Photovoltaic power station unmanned aerial vehicle intelligent inspection and fault diagnosis integrated system and method

Abstract

The application provides a photovoltaic station unmanned aerial vehicle intelligent inspection and fault diagnosis integrated system and method, comprising: a building module, which is used for collecting photovoltaic station image data, and establishing a three-dimensional panorama of the photovoltaic station according to the photovoltaic station image data; a strategy module, which is used for determining multi-dimensional fault features of photovoltaic components in the station according to the three-dimensional panorama, and formulating a station inspection strategy according to the multi-dimensional fault features; and an inspection module, which is used for determining an optimal inspection route according to the station inspection strategy, and performing unmanned aerial vehicle intelligent inspection according to the optimal inspection route. Efficient, accurate and intelligent inspection and fault diagnosis of the photovoltaic station are realized.

Description

Technical Field

[0001] This invention relates to the field of photovoltaic power plant operation and maintenance technology, and in particular to an integrated system and method for intelligent inspection and fault diagnosis of photovoltaic power plants using unmanned aerial vehicles (UAVs). Background Technology

[0002] With the transformation of the global energy structure, photovoltaic (PV) power generation, as a clean and renewable energy source, has experienced explosive growth in installed capacity. Large-scale PV power plants typically occupy vast areas with complex geographical environments and consist of tens of thousands of PV modules. To ensure power generation efficiency and system safety, regular inspections of PV modules to promptly detect and eliminate abnormal conditions such as hot spots, microcracks, shading, and diode failures have become a core aspect of power plant operation and maintenance. Traditional manual inspection methods suffer from low efficiency, high labor intensity, numerous blind spots, and strong subjectivity in fault detection, and can no longer meet the operation and maintenance needs of modern large-scale PV power plants.

[0003] In recent years, intelligent inspection technology using drones has gradually become a research hotspot in the industry. Drones equipped with visible light cameras and infrared thermal imagers can quickly acquire images and temperature information of photovoltaic module surfaces, providing a data foundation for automated fault diagnosis. However, existing technical solutions mostly focus on optimizing single aspects, failing to form an integrated closed-loop system from data acquisition, scene modeling, fault analysis to inspection execution, resulting in low efficiency in photovoltaic power plant inspections. Summary of the Invention

[0004] This invention provides an integrated system and method for intelligent inspection and fault diagnosis of photovoltaic power plants using unmanned aerial vehicles (UAVs), which solves the problem of low inspection efficiency of photovoltaic power plants in the prior art.

[0005] On the one hand, the present invention provides an integrated system for intelligent inspection and fault diagnosis of photovoltaic power plants using unmanned aerial vehicles (UAVs), which includes: The system includes a module for collecting image data of photovoltaic power plants and creating a 3D panoramic view of the photovoltaic power plants based on the image data; a strategy module for determining the multi-dimensional fault characteristics of photovoltaic modules within the power plant based on the 3D panoramic view and formulating a power plant inspection strategy based on the multi-dimensional fault characteristics; and an inspection module for determining the optimal inspection route based on the power plant inspection strategy and conducting intelligent drone inspections based on the optimal inspection route.

[0006] Furthermore, the establishment module establishes a three-dimensional panoramic image of the photovoltaic power station based on the photovoltaic power station image data, including: Edge detection is performed on the photovoltaic power station image data, and multiple image blocks are extracted based on the edge detection results, and the pixel features of the image blocks are extracted; Infrared images of photovoltaic power plants are acquired, infrared features of the image blocks are extracted, and image blending features of the image blocks are determined based on the pixel features and the infrared features. The image blending features are matched with image blocks to obtain the image block type of each image block; The image patch types of all image patches are identified, and a 3D panoramic image is created based on the image patch types of all image patches.

