A solar power station automatic inspection system and method based on artificial intelligence

Abstract

The invention discloses an automatic inspection system and method for a solar power station based on artificial intelligence, and relates to the technical field of inspection of solar power stations. The system comprises: a database module, a remote monitoring module, a path planning module, a drone platform module and a data analysis module; the database module is used to store image data collected during drone inspection; the remote monitoring module is used to remotely monitor the drone; the path planning module plans the inspection path of the drone; the drone platform module is used to inspect the solar power station; the data analysis module is used to analyze the image data collected by the drone to determine the working status of the photovoltaic panel; the invention performs automated inspection by the drone to improve the inspection efficiency; the drone inspection path is planned according to the change of shadow occlusion over time, so as to minimize the influence of shadow occlusion on the image acquisition of the drone during the inspection.

Description

Technical Field

[0001] The present invention relates to the field of solar power station inspection technology, and in particular to an artificial intelligence-based solar power station automatic inspection system and method. Background Art

[0002] In solar power stations, drone inspection is a very important inspection method. Solar power stations are usually located in remote areas with complex terrain and harsh environment. Manual inspection is difficult, costly and inefficient. Drone inspection can quickly and accurately detect the equipment status and environmental conditions of solar power stations, discover faults and safety hazards in time, improve operation and maintenance efficiency, and reduce operation and maintenance costs. At the same time, drone inspection can also provide important data support for the planning, design and optimization of solar power stations. Therefore, in solar power stations, drone inspection has great application value and potential. When drones are inspecting, there will be situations where photovoltaic panels are blocked by shadows of nearby objects, which will affect the photovoltaic panel image data collected by drones, and thus affect the inspection effect of solar power stations. The time for drone inspections is usually several hours. During the drone inspection process, the shadows of objects near the photovoltaic panels will also change over time. Therefore, how to consider the changes in shadows and plan the inspection path of drones has become a problem that needs to be solved. Summary of the invention

[0003] The purpose of the present invention is to provide an artificial intelligence-based solar power station automatic inspection system and method to solve the problems raised in the above background technology.

[0004] In one aspect of the present invention, there is provided an artificial intelligence-based automatic inspection system for solar power stations, comprising a database module, a remote monitoring module, a path planning module, a drone platform module and a data analysis module; the output end of the database module is interconnected with the input end of the data analysis module, and is used to store image data collected during drone inspection; the remote monitoring module is used to remotely monitor the operating status and data transmission of the drone; the output end of the path planning module is interconnected with the input end of the drone platform module, and the inspection path of the drone is planned based on the impact of shadow occlusion on the photovoltaic panel; the output end of the drone platform module is interconnected with the input end of the database module, and is used to inspect the solar power station; the output end of the data analysis module is interconnected with the input end of the path planning module, and is used to analyze the image data collected by the drone and determine the working status of the photovoltaic panel.

[0005] Specifically, the UAV platform module also includes a communication unit, a navigation unit, a control unit, a positioning unit and an image acquisition unit; the communication unit is used to receive instructions from the path planning module and transmit the image data collected by the UAV to the database module; the navigation unit navigates the UAV according to the moving path received by the UAV; the control unit is used to control the flight of the UAV; the positioning unit is used to position the UAV; and the image acquisition unit is used to collect image data of the photovoltaic panel.

[0006] Specifically, the UAV platform module also includes a communication unit, a navigation unit, a control unit, a positioning unit and an image acquisition unit; the communication unit is used to receive instructions from the path planning module and transmit the image data collected by the UAV to the database module; the navigation unit navigates the UAV according to the moving path received by the UAV; the control unit is used to control the flight of the UAV; the positioning unit is used to position the UAV; and the image acquisition unit is used to collect image data of the photovoltaic panel.

[0007] Specifically, the data analysis module also establishes a relationship between shadow occlusion and the time when the drone acquires image data based on the image of the first type of photovoltaic panel when the shadow is blocked and the time when the drone acquires the image data, determines the solar altitude angle and the solar azimuth angle according to the time when the drone acquires the image data, and obtains the relationship between the solar altitude angle and the solar azimuth angle and the shadow occlusion; the first type of photovoltaic panel refers to the photovoltaic panel in the solar power station that will be blocked by shadows.

