Method, system and equipment for detecting abnormal points of photovoltaic panel and medium

Abstract

The invention relates to a photovoltaic panel abnormal point detection method, a photovoltaic panel abnormal point detection system, photovoltaic panel abnormal point detection equipment and a medium, and relates to the field of photovoltaic power generation. Executing the preset steps of identifying the foreign matters and the hot spots, determining the water spraying duration and the water spraying distance of the foreign matters and the hot spots, controlling the current photovoltaic panel to stop running and the like; and cyclically executing the acquisition of the first color image and the first infrared image of the new photovoltaic panel acquired by the first unmanned aerial vehicle, and executing the preset step on the new photovoltaic panel until the new photovoltaic panel is the last photovoltaic panel. According to the invention, the power generation efficiency of the photovoltaic panel is improved during the cleaning period.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of photovoltaic power generation, in particular to a photovoltaic panel abnormal point detection method, system, device and medium. BACKGROUND

[0002] With the development of green energy, the scale of photovoltaic power generation is getting larger and larger, and at present, large-scale photovoltaic panels are mainly laid in areas with sufficient light conditions for photovoltaic power generation. However, after long-term operation of the open laying of photovoltaic panels, dust will cover the photovoltaic panels, resulting in a decrease in the power generation capacity of the photovoltaic panels. In addition, bird droppings, leaves and other debris will adhere to the photovoltaic panels, resulting in the inability of some areas of the photovoltaic panels to generate electricity. The bird droppings and leaves and other debris will block some areas of the photovoltaic panels from generating electricity, resulting in hot spots on the photovoltaic panels, which will further cause damage to the photovoltaic panels and even cause a fire.

[0003] At present, artificial or related cleaning robots are mainly used to clean the photovoltaic panels regularly. However, due to the fact that some photovoltaic panels may be damaged, the lines may be aged, and the connections may be loose, etc., it is necessary to power off all or most of the photovoltaic panels before cleaning to prevent water from entering the photovoltaic panels or causing short circuit damage to the photovoltaic panels. However, since the scale of the photovoltaic panel deployment is large, the cleaning time is too long, and the photovoltaic panels cannot generate electricity during the cleaning period, resulting in low photovoltaic panel power generation efficiency. Therefore, how to improve the power generation efficiency of the photovoltaic panels during cleaning has become a problem. SUMMARY

[0004] In order to improve the power generation efficiency of the photovoltaic panels during cleaning, the present application provides a photovoltaic panel abnormal point detection method, system, device and medium.

[0005] In a first aspect, the present application provides a photovoltaic panel abnormal point detection method, which adopts the following technical solution: A photovoltaic panel abnormal point detection method, comprising: When the first color image and the first infrared image of the current photovoltaic panel collected by the first unmanned aerial vehicle are obtained, a preset step is performed; When the detection of all foreign matter features and hot spot features in the preset step is completed, the first color image and the first infrared image of the new photovoltaic panel collected by the first unmanned aerial vehicle are cyclically executed, and the preset step is executed on the new photovoltaic panel until the new photovoltaic panel is the last photovoltaic panel; The preset step comprises: Based on the first color image, the foreign matter features of the current photovoltaic panel are determined, and based on the first infrared image, the hot spot features of the current photovoltaic panel are determined; determine a water spraying time and a water spraying distance of the second spray head on the second unmanned aerial vehicle on each foreign matter feature and hot spot feature, the second unmanned aerial vehicle being provided with a first spray head for overall cleaning of the photovoltaic panel and a second spray head for cleaning of the foreign matter feature or the hot spot feature; control the current photovoltaic panel to stop running, and control the second unmanned aerial vehicle to clean the foreign matter feature and the hot spot feature on the current photovoltaic panel according to the water spraying time and the water spraying distance; when it is detected that the cleaning of all the foreign matter features and the hot spot features is completed, and a preset time length corresponding to the current photovoltaic panel is reached, control the current photovoltaic panel to run.

[0006] By adopting the above technical solution, the first unmanned aerial vehicle sequentially collects the first color image and the infrared image of each photovoltaic panel, sequentially obtains the first color image and the first infrared image of each photovoltaic panel in order, facilitates the identification of the foreign matter feature on each photovoltaic panel from the first color image, and the identification of the hot spot feature on each photovoltaic panel from the first infrared image, and then the second unmanned aerial vehicle sequentially cleans each photovoltaic panel. The foreign matter feature and the hot spot feature are both features that need to be cleaned in priority. Therefore, the water spraying time and the water spraying distance corresponding to each foreign matter feature and hot spot feature on the current photovoltaic panel are determined. The more serious the foreign matter feature or the hot spot feature, the longer the water spraying time and the closer the water spraying distance. After the appropriate water spraying time and water spraying distance of each foreign matter feature and hot spot feature are determined and the cleaning of the previous photovoltaic panel is completed, the current photovoltaic panel is controlled to stop running, and then the second unmanned aerial vehicle is controlled to clean the current photovoltaic panel according to the water spraying time and the water spraying distance of each foreign matter feature and hot spot feature on the current photovoltaic panel. After the cleaning is completed and the preset time length corresponding to the current photovoltaic panel is reached, the current photovoltaic panel is controlled to continue running. The above process is sequentially repeated, so that only a small number of photovoltaic panels stop running while the other photovoltaic panels normally generate electricity when one photovoltaic panel is cleaned. Thus, the normal generation of large-scale photovoltaic panels is ensured when the photovoltaic panels are cleaned, and the possibility of water erosion and damage to the photovoltaic panels during cleaning is reduced, thereby improving the generation efficiency of the photovoltaic panels during cleaning.

