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

Abstract

The invention provides a photovoltaic module cleaning method and cleaning system based on unmanned aerial vehicle inspection, and the method comprises the following steps: S10, calling the coordinates and numbers of all ideal photovoltaic arrays; s20, segmenting the image of the photovoltaic panel; step S30, photovoltaic array matching is carried out; step S40, photovoltaic array segmentation is carried out; step S50, mapping the photovoltaic module; and S60, receiving data and sending an instruction. According to the technical scheme, the problems that in the prior art, the positioning precision of stains is low, and rapid and accurate cleaning cannot be achieved are effectively solved.

Description

Technical Field

[0001] The present application relates to the technical field of photovoltaic solar panels, and in particular to a photovoltaic component cleaning method and cleaning system based on drone inspection. Background Art

[0002] With the rapid development of the global solar photovoltaic industry, the number of large-scale photovoltaic panels in countries around the world is increasing. Photovoltaic power stations are mostly built in places with long sunshine hours, abundant sunshine, open areas and no obstructions. However, this also makes the photovoltaic array exposed to the outdoors for a long time. After a long period of wind and rain, temperature difference between day and night, sun exposure and sand erosion, as well as extreme weather such as hail, the photovoltaic array will accelerate aging and damage. At the same time, the relatively complex internal factors of photovoltaic panel production process and the external factors of uncertainty during installation will also affect the quality of photovoltaic arrays. Foreign objects such as buildings, leaves, and bird droppings will also have an impact on the photovoltaic system that cannot be ignored. If not cleaned in time, it will cause local hot spots and damage the photovoltaic panels.

[0003] Some existing cleaning methods and systems (for example, application number: 202310865517.6, named as a method and system for automatically cleaning photovoltaic electric fields based on drones) determine the working space according to the cleaning and inspection task area of ​​the photovoltaic electric field, and use drones to automatically photograph the working space globally; extract the photovoltaic array from the global image, accurately calculate the GPS value of the photovoltaic array in the working area, and generate a secondary high-precision shooting route for the drone; according to the secondary high-precision shooting route of the drone, use the drone to take secondary images of the working space, and perform stain and defect detection based on the deep learning defect detection model, and generate a cleaning report and cleaning instructions; based on the cleaning instructions, use a drone with a high-pressure nozzle to automatically navigate, identify, and clean according to the detected stain location. In this cleaning system, the drone takes a global photograph to determine the coordinates of the photovoltaic array once, and takes a second shot to determine the position to be cleaned. However, in actual applications, there will be a situation where only part of the photovoltaic array has stains, and the photovoltaic array still needs to be cleaned as a whole. Therefore, this cleaning method has low positioning accuracy for stains and high cleaning costs. Summary of the invention

[0004] A technical problem to be solved by the present application is to improve the accuracy of stain positioning and achieve fast and accurate cleaning.

[0005] In order to solve the above technical problems, the present application provides a photovoltaic module cleaning method based on drone inspection, comprising the following steps:

[0006] Step S10: Retrieving the coordinates and numbers of all ideal photovoltaic arrays;

[0007] Step S20: Photovoltaic panel image segmentation, using a drone to perform image segmentation on all photovoltaic panel areas, and after segmentation, multiple groups of actual photovoltaic arrays are obtained;

[0008] Step S30: Photovoltaic array matching, matching all actual photovoltaic arrays with all ideal photovoltaic arrays, and determining the coordinates and numbers of all actual photovoltaic arrays;

[0009] Step S40: Photovoltaic array segmentation, using a drone to perform image segmentation on a group of actual photovoltaic arrays, obtaining multiple photovoltaic modules after segmentation, and determining the coordinates of the center points of the multiple photovoltaic modules;

[0010] Step S50: Photovoltaic component mapping, by establishing a logical mapping relationship with the photovoltaic power station layout planning map, to obtain the center point coordinates of all photovoltaic components segmented from all actual photovoltaic arrays;

[0011] Step S60: Data reception and command sending. The central control module receives the collected data sent back from the drone in real time or periodically, and sends cleaning instructions to the cleaning robot according to the collected data.