[0008] Specifically, the path planning module performs local path planning between the first type of photovoltaic panels under the relative position relationship of the first type of photovoltaic panels, and obtains the overall path planning within the solar power station based on the local path planning; the path planning module plans the drone inspection path through a simulated annealing algorithm without considering shadow occlusion, and obtains the relative position of the first type of photovoltaic panels.

[0009] In another aspect of the present invention, there is provided a method for automatic inspection of a solar power station based on artificial intelligence, comprising the following steps:

[0010] S5-1, obtaining historical inspection data of the solar power station; determining the location of each photovoltaic panel in the solar power station;

[0011] S5-2, based on the historical inspection data of the solar power station, determine that the photovoltaic panels that will be blocked by shadows are recorded as the first type of photovoltaic panels, and the photovoltaic panels that will not be blocked by shadows are recorded as the second type of photovoltaic panels, and establish a connection between the shadow blocking situation and the inspection time;

[0012] S5-3, for each first-class photovoltaic panel that will be blocked by shadows, perform local path planning between any two first-class photovoltaic panels;

[0013] S5-4, based on all local path planning, the path planning of the entire solar power station is obtained; directly performing path planning for the entire solar power station is difficult to achieve the purpose of minimizing the impact of shadow occlusion on the first-class photovoltaic panels. Local path planning is performed between the first-class photovoltaic panels, and then path planning is performed for the second-class photovoltaic panels that are not involved in the local path planning, the first first-class photovoltaic panel, and the last photovoltaic panel, so that the path planning of the entire solar power station can be obtained;

[0014] S5-5, inspecting the solar power station according to the path planned by the solar power station in step S5-4, and generating an inspection report.

[0015] Specifically, in step S5-2, establishing the relationship between the shadow blocking situation and the inspection time further includes the following steps:

[0016] S6-1, for each first-type photovoltaic panel, determining an image of the first-type photovoltaic panel when it is shaded and a time when the image is captured;

[0017] S6-2, determining the solar altitude angle and solar azimuth angle at the time of image capture based on the time when the image was captured;

[0018] S6-3, according to the set date for the drone to inspect the solar power station, calculate the solar altitude angle and solar azimuth angle at different times on the inspection date;

[0019] S6-4, based on the solar altitude angle and solar azimuth angle at different times on the inspection date, obtaining the shading conditions of the first type of photovoltaic panels at different times on the inspection date;

[0020] By continuously acquiring images of the first type of photovoltaic panels within a day, it is possible to obtain periodic images of the first type of photovoltaic panels affected by shadow occlusion; the size and direction of the shadow of the obstruction change periodically with time because the position, direction, altitude and angle of the sun change with time; at the same time in different seasons, the position, direction, altitude and angle of the sun are not exactly the same, while the solar altitude angle and solar azimuth angle can reflect the position, direction, altitude and angle of the sun, and the solar altitude angle and solar azimuth angle can be calculated according to time, thereby obtaining the occlusion conditions of the first type of photovoltaic panels at different times on the inspection date.

[0021] Specifically, the local path planning between any two first-type photovoltaic panels also includes the step of determining the second-type photovoltaic panel between the two first-type photovoltaic panels, and the specific steps include:

[0022] S7-1, number all photovoltaic panel positions, use simulated annealing algorithm to perform shortest path planning, the path planning requires traversing each photovoltaic panel without repetition, by adjusting the initial temperature of the simulated annealing algorithm, changing the output result of the simulated annealing algorithm, increasing the number of shortest paths obtained by the simulated annealing algorithm, and k represents the number of relative position relationships of the first type of photovoltaic panels corresponding to the shortest path;

[0023] In the relative position relationship of the first type of photovoltaic panels corresponding to different shortest paths obtained by the simulated annealing algorithm, the relative position relationship of the first type of photovoltaic panels may be the same, and k is less than or equal to the number of shortest paths; the simulated annealing algorithm is a heuristic algorithm and cannot guarantee a global optimal solution, that is, the true shortest path. Therefore, by adjusting the initial temperature of the simulated annealing algorithm, different path planning results can be obtained;

[0024] S7-2, for each shortest path obtained in step S7-1, the relative position relationship path of the first type of photovoltaic panels corresponding to the shortest path i , i is a positive integer between [1, k], and the first type of photovoltaic panels are recorded as n is the number of photovoltaic panels of the first type;

[0025] S7-3, for i=1, 2, ..., k, execute step S7-4;

[0026] S7-4, for counting unit j=1, 2, ..., n-1, execute step S7-5;

[0027] S7-5, the relative position relationship of the first type of photovoltaic panels corresponding to the shortest path path1, path2, ..., path n Find the first type of photovoltaic panel and The second type of photovoltaic panels between, as elements stored in the collection middle; Indicates the first type of photovoltaic panels and and and The second type of photovoltaic panels between the formation of the collection, different collection The elements in the set are allowed to be repeated. There are no identical elements in the same set. Delete identical elements in the same set and record the number of repetitions of the identical elements in the same set.