[0007] In another possible implementation manner, the determination of the water spraying time and the water spraying distance of the second spray head on the second unmanned aerial vehicle on each foreign matter feature and hot spot feature comprises: determining a first contour area of each foreign matter feature and a corresponding foreign matter type; determining a first abnormality score of each foreign matter feature based on the first contour area and a preset score corresponding to the foreign matter type; determining the water spraying time and the water spraying distance of each foreign matter feature based on the first abnormality score; determining a second contour area of each hot spot feature and a temperature; determining a second abnormality score of each hot spot feature based on the second contour area and the temperature; If the target hot spot feature exists, a first number of all foreign object features in a range of the target hot spot feature is determined, and a first sum of first profile areas of all foreign object features in the range of the target hot spot feature is determined, the target hot spot feature being a hot spot feature with a foreign object feature; A second abnormality score of the target hot spot feature is corrected based on the first number and the first sum to obtain a corrected second abnormality score. The water spraying time and distance of each hot spot feature are determined based on the second abnormality score and the corrected second abnormality score.

[0008] In another possible implementation manner, the water spraying time and distance of each foreign object feature are determined based on the first abnormality score, the water spraying time and distance of each hot spot feature are determined based on the second abnormality score and the corrected second abnormality score, and the method comprises the following steps of: The water spraying time of each foreign object feature is determined based on the first abnormality score and a first preset coefficient, and the water spraying distance of each foreign object feature is determined based on the first abnormality score and a second preset coefficient. The water spraying time of each hot spot feature is determined based on the second abnormality score, the corrected second abnormality score and the first preset coefficient, and the water spraying distance of each hot spot feature is determined based on the second abnormality score, the corrected second abnormality score and the second preset coefficient.

[0009] In another possible implementation manner, the second abnormality score of the target hot spot feature is corrected based on the first number and the first sum to obtain the corrected second abnormality score, and the method comprises the following steps of: A correction value is determined based on the first number, the first sum and respective corresponding weights. The correction value and the second abnormality score are summed to obtain the corrected second abnormality score of each target hot spot feature.

[0010] In another possible implementation manner, a preset time length corresponding to the current photovoltaic panel is determined, and the method further comprises the following steps of: A second sum of water spraying time of each foreign object feature and hot spot feature of the current photovoltaic panel is determined, and a total number of all foreign object features and hot spot features of the current photovoltaic panel is determined. Color saturation of the first color image of the current photovoltaic panel and color saturation of the reference photovoltaic panel image are determined, and a saturation difference value is determined. The preset time length corresponding to the current photovoltaic panel is determined based on the second sum, the total number and the saturation difference value.

[0011] In another possible implementation manner, the method further comprises the following steps of: The first drone is controlled to sequentially acquire the second color image and the second infrared image of each photovoltaic panel; Based on the second color image, a second number of foreign object features on each photovoltaic panel is determined, and based on the second infrared image, a third number of hot spot features on each photovoltaic panel is determined. Determine the similarity between each hot spot feature in the second infrared image and the hot spot feature at the corresponding location in the first infrared image; Based on the similarity, permanent hot spots are determined, and the difference is obtained by subtracting the fourth number of all permanent hot spots from the third number. The cleaning quality of each photovoltaic panel is determined based on the second quantity and the difference.

[0012] In another possible implementation, the method further includes: The temperature corresponding to each permanent hot spot and the third contour area of ​​each permanent hot spot are determined from the first infrared image. The severity score for each permanent hot spot is determined based on the temperature and the third profile area. Determine the third summation of the third profile area of ​​all permanent hot spots on each photovoltaic panel; The degree of damage to each photovoltaic panel is determined based on the severity score of each hot spot feature, the fourth number of all hot spot features, and the third sum.

[0013] Secondly, this application provides a photovoltaic panel anomaly detection system, which adopts the following technical solution: A photovoltaic panel anomaly detection system includes: The first drone is used to acquire the first color image and the first infrared image of each photovoltaic panel; The second drone is equipped with a first nozzle for thoroughly cleaning the photovoltaic panel and a second nozzle for cleaning foreign object features or hot spot features, and is used to clean the foreign object features and hot spot features on the current photovoltaic panel according to the spraying time and spraying distance. Multiple controllers are installed on the power circuit of each photovoltaic panel to control the operation or shutdown of each photovoltaic panel; An electronic device, communicatively connected to a first UAV, communicatively connected to a second UAV, and communicatively connected to multiple controllers, is used to execute a photovoltaic panel anomaly detection method according to any one of claims 1 to 7.

[0014] By adopting the above technical solution, the first drone sequentially acquires first color images and infrared images of each photovoltaic panel. This sequential acquisition of the first color and infrared images of each photovoltaic panel facilitates the electronic equipment's identification of foreign object features from the first color images and hot spot features from the first infrared images. Then, the second drone sequentially cleans each photovoltaic panel. Foreign object features and hot spot features are key areas requiring cleaning; therefore, the electronic equipment determines the corresponding water spraying duration and distance for each foreign object feature and hot spot feature on the current photovoltaic panel. More severe foreign object features or hot spot features require longer spraying durations and shorter spraying distances. The electronic equipment determines the appropriate water spraying duration and distance for each foreign object feature and hot spot feature. After the water spraying time and distance are set and the previous photovoltaic panel is cleaned, the electronic device sends a control signal to the controller corresponding to the current photovoltaic panel to stop the current photovoltaic panel from operating. Then, it controls the second drone to clean the current photovoltaic panel according to the water spraying time and distance for each foreign object feature and hot spot feature on the current photovoltaic panel. After the electronic device detects that the cleaning is complete and the preset time corresponding to the current photovoltaic panel has been reached, it can control the current photovoltaic panel to continue operating. This cycle is repeated, so that only a few photovoltaic panels stop operating while the other photovoltaic panels generate electricity normally when cleaning a photovoltaic panel. This ensures the normal power generation of a large number of photovoltaic panels during the photovoltaic panel cleaning operation, reduces the possibility of water erosion and damage to the photovoltaic panels during cleaning, and improves the power generation efficiency of photovoltaic panel cleaning.