[0012] In some embodiments, photovoltaic array matching includes the following steps:

[0013] Step S301: the drone determines a group of actual photovoltaic arrays, and calculates the GPS coordinates of four valid vertices of the group of actual photovoltaic arrays according to the drone navigation information and camera internal parameters;

[0014] Step S302: determining an ideal photovoltaic array corresponding to the group of actual photovoltaic arrays through the GPS coordinates of the four effective vertices of the group of actual photovoltaic arrays;

[0015] Step S303: Match all actual photovoltaic arrays with all ideal photovoltaic arrays, determine the GPS coordinates and row and column information of the four vertices of all actual photovoltaic arrays, and obtain matching results.

[0016] In some embodiments, the photovoltaic array segmentation specifically performs uniform segmentation and post-processing on the group of actual photovoltaic arrays, obtains multiple photovoltaic components in the infrared image, and determines the coordinates of the center points of the multiple photovoltaic components.

[0017] In some embodiments, photovoltaic assembly mapping includes the following steps:

[0018] Step S501: along the straight line direction of the group of actual photovoltaic arrays, obtain the center point coordinates of a plurality of photovoltaic modules segmented from other actual photovoltaic arrays in the straight line direction;

[0019] Step S502: Based on multiple actual photovoltaic arrays in a straight line direction, a logical mapping relationship with the photovoltaic power station layout planning map is established according to the matching results to obtain the center point coordinates of all photovoltaic modules segmented from all actual photovoltaic arrays.

[0020] In some embodiments, data receiving and command sending include the following steps:

[0021] Step S601: The drone automatically patrols along a preset route, takes photos and videos, and uses image recognition technology to analyze the surface condition of the photovoltaic modules and identify abnormal conditions such as stains or faults;

[0022] Step S602: The drone transmits the collected data back to the central control module in real time or at a fixed time;

[0023] Step S603: The central control module determines whether the photovoltaic module needs to be cleaned through an algorithm, and based on the determination result, sends the cleaning instruction, the precise coordinates of the cleaning area, the cleaning degree requirement, and the planned cleaning path to the corresponding cleaning robot;

[0024] Step S604: The cleaning robot performs the cleaning task and can adjust the cleaning strategy according to actual conditions during the cleaning process.

[0025] In some embodiments, retrieving the coordinates and numbers of all ideal photovoltaic arrays specifically involves retrieving the GPS coordinates and row and column numbers of four vertices of each ideal photovoltaic array from a constructed photovoltaic power station laying planning map database.

[0026] In some embodiments, photovoltaic panel image segmentation is specifically to segment and post-process all photovoltaic panel areas based on a deep learning-based photovoltaic panel segmentation network model to obtain multiple groups of actual photovoltaic arrays in the infrared image.

[0027] The present application also provides a photovoltaic module cleaning system based on drone inspection, including: a central control module, a drone, a cleaning robot and a path planning and optimization module. The central control module is used to adaptively adjust the cleaning path and parameters according to real-time environmental changes and task requirements. The drone is equipped with a visual recognition module, which is in communication connection with the central control module to transmit the detected information. The cleaning robot is in communication connection with the central control module, and the cleaning robot makes adjustments according to the designation issued by the central control module. The path planning and optimization module is in communication connection with the central control module, and the path planning and optimization module makes optimal path adjustments according to the information transmitted by the central control module. The path planning and optimization module is in communication connection with the drone and the cleaning robot, and plans the navigation path of the drone and the cleaning path of the cleaning robot.

[0028] In some embodiments, the path planning and optimization module includes a cleaning planning unit, an intelligent optimization unit and an autonomous positioning unit. The cleaning planning unit realizes cleaning path planning through a neural network model. The intelligent optimization unit can realize intelligent control of the cleaning robot based on a fuzzy logic control algorithm of deep reinforcement learning. The autonomous positioning unit realizes the positioning of photovoltaic modules through a satellite system.