[0028] In order to perform local path planning between adjacent first-type photovoltaic panels, it is necessary to determine the second-type photovoltaic panels included between the adjacent first-type photovoltaic panels; since there are multiple situations of relative position relationships between the first-type photovoltaic panels, it is necessary to determine the second-type photovoltaic panels included between the adjacent first-type photovoltaic panels under each relative position relationship;

[0029] Specifically, performing local path planning between two first-type photovoltaic panels further includes the following steps:

[0030] S8-1, for i=1, 2, ..., k, execute steps S8-2 to S8-5;

[0031] S8-2, for counting unit j=1, 2, ..., n-1, execute step S8-3;

[0032] S8-3, for the set For each element A in , calculate whether the element A belongs to the set The probability P is Indicates that the element is in the set The number of repetitions in represents the number of repetitions of element A in all sets, Determine whether element A belongs to the set by generating random numbers If so, then element A belongs to the set In addition to the collection Delete element A from other sets except ; if not, element A does not belong to the set Update the number of repetitions of element A in all collections to

[0033] Under a relative position relationship between the first-class photovoltaic panels, the second-class photovoltaic panels are allocated between the adjacent first-class photovoltaic panels according to probability, and the second-class photovoltaic panels are allocated according to the relative position relationship between the first-class photovoltaic panels. In order to avoid duplication of inspection paths, the allocated second-class photovoltaic panels will not participate in subsequent allocations; in order to ensure that all second-class photovoltaic panels can be inspected, the probability of the second-class photovoltaic panels being allocated increases with the increase in the number of sets that have completed allocation, and the maximum value is 1;

[0034] S8-4, after completing step S8-1, the second type of photovoltaic panels included between two adjacent first type photovoltaic panels under the relative position relationship corresponding to different shortest paths are obtained, and the inspection order of the first type photovoltaic panels under the relative position relationship of the first type photovoltaic panels corresponding to the k shortest paths is used, and the simulated annealing algorithm is used to plan the paths of the two adjacent first type photovoltaic panels and the second type photovoltaic panels included between the two adjacent first type photovoltaic panels; the previous photovoltaic panel and the next photovoltaic panel are used to distinguish the two adjacent first type photovoltaic panels, assuming that the previous photovoltaic panel is not blocked by the shadow when the drone arrives at the previous photovoltaic panel, according to the planned path, the unmanned The occlusion of the first type of photovoltaic panels at different times of the drone speed and the inspection date is used to determine whether the next photovoltaic panel is blocked by the shadow when the drone arrives at the next photovoltaic panel. If all the first type of photovoltaic panels are not blocked by the shadow, the local path planning is completed and the initial temperature cycle is exited; if there is a first type of photovoltaic panel that is blocked, the path is replanned by modifying the initial temperature of the simulated annealing algorithm. If the initial temperature of the simulated annealing algorithm is continuously modified and there is still a first type of photovoltaic panel that is blocked, then step S8-5 is entered; if all the first type of photovoltaic panels are not blocked by the shadow after the initial temperature is modified, the local path planning is completed and the initial temperature cycle is exited;

[0035] The simulated annealing algorithm performs path planning for the second-class photovoltaic panels and the two adjacent first-class photovoltaic panels included between the adjacent first-class photovoltaic panels. Under different path planning, the inspection time of the drone is different, and the time taken to reach the next photovoltaic panel from the previous photovoltaic panel is different. If the next photovoltaic panel is blocked by a shadow when the drone reaches it, the initial temperature of the simulated annealing algorithm is changed and the path planning is performed again; if the next photovoltaic panel is still blocked by a shadow after the initial temperature is modified, it means that under the current allocation method of the second-class photovoltaic panels, the next photovoltaic panel is inevitably blocked by a shadow, and the allocation method of the second-class photovoltaic panels needs to be adjusted; if the adjusted allocation method still cannot meet the requirements, it means that under the current relative position relationship, the shadow is inevitable, and the loop of step S8-2 is jumped out to perform local path planning for the next relative position relationship;

[0036] S8-5, return to step S8-2, adjust the second type photovoltaic panel included between two adjacent first type photovoltaic panels, and record the number of times of returning to step S8-2, when the number reaches a threshold, exit the loop of step S8-2.