[0015] Thirdly, this application provides an electronic device that adopts the following technical solution: An electronic device comprising: At least one processor; Memory; At least one application, wherein the at least one application is stored in memory and configured to be executed by at least one processor, the at least one configuration being for: executing a photovoltaic panel anomaly detection method according to any possible implementation of the first aspect.

[0016] Fourthly, this application provides a computer-readable storage medium, which adopts the following technical solution: A computer-readable storage medium, when the computer program is executed in a computer, causes the computer to perform a photovoltaic panel anomaly detection method as described in any of the first aspects.

[0017] In summary, this application includes at least one of the following beneficial technical effects: The first drone sequentially acquires first color and infrared images of each photovoltaic panel. This sequential acquisition of the first color and infrared images of each photovoltaic panel facilitates the identification of foreign object features from the first color image and hot spot features from the first infrared image. Then, the second drone sequentially cleans each photovoltaic panel. Foreign object and hot spot features are the key features requiring cleaning; therefore, the spraying duration and spraying distance are determined for each foreign object and hot spot feature on the current photovoltaic panel. More severe foreign object or hot spot features require longer spraying durations and shorter spraying distances. After the appropriate water spraying duration and distance are set and the previous photovoltaic panel has been cleaned, the current photovoltaic panel is stopped. Then, the second drone is controlled to clean the current photovoltaic panel according to the water spraying duration and distance for each foreign object feature and hot spot feature. After cleaning is completed and the preset time corresponding to the current photovoltaic panel is reached, the current photovoltaic panel can be controlled to continue operating. This cycle is repeated, so that when cleaning a photovoltaic panel, only a few photovoltaic panels stop operating while the others generate electricity normally. This ensures the normal power generation of a large number of photovoltaic panels during the photovoltaic panel cleaning operation, reduces the possibility of water erosion and damage to the photovoltaic panels during cleaning, and improves the power generation efficiency of photovoltaic panel cleaning. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating a photovoltaic panel anomaly detection method according to an embodiment of this application.

[0019] Figure 2 This is a flowchart illustrating the preset steps in the embodiments of this application.

[0020] Figure 3 This is a schematic diagram of the structure of a photovoltaic panel anomaly detection system according to an embodiment of this application.

[0021] Figure 4 This is a schematic diagram of the structure of an electronic device according to an embodiment of this application.

[0022] Reference numerals: 1. First UAV; 11. Color camera; 12. Infrared camera; 2. Second UAV; 3. Controller; 4. Electronic device; 41. Processor; 42. Bus; 43. Memory; 44. Transceiver. Detailed Implementation

[0023] The present application will be further described in detail below with reference to the accompanying drawings.

[0024] After reading this specification, those skilled in the art may make modifications to this embodiment without contributing any inventive step, but such modifications are protected by patent law as long as they fall within the scope of the claims of this application.

[0025] 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 only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0026] Furthermore, the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article, unless otherwise specified, generally indicates that the preceding and following related objects have an "or" relationship.

[0027] The embodiments of this application will now be described in further detail with reference to the accompanying drawings.

[0028] This application provides a method for detecting abnormal points on photovoltaic panels, executed by an electronic device. This electronic device can be a server or a terminal device. The server can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing cloud computing services. The terminal device can be a smartphone, tablet, laptop, desktop computer, etc., but is not limited to these. The terminal device and the server can be directly or indirectly connected via wired or wireless communication. This application does not impose any limitations on this. Figure 1 As shown, the method includes steps S101 and S102, wherein, S101, when the first color image and the first infrared image of the current photovoltaic panel are acquired by the first drone, the preset steps are executed.

[0029] In this embodiment, the first drone is equipped with a color camera and an infrared camera to capture a first color image and a first infrared image of each photovoltaic panel. The first drone connects to an electronic device via Wi-Fi or Bluetooth, enabling the electronic device to acquire the first color image and the first infrared image of each photovoltaic panel. Multiple photovoltaic panels are arranged in a regular pattern to form a photovoltaic panel array consisting of multiple rows and columns. The first drone can sequentially fly and collect images from multiple photovoltaic panels in each row. The first color image records the specific condition of the photovoltaic panel, and the first infrared image records the temperature of the photovoltaic panel. The electronic device uses the first color image and the first infrared image to facilitate subsequent analysis and identification of foreign objects and hot spots on each photovoltaic panel. After acquiring the first color image and the first infrared image of the current photovoltaic panel, the electronic device can perform preset steps to clean the current photovoltaic panel and the foreign objects and hot spots on it.

[0030] Among them, the preset steps are as follows Figure 2 As shown, the preset steps include step Sa, step Sb, step Sc, and step Sd, wherein, Sa determines the foreign object features of the current photovoltaic panel based on the first color image and the hot spot features of the current photovoltaic panel based on the first infrared image.

