[0092] 2. The role of cloud service platform

[0093] Data Processing Center: The cloud service platform processes large amounts of data from drones, including images and video streams.

[0094] AI model hosting: The cloud service platform hosts the models trained by the AI ​​training system, which are used to identify defects in photovoltaic panels in real time.

[0095] Data storage: All inspection data and identification results are stored in the cloud to ensure data security and accessibility.

[0096] 3. Integration of merchant platform, annotation platform and AI training system

[0097] Merchant platform: Users access the merchant platform through application software on smart devices to manage photovoltaic project management, surveying and mapping management, map management, drone management, task management, etc.

[0098] Labeling platform: Users can participate in data labeling work remotely and provide high-quality labeled data for AI training systems.

[0099] AI training system: Users can monitor the training progress and performance of AI models through application software to ensure the accuracy and reliability of the models.

[0100] 4. Execution of drone inspections

[0101] Flight path planning: Users can set the flight path of the drone to ensure that all PV areas that need to be inspected are covered.

[0102] Defect identification: The images taken by the drone are analyzed in real time through the AI ​​training system to identify defects in the photovoltaic panels.

[0103] 5. Security and Privacy

[0104] Data encryption: The communication between smart devices and cloud service platforms uses encryption technology to protect the security of data transmission.

[0105] User authentication: The application software requires users to authenticate themselves to ensure that only authorized users can access and control the drone inspection system.

[0106] Through this integrated system, the photovoltaic smart drone inspection system can realize remote, efficient and safe drone inspection operations, and improve the operation and maintenance efficiency and quality of photovoltaic power stations.

[0107] In addition, an embodiment of the present application may also be a computer-readable storage medium on which computer program instructions are stored. When the computer program instructions are executed by a processor, the processor executes the steps of the data security analysis method according to various embodiments of the present application described in the above "Exemplary Method" section of this specification.

[0108] The computer readable storage medium can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can include, for example, but is not limited to, a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0109] The above description has been given for the purpose of illustration and description. In addition, this description is not intended to limit the embodiments of the present application to the forms disclosed herein. Although multiple example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, changes, additions and sub-combinations thereof.

Claims

1. A photovoltaic intelligent drone inspection system, characterized in that: include: The merchant platform includes photovoltaic project management, surveying and mapping management, map management, drone management, task management, infrared image AI recognition function and photovoltaic defect report generation function. Among them, photovoltaic project management controls the basic information and status of photovoltaic power stations, surveying and mapping management supports drone surveying and mapping to generate photovoltaic area maps, map management realizes map generation and display, drone management provides map drawing area, waypoint setting and route setting functions to plan drone flight paths, task management coordinates drone inspection tasks, infrared image AI recognition function identifies photovoltaic defects in infrared images taken by drones based on AI algorithms, and photovoltaic defect report generation function automatically generates reports based on recognition results for reference by operation and maintenance personnel; The AI ​​training system includes a training system and a recognition system. The training system optimizes the model performance by using the labeled data. The recognition system is used to identify defects in the photovoltaic images taken by the drone based on the model. The defects include spots, multiple spots, strips, missing, empty and occlusion, and is used to provide support for the AI ​​recognition function of infrared images; A labeling platform, which has project management, task management, employee management, role management and image labeling functions to carry out data labeling and preprocessing work, and provide support for the merchant platform and the AI ​​training system; The cloud service platform is used to provide preset application software for users to access the merchant platform, the AI ​​training system and the annotation platform based on the application software to manage the annotation platform, the merchant platform, platform setting functions and content management.

2. The photovoltaic intelligent drone inspection system according to claim 1 is characterized in that: The training system is used to periodically retrain the model.

3. The photovoltaic intelligent drone inspection system according to claim 2 is characterized in that: The model is a deep learning model.

4. The photovoltaic intelligent drone inspection system according to claim 1 is characterized in that: The drone management includes: Control the drone to perform inspections.

5. The photovoltaic intelligent drone inspection system according to claim 1 is characterized in that: The preset application software is used to run on the user's smart device; The smart device has a wireless communication function, which is used to communicate with the cloud service platform and perform drone inspections through the merchant platform, the labeling platform and the AI ​​training system.

Images