[0007] Further, the step of matching the image blending features with image patches to obtain the image patch type of each image patch includes: Obtain the geometric structure of any target image patch, and determine the initial image patch type of the target image patch based on the geometric structure; Obtain the standard pixel features and standard infrared features of the initial image patch type, and calculate the deviation between the pixel features and infrared features of the target image patch and the standard pixel features and standard infrared features; The similarity between the target image block and the initial image block type is determined based on the deviation between the pixel features and infrared features of the target image block and the standard pixel features and infrared features. The image block type of the target image block is then determined based on the similarity between the target image block and the initial image block type.

[0008] Further, determining the image patch type of the target image patch based on the similarity between the target image patch and the initial image patch type includes: If the similarity between the target image patch and the initial image patch type is greater than a preset similarity threshold, then the initial image patch type is determined as the image patch type of the target image patch; If the similarity between the target image block and the initial image block type is less than or equal to a preset similarity threshold, then the image block types of the remaining image blocks in the preset neighborhood of the target image block are obtained; The similarity between the target image patch and the image patch types of all other image patches in a preset neighborhood is statistically analyzed, and the image patch type of the other image patch with the highest similarity is determined as the image patch type of the target image patch.

[0009] Furthermore, the strategy module determines the multi-dimensional fault characteristics of photovoltaic modules within the site based on the three-dimensional panoramic image, including: The defect distribution of the photovoltaic module is determined based on the pixel features and infrared features of the photovoltaic module image blocks, and the defect features are determined based on the defect distribution of the photovoltaic module. Obtain the current change of the photovoltaic module corresponding to the image block of the photovoltaic module, and determine the power loss characteristics based on the current change; Multidimensional fault characteristics are determined based on defect characteristics and power loss characteristics.

[0010] Furthermore, the strategy module formulates a site inspection strategy based on multi-dimensional fault characteristics, including: Obtain historical inspection logs of the site, and determine the historical multi-dimensional fault characteristics of photovoltaic modules and the corresponding site inspection strategies based on the historical inspection logs. A training sample set is established based on the historical multidimensional fault characteristics of the photovoltaic modules and the corresponding site inspection strategies. An inspection strategy generation model is then established and trained based on the training sample set. The multi-dimensional fault characteristics of the photovoltaic modules at the current site are input into the trained inspection strategy generation model to obtain the inspection strategy for the current site.

[0011] Furthermore, the inspection module determines the optimal inspection route based on the station inspection strategy, including: Based on the current site inspection strategy, the photovoltaic modules to be inspected are identified, and multiple inspection routes are generated in the three-dimensional panoramic map based on the photovoltaic modules to be inspected. Efficiency assessments are conducted for each inspection route, and the optimal inspection route is determined based on the results.

[0012] Furthermore, the efficiency evaluation of each inspection route includes: The inspection time and fault type of each photovoltaic module in each inspection route are statistically analyzed, and the photovoltaic modules are clustered according to the inspection time and fault type. Based on the clustering results, the drone loss type of each photovoltaic module is determined, an efficiency coefficient is assigned to each drone loss type, and the efficiency of each inspection route is evaluated based on the efficiency coefficient.

[0013] Furthermore, the efficiency evaluation of each inspection route based on the efficiency coefficient includes: The inspection route is divided into several inspection sections, and an efficiency requirement coefficient is assigned to each inspection section according to the order of inspection. Calculate the deviation between the efficiency coefficient and the efficiency requirement coefficient of each photovoltaic module to be inspected within the inspection interval, and determine the comprehensive efficiency coefficient of the inspection route based on the deviation between the efficiency coefficient and the efficiency requirement coefficient of all photovoltaic modules to be inspected.

[0014] On the other hand, the present invention also provides an integrated method for intelligent inspection and fault diagnosis of photovoltaic power plants using unmanned aerial vehicles (UAVs), comprising: Collect image data of photovoltaic power stations, and create a three-dimensional panoramic image of the photovoltaic power stations based on the image data; Based on the 3D panoramic image, determine the multi-dimensional fault characteristics of the photovoltaic modules in the site, and formulate a site inspection strategy based on the multi-dimensional fault characteristics. The optimal inspection route is determined based on the site inspection strategy, and then intelligent drone inspection is carried out based on the optimal inspection route.