[0031] In this embodiment, the electronic device can first perform preprocessing on the first color image, including denoising, and then input the preprocessed first color image into a trained network model for foreign object recognition, thereby identifying foreign object features on the photovoltaic panel. Common foreign object features on photovoltaic panels include leaves, bird droppings, and dust accumulations. The network model can be a convolutional neural network model. The operator first creates a training sample set consisting of a large number of training samples. Each training sample includes an image and the name of the corresponding foreign object feature, for example, one training sample might be "Image 1, Bird Droppings". The electronic device inputs the training sample set into the initial convolutional neural network model for training to obtain a trained convolutional neural network model. Similarly, the electronic device can input the first infrared image into another trained convolutional neural network model for hot spot recognition. Alternatively, the electronic device can perform grayscale transformation on the denoised preprocessed first infrared image to obtain a grayscale image. Abnormal grayscale values ​​in the grayscale image, i.e., areas with excessively high temperatures, are segmented to obtain the contours of these areas; these contours are the hot spot features. Foreign objects obstructing the view of a photovoltaic (PV) panel can prevent the obstructed area from generating electricity. In this case, the obstructed area acts as a resistor, generating heat and affecting the panel's power generation efficiency, or even damaging it. Therefore, areas with foreign object characteristics require focused cleaning. Hot spot characteristics include both emerging and permanent hot spots. Emerging hot spots can recover after removing foreign objects from their characteristics. Permanent hot spots, however, are severely damaged due to prolonged obstruction and cannot be restored to normal after removing the foreign objects. Therefore, areas with hot spot characteristics also require focused cleaning.

[0032] Sb, determine the spraying duration and spraying distance of the second nozzle on the second UAV for each foreign object feature and hot spot feature.

[0033] The second drone is equipped with a first nozzle for thoroughly cleaning the photovoltaic panels and a second nozzle for cleaning foreign objects or hot spots.

[0034] In this embodiment, the second nozzle is equipped with a first nozzle and a second nozzle, both connected to a water source via flexible water pipes. The first nozzle is used for comprehensive cleaning of each photovoltaic panel. The first nozzle has a large coverage area and can be a planar spray nozzle capable of rotation, thus achieving overall cleaning of the photovoltaic panel. The second nozzle is used to clean foreign matter and hot spots on the photovoltaic panel. The second nozzle sprays a water jet, concentrating pressure to focus on cleaning these areas. Different foreign matter and hot spot characteristics exhibit different behaviors; therefore, the electronic equipment determines the appropriate spraying duration and distance for each foreign matter and hot spot based on its characteristics. For example, larger areas and more stubborn foreign matter or hot spot characteristics require longer spraying times and shorter spraying distances for better cleaning results.

[0035] Sc controls the current photovoltaic panel to stop operating and controls the second drone to clean the foreign object features and hot spot features on the current photovoltaic panel according to the water spraying duration and spraying distance.

[0036] In this embodiment, after prolonged operation, photovoltaic panels may develop cracks, aged and damaged wiring, and loose connections. If water seeps into the photovoltaic panel or its wiring during cleaning, it can cause a short circuit and damage. Therefore, the electronic equipment determines the spraying duration and distance for each foreign object and hot spot feature on the photovoltaic panel and then controls the panel to stop operating, i.e., stop power generation. When deploying the photovoltaic panels, operators can install a controller on the power circuit of each panel to control the on / off state of the circuit, thereby controlling the operation or shutdown of the photovoltaic system. The controller can be a relay or a PLC controller. The electronic equipment is connected to the controller of each photovoltaic panel via wires, enabling it to send control signals to each panel's controller to control the on / off state of the photovoltaic panel. The electronic equipment sends a control signal to the current photovoltaic panel's controller to disconnect the power circuit, thus stopping power generation. Then, the electronic equipment controls a second drone to clean each foreign object and hot spot feature on the photovoltaic panel according to its corresponding spraying duration and distance.

[0037] Sd controls the operation of the current photovoltaic panel after detecting that all foreign object features and hot spot features have been cleaned and the preset time corresponding to the current photovoltaic panel has been reached.

[0038] In this embodiment, the electronic device identifies the foreign object features and hot spot features of each photovoltaic panel. Based on the top-to-bottom distribution order of these features, it connects them sequentially to determine the movement path of the second drone on each photovoltaic panel, thus achieving orderly cleaning of each foreign object feature and hot spot feature from top to bottom. When the electronic device detects that the second drone has cleaned the bottommost foreign object feature or hot spot feature for the corresponding spraying time, it determines that cleaning of all foreign object features and hot spot features is complete. When the current photovoltaic panel is powered off, the power-off time is recorded. When the power-off time reaches the preset time corresponding to the current photovoltaic panel, the electronic device sends a control signal to the controller of the current photovoltaic panel to reconnect the power circuit, allowing the photovoltaic panel to continue generating electricity. The preset time allows for the evaporation of residual moisture on the photovoltaic panel, preventing damage to the photovoltaic panel from residual moisture caused by short-term power restoration after cleaning.

[0039] S102, when the preset steps of detecting all foreign object features and hot spot features have been completed and cleaning is finished, the first color image and the first infrared image of the new photovoltaic panel collected by the first UAV are obtained in a loop, and the preset steps are performed on the new photovoltaic panel until the new photovoltaic panel is the last photovoltaic panel.