[0037] Specifically, obtaining the path planning of the entire solar power station based on all local path planning also includes the following steps:

[0038] For each shortest path obtained in step S8-4, the relative position relationship path iFor each local path planning method under the above conditions, find the second type of photovoltaic panels that do not participate in the local path planning, and use the simulated annealing algorithm to calculate the second type of photovoltaic panels and photovoltaic panels that do not participate in the local path planning. and photovoltaic panels Perform path planning to obtain a complete path plan for the entire solar power station;

[0039] The first type of photovoltaic panels have completed path planning based on their relative position relationships. It is only necessary to perform path planning on the remaining second type of photovoltaic panels, the first first type of photovoltaic panel and the last first type of photovoltaic panel to obtain a complete path planning for the entire solar power station.

[0040] The planned path for drone inspection is determined by:

[0041] The relative position relationship path of the first type of photovoltaic panels corresponding to the same shortest path i For each complete path planning, adjust the time when the drone starts inspection, so that when the drone arrives at the first type of photovoltaic panel, the number of times the first type of photovoltaic panel is blocked by shadows is the least; select the complete path planning with the least number of times the first type of photovoltaic panel is blocked by shadows as the alternative path planning. If there is a situation where the number of blockages is the same, take the one with the shortest drone inspection mileage as the alternative path planning;

[0042] For the alternative path planning under the relative position relationship of the first-class photovoltaic panels corresponding to different shortest paths, the alternative path planning with the least number of shadow occlusions of the first-class photovoltaic panels is selected as the planned path for UAV inspection. If the number of occlusions is the same, the one with the shortest UAV inspection mileage is taken as the planned path for UAV inspection.

[0043] Compared with the prior art, the beneficial effects achieved by the present invention are: automated inspections through drones can reduce the time and cost required for manual inspections and improve the efficiency of inspections; in some dangerous environments, manual inspections involve certain safety risks, and automated inspections through drones can reduce manual inspections, reduce the work risks of inspectors, and improve safety; the drone inspection path is planned based on how shadow occlusion changes over time, so that the impact of shadow occlusion on drone image acquisition is minimized during inspections. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0045] Figure 1 It is a structural schematic diagram of an automatic inspection system for solar power stations based on artificial intelligence according to an embodiment of the present invention;

[0046] Figure 2 It is a flow chart of local path planning according to an embodiment of the present invention. DETAILED DESCRIPTION

[0047] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0048] In the embodiments of the present invention, please refer to Figure 1 , provides an artificial intelligence-based solar power station automatic inspection system, including a database module, a remote monitoring module, a path planning module, a drone platform module and a data analysis module; the output end of the database module is interconnected with the input end of the data analysis module, and is used to store image data collected during drone inspection; the remote monitoring module is used to remotely monitor the operating status and data transmission of the drone; the output end of the path planning module is interconnected with the input end of the drone platform module, and the inspection path of the drone is planned based on the influence of shadow occlusion on the photovoltaic panel; the output end of the drone platform module is interconnected with the input end of the database module, and is used to inspect the solar power station; the output end of the data analysis module is interconnected with the input end of the path planning module, and is used to analyze the image data collected by the drone and determine the working status of the photovoltaic panel.

[0049] The UAV platform module also includes a communication unit, a navigation unit, a control unit, a positioning unit and an image acquisition unit; the communication unit is used to receive instructions from the path planning module and transmit the image data collected by the UAV to the database module; the navigation unit navigates the UAV according to the moving path received by the UAV; the control unit is used to control the flight of the UAV; the positioning unit is used to position the UAV; and the image acquisition unit is used to collect image data of the photovoltaic panel.

[0050] The data analysis module also establishes a relationship between shadow occlusion and the time when the drone acquires image data based on the image of the first type of photovoltaic panel when the shadow is blocked and the time when the drone acquires the image data, determines the solar altitude angle and the solar azimuth angle based on the time when the drone acquires the image data, and obtains the relationship between the solar altitude angle and the solar azimuth angle and the shadow occlusion; the first type of photovoltaic panel refers to the photovoltaic panel in the solar power station that will be blocked by shadows.