[0029] In some embodiments, the visual recognition module includes a camera and a multispectral imaging system, the camera is used to identify the position, shape and state of the photovoltaic module, and the multispectral imaging system is used to detect the type and degree of dirt on the surface of the photovoltaic module in real time.

[0030] Through the above technical scheme, the photovoltaic module cleaning method based on drone inspection provided by the present application includes multiple steps of retrieving the coordinates and numbers of all ideal photovoltaic arrays, photovoltaic panel image segmentation, photovoltaic array matching, photovoltaic array segmentation, photovoltaic module mapping, and data reception and instruction sending, wherein the photovoltaic panel area image segmentation can obtain the photovoltaic array, and the photovoltaic array segmentation can obtain multiple photovoltaic modules. The central control module sends cleaning instructions to the cleaning robot based on the collected data transmitted by the drone. The cleaning robot uses the photovoltaic module as the cleaning basis during cleaning, can achieve accurate positioning of the cleaning position, and can also reduce the cleaning area and reduce the cleaning cost. The technical scheme of the present application effectively solves the problem of low positioning accuracy of stains in the prior art and the inability to achieve fast and accurate cleaning. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0032] Figure 1 It is a flow chart of a photovoltaic module cleaning method based on drone inspection according to an embodiment of the present application;

[0033] Figure 2 This is a schematic diagram of a specific process of photovoltaic array matching in an embodiment of the present application;

[0034] Figure 3 This is a schematic diagram of a specific process of photovoltaic component mapping according to an embodiment of the present application;

[0035] Figure 4 It is a schematic diagram of a specific flow of data receiving and command sending in an embodiment of the present application;

[0036] Figure 5It is a structural schematic diagram of a photovoltaic module cleaning system based on drone inspection according to an embodiment of the present application.

[0037] The above drawings include the following reference numerals:

[0038] 10. Central control module; 20. Drone; 21. Visual recognition module; 30. Cleaning robot; 40. Path planning and optimization module. DETAILED DESCRIPTION

[0039] The following is a further detailed description of the implementation methods of the present application in conjunction with the accompanying drawings and examples. The detailed descriptions of the following examples and the accompanying drawings are used to exemplarily illustrate the principles of the present application, but cannot be used to limit the scope of the present application. The present application can be implemented in many different forms and is not limited to the specific embodiments disclosed herein, but includes all technical solutions that fall within the scope of the claims.

[0040] The present application provides these embodiments to make the present application thorough and complete, and to fully express the scope of the present application to those skilled in the art. It should be noted that unless otherwise specifically stated, the relative arrangement of the parts and steps, the composition of the materials, the numerical expressions and the numerical values ​​set forth in these embodiments should be interpreted as being merely exemplary, and not as limiting.

[0041] It should be noted that, in the description of this application, unless otherwise specified, the meaning of "multiple" is greater than or equal to two; the terms "upper", "lower", "left", "right", "inner", "outer", etc., indicating the orientation or positional relationship, are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on this application. When the absolute position of the object being described changes, the relative positional relationship may also change accordingly.

[0042] In addition, the words "first", "second" and similar words used in this application do not indicate any order, quantity or importance, but are only used to distinguish different parts. "Vertical" does not mean vertical in the strict sense, but is within the tolerance range. "Parallel" does not mean parallel in the strict sense, but is within the tolerance range. "Include" or "comprising" and similar words mean that the elements before the word include the elements listed after the word, and do not exclude the possibility of including other elements.

[0043] It should also be noted that in the description of this application, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be directly connected or indirectly connected through an intermediate medium. For ordinary technicians in this field, the specific meanings of the above terms in this application can be understood according to the specific circumstances. When a specific device is described as being located between a first device and a second device, there may or may not be an intermediate device between the specific device and the first device or the second device.