[0040] In this embodiment, the electronic device performs preset steps on the current photovoltaic panel and detects that all foreign matter features and hot spot features on the current photovoltaic panel have been cleaned before starting to clean the next photovoltaic panel. Therefore, the electronic device acquires the first color image and the first infrared image of the next photovoltaic panel (the new photovoltaic panel), and then cleans the new photovoltaic panel according to the method disclosed in steps Sa to Sd above. After detecting that all foreign matter features and hot spot features on the new photovoltaic panel have been cleaned, the electronic device acquires the first color image and the first infrared image of the next photovoltaic panel again and performs the preset steps, and so on, until the last photovoltaic panel is cleaned. Ultimately, this achieves the goal of only a few photovoltaic panels stopping operation while others continue to generate electricity during the cleaning process, ensuring normal power generation of a large number of photovoltaic panels during cleaning operations, reducing the possibility of water erosion and damage to the photovoltaic panels during cleaning, and improving the power generation efficiency during photovoltaic panel cleaning.

[0041] One possible implementation of this application embodiment involves determining the spraying duration and spraying distance of the second nozzle on the second UAV for each foreign object feature and hot spot feature in step Sb. Specifically, this includes steps Sb1 (not shown in the figure), Sb2 (not shown in the figure), Sb3 (not shown in the figure), Sb4 (not shown in the figure), Sb5 (not shown in the figure), Sb6 (not shown in the figure), Sb7 (not shown in the figure), and Sb8 (not shown in the figure). Sb1 determines the first contour area of ​​each foreign object feature and the corresponding foreign object type.

[0042] In this embodiment, the electronic device performs grayscale transformation on the first color image after denoising the photovoltaic panel to obtain a grayscale image. Then, based on the position where the grayscale value of the foreign object feature changes stepwise, the contour of each foreign object feature is determined. The area of ​​the first contour is obtained by counting the number of pixels within the contour range. Foreign object types include dust accumulation, bird droppings, and leaves. After identifying the foreign object features through a network model, the electronic device can determine the type of foreign object.

[0043] Sb2 determines the first anomaly score for each foreign object feature based on the first contour area and the preset score corresponding to the foreign object type.

[0044] In this embodiment, different types of foreign objects have varying degrees of cleaning difficulty. Leaves falling onto the photovoltaic panel can be removed with just a short period of water impact. Bird droppings tend to adhere to the photovoltaic panel, making them more difficult to clean and requiring a longer water flow. Dust accumulates continuously, ranking between leaves and bird droppings in terms of cleaning difficulty. Therefore, different types of foreign objects correspond to different preset scores. The preset scores represent the degree of difficulty, and the preset scores for different types of foreign objects can be set by the staff according to the actual situation. A larger first contour area indicates greater cleaning difficulty. In summary, the first contour area and the foreign object feature are key factors representing the difficulty of cleaning the foreign object feature, i.e., the degree of foreign object anomaly. Therefore, the staff sets corresponding weights for the first contour area and the preset score, and then the electronic device normalizes the first contour area and the preset score respectively to obtain their corresponding normalized values. Then, the weights of the normalized values ​​are applied to calculate the first anomaly score for each foreign object feature. The first anomaly score represents the difficulty of cleaning the foreign object feature. For example, the first contour area of ​​a foreign object feature is 15 pixels, and the preset score is 3. The weights corresponding to the first contour area and the preset score are 0.7 and 0.3, respectively. Therefore, the electronic device calculates the first anomaly score of the foreign object feature as 15 × 0.7 + 3 × 0.3 = 11.4.

[0045] Sb3 determines the spraying duration and spraying distance for each foreign object feature based on the first anomaly score.

[0046] In the embodiments of this application, after the electronic device determines the first abnormal score characterizing the difficulty of cleaning foreign objects, it can accurately determine the appropriate water spraying duration and spraying distance for each foreign object feature based on the score.

[0047] Sb4 determines the second contour area and temperature of each hot spot feature.

[0048] In this embodiment of the application, the electronic device can obtain the contour of each hot spot feature by performing edge detection from the first infrared image, and count the pixels within the contour range to obtain the total number of pixels. The total number of pixels is used to represent the second contour area of ​​each hot spot feature. The temperature of each hot spot feature is recorded in the first infrared image, and the temperature of each hot spot feature can be directly determined based on the first infrared image.

[0049] Sb5, based on the second contour area and temperature, determines the second anomaly score for each hot spot feature.

[0050] In this embodiment, some hot spot features may have small foreign matter characteristics, or may be hot spots caused by previous foreign matter residue. Therefore, these foreign matter characteristics cannot be identified by the first color image, requiring cleaning of the hot spot features. A larger second contour area indicates a larger hot spot feature, making cleaning more difficult. Higher temperatures indicate a higher degree of hot spot anomaly, requiring thorough and careful cleaning. In summary, the second contour area and temperature are key factors affecting the ease of cleaning each hot spot. Therefore, operators can assign corresponding weights to the second contour area and temperature and store them in the electronic device. The electronic device normalizes the second contour area and temperature to obtain normalized values. Then, the electronic device uses the corresponding weights to perform a weighted calculation on these normalized values ​​to obtain the second anomaly score for each hot spot feature. A larger second anomaly score corresponds to a higher degree of cleaning difficulty.

[0051] Sb6, if target hot spot features exist, determine the first number of all foreign object features within the target hot spot feature range, and the first sum of the first contour areas of all foreign object features within the target hot spot feature range.