[0051] The path planning module performs local path planning between the first type of photovoltaic panels under the relative position relationship of the first type of photovoltaic panels, and obtains the overall path planning within the solar power station based on the local path planning; the path planning module plans the drone inspection path through a simulated annealing algorithm without considering shadow occlusion, and obtains the relative position of the first type of photovoltaic panels.

[0052] In an embodiment of the present invention, a solar power station automatic inspection method based on artificial intelligence is provided, comprising the following steps:

[0053] S5-1, obtaining historical inspection data of the solar power station; determining the location of each photovoltaic panel in the solar power station;

[0054] S5-2, based on the historical inspection data of the solar power station, determine that the photovoltaic panels that will be blocked by shadows are recorded as the first type of photovoltaic panels, and the photovoltaic panels that will not be blocked by shadows are recorded as the second type of photovoltaic panels, and establish the connection between the shadow blocking situation and the inspection time:

[0055] For each first-type photovoltaic panel, determine an image of the first-type photovoltaic panel when shadowed and a time when the image was taken;

[0056] Based on the time when the image was taken, determine the solar altitude angle and solar azimuth angle at the time when the image was taken;

[0057] According to the set date for the drone to inspect the solar power station, the solar altitude angle and solar azimuth angle at different times of the inspection date are calculated;

[0058] Based on the solar altitude angle and solar azimuth angle at different times on the inspection date, the shading conditions of the first type of photovoltaic panels at different times on the inspection date are obtained;

[0059] S5-3, for each first-class photovoltaic panel that will be blocked by shadows, perform local path planning between any two first-class photovoltaic panels; first determine the second-class photovoltaic panel between the two first-class photovoltaic panels:

[0060] S7-1, number all photovoltaic panel positions, use simulated annealing algorithm to perform shortest path planning, the path planning requires traversing each photovoltaic panel without repetition, by adjusting the initial temperature of the simulated annealing algorithm, changing the output result of the simulated annealing algorithm, increasing the number of shortest paths obtained by the simulated annealing algorithm, and k represents the number of relative position relationships of the first type of photovoltaic panels corresponding to the shortest path;

[0061] S7-2, for each shortest path obtained in step S7-1, the relative position relationship path of the first type of photovoltaic panels corresponding to the shortest path i , i is a positive integer between [1, k], and the first type of photovoltaic panels are recorded as n is the number of photovoltaic panels of the first type;

[0062] S7-3, for i=1, 2, ..., k, execute step S7-4;

[0063] S7-4, for counting unit j=1, 2, ..., n-1, execute step S7-5;

[0064] S7-5, the relative position relationship of the first type of photovoltaic panels corresponding to the shortest path path1, path2, ..., path n Find the first type of photovoltaic panel and The second type of photovoltaic panels between, as elements stored in the collection middle; Indicates the first type of photovoltaic panels and and and The second type of photovoltaic panels between the formation of the collection, different collection The elements in the set are allowed to be repeated. There are no identical elements in the same set. Delete identical elements in the same set and record the number of repetitions of the identical elements in the same set.

[0065] Then, local path planning is performed between two first-class photovoltaic panels. Figure 2 , comprising the following steps S8-1 to S8-5:

[0066] S8-1, for i=1, 2, ..., k, execute steps S8-2 to S8-5;

[0067] S8-2, for counting unit j=1, 2, ..., n-1, execute step S8-3;

[0068] S8-3, for the set For each element A in , calculate whether the element A belongs to the set The probability P is Indicates that the element is in the set The number of repetitions in represents the number of repetitions of element A in all sets, Determine whether element A belongs to the set by generating random numbers If so, then element A belongs to the set In addition to the collection Delete element A from other sets except ; if not, element A does not belong to the set Update the number of repetitions of element A in all collections to

[0069] For example, under the relative position relationship path1 of the first type of photovoltaic panels corresponding to the shortest path, the first type of photovoltaic panels are recorded as First determine and The second type of photovoltaic panels included between; under the relative position relationship path1, including one or more path plans generated in step S7-1, if the second type of photovoltaic panel B is and There was one occurrence between and There was one occurrence between There are 6 occurrences between them, so the second type of photovoltaic panel B has a one-sixth probability of being assigned to and If the second type of photovoltaic panel B is assigned to and In between, not If the second type of photovoltaic panel B is not allocated and There is a one-fifth probability of being assigned to and between;