[0015] The integrated system and method for intelligent inspection and fault diagnosis of photovoltaic power plants provided by this invention establishes a three-dimensional panoramic map as a spatial base, and integrates visible light pixel features and infrared temperature features on this basis to achieve accurate classification of various areas within the photovoltaic power plant and multi-dimensional feature extraction of photovoltaic module faults. Based on this, the optimal inspection strategy and route with the best efficiency are dynamically generated, thereby achieving efficient, accurate and intelligent inspection and fault diagnosis of photovoltaic power plants. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in this 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 some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0017] Figure 1 This is a schematic diagram of the integrated intelligent inspection and fault diagnosis system for photovoltaic power plants provided in an embodiment of the present invention; Figure 2 This is a flowchart illustrating the integrated method for intelligent inspection and fault diagnosis of photovoltaic power plants using unmanned aerial vehicles (UAVs) provided in this embodiment of the invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0019] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this invention, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0020] Figure 1 This is a schematic diagram of the integrated intelligent inspection and fault diagnosis system for photovoltaic power plants provided in this embodiment of the invention, including: The system includes a module for collecting image data of photovoltaic power plants and creating a 3D panoramic view of the photovoltaic power plants based on the image data; a strategy module for determining the multi-dimensional fault characteristics of photovoltaic modules within the power plant based on the 3D panoramic view and formulating a power plant inspection strategy based on the multi-dimensional fault characteristics; and an inspection module for determining the optimal inspection route based on the power plant inspection strategy and conducting intelligent drone inspections based on the optimal inspection route.

[0021] In this embodiment, high-definition cameras inside the photovoltaic power station collect image data of the photovoltaic power station. A three-dimensional panoramic image containing the geometric structure and visual texture of each area within the photovoltaic power station is established using the photovoltaic power station image data. Multi-dimensional fault features are extracted from the photovoltaic module features in the three-dimensional panoramic image. Based on the multi-dimensional fault features, an inspection strategy for each module is formulated and the optimal inspection route is selected. Then, the drone is controlled to perform inspections according to the optimal inspection route.

[0022] In some embodiments of this application, the establishment module establishes a three-dimensional panoramic image of the photovoltaic power station based on the photovoltaic power station image data, including: performing edge detection on the photovoltaic power station image data, extracting multiple image blocks based on the edge detection results and extracting pixel features of the image blocks; acquiring infrared images of the photovoltaic power station, extracting infrared features of the image blocks, determining image blending features of the image blocks based on the pixel features and the infrared features; matching the image blending features with the image blocks to obtain the image block type of each image block; counting the image block types of all image blocks, and establishing a three-dimensional panoramic image based on the image block types of all image blocks.

[0023] In this embodiment, the Canny operator is used to perform edge detection on the photovoltaic power station image data, segmenting and extracting the image into multiple independent, physically meaningful image blocks. Each image block corresponds to a photovoltaic module or vegetation or equipment within the photovoltaic power station. Simultaneously, pixel features of each image block are extracted. Pixel features include local binary pattern texture features and color histogram statistical features. Infrared features of each image block are extracted by acquiring infrared images of the photovoltaic power station. Infrared features include mean temperature, standard deviation of temperature, and histogram of temperature gradient direction. Furthermore, the pixel features and infrared features of the same image block are combined to obtain image blending features. Based on the image blending features, the image block type of the image block is determined, and the image block types and their positions in three-dimensional space of all image blocks within the power station area are statistically analyzed. A three-dimensional sparse point cloud is obtained based on a multi-view stereo vision algorithm. Each image block is projected and fused into this three-dimensional point cloud, thereby establishing a three-dimensional panoramic image of the photovoltaic power station. This panoramic image is a semantic map, providing high-dimensional environmental cognition for subsequent strategy generation.