[0052] Among them, the target hot spot feature is the hot spot feature with foreign matter characteristics.

[0053] In this embodiment, some hot spot features are caused by foreign matter coverage. Therefore, foreign matter features coexist on the hot spot features, and the hot spot features with foreign matter features are the target hot spot features. The target foreign matter features require simultaneous cleaning of both the foreign matter and the hot spot features. Therefore, the analysis of the cleaning difficulty of the target hot spot features must consider the superimposed effects caused by the foreign matter features. The electronic device determines a first number of all foreign matter features within the range of each target hot spot feature, and then sums the first contour areas of all foreign matter features within the range to obtain a first sum. That is, the number of foreign matter features within the range of the target hot spot features and the total area of ​​all foreign matter are key factors affecting the cleaning difficulty of the target hot spot features.

[0054] Sb7, based on the first quantity and the first sum, corrects the second anomaly score of the target hot spot feature to obtain the corrected second anomaly score.

[0055] In the embodiments of this application, after the electronic device determines the first quantity and the first sum, it can correct the second abnormal score of the target hot spot feature based on these two key factors affecting the difficulty of cleaning the target hot spot feature, thereby obtaining the corrected second abnormal score, so that the second abnormal score is more consistent with the actual situation of the target hot spot feature.

[0056] Sb8 determines the spraying duration and spraying distance for each hot spot feature based on the second anomaly score and the corrected second anomaly score.

[0057] In the embodiments of this application, after the electronic device determines the first abnormal score of the hot spot feature and the second abnormal score of the target hot spot feature, it can accurately determine the appropriate water spraying duration and spraying distance for each hot spot feature and the target hot spot feature.

[0058] One possible implementation of this application embodiment includes step Sb2 determining the spraying duration and spraying distance of each foreign object feature based on a first anomaly score, and step Sb8 determining the spraying duration and spraying distance of each hot spot feature based on a second anomaly score and a corrected second anomaly score. Specifically, this includes steps one and two, wherein... Step 1: Determine the water spraying duration for each foreign object feature based on the first anomaly score and the first preset coefficient, and determine the water spraying distance for each foreign object feature based on the first anomaly score and the second preset coefficient.

[0059] In this embodiment, a first preset coefficient is used to calculate the appropriate water spraying duration for each foreign object feature, and a second preset coefficient is used to calculate the appropriate water spraying distance for each foreign object feature. Assuming the first preset coefficient is 0.5, taking step Sb2 as an example, if the first anomaly score of the foreign object feature is 11.4, then the water spraying duration for that foreign object feature is 11.4 × 0.5 = 5.7 seconds. Assuming the second preset coefficient is 3, then the electronic device uses 11.4 × 3 = 34.2, which is the descent distance of the second drone at the preset height. That is, the larger the first anomaly score, the greater the descent distance of the second drone during cleaning, meaning it gets closer to the foreign object feature for more thorough rinsing, and the closer it is, the greater the water impact force. The preset height is the reference height of the second drone when cleaning above the photovoltaic panel, for example, a preset height of 1 meter.

[0060] Step 2: Determine the spraying duration for each hot spot feature based on the second anomaly score, the corrected second anomaly score, and the first preset coefficient; determine the spraying distance for each hot spot feature based on the second anomaly score, the corrected second anomaly score, and the second preset coefficient.

[0061] In the embodiments of this application, the electronic device also multiplies the second abnormal score and the corrected second abnormal score of each hot spot feature with the first preset coefficient and the second preset coefficient respectively, thereby calculating the appropriate water spraying time and water spraying distance for each hot spot feature. The method described in step one can be referred to, and will not be repeated here.

[0062] One possible implementation of this application embodiment is that step Sb7 corrects the second anomaly score of the target hotspot feature based on the first quantity and the first sum to obtain the corrected second anomaly score, specifically including steps three and four, wherein... Step 3: Determine the correction value based on the first quantity, the first sum, and their respective weights.

[0063] Step four: Sum the corrected value and the second anomaly score to obtain the corrected second anomaly score for each target hotspot feature.

[0064] In this embodiment, the first number of all foreign matter features within the target hot spot feature range and the first sum of the areas of all foreign matter features are key factors affecting the ease of cleaning the target hot spot features. Therefore, the second anomaly score of the target hot spot feature is corrected based on the first number and the first sum to obtain the corrected second anomaly score. The operator can assign corresponding weights to the first number and the first sum and store them in the electronic device. The electronic device normalizes the first number and the first sum for each target hot spot feature to obtain their respective normalized values. Then, the electronic device performs a weighted calculation on each normalized value using its corresponding weight to obtain the corrected second anomaly score for each target hot spot feature. Correcting the second anomaly score using the first number and the first sum is more accurate.

[0065] One possible implementation of this application embodiment involves determining the preset duration corresponding to the current photovoltaic panel. The method further includes steps S1 (not shown in the figure), S2 (not shown in the figure), and S3 (not shown in the figure), wherein... S1, determine the second sum of the water spray duration for each foreign object feature and hot spot feature of the current photovoltaic panel, and determine the total number of all foreign object features and hot spot features of the current photovoltaic panel.

