Photovoltaic module inspection method and inspection system based on unmanned aerial vehicle

Abstract

The invention discloses an unmanned aerial vehicle-based photovoltaic module inspection method and inspection system, and belongs to the technical field of photovoltaic equipment monitoring, and the method comprises the steps: determining a photovoltaic power generation region to be inspected, dividing the photovoltaic power generation region into a plurality of optimal inspection sub-regions, and distributing an inspection sub-region for each unmanned aerial vehicle, the method comprises the following steps: collecting inspection sub-areas of photovoltaic equipment, generating an optimal inspection path corresponding to the inspection sub-areas, installing a visible light camera and an infrared thermal imager on an inspection unmanned aerial vehicle, shooting the photovoltaic equipment in the inspection sub-areas, collecting visible light images and infrared images, carrying out multi-source data feature collection and fusion on the shot visible light images and infrared images, and carrying out multi-source data feature fusion on the photovoltaic equipment in the inspection sub-areas; according to the method, the multi-dimensional feature map is generated, the multi-dimensional feature map is analyzed, the fault type is identified, the fault report and the maintenance site moving route are generated according to the fault type, the maintenance personnel maintain the target fault equipment according to the maintenance work order and the maintenance site moving route, the inspection efficiency can be improved, the fault judgment accuracy can be improved, and the operation and maintenance cost can be reduced.

Description

Technical Field

[0001] This application belongs to the field of photovoltaic equipment monitoring technology, specifically, it relates to a method and system for inspecting photovoltaic modules based on unmanned aerial vehicles (UAVs). Background Technology

[0002] To address climate change, countries are vigorously developing renewable energy. Photovoltaic energy, as a clean and renewable energy source, has been widely applied and developed globally, with the scale of photovoltaic power plants and the number of photovoltaic modules continuously increasing. Photovoltaic modules are exposed to the outdoors for extended periods, making them susceptible to environmental factors such as sunlight, temperature, and wind erosion, leading to faults such as hot spots, microcracks, and shading. Regular inspections are necessary to ensure the power plant's efficiency and operational safety. Traditional manual inspections are limited by terrain and the number of modules, resulting in low efficiency, high cost, and numerous blind spots. With the development of deep learning technologies and drone hardware, drone inspection technology is gradually becoming the mainstream approach for photovoltaic module inspection. Drones replace manual inspections, automatically inspecting modules by carrying infrared imagers and collecting infrared images for fault detection.

[0003] Current technology involves controlling drones to conduct inspections and take infrared or electroluminescent photographs along pre-planned routes. However, due to limitations in the number of drones and battery capacity, the total range of all drones has an upper limit. To ensure that no equipment is missed during inspections, the drones' flight altitude is increased. However, increasing the drones' flight altitude affects the quality of the captured images, resulting in low accuracy in fault identification. At the same time, existing inspection routes are not comprehensive and are prone to route conflicts or duplicate inspections, further reducing the total range of the drones. Increasing the drones' flight altitude makes it difficult to complete the inspection task. Summary of the Invention

[0004] To address the aforementioned problems and technical deficiencies, this application adopts the following technical solution: a method for inspecting photovoltaic modules based on unmanned aerial vehicles (UAVs), comprising the following steps: The photovoltaic power generation area to be inspected is determined and divided into multiple optimal inspection sub-areas. Each drone is assigned an inspection sub-area and the optimal inspection path corresponding to the inspection sub-area is generated. Visible light cameras and infrared thermal imagers are installed on inspection drones to photograph photovoltaic equipment within their respective inspection sub-areas, collecting visible light and infrared images. Multi-source data feature acquisition and fusion are performed on captured visible light and infrared images to generate multi-dimensional feature maps. The multi-dimensional feature maps are then analyzed to identify fault types. A fault report and a repair site movement route are generated based on the fault type. Repair personnel then repair the target faulty equipment according to the repair work order and the repair site movement route.

[0005] Preferably, the process of generating the optimal inspection path includes: Obtain maps of photovoltaic power plants and layout information of photovoltaic equipment, and collect geometric features of photovoltaic equipment arrays, including row and column coordinates, tilt angle, azimuth angle and shading status of photovoltaic equipment; The improved genetic algorithm is used to divide the region, and the geometric features of the photovoltaic equipment array are used as optimization parameters to output multiple optimal inspection sub-regions. Then, taking the coverage of all photovoltaic devices and minimizing the path as the objective function, the flight time and inspection speed of the inspection drone as constraints, and the row and column coordinates of the photovoltaic devices as coordinate points, the optimal inspection path is output.