[0024] In some embodiments of this application, matching the image blending features with image blocks to obtain the image block type of each image block includes: obtaining the geometric structure of any target image block; determining the initial image block type of the target image block based on the geometric structure; obtaining standard pixel features and standard infrared features of the initial image block type; calculating the deviation between the pixel features and infrared features of the target image block and the standard pixel features and standard infrared features; determining the similarity between the target image block and the initial image block type based on the deviation between the pixel features and infrared features of the target image block and the standard pixel features and standard infrared features; and determining the image block type of the target image block based on the similarity between the target image block and the initial image block type.

[0025] In this embodiment, for any target image patch, its geometric structure obtained from edge detection, such as rectangularity, aspect ratio, and internal grid line density, is analyzed to classify it into an initial image patch type. For example, an image patch with high rectangularity and regular internal grid lines is a priori determined to be a "component class," rather than vegetation or other equipment. Then, from a pre-built standard feature library, the standard pixel feature vectors and standard infrared feature vectors corresponding to all candidate image patch types under this "component class" are retrieved. The deviation is obtained by calculating the Euclidean distance between the pixel features and infrared features of the target image patch and the standard features of each candidate type, and the similarity is determined based on the deviation.

[0026] In some embodiments of this application, determining the image block type of the target image block based on the similarity between the target image block and the initial image block type includes: if the similarity between the target image block and the initial image block type is greater than a preset similarity threshold, then determining the initial image block type as the image block type of the target image block; if the similarity between the target image block and the initial image block type is less than or equal to the preset similarity threshold, then obtaining the image block types of the remaining image blocks in a preset neighborhood of the target image block; calculating the similarity between the target image block and the image block types of each of the remaining image blocks in the preset neighborhood, and determining the image block type of the remaining image block with the highest similarity as the image block type of the target image block.

[0027] In this embodiment, when the similarity between the target image block and the initial image block type is less than or equal to a preset similarity threshold, the target image block type is set by collecting the remaining image blocks within a preset neighborhood range in the 3D panoramic image and using the similarity between the remaining image blocks and the target image block.

[0028] In some embodiments of this application, the strategy module determines the multidimensional fault characteristics of photovoltaic modules in the power station based on the three-dimensional panoramic image, including: determining the defect distribution of photovoltaic modules based on the pixel features and infrared features of photovoltaic module image blocks, and determining defect characteristics based on the defect distribution of photovoltaic modules; obtaining the current change of photovoltaic modules corresponding to the photovoltaic module image blocks, and determining the power loss characteristics based on the current change; and determining multidimensional fault characteristics based on the defect characteristics and power loss characteristics.

[0029] In this embodiment, the distribution of defects is further extracted from the pixel and infrared features of the image blocks. Pixel features are used to determine the shadow area and crack area of ​​the photovoltaic module, while infrared features are used to determine the area of ​​the hot spot region and the difference between the highest and average temperatures, thus obtaining defect features describing the physical morphology of the fault. Real-time IV characteristic curves and maximum power point tracking data of the photovoltaic module are obtained through an interface. The system analyzes current changes, such as calculating the decay rate of the fill factor and the degree of mismatch loss indicated by the inflection point in the current-voltage curve, to obtain power loss features. The defect features and power loss features are then combined to obtain multi-dimensional fault features.

[0030] In some embodiments of this application, the strategy module formulates a site inspection strategy based on multi-dimensional fault characteristics, including: obtaining historical inspection logs of the site; determining historical multi-dimensional fault characteristics of photovoltaic modules and corresponding site inspection strategies based on the historical inspection logs; establishing a training sample set based on the historical multi-dimensional fault characteristics of photovoltaic modules and corresponding site inspection strategies; establishing and training an inspection strategy generation model based on the training sample set; and inputting the multi-dimensional fault characteristics of photovoltaic modules at the current site into the trained inspection strategy generation model to obtain the current site inspection strategy.

[0031] In this embodiment, a neural network model is constructed as the inspection strategy generation model. This model is a multilayer perceptron model. The input layer of the model is a multidimensional fault feature vector, and the output layer is the probability distribution of all possible inspection strategy categories. The model is trained using the backpropagation algorithm, enabling it to learn the mapping relationship between fault severity and type to the optimal response strategy.