[0044] All terms used in this application have the same meaning as those understood by those of ordinary skill in the art to which this application belongs, unless otherwise specifically defined. It should also be understood that terms defined in general dictionaries, such as general dictionaries, should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and should not be interpreted in an idealized or extremely formal sense, unless explicitly defined herein.

[0045] Technologies, methods, and equipment known to ordinary technicians in the relevant art may not be discussed in detail, but where appropriate, the technologies, methods, and equipment should be considered part of the specification.

[0046] like Figure 1 As shown, the embodiment relates to a photovoltaic module cleaning method based on drone inspection, comprising the following steps:

[0047] Step S10: Retrieving the coordinates and numbers of all ideal photovoltaic arrays;

[0048] Step S20: Photovoltaic panel image segmentation, using the drone 20 to perform image segmentation on all photovoltaic panel areas, and after segmentation, multiple groups of actual photovoltaic arrays are obtained;

[0049] Step S30: Photovoltaic array matching, matching all actual photovoltaic arrays with all ideal photovoltaic arrays, and determining the coordinates and numbers of all actual photovoltaic arrays;

[0050] Step S40: Photovoltaic array segmentation, using the drone 20 to perform image segmentation on a group of actual photovoltaic arrays, obtaining multiple photovoltaic modules after segmentation, and determining the coordinates of the center points of the multiple photovoltaic modules;

[0051] Step S50: Photovoltaic component mapping, by establishing a logical mapping relationship with the photovoltaic power station layout planning map, to obtain the center point coordinates of all photovoltaic components segmented from all actual photovoltaic arrays;

[0052] Step S60: Data reception and command sending. The central control module 10 receives the collected data sent back from the drone 20 in real time or periodically, and sends a cleaning command to the cleaning robot 30 according to the collected data.

[0053] Through the above technical scheme, the photovoltaic module cleaning method based on drone inspection provided in this embodiment includes multiple steps of retrieving the coordinates and numbers of all ideal photovoltaic arrays, photovoltaic panel image segmentation, photovoltaic array matching, photovoltaic array segmentation, photovoltaic module mapping, and data reception and instruction sending, wherein the photovoltaic panel area image segmentation can obtain the photovoltaic array, and the photovoltaic array segmentation can obtain multiple photovoltaic modules. The central control module sends cleaning instructions to the cleaning robot based on the collected data transmitted by the drone. The cleaning robot uses the photovoltaic module as the cleaning basis during cleaning, can achieve accurate positioning of the cleaning position, and can also reduce the cleaning area and reduce the cleaning cost. The technical scheme of this embodiment effectively solves the problem of low positioning accuracy of stains in the prior art and the inability to achieve fast and accurate cleaning.

[0054] It should be noted that when photovoltaic panels are initially laid, they need to be planned according to the ground area of ​​the actual environment. After the planning is completed, a photovoltaic power station laying planning map will be constructed and stored in the database. However, in actual laying, the ground may be potholes or the soil conditions may not be suitable for laying photovoltaic panels, so it is impossible to be completely consistent with the photovoltaic power station laying planning map. The photovoltaic power station laying planning map includes the coordinates and numbers of all ideal photovoltaic arrays. All ideal photovoltaic arrays can fully cover all actual photovoltaic arrays. Due to the influence of actual laying conditions, some ideal photovoltaic arrays may not be actually laid. Retrieving the coordinates and numbers of all ideal photovoltaic arrays is specifically to retrieve the four vertex GPS coordinates and row and column numbers of each ideal photovoltaic array from the constructed photovoltaic power station laying planning map database.

[0055] In some embodiments, the photovoltaic panel image segmentation is specifically to segment and post-process all photovoltaic panel areas based on a deep learning photovoltaic panel segmentation network model to obtain multiple groups of actual photovoltaic arrays in the infrared image. The drone 20 collects infrared images along the actual photovoltaic arrays and divides the photovoltaic panel area into multiple actual photovoltaic arrays through a deep learning photovoltaic panel segmentation network model. The collected infrared images are post-processed, for example, image enhancement is performed to improve the clarity of the infrared images and image segmentation is performed to distinguish between the photovoltaic panel area and the ground. The drone 20 transmits the collected information to the central control module 10.