[0006] Furthermore, the inspection drone is equipped with a dynamic path adjustment mechanism, which includes: Inspection drones build a self-organizing network based on a distributed communication protocol, sharing location and status information. When replanning their routes, they maintain a safe distance from each other and avoid flight path conflicts. When a drone is unable to complete the remaining route of the inspection path, it is instructed to return to base, and its unfinished inspection tasks are dynamically assigned to other drones in the ad hoc network that have completed their tasks or have sufficient power. At the same time, the coordinates of the uninspected equipment and the images already collected are handed over.

[0007] Furthermore, during the inspection process, the inspection drone uploads its location in real time. When the inspection drone deviates from the inspection path due to an accident and the deviation distance reaches a threshold, it can replan the optimal inspection path of the inspection drone in real time based on real-time obstacle and wind speed data.

[0008] Preferably, the fault type identification includes using an improved semantic segmentation network to enhance the extraction of multi-dimensional feature maps, achieving feature map segmentation, and inputting the segmented feature image segments into a neural network classifier for diagnosis, thereby providing the fault type and realizing fault identification and classification.

[0009] Furthermore, the generation of the multi-dimensional feature map includes: Local binary mode and HSV color space analysis were used to extract component texture entropy and color moment features from visible light images; Temperature gradient and hotspot area features in infrared images were extracted using adaptive thresholding segmentation and morphological operations. Based on the time stamp synchronization of two types of images, the SIFT feature point matching algorithm is then used to align the component outlines of the infrared image and the visible light image. The two types of image features, after spatiotemporal alignment, are fused using a weighted fusion algorithm to generate a multi-dimensional feature map.

[0010] Preferably, the fault report includes the fault type, severity, maintenance work order, precise latitude and longitude coordinates of the target faulty device, and its row and column position in the array; Based on Dijkstra's algorithm, and taking into account the current location of maintenance personnel, the distance between faulty equipment, the distribution of on-site passages, and the severity of the fault, the optimal on-site movement route for maintenance personnel is planned on the photovoltaic power station map. Maintenance personnel then carry out maintenance on the target faulty equipment according to the maintenance work order and the on-site movement route.

[0011] An inspection system for photovoltaic modules based on unmanned aerial vehicles (UAVs) includes: The route generation module is used to determine the photovoltaic power generation area to be inspected and divide it into multiple optimal inspection sub-areas. It assigns an inspection sub-area to each drone and generates the optimal inspection path corresponding to the inspection sub-area. The inspection and imaging module is installed on the inspection drone to photograph the photovoltaic equipment in the designated inspection sub-area and collect visible light and infrared images. The fault identification module is used to acquire and fuse multi-source data features from captured visible light and infrared images to generate multi-dimensional feature maps, analyze the multi-dimensional feature maps, and identify fault types. The fault repair module generates fault reports and repair site movement routes based on the fault type. Repair personnel then repair the target faulty equipment according to the repair work order and the repair site movement route.

[0012] An electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the program to implement the content of an inspection method for photovoltaic modules based on a drone as described above.

[0013] A computer-readable storage medium storing a computer program thereon, characterized in that, when executed by a processor, the computer program implements the content of an inspection method for photovoltaic modules based on a drone, as described above.

[0014] Compared to existing technologies, the beneficial effects of this application are as follows: (1) This application improves the genetic algorithm, divides the optimal sub-regions by using the geometric features of photovoltaic equipment as optimization parameters, and generates paths by combining the drone's endurance and speed constraints, thereby reducing path redundancy. It also avoids route conflicts by sharing location and status information through a self-organizing network. When the drone cannot complete the task, it can dynamically allocate unfinished tasks and transfer data to ensure that the inspection is not interrupted, ensure full coverage of equipment inspection, and improve inspection efficiency. (2) This application generates multi-dimensional feature maps by extracting features from multiple sources, SIFT alignment and weighted fusion, and combined with improved semantic segmentation network and neural network classifier, which can quickly diagnose a variety of faults, greatly reduce the rate of missed faults and false faults, and then carry out on-site mobile route planning, integrate the severity level of faults and the distribution of on-site channels, provide the optimal path for maintenance personnel, shorten maintenance response time and reduce operation and maintenance costs. Attached Figure Description