[0032] In some embodiments of this application, the inspection module determines the optimal inspection route according to the site inspection strategy, including: determining the photovoltaic modules to be inspected according to the current site inspection strategy; generating multiple inspection routes based on the photovoltaic modules to be inspected in a three-dimensional panoramic image; evaluating the efficiency of each inspection route; and determining the optimal inspection route based on the efficiency evaluation results.

[0033] In this embodiment, multiple initial inspection routes are generated based on the photovoltaic modules to be inspected according to the inspection strategy, and the optimal inspection route is obtained through efficiency evaluation.

[0034] In some embodiments of this application, the efficiency evaluation of each inspection route includes: statistically analyzing the inspection time and fault type of each photovoltaic module in each inspection route, clustering the photovoltaic modules according to the inspection time and fault type; determining the drone loss type of each photovoltaic module according to the clustering results, assigning an efficiency coefficient to each drone loss type, and evaluating the efficiency of each inspection route according to the efficiency coefficient.

[0035] In this embodiment, by statistically analyzing the required inspection time and fault types for each photovoltaic module to be inspected along the inspection route, all photovoltaic modules to be inspected are clustered into several task clusters based on the K-Means clustering algorithm. For example, there are long-duration-high-severity clusters, medium-duration-medium-severity clusters, and short-duration-low-severity clusters. Based on the clustering results, a drone loss type is defined for each cluster. The drone loss type is a quantitative label measuring the consumption of flight resources, such as high-power hovering loss, medium-power cruise loss, and low-power inspection loss. An efficiency coefficient is assigned to each drone loss type. This efficiency coefficient is a positive value; the larger the value, the higher the loss of drone endurance, time, and other resources per unit time when performing this type of task. The efficiency of each inspection route is then evaluated based on the efficiency coefficient.

[0036] In some embodiments of this application, the step of evaluating the efficiency of each inspection route based on the efficiency coefficient includes: dividing the inspection route into several inspection intervals, assigning an efficiency requirement coefficient to each inspection interval according to the order of the inspection intervals; calculating the deviation between the efficiency coefficient and the efficiency requirement coefficient of each photovoltaic module to be inspected within the inspection interval, and determining the comprehensive efficiency coefficient of the inspection route based on the deviation between the efficiency coefficient and the efficiency requirement coefficient of all photovoltaic modules to be inspected.

[0037] In this embodiment, the inspection route to be evaluated is divided into several consecutive inspection intervals. For example, every five consecutive inspection points constitute one interval. According to the execution logic of the inspection tasks, tasks with high efficiency coefficients should generally be prioritized to achieve early detection, early diagnosis, and reduce the risk of accident escalation. Therefore, based on the order of the inspection intervals in the entire route, higher efficiency requirement coefficients are assigned to earlier inspection intervals, and lower efficiency requirement coefficients are assigned to later intervals. For each photovoltaic module to be inspected in the inspection route, the absolute difference between the efficiency coefficient and the efficiency requirement coefficient is calculated. If a task with a high efficiency coefficient is arranged in an earlier interval with a high efficiency requirement coefficient, the deviation is small and the efficiency is high; conversely, if a high efficiency coefficient task is arranged in a later interval, the deviation is large and the efficiency is low. Finally, by summing up the deviations of all components to be inspected along the entire route, the overall efficiency coefficient of the route is obtained. The smaller the overall efficiency coefficient, the higher the degree of matching between the task execution order and the resource optimization allocation principle, the earlier serious faults can be handled, and the overall efficiency and safety are optimal. Therefore, the inspection route with the smallest overall efficiency coefficient is selected as the optimal inspection route.

[0038] Based on the same general inventive concept, this invention also protects an integrated method for intelligent inspection and fault diagnosis of photovoltaic power plants using unmanned aerial vehicles (UAVs).