[0070] S8-4, after completing step S8-1, the second type of photovoltaic panels included between two adjacent first type photovoltaic panels under the relative position relationship corresponding to different shortest paths are obtained, and the inspection order of the first type photovoltaic panels under the relative position relationship of the first type photovoltaic panels corresponding to the k shortest paths is used, and the simulated annealing algorithm is used to plan the paths of the two adjacent first type photovoltaic panels and the second type photovoltaic panels included between the two adjacent first type photovoltaic panels; the previous photovoltaic panel and the next photovoltaic panel are used to distinguish the two adjacent first type photovoltaic panels, assuming that the previous photovoltaic panel is not blocked by the shadow when the drone arrives at the previous photovoltaic panel, according to the planned path, the unmanned The occlusion of the first type of photovoltaic panels at different times of the drone speed and the inspection date is used to determine whether the next photovoltaic panel is blocked by the shadow when the drone arrives at the next photovoltaic panel. If all the first type of photovoltaic panels are not blocked by the shadow, the local path planning is completed and the initial temperature cycle is exited; if there is a first type of photovoltaic panel that is blocked, the path is replanned by modifying the initial temperature of the simulated annealing algorithm. If the initial temperature of the simulated annealing algorithm is continuously modified and there is still a first type of photovoltaic panel that is blocked, then step S8-5 is entered; if all the first type of photovoltaic panels are not blocked by the shadow after the initial temperature is modified, the local path planning is completed and the initial temperature cycle is exited;

[0071] For example, for the position relationship path1, the simulated annealing algorithm and The second type of photovoltaic panels included in and Path planning is performed. Under different path planning, the inspection time of the drone is different. arrive The time taken is different, if the drone arrives hour, If the simulated annealing algorithm is blocked by shadows, the initial temperature of the simulated annealing algorithm should be changed. and The second type of photovoltaic panels included in and Perform path planning; if the initial temperature is modified, If it is still blocked by shadows, it means that under the current second-type photovoltaic panel allocation method, Being obscured by shadows is unavoidable and needs to be adjusted and The second type of photovoltaic panels included in After completing the local path planning, adjust The second type of photovoltaic panels included in the distribution; if the adjustment of the allocation method still cannot meet the requirements, it means that under the current relative position relationship path1, shadow occlusion is unavoidable. After completing the local path planning between the two, jump out of the loop of step S8-2 and perform local path planning for the next relative position relationship path2; when the allocation method of the second type of photovoltaic panels is not changed, the local path planning can be performed at any time. and Path planning is performed between the two groups; when the allocation method of the second type of photovoltaic panels needs to be changed, After the local path planning is completed, the allocation method of the second type of photovoltaic panels is adjusted;

[0072] S8-5, return to step S8-2, adjust the second type of photovoltaic panels included between two adjacent first type photovoltaic panels, and record the number of times of returning to step S8-2, and when the number reaches a threshold, exit the loop of step S8-2;

[0073] S5-4, based on all local path planning, obtain the path planning of the entire solar power station: for each shortest path obtained in step S8-4, the relative position relationship path of the first type of photovoltaic panels corresponding to the path i For each local path planning method under the above conditions, find the second type of photovoltaic panels that do not participate in the local path planning, and use the simulated annealing algorithm to calculate the second type of photovoltaic panels and photovoltaic panels that do not participate in the local path planning. and photovoltaic panels Perform path planning to obtain a complete path plan for the entire solar power station;

[0074] The relative position relationship path of the first type of photovoltaic panels corresponding to the same shortest pathi For each complete path planning, adjust the time when the drone starts inspection, so that when the drone arrives at the first type of photovoltaic panel, the number of times the first type of photovoltaic panel is blocked by shadows is the least; select the complete path planning with the least number of times the first type of photovoltaic panel is blocked by shadows as the alternative path planning. If there is a situation where the number of blockages is the same, take the one with the shortest drone inspection mileage as the alternative path planning;

[0075] For the alternative path planning under the relative position relationship of the first-class photovoltaic panels corresponding to different shortest paths, the alternative path planning with the least number of shadow occlusions of the first-class photovoltaic panels is selected as the planning path for drone inspection. If the number of occlusions is the same, the one with the shortest drone inspection mileage is selected as the planning path for drone inspection.

[0076] S5-5, inspect the solar power station according to the planned inspection path of the drone and generate an inspection report.