[0015] In the attached diagram: Figure 1 This is a schematic diagram of the method steps in an embodiment of this application; Figure 2 This is a schematic diagram of the system structure according to an embodiment of this application. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some embodiments of this application, but not all embodiments. Generally, the components of the embodiments of this application described and shown in the accompanying drawings can be arranged and designed in various different configurations. Example

[0017] like Figure 1 As shown, a method for inspecting photovoltaic modules based on drones includes the following steps: The photovoltaic power generation area to be inspected is determined and divided into multiple optimal inspection sub-areas. Each drone is assigned an inspection sub-area and the optimal inspection path corresponding to the inspection sub-area is generated. The process of generating the optimal inspection path includes: Obtain maps of photovoltaic power plants and layout information of photovoltaic equipment, and collect geometric features of photovoltaic equipment arrays, including row and column coordinates, tilt angle, azimuth angle and shading status of photovoltaic equipment; The improved genetic algorithm is used to divide the region, and the geometric features of the photovoltaic equipment array are used as optimization parameters to output multiple optimal inspection sub-regions. Then, taking the coverage of all photovoltaic devices and minimizing the path as the objective function, the flight time and inspection speed of the inspection drone as constraints, and the row and column coordinates of the photovoltaic devices as coordinate points, the optimal inspection path is output.

[0018] The inspection drone is equipped with a dynamic path adjustment mechanism, which includes: Inspection drones build a self-organizing network based on a distributed communication protocol, sharing location and status information. When replanning their routes, they maintain a safe distance from each other and avoid flight path conflicts. When a drone is unable to complete the remaining route of the inspection path, it is instructed to return to base, and its unfinished inspection tasks are dynamically assigned to other drones in the ad hoc network that have completed their tasks or have sufficient power. At the same time, the coordinates of the uninspected equipment and the images already collected are handed over.

[0019] During the inspection process, the inspection drone uploads its location in real time. When the inspection drone deviates from the inspection path due to an accident and the deviation distance reaches a threshold, the drone can replan the optimal inspection path in real time based on real-time obstacle and wind speed data.

[0020] Visible light cameras and infrared thermal imagers are installed on inspection drones to photograph photovoltaic equipment within their respective inspection sub-areas, collecting visible light and infrared images. Multi-source data feature acquisition and fusion are performed on captured visible light and infrared images to generate multi-dimensional feature maps. The multi-dimensional feature maps are then analyzed to identify fault types. Fault type identification involves using an improved semantic segmentation network to enhance the extraction of multi-dimensional feature maps, achieving feature map segmentation. The segmented feature image segments are then input into a neural network classifier for diagnosis, which identifies the fault type and classifies the fault.

[0021] Generating multi-dimensional feature maps includes: Local binary mode and HSV color space analysis were used to extract component texture entropy and color moment features from visible light images; Temperature gradient and hotspot area features in infrared images were extracted using adaptive thresholding segmentation and morphological operations. Based on the time stamp synchronization of two types of images, the SIFT feature point matching algorithm is then used to align the component outlines of the infrared image and the visible light image. The two types of image features, after spatiotemporal alignment, are fused using a weighted fusion algorithm to generate a multi-dimensional feature map.

[0022] A fault report and a repair site movement route are generated based on the fault type. Repair personnel then repair the target faulty equipment according to the repair work order and the repair site movement route.

[0023] The fault report includes the fault type, severity, repair work order, precise latitude and longitude coordinates of the target faulty device, and its row and column position in the array; Based on Dijkstra's algorithm, and taking into account the current location of maintenance personnel, the distance between faulty equipment, the distribution of on-site passages, and the severity of the fault, the optimal on-site movement route for maintenance personnel is planned on the photovoltaic power station map. Maintenance personnel then carry out maintenance on the target faulty equipment according to the maintenance work order and the on-site movement route. Example

[0024] like Figure 2 As shown, a drone-based photovoltaic module inspection system includes: The route generation module is used to determine the photovoltaic power generation area to be inspected and divide it into multiple optimal inspection sub-areas. It assigns an inspection sub-area to each drone and generates the optimal inspection path corresponding to the inspection sub-area. The process of generating the optimal inspection path includes: Obtain maps of photovoltaic power plants and layout information of photovoltaic equipment, and collect geometric features of photovoltaic equipment arrays, including row and column coordinates, tilt angle, azimuth angle and shading status of photovoltaic equipment; The improved genetic algorithm is used to divide the region, and the geometric features of the photovoltaic equipment array are used as optimization parameters to output multiple optimal inspection sub-regions. Then, taking the coverage of all photovoltaic devices and minimizing the path as the objective function, the flight time and inspection speed of the inspection drone as constraints, and the row and column coordinates of the photovoltaic devices as coordinate points, the optimal inspection path is output.