[0066] In this embodiment, since the first nozzle on the second drone is always spraying water, the more foreign object features and hot spot features there are on the photovoltaic panel, the longer the second drone stays above the photovoltaic panel. Consequently, the first nozzle sprays water continuously above the photovoltaic panel for a longer period, and the larger the volume of water sprayed by the first nozzle. Similarly, the electronic device sums the spraying times of all foreign object features and hot spot features on the photovoltaic panel to obtain a second sum. The larger the second sum, the larger the volume of water sprayed by the second nozzle. Therefore, the larger the second sum and the larger the total number, the more water the second drone sprays above the photovoltaic panel, resulting in more thorough contact erosion of the photovoltaic panel and its associated circuitry. This leads to more water residue on the photovoltaic panel after cleaning and a longer time required for water evaporation.

[0067] S2, determine the color saturation of the first color image of the current photovoltaic panel and the color saturation of the reference photovoltaic panel image, and determine the saturation difference.

[0068] In this embodiment, the reference photovoltaic panel image is an image of the photovoltaic panel when there is no dust, hot spots, or foreign objects. The electronic device first preprocesses the reference photovoltaic panel image (unifying it to RGB format, removing noise through Gaussian filtering, and then segmenting the effective area of ​​the photovoltaic panel using the OTSU thresholding method or Canny edge detection to eliminate interference from borders, backgrounds, etc.). Then, the preprocessed RGB image is standardized to the [0, 1] interval, converted to the HSV color space, and the saturation of a single pixel is obtained by calculating the difference between the maximum and minimum values ​​of the three RGB channels of the pixel, and the maximum value (when the maximum value is not 0). When the maximum value is 0, the pixel saturation is 0. Finally, the saturation of all pixels in the effective area of ​​the photovoltaic panel is counted, and the average saturation is calculated, which is the core color saturation index of the photovoltaic panel image. The electronic device determines the color saturation of the first color image in the same way, and then the electronic device takes the absolute value of the worst color saturation of the two images to obtain the saturation difference. The greater the saturation difference, the more dust covers the photovoltaic panel. The longer the first nozzle needs to spray water to clean it, that is, the more water is used. Similarly, the more water the second drone sprays above the photovoltaic panel, the more thorough the contact corrosion of the photovoltaic panel and its associated circuitry. After cleaning, there will be more water left on the photovoltaic panel, and the longer it will take for the water to evaporate.

[0069] Furthermore, the electronic equipment can calculate the spraying duration of the first nozzle based on the color saturation difference and the preset coefficients set by the staff. After determining the spraying duration of the first nozzle, the moving speed of the first nozzle from top to bottom can be obtained by dividing the distance from the top to the bottom of the photovoltaic panel by the spraying duration. The second drone then moves and sprays water according to this speed to thoroughly clean the dust on the photovoltaic panel.

[0070] The second drone can first perform a comprehensive cleaning of the photovoltaic panel using the first nozzle at a predetermined speed. Then, starting from the top of the photovoltaic panel, it can sequentially clean foreign object features and hot spot features using the second nozzle. If the first nozzle washes away some foreign object features, the second nozzle will still clean each identified foreign object feature and hot spot feature to ensure a more thorough cleaning. Alternatively, while the second drone is cleaning the photovoltaic panel from top to bottom using the first nozzle at a predetermined speed, when the second drone reaches a location with foreign object features or hot spot features, the first nozzle will clean that location first, and then the second drone will be controlled to move so that the second nozzle sequentially cleans each foreign object feature and hot spot feature. This is not limited to a single method.

[0071] S3. Determine the preset duration corresponding to the current photovoltaic panel based on the second sum, the total quantity, and the saturation difference.

[0072] In the embodiments of this application, as summarized above, the second sum, the total quantity, and the saturation difference are all key factors affecting the amount of water sprayed by the second drone onto the photovoltaic panels, i.e., the amount of water remaining on the photovoltaic panels after cleaning. Therefore, the operator assigns corresponding weights to the second sum, the total quantity, and the saturation difference and stores them in the electronic device. The electronic device normalizes the second sum, the total quantity, and the saturation difference, and then uses the corresponding weights to perform a weighted calculation on the normalized values ​​to obtain a score. This score is used to calculate the preset duration corresponding to the photovoltaic panel, i.e., the time from stopping operation to resuming operation. The operator can set a coefficient or function to calculate the preset duration based on this score and store it in the electronic device. Then, the preset duration corresponding to each photovoltaic panel can be obtained by multiplying the score by the corresponding coefficient or substituting it into the corresponding function.

[0073] In one possible implementation of this application embodiment, step S102, after the new photovoltaic panel becomes the last photovoltaic panel, further includes steps S103 (not shown in the figure), S104 (not shown in the figure), S105 (not shown in the figure), S106 (not shown in the figure), and S107 (not shown in the figure), wherein, S103, control the first drone to sequentially collect the second color image and the second infrared image of each photovoltaic panel.

[0074] In this embodiment of the application, after all photovoltaic panels have been cleaned and are resuming power generation, the first UAV sequentially acquires a second color image and a second infrared image of each photovoltaic panel. The second color image and the second infrared image record the appearance and specific condition of each photovoltaic panel after cleaning.

[0075] S104, determine the second number of foreign object features on each photovoltaic panel based on the second color image, and determine the third number of hot spot features on each photovoltaic panel based on the second infrared image.

[0076] In this embodiment of the application, the electronic device inputs a second color image of each photovoltaic panel into a trained convolutional neural network model to identify foreign objects and obtain a second number of foreign objects on each photovoltaic panel after cleaning. The electronic device then uses a trained network model or grayscale transformation edge detection on the second infrared image to obtain a third number of hot spot features on each photovoltaic panel after cleaning.

[0077] S105, determine the similarity between each hot spot feature on the second infrared image and the hot spot feature at the corresponding position on the first infrared image.