[0056] like Figure 2 As shown, in some embodiments, photovoltaic array matching includes the following steps:

[0057] Step S301: the drone 20 determines a group of actual photovoltaic arrays, and calculates the GPS coordinates of four valid vertices of the group of actual photovoltaic arrays according to the navigation information of the drone 20 and the camera internal parameters;

[0058] Step S302: determining an ideal photovoltaic array corresponding to the group of actual photovoltaic arrays through the GPS coordinates of the four effective vertices of the group of actual photovoltaic arrays;

[0059] Step S303: Match all actual photovoltaic arrays with all ideal photovoltaic arrays, determine the GPS coordinates and row and column information of the four vertices of all actual photovoltaic arrays, and obtain matching results.

[0060] The initial position of the drone 20 is a determined position in the photovoltaic power station laying planning map, and the determined position has coordinates and row and column information. The drone 20 navigates from the initial position to a group of actual photovoltaic arrays. The GPS coordinates of the four effective vertices of the actual photovoltaic array can be calculated based on the navigation distance and direction (i.e., the navigation information of the drone 20) and the camera internal parameters. Therefore, the ideal photovoltaic array corresponding to the actual photovoltaic array can be found in the photovoltaic power station laying planning map according to the GPS coordinates of the four effective vertices of the actual photovoltaic array. After the drone 20 collects all the actual photovoltaic arrays, it generates an actual photovoltaic power station laying map, compares the photovoltaic power station laying planning map with the photovoltaic power station laying actual map, and overlaps the group of actual photovoltaic arrays with the corresponding group of ideal photovoltaic arrays to complete a logical mapping. At this time, other actual photovoltaic arrays can find the corresponding ideal photovoltaic arrays, thereby obtaining the GPS coordinates and row and column information of the four vertices of all actual photovoltaic arrays, that is, obtaining the matching result.

[0061] In some embodiments, the photovoltaic array segmentation specifically performs uniform segmentation and post-processing on the group of actual photovoltaic arrays, obtains multiple photovoltaic modules in the infrared image, and determines the center point coordinates of the multiple photovoltaic modules. The infrared acquisition image of a group of actual photovoltaic arrays is post-processed such as image enhancement, and uniformly segmented to obtain multiple photovoltaic modules. Since the GPS coordinates and row and column information of the four vertices of the group of actual photovoltaic arrays have been determined, the center point coordinates of the multiple photovoltaic modules obtained after uniform segmentation can be determined.

[0062] like Figure 3 As shown, in some embodiments, photovoltaic component mapping includes the following steps:

[0063] Step S501: along the straight line direction of the group of actual photovoltaic arrays, obtain the center point coordinates of a plurality of photovoltaic modules segmented from other actual photovoltaic arrays in the straight line direction;

[0064] Step S502: Based on multiple actual photovoltaic arrays in a straight line direction, a logical mapping relationship with the photovoltaic power station layout planning map is established according to the matching results to obtain the center point coordinates of all photovoltaic modules segmented from all actual photovoltaic arrays.

[0065] According to the GPS coordinates and row and column information of the four vertices of the group of actual photovoltaic arrays, the center point coordinates of the photovoltaic components segmented from the actual photovoltaic array can be obtained. Along the straight line direction of the group of actual photovoltaic arrays, the center point coordinates of multiple photovoltaic components segmented from other actual photovoltaic arrays in the straight line direction can be obtained. Based on the multiple actual photovoltaic arrays in the straight line direction, the photovoltaic power station layout planning map and the photovoltaic power station layout actual map are compared to complete the secondary logical mapping. At this time, the center point coordinates of all photovoltaic components segmented from all actual photovoltaic arrays can be obtained.