[0039] Figure 2 This is a flowchart illustrating the integrated method for intelligent inspection and fault diagnosis of photovoltaic power plants using unmanned aerial vehicles (UAVs) provided in this embodiment of the invention. Figure 2 As shown in this embodiment, the integrated method for intelligent inspection and fault diagnosis of photovoltaic power plants using unmanned aerial vehicles includes: S101, Collect photovoltaic power station image data, and establish a three-dimensional panoramic view of the photovoltaic power station based on the photovoltaic power station image data; S102, determine the multi-dimensional fault characteristics of photovoltaic modules in the site based on the three-dimensional panoramic image, and formulate a site inspection strategy based on the multi-dimensional fault characteristics; S103 determines the optimal inspection route based on the site inspection strategy and conducts intelligent drone inspections based on the optimal inspection route.

[0040] It should be noted that all relevant information that may be involved in the various embodiments of the present invention is processed in strict accordance with the requirements of laws and regulations, following the principles of legality, legitimacy, and necessity, based on the reasonable purpose of the business scenario, and is information that users actively provide or generate during the use of the product / service, as well as information obtained with user authorization.

[0041] The information processed by this invention may vary depending on the specific product / service scenario and should be based on the specific scenario in which the user uses the product / service. This may involve user account information, device information, or other related information. This invention will treat the relevant information and its processing with the utmost diligence.

[0042] This invention places great emphasis on the security of relevant information and has adopted reasonable and feasible security protection measures that comply with industry standards to protect user information and prevent unauthorized access, public disclosure, use, modification, damage or loss of relevant information.

[0043] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0044] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0045] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An integrated system for intelligent unmanned aerial vehicle (UAV) inspection and fault diagnosis of photovoltaic power plants, characterized in that, include: A module is established to collect image data of photovoltaic power stations and to create a three-dimensional panoramic image of the photovoltaic power station based on the image data. The strategy module is used to determine the multi-dimensional fault characteristics of photovoltaic modules in the site based on the 3D panoramic image, and to formulate site inspection strategies based on the multi-dimensional fault characteristics. The inspection module is used to determine the optimal inspection route based on the site inspection strategy, and to perform intelligent drone inspections based on the optimal inspection route.

2. The integrated system for intelligent inspection and fault diagnosis of photovoltaic power plants using unmanned aerial vehicles (UAVs) according to claim 1, characterized in that, The module establishes a three-dimensional panoramic view of the photovoltaic power station based on the photovoltaic power station image data, including: Edge detection is performed on the photovoltaic power station image data, and multiple image blocks are extracted based on the edge detection results, and the pixel features of the image blocks are extracted; Infrared images of photovoltaic power plants are acquired, infrared features of the image blocks are extracted, and image blending features of the image blocks are determined based on the pixel features and the infrared features. The image blending features are matched with image blocks to obtain the image block type of each image block; The image patch types of all image patches are identified, and a 3D panoramic image is created based on the image patch types of all image patches.

3. The integrated system for intelligent inspection and fault diagnosis of photovoltaic power plants using unmanned aerial vehicles (UAVs) according to claim 2, characterized in that, The process of matching the image blending features with image patches to obtain the image patch type for each image patch includes: Obtain the geometric structure of any target image patch, and determine the initial image patch type of the target image patch based on the geometric structure; Obtain the standard pixel features and standard infrared features of the initial image patch type, and calculate the deviation between the pixel features and infrared features of the target image patch and the standard pixel features and standard infrared features; The similarity between the target image block and the initial image block type is determined based on the deviation between the pixel features and infrared features of the target image block and the standard pixel features and infrared features. The image block type of the target image block is then determined based on the similarity between the target image block and the initial image block type.

4. The integrated intelligent inspection and fault diagnosis system for photovoltaic power plants using unmanned aerial vehicles (UAVs) according to claim 3, characterized in that, The step of determining the image patch type of the target image patch based on the similarity between the target image patch and the initial image patch type includes: If the similarity between the target image patch and the initial image patch type is greater than a preset similarity threshold, then the initial image patch type is determined as the image patch type of the target image patch; If the similarity between the target image block and the initial image block type is less than or equal to a preset similarity threshold, then the image block types of the remaining image blocks in the preset neighborhood of the target image block are obtained; The similarity between the target image patch and the image patch types of all other image patches in a preset neighborhood is statistically analyzed, and the image patch type of the other image patch with the highest similarity is determined as the image patch type of the target image patch.