[0025] The inspection drone is equipped with a dynamic path adjustment mechanism, which includes: Inspection drones build a self-organizing network based on a distributed communication protocol, sharing location and status information. When replanning their routes, they maintain a safe distance from each other and avoid flight path conflicts. When a drone is unable to complete the remaining route of the inspection path, it is instructed to return to base, and its unfinished inspection tasks are dynamically assigned to other drones in the ad hoc network that have completed their tasks or have sufficient power. At the same time, the coordinates of the uninspected equipment and the images already collected are handed over.

[0026] During the inspection process, the inspection drone uploads its location in real time. When the inspection drone deviates from the inspection path due to an accident and the deviation distance reaches a threshold, the drone can replan the optimal inspection path in real time based on real-time obstacle and wind speed data.

[0027] The inspection and imaging module is installed on the inspection drone to photograph the photovoltaic equipment in the designated inspection sub-area and collect visible light and infrared images. The fault identification module is used to acquire and fuse multi-source data features from captured visible light and infrared images to generate multi-dimensional feature maps, analyze the multi-dimensional feature maps, and identify fault types. Fault type identification involves using an improved semantic segmentation network to enhance the extraction of multi-dimensional feature maps, achieving feature map segmentation. The segmented feature image segments are then input into a neural network classifier for diagnosis, which identifies the fault type and classifies the fault.

[0028] Generating multi-dimensional feature maps includes: Local binary mode and HSV color space analysis were used to extract component texture entropy and color moment features from visible light images; Temperature gradient and hotspot area features in infrared images were extracted using adaptive thresholding segmentation and morphological operations. Based on the time stamp synchronization of two types of images, the SIFT feature point matching algorithm is then used to align the component outlines of the infrared image and the visible light image. The two types of image features, after spatiotemporal alignment, are fused using a weighted fusion algorithm to generate a multi-dimensional feature map.

[0029] The fault repair module generates fault reports and repair site movement routes based on the fault type. Repair personnel then repair the target faulty equipment according to the repair work order and the repair site movement route.

[0030] The fault report includes the fault type, severity, repair work order, precise latitude and longitude coordinates of the target faulty device, and its row and column position in the array; Based on Dijkstra's algorithm, and taking into account the current location of maintenance personnel, the distance between faulty equipment, the distribution of on-site passages, and the severity of the fault, the optimal on-site movement route for maintenance personnel is planned on the photovoltaic power station map. Maintenance personnel then carry out maintenance on the target faulty equipment according to the maintenance work order and the on-site movement route. Example

[0031] From a hardware perspective, this application provides an embodiment of an electronic device comprising all or part of a method for inspecting photovoltaic modules based on unmanned aerial vehicles (UAVs). The electronic device includes a service processor and a distributed memory. The service processor is connected to the memory, and the distributed memory stores a service self-management program configured to store machine-readable instructions. The service processor executes the service self-management program, and the instructions, when executed by the processor, implement the method for inspecting photovoltaic modules based on UAVs as described above. Example

[0032] The embodiments of this application also provide a computer-readable storage medium capable of implementing a drone-based photovoltaic module inspection method with a server or client as the execution subject in the above embodiments. The computer-readable storage medium stores a computer program that, when executed by a processor, implements all the contents of the drone-based photovoltaic module inspection method with a server or client as the execution subject in the above embodiments.

[0033] The embodiments described above are merely preferred embodiments of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this application. It should be noted that those skilled in the art can make various modifications, improvements, and substitutions without departing from the concept of this application, and these all fall within the protection scope of this application.

Claims

1. A method for inspecting photovoltaic modules based on unmanned aerial vehicles (UAVs), characterized in that, Includes the following steps: The photovoltaic power generation area to be inspected is determined and divided into multiple optimal inspection sub-areas. Each drone is assigned an inspection sub-area and the optimal inspection path corresponding to the inspection sub-area is generated. Visible light cameras and infrared thermal imagers are installed on inspection drones to photograph photovoltaic equipment within their respective inspection sub-areas, collecting visible light and infrared images. Multi-source data feature acquisition and fusion are performed on captured visible light and infrared images to generate multi-dimensional feature maps. The multi-dimensional feature maps are then analyzed to identify fault types. A fault report and a repair site movement route are generated based on the fault type. Repair personnel then repair the target faulty equipment according to the repair work order and the repair site movement route.