[0066] like Figure 4 As shown, in some embodiments, data receiving and instruction sending include the following steps:

[0067] Step S601: The drone 20 automatically patrols along a preset route, takes photos and videos, and uses image recognition technology to analyze the surface condition of the photovoltaic module and identify abnormal conditions such as stains or faults;

[0068] Step S602: The drone 20 transmits the collected data back to the central control module 10 in real time or at a fixed time;

[0069] Step S603: the central control module 10 determines whether the photovoltaic module needs to be cleaned through an algorithm, and based on the determination result, sends the cleaning instruction, the precise coordinates of the cleaning area, the cleaning degree requirement and the planned cleaning path to the corresponding cleaning robot 30;

[0070] Step S604: the cleaning robot 30 performs the cleaning task and can adjust the cleaning strategy according to actual conditions during the cleaning process.

[0071] The drone 20 transmits the collected information to the central control module 10. The central control module 10 can find the coordinates of the center point of the nearest photovoltaic component based on the pixel position of the photovoltaic component with stains or failures displayed on the infrared image, and then accurately locate the photovoltaic component with stains or failures. Based on the analysis results, the central control module 10 calls the nearest cleaning robot 30 for cleaning, and sends the cleaning instructions, the precise coordinates of the cleaning area, the cleaning degree requirements, and the planned cleaning path to the corresponding cleaning robot 30. The cleaning robot 30 automatically identifies the cleaning degree, and when the cleaning does not meet the standard, it goes back to clean again, and then returns according to the original planned path. The cleaning robot 30 is a ground cleaning robot.

[0072] It should be noted that for the same photovoltaic power station, there is no need to repeat the steps of photovoltaic panel image segmentation, photovoltaic array matching, photovoltaic array segmentation and photovoltaic component mapping. When the drone 20 collects a photovoltaic component that needs to be cleaned, the precise coordinates of the photovoltaic component can be directly obtained. Because the navigation path of the drone 20 is fixed, the photovoltaic component that arrives at a certain time is also fixed, and the coordinates of the photovoltaic component can be directly retrieved from the database.

[0073] like Figure 5 As shown, the embodiment also relates to a photovoltaic module cleaning system based on drone inspection. The photovoltaic module cleaning system based on drone inspection includes: a central control module 10, a drone 20, a cleaning robot 30 and a path planning and optimization module 40. The central control module 10 is used to adaptively adjust the cleaning path and parameters according to real-time environmental changes and task requirements. The central control module 10 calls the nearest cleaning robot 30 for cleaning. If the image analysis determines that the stain range is large, multiple cleaning robots 30 can be called for cleaning. The drone 20 is equipped with a visual recognition module 21, and the visual recognition module 21 is connected to the central control module 10 for communication and transmission of detected information. The cleaning robot 30 is connected to the central control module 10 for communication, and the cleaning robot 30 is adjusted according to the designation issued by the central control module 10. The path planning and optimization module 40 is connected to the central control module 10 for communication, and the path planning and optimization module 40 adjusts the best path according to the information transmitted by the central control module 10. The path planning and optimization module 40 is connected to the drone 20 and the cleaning robot 30 for communication, and the navigation path of the drone 20 and the cleaning path of the cleaning robot 30 are planned.

[0074] It should be noted that the cleaning robot 30 is provided with a detection sensor, which can detect whether there are obstacles on the cleaning path, such as animals staying there, and transmit the detection results to the path planning and optimization module 40 and the central control module 10. The path planning and optimization module 40 re-plans the path or the central control module 10 calls the cleaning robot at other locations.

[0075] It should also be noted that the path planning and optimization module 40 plans the navigation path of the drone 20 according to the locations of the photovoltaic components that are cleaned more frequently, so that information can be collected with emphasis on the photovoltaic components that are cleaned more frequently.