5. The integrated system for intelligent inspection and fault diagnosis of photovoltaic power plants using unmanned aerial vehicles (UAVs) according to claim 1, characterized in that, The strategy module determines the multi-dimensional fault characteristics of photovoltaic modules within the site based on the 3D panoramic image, including: The defect distribution of the photovoltaic module is determined based on the pixel features and infrared features of the photovoltaic module image blocks, and the defect features are determined based on the defect distribution of the photovoltaic module. Obtain the current change of the photovoltaic module corresponding to the image block of the photovoltaic module, and determine the power loss characteristics based on the current change; Multidimensional fault characteristics are determined based on defect characteristics and power loss characteristics.

6. The integrated system for intelligent inspection and fault diagnosis of photovoltaic power plants using unmanned aerial vehicles (UAVs) according to claim 5, characterized in that, The strategy module formulates a site inspection strategy based on multi-dimensional fault characteristics, including: Obtain historical inspection logs of the site, and determine the historical multi-dimensional fault characteristics of photovoltaic modules and the corresponding site inspection strategies based on the historical inspection logs. A training sample set is established based on the historical multidimensional fault characteristics of the photovoltaic modules and the corresponding site inspection strategies. An inspection strategy generation model is then established and trained based on the training sample set. The multi-dimensional fault characteristics of the photovoltaic modules at the current site are input into the trained inspection strategy generation model to obtain the inspection strategy for the current site.

7. The integrated system for intelligent inspection and fault diagnosis of photovoltaic power plants using unmanned aerial vehicles (UAVs) according to claim 1, characterized in that, The inspection module determines the optimal inspection route based on the station inspection strategy, including: Based on the current site inspection strategy, the photovoltaic modules to be inspected are identified, and multiple inspection routes are generated in the three-dimensional panoramic map based on the photovoltaic modules to be inspected. Efficiency assessments are conducted for each inspection route, and the optimal inspection route is determined based on the results.

8. The integrated system for intelligent inspection and fault diagnosis of photovoltaic power plants using unmanned aerial vehicles (UAVs) according to claim 7, characterized in that, The efficiency evaluation of each inspection route includes: The inspection time and fault type of each photovoltaic module in each inspection route are statistically analyzed, and the photovoltaic modules are clustered according to the inspection time and fault type. Based on the clustering results, the drone loss type of each photovoltaic module is determined, an efficiency coefficient is assigned to each drone loss type, and the efficiency of each inspection route is evaluated based on the efficiency coefficient.

9. The integrated system for intelligent inspection and fault diagnosis of photovoltaic power plants using unmanned aerial vehicles (UAVs) according to claim 8, characterized in that, The efficiency evaluation of each inspection route based on the efficiency coefficient includes: The inspection route is divided into several inspection sections, and an efficiency requirement coefficient is assigned to each inspection section according to the order of inspection. Calculate the deviation between the efficiency coefficient and the efficiency requirement coefficient of each photovoltaic module to be inspected within the inspection interval, and determine the comprehensive efficiency coefficient of the inspection route based on the deviation between the efficiency coefficient and the efficiency requirement coefficient of all photovoltaic modules to be inspected.

10. A method for integrating unmanned aerial vehicle (UAV) intelligent inspection and fault diagnosis of photovoltaic power plants, characterized in that, include: Collect image data of photovoltaic power stations, and create a three-dimensional panoramic image of the photovoltaic power stations based on the image data; Based on the 3D panoramic image, determine the multi-dimensional fault characteristics of the photovoltaic modules in the site, and formulate a site inspection strategy based on the multi-dimensional fault characteristics. The optimal inspection route is determined based on the site inspection strategy, and then intelligent drone inspection is carried out based on the optimal inspection route.

Images