[0077] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0078] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. An automatic inspection method for solar power stations based on artificial intelligence, characterized in that: The following steps are involved: S5-1, obtaining historical inspection data of the solar power station; determining the location of each photovoltaic panel in the solar power station; S5-2, based on the historical inspection data of the solar power station, determine that the photovoltaic panels that will be blocked by shadows are recorded as the first type of photovoltaic panels, and the photovoltaic panels that will not be blocked by shadows are recorded as the second type of photovoltaic panels, and establish a connection between the shadow blocking situation and the inspection time; Comprising steps S6-1 to S6-4; S6-1, for each first-type photovoltaic panel, determining an image of the first-type photovoltaic panel when it is shaded and a time when the image is captured; S6-2, determining the solar altitude angle and solar azimuth angle at the time of image capture based on the time when the image was captured; S6-3, according to the set date for the drone to inspect the solar power station, calculate the solar altitude angle and solar azimuth angle at different times on the inspection date; S6-4, based on the solar altitude angle and solar azimuth angle at different times on the inspection date, obtaining the shading conditions of the first type of photovoltaic panels at different times on the inspection date; S5-3, for each first-class photovoltaic panel that will be blocked by shadows, perform local path planning between any two first-class photovoltaic panels; Determine the second type of photovoltaic panel between two first type photovoltaic panels through steps S7-1 to S7-5: S7-1, number all photovoltaic panel positions, use simulated annealing algorithm to perform shortest path planning, the path planning requires traversing each photovoltaic panel without repetition, by adjusting the initial temperature of the simulated annealing algorithm, changing the output result of the simulated annealing algorithm, increasing the number of shortest paths obtained by the simulated annealing algorithm, and k represents the number of relative position relationships of the first type of photovoltaic panels corresponding to the shortest path; S7-2, for each shortest path obtained in step S7-1, the relative position relationship path of the first type of photovoltaic panels corresponding to the shortest path i , i is a positive integer between [1, k], and the first type of photovoltaic panels are recorded as n is the number of photovoltaic panels of the first type; S7-3, for i=1, 2, ..., k, execute step S7-4; S7-4, for counting unit j=1, 2, ..., n-1, execute step S7-5; S7-5, the relative position relationship of the first type of photovoltaic panels corresponding to the shortest path path1, path2, ..., path n Find the first type of photovoltaic panel and The second type of photovoltaic panels between, as elements stored in the collection middle; Indicates the first type of photovoltaic panels and and and The second type of photovoltaic panels between the formation of the collection, different collection The elements in the set are allowed to be repeated. There are no identical elements in the same set. Delete identical elements in the same set and record the number of repetitions of the identical elements in the same set. Perform local path planning between two first-type photovoltaic panels through steps S8-1 to S8-5: S8-1, for i=1, 2, ..., k, execute steps S8-2 to S8-5; S8-2, for counting unit j=1, 2, ..., n-1, execute step S8-3; S8-3, for the set For each element A in , calculate whether the element A belongs to the set The probability P is Indicates that the element is in the set The number of repetitions in represents the number of repetitions of element A in all sets, Determine whether element A belongs to the set by generating random numbers If so, then element A belongs to the set In addition to the collection Delete element A from other sets except ; if not, element A does not belong to the set Update the number of repetitions of element A in all collections to S8-4, after completing step S8-1, the second type of photovoltaic panels included between two adjacent first type photovoltaic panels under the relative position relationship corresponding to different shortest paths are obtained, and the inspection order of the first type photovoltaic panels under the relative position relationship of the first type photovoltaic panels corresponding to the k shortest paths is used, and the simulated annealing algorithm is used to plan the paths of the two adjacent first type photovoltaic panels and the second type photovoltaic panels included between the two adjacent first type photovoltaic panels; the previous photovoltaic panel and the next photovoltaic panel are used to distinguish the two adjacent first type photovoltaic panels, assuming that the previous photovoltaic panel is not blocked by the shadow when the drone arrives at the previous photovoltaic panel, according to the planned path, the unmanned The occlusion of the first type of photovoltaic panels at different times of the drone speed and the inspection date is used to determine whether the next photovoltaic panel is blocked by the shadow when the drone arrives at the next photovoltaic panel. If all the first type of photovoltaic panels are not blocked by the shadow, the local path planning is completed and the initial temperature cycle is exited; if there is a first type of photovoltaic panel that is blocked, the path is replanned by modifying the initial temperature of the simulated annealing algorithm. If the initial temperature of the simulated annealing algorithm is continuously modified and there is still a first type of photovoltaic panel that is blocked, then step S8-5 is entered; if all the first type of photovoltaic panels are not blocked by the shadow after the initial temperature is modified, the local path planning is completed and the initial temperature cycle is exited; S8-5, return to step S8-2, adjust the second type of photovoltaic panels included between two adjacent first type photovoltaic panels, and record the number of times of returning to step S8-2, and when the number reaches a threshold, exit the loop of step S8-2; S5-4, obtaining the path planning of the entire solar power station based on all local path planning: The relative position relationship path of the first type of photovoltaic panels corresponding to each shortest path obtained in step S8-4 i For each local path planning method under the above conditions, find the second type of photovoltaic panels that do not participate in the local path planning, and use the simulated annealing algorithm to calculate the second type of photovoltaic panels and photovoltaic panels that do not participate in the local path planning. and photovoltaic panels Perform path planning to obtain a complete path plan for the entire solar power station; S5-5, inspecting the solar power station according to the path planned by the solar power station in step S5-4, and generating an inspection report.