[0076] In some embodiments, the path planning and optimization module 40 includes a cleaning planning unit, an intelligent optimization unit and an autonomous positioning unit. The cleaning planning unit implements cleaning path planning through a neural network model, the intelligent optimization unit can realize intelligent control of the cleaning robot 30 based on a fuzzy logic control algorithm of deep reinforcement learning, and the autonomous positioning unit realizes the positioning of the photovoltaic module through a satellite system. The cleaning robot 30 is intelligently called according to the cleaning path planning, and the working experience and environmental data of the cleaning robot 30 are continuously learned and optimized through reinforcement learning and data-driven optimization methods.

[0077] In some embodiments, the visual recognition module 21 includes a camera and a multispectral imaging system, and the camera is used to identify the position, shape and state of the photovoltaic module. Therefore, it is not only possible to clean the photovoltaic module, but also to determine whether the photovoltaic module is in a faulty state, which is convenient for timely maintenance. The multispectral imaging system is used to detect the type and degree of dirt on the surface of the photovoltaic module in real time. During the navigation of the drone 20, the multispectral imaging system performs real-time detection.

[0078] So far, various embodiments of the present application have been described in detail. In order to avoid obscuring the concept of the present application, some details known in the art are not described. Based on the above description, those skilled in the art can fully understand how to implement the technical solution disclosed herein.

[0079] Although some specific embodiments of the present application have been described in detail by way of examples, it should be understood by those skilled in the art that the above examples are for illustration only and are not intended to limit the scope of the present application. It should be understood by those skilled in the art that the above embodiments may be modified or some technical features may be replaced by equivalents without departing from the scope and spirit of the present application. In particular, the various technical features mentioned in the various embodiments may be combined in any manner as long as there is no structural conflict.

Claims

1. A photovoltaic module cleaning method based on drone inspection, characterized in that: The following steps are involved: step S10: Retrieve the coordinates and numbers of all ideal photovoltaic arrays; Step S20: Photovoltaic panel image segmentation, using the drone (20) to perform image segmentation on all photovoltaic panel areas, and after segmentation, multiple groups of actual photovoltaic arrays are obtained; Step S30: Photovoltaic array matching, matching all the actual photovoltaic arrays with all the ideal photovoltaic arrays, and determining the coordinates and numbers of all the actual photovoltaic arrays; Step S40: Photovoltaic array segmentation, using the drone (20) to perform image segmentation on a group of actual photovoltaic arrays, obtaining a plurality of photovoltaic modules after segmentation, and determining the coordinates of the center points of the plurality of photovoltaic modules; Step S50: Photovoltaic component mapping, by establishing a logical mapping relationship with the photovoltaic power station layout planning map, to obtain the center point coordinates of all the photovoltaic components segmented from all the actual photovoltaic arrays; Step S60: Data reception and command sending, the central control module (10) receives the collected data transmitted back from the drone (20) in real time or periodically, and sends a cleaning command to the cleaning robot (30) based on the collected data.

2. The photovoltaic module cleaning method based on drone inspection according to claim 1 is characterized in that: The photovoltaic array matching comprises the following steps: Step S301: the drone (20) determines a group of actual photovoltaic arrays, and calculates the GPS coordinates of four valid vertices of the group of actual photovoltaic arrays according to the navigation information of the drone (20) and the camera internal parameters; Step S302: determining the ideal photovoltaic array corresponding to the actual photovoltaic array of the group through the GPS coordinates of the four valid vertices of the actual photovoltaic array of the group; Step S303: matching all the actual photovoltaic arrays with all the ideal photovoltaic arrays, determining the GPS coordinates and row and column information of the four vertices of all the actual photovoltaic arrays, and obtaining a matching result.

3. The photovoltaic module cleaning method based on drone inspection according to claim 2 is characterized in that: The photovoltaic array segmentation specifically includes uniformly segmenting and post-processing the group of actual photovoltaic arrays, acquiring multiple photovoltaic components in the infrared image, and determining the coordinates of the center points of the multiple photovoltaic components.