2. The method for inspecting photovoltaic modules based on unmanned aerial vehicles (UAVs) according to claim 1, characterized in that, The process of generating the optimal inspection path includes: Obtain maps of photovoltaic power plants and layout information of photovoltaic equipment, and collect geometric features of photovoltaic equipment arrays, including row and column coordinates, tilt angle, azimuth angle and shading status of photovoltaic equipment; The improved genetic algorithm is used to divide the region, and the geometric features of the photovoltaic equipment array are used as optimization parameters to output multiple optimal inspection sub-regions. Then, taking the coverage of all photovoltaic devices and minimizing the path as the objective function, the flight time and inspection speed of the inspection drone as constraints, and the row and column coordinates of the photovoltaic devices as coordinate points, the optimal inspection path is output.

3. The method for inspecting photovoltaic modules based on unmanned aerial vehicles (UAVs) according to claim 2, characterized in that, The inspection drone is equipped with a dynamic path adjustment mechanism, which includes: Inspection drones build a self-organizing network based on a distributed communication protocol, sharing location and status information. When replanning their routes, they maintain a safe distance from each other and avoid flight path conflicts. When a drone is unable to complete the remaining route of the inspection path, it is instructed to return to base, and its unfinished inspection tasks are dynamically assigned to other drones in the ad hoc network that have completed their tasks or have sufficient power. At the same time, the coordinates of the uninspected equipment and the images already collected are handed over.

4. The method for inspecting photovoltaic modules based on unmanned aerial vehicles (UAVs) according to claim 3, characterized in that, During the inspection process, the inspection drone uploads its location in real time. When the inspection drone deviates from the inspection path due to an accident and the deviation distance reaches a threshold, it can replan the optimal inspection path of the inspection drone in real time based on real-time obstacle and wind speed data.

5. The method for inspecting photovoltaic modules based on unmanned aerial vehicles (UAVs) according to claim 1, characterized in that, The fault type identification includes using an improved semantic segmentation network to enhance the extraction of multi-dimensional feature maps, achieving feature map segmentation, and inputting the segmented feature image segments into a neural network classifier for diagnosis, thereby identifying and classifying the fault type.

6. The method for inspecting photovoltaic modules based on unmanned aerial vehicles (UAVs) according to claim 5, characterized in that, The generation of the multi-dimensional feature map includes: Local binary mode and HSV color space analysis were used to extract component texture entropy and color moment features from visible light images; Temperature gradient and hotspot area features in infrared images were extracted using adaptive thresholding segmentation and morphological operations. Based on the time stamp synchronization of two types of images, the SIFT feature point matching algorithm is then used to align the component outlines of the infrared image and the visible light image. The two types of image features, after spatiotemporal alignment, are fused using a weighted fusion algorithm to generate a multi-dimensional feature map.

7. The method for inspecting photovoltaic modules based on unmanned aerial vehicles (UAVs) according to claim 1, characterized in that, The fault report includes the fault type, severity, repair work order, precise latitude and longitude coordinates of the target faulty device, and its row and column position in the array; Based on Dijkstra's algorithm, and taking into account the current location of maintenance personnel, the distance between faulty equipment, the distribution of on-site passages, and the severity of the fault, the optimal on-site movement route for maintenance personnel is planned on the photovoltaic power station map. Maintenance personnel then carry out maintenance on the target faulty equipment according to the maintenance work order and the on-site movement route.

8. A photovoltaic module inspection system based on unmanned aerial vehicles (UAVs), characterized in that, include: The route generation module is used to determine the photovoltaic power generation area to be inspected and divide it into multiple optimal inspection sub-areas. It assigns an inspection sub-area to each drone and generates the optimal inspection path corresponding to the inspection sub-area. The inspection and imaging module is installed on the inspection drone to photograph the photovoltaic equipment in the designated inspection sub-area and collect visible light and infrared images. The fault identification module is used to acquire and fuse multi-source data features from captured visible light and infrared images to generate multi-dimensional feature maps, analyze the multi-dimensional feature maps, and identify fault types. The fault repair module generates fault reports and repair site movement routes based on the fault type. Repair personnel then repair the target faulty equipment according to the repair work order and the repair site movement route.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the content of the inspection method for photovoltaic modules based on unmanned aerial vehicles as described in claim 1.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the content of the inspection method for photovoltaic modules based on drones as described in claim 1.

Images