2. The method for automatic inspection of solar power stations based on artificial intelligence according to claim 1, characterized in that: The following steps are involved: The relative position relationship path of the first type of photovoltaic panels corresponding to the same shortest path i Under each complete path planning, adjust the time when the drone starts inspection, so that when the drone reaches the first type of photovoltaic panel, the number of times the first type of photovoltaic panel is shaded is the least; Select the complete path planning with the least number of shadow occlusions of the first type of photovoltaic panels as the alternative path planning. If the number of shadow occlusions is the same, take the one with the shortest inspection mileage of the drone as the alternative path planning; For the alternative path planning under the relative position relationship of the first-class photovoltaic panels corresponding to different shortest paths, the alternative path planning with the least number of shadow occlusions of the first-class photovoltaic panels is selected as the planned path for UAV inspection. If the number of occlusions is the same, the one with the shortest UAV inspection mileage is taken as the planned path for UAV inspection.

3. An artificial intelligence-based solar power station automatic inspection system, using an artificial intelligence-based solar power station automatic inspection method as described in any one of claims 1-2, characterized in that: It includes a database module, a remote monitoring module, a path planning module, a UAV platform module and a data analysis module; the output end of the database module is interconnected with the input end of the data analysis module, and is used to store image data collected during the UAV inspection; the remote monitoring module is used to remotely monitor the operation status and data transmission status of the UAV; the output end of the path planning module is interconnected with the input end of the UAV platform module, and the inspection path of the UAV is planned based on the influence of shadow occlusion on the photovoltaic panel; the output end of the UAV platform module is interconnected with the input end of the database module, and is used to inspect the solar power station; The output end of the data analysis module is connected to the input end of the path planning module, and is used to analyze the image data collected by the drone to determine the working status of the photovoltaic panel.

4. The solar power station automatic inspection system based on artificial intelligence according to claim 3 is characterized in that: The UAV platform module also includes a communication unit, a navigation unit, a control unit, a positioning unit and an image acquisition unit; the communication unit is used to receive instructions from the path planning module and transmit the image data collected by the UAV to the database module; the navigation unit navigates the UAV according to the moving path received by the UAV; the control unit is used to control the flight of the UAV; the positioning unit is used to position the UAV; and the image acquisition unit is used to collect image data of the photovoltaic panel.

5. The solar power station automatic inspection system based on artificial intelligence according to claim 3 is characterized in that: The data analysis module also establishes a relationship between shadow occlusion and the time when the drone acquires image data based on the image of the first type of photovoltaic panel when the shadow is blocked and the time when the drone acquires the image data, determines the solar altitude angle and the solar azimuth angle based on the time when the drone acquires the image data, and obtains the relationship between the solar altitude angle and the solar azimuth angle and the shadow occlusion; the first type of photovoltaic panel refers to the photovoltaic panel in the solar power station that will be blocked by shadows.

6. The solar power station automatic inspection system based on artificial intelligence according to claim 3 is characterized in that: The path planning module performs local path planning between the first type of photovoltaic panels under the relative position relationship of the first type of photovoltaic panels, and obtains the overall path planning within the solar power station based on the local path planning; the path planning module plans the drone inspection path through a simulated annealing algorithm without considering shadow occlusion, and obtains the relative position of the first type of photovoltaic panels.

Images