4. The photovoltaic module cleaning method based on drone inspection according to claim 3 is characterized in that: The photovoltaic module mapping comprises the following steps: Step S501: along the straight line direction of the group of actual photovoltaic arrays, obtaining the center point coordinates of a plurality of photovoltaic modules segmented from other actual photovoltaic arrays in the straight line direction; Step S502: Based on the multiple actual photovoltaic arrays in the straight line direction, and according to the matching results, a logical mapping relationship with the photovoltaic power station layout planning map is established to obtain the center point coordinates of all the photovoltaic components segmented from all the actual photovoltaic arrays.

5. The photovoltaic module cleaning method based on drone inspection according to claim 1 is characterized in that: The data receiving and instruction sending comprises the following steps: Step S601: the drone (20) automatically patrols along a preset route, takes photos and videos, and uses image recognition technology to analyze the surface condition of the photovoltaic module to identify abnormal conditions such as stains or faults; Step S602: the drone (20) transmits the collected data back to the central control module (10) in real time or at a fixed time; Step S603: the central control module (10) determines whether the photovoltaic module needs to be cleaned through an algorithm, and based on the determination result, sends the cleaning instruction, the precise coordinates of the cleaning area, the cleaning degree requirement and the planned cleaning path to the corresponding cleaning robot (30); Step S604: The cleaning robot (30) performs the cleaning task and can adjust the cleaning strategy according to actual conditions during the cleaning process.

6. The photovoltaic module cleaning method based on drone inspection according to claim 1 is characterized in that: The method of retrieving the coordinates and numbers of all the ideal photovoltaic arrays is specifically to retrieve the GPS coordinates and row and column numbers of the four vertices of each of the ideal photovoltaic arrays from the constructed photovoltaic power station laying planning map database.

7. The photovoltaic module cleaning method based on drone inspection according to claim 1 is characterized in that: The photovoltaic panel image segmentation is specifically to segment and post-process all the photovoltaic panel areas based on a deep learning photovoltaic panel segmentation network model to obtain multiple groups of actual photovoltaic arrays in the infrared image.

8. A photovoltaic module cleaning system based on drone inspection, characterized in that: The photovoltaic module cleaning system based on drone inspection is applied to the photovoltaic module cleaning method based on drone inspection according to any one of claims 1 to 7, and the photovoltaic module cleaning system based on drone inspection comprises: A central control module (10), the central control module (10) being used to adaptively adjust the cleaning path and parameters according to real-time environmental changes and task requirements; A drone (20), wherein the drone (20) is equipped with a visual recognition module (21), and the visual recognition module (21) is communicatively connected to the central control module (10) to transmit detected information; A cleaning robot (30), the cleaning robot (30) being in communication connection with the central control module (10), the cleaning robot (30) being adjusted according to the designation issued by the central control module (10); A path planning and optimization module (40), wherein the path planning and optimization module (40) is communicatively connected to the central control module (10), wherein the path planning and optimization module (40) performs optimal path adjustment according to information transmitted by the central control module (10), and wherein the path planning and optimization module (40) is communicatively connected to the drone (20) and the cleaning robot (30), and plans the navigation path of the drone (20) and the cleaning path of the cleaning robot (30).

9. The photovoltaic module cleaning system based on drone inspection according to claim 8, characterized in that: The path planning and optimization module (40) comprises a cleaning planning unit, an intelligent optimization unit and an autonomous positioning unit. The cleaning planning unit implements cleaning path planning through a neural network model. The intelligent optimization unit can realize intelligent control of the cleaning robot (30) based on a fuzzy logic control algorithm of deep reinforcement learning. The autonomous positioning unit realizes the positioning of the photovoltaic module through a satellite system.

10. The photovoltaic module cleaning system based on drone inspection according to claim 8, characterized in that: The visual recognition module (21) comprises a camera and a multispectral imaging system, wherein the camera is used to identify the position, shape and state of the photovoltaic module, and the multispectral imaging system is used to detect the type and degree of dirt on the surface of the photovoltaic module in real time.

Images