[0078] In this embodiment, the electronic device segments each hot spot feature in the second infrared image to obtain a region image of each hot spot feature in the second infrared image. Based on the position of the region image of each hot spot feature, the corresponding region image of the hot spot feature before cleaning is determined from the first infrared image. Then, the electronic device calculates the similarity between the two region images by calculating cosine distance or using a neural network model. The greater the similarity, the greater the likelihood that the hot spot has not changed before and after cleaning, and thus the greater the likelihood that it is a permanent hot spot.

[0079] S106. Based on similarity, permanent hot spots are determined, and the difference is obtained by subtracting the fourth number of all permanent hot spots from the third number.

[0080] In this embodiment, the electronic device compares the similarity of each hot spot feature in the second infrared image with a preset similarity. The preset similarity serves as a dividing point for high similarity, for example, a preset similarity of 92%. If a hot spot has a similarity reaching the preset similarity, it indicates that such a hot spot has undergone minimal change before and after cleaning, and therefore belongs to permanent hot spots. Then, the electronic device subtracts the number of permanent hot spots from the third number of all hot spot features on the cleaned photovoltaic panel to obtain the number of residual hot spots after cleaning, i.e., the difference between the third and fourth numbers. Residual hot spots indicate that some foreign matter was not thoroughly cleaned, and the more residual hot spots, the worse the cleaning quality.

[0081] S107, determine the cleaning quality of each photovoltaic panel based on the second quantity and the difference.

[0082] In summary, for the embodiments of this application, the second number of foreign matter features on the photovoltaic panel after cleaning and the number of residual hot spots are key factors characterizing the cleaning quality of each photovoltaic panel. The operator assigns corresponding weights to the second number and the difference, storing them in an electronic device. The electronic device normalizes the second number and the difference to obtain normalized values, and then uses their respective weights to perform a weighted calculation to obtain a score. This score represents the cleaning quality of each photovoltaic panel. The higher the score, the worse the cleaning quality.

[0083] In one possible implementation of this application embodiment, after step S102, steps five, six, seven, and eight are further included, wherein, Step 5: Determine the temperature corresponding to each permanent hot spot and the third contour area of ​​each permanent hot spot from the first infrared image.

[0084] In this embodiment of the application, the electronic device assesses the anomaly of each photovoltaic panel after each cleaning. Therefore, after identifying permanent hot spots on the photovoltaic panel, the electronic device determines the temperature of the permanent hot spots from the first infrared image of the photovoltaic panel. A higher temperature of the permanent hot spot indicates greater damage to the photovoltaic panel and a higher degree of anomaly. Then, the third contour area of ​​each permanent hot spot is determined. A larger third contour area means a larger area of ​​the photovoltaic panel occupied by the permanent hot spot, resulting in lower power generation efficiency and a higher degree of anomaly.

[0085] Step six: Determine the severity score for each permanent hot spot based on temperature and third profile area.

[0086] In summary, for the embodiments of this application, the temperature and third contour area of ​​each permanent hot spot are key factors affecting the severity of each permanent hot spot. The operator assigns corresponding weights to the temperature and third contour area and stores them in the electronic device. The electronic device normalizes the temperature and third contour area of ​​each permanent hot spot to eliminate the influence of different dimensions, and then uses the corresponding weights to perform a weighted calculation to obtain the severity score of each permanent hot spot.

[0087] Step 7: Determine the third sum of the third profile areas of all permanent hot spots on each photovoltaic panel.

[0088] In the embodiments of this application, the electronic device sums the third contour areas of all permanent hot spots on each photovoltaic panel to obtain a third sum. The larger the third sum, the larger the area occupied by the permanent hot spots on the photovoltaic panel, the lower the power generation efficiency of the photovoltaic panel, and the higher the degree of damage to the photovoltaic panel.

[0089] Step 8: Determine the degree of damage to each photovoltaic panel based on the severity score of each hot spot feature, the fourth number of all hot spot features, and the third sum.

[0090] In this embodiment, the electronic device sums the severity scores of each hot spot feature to obtain a total severity score. A higher number of fourth severity scores across all hot spots on the photovoltaic panel indicates a higher degree of damage. In summary, the total severity score, the number of fourth severity scores, and the third sum are all key factors characterizing the degree of damage to the photovoltaic panel. The operator assigns weights to the total severity score, the number of fourth severity scores, and the third sum and stores them in the electronic device. The electronic device normalizes the total severity score, the number of fourth severity scores, and the third sum, and then uses the corresponding weights to perform a weighted calculation to obtain the degree of damage to the photovoltaic panel. Furthermore, the electronic device stores the serial number of each photovoltaic panel and determines the association between the serial number and the corresponding degree of damage, storing this information for later retrieval by the operator. If a photovoltaic panel's damage reaches a preset damage threshold, the electronic device sends the photovoltaic panel's serial number and damage level to the operator's mobile phone or other terminal device via SMS, allowing the operator to promptly identify and repair or replace the severely damaged photovoltaic panels. If there are no permanent hot spots on some photovoltaic panels, the electronic equipment does not calculate the extent of damage to these photovoltaic panels.

[0091] The above embodiments describe a method for detecting abnormal points on a photovoltaic panel from the perspective of the process flow. The following embodiments describe a system for detecting abnormal points on a photovoltaic panel from the perspective of a virtual module or virtual unit. For details, please refer to the following embodiments.

[0092] This application provides a photovoltaic panel anomaly detection system, such as... Figure 3 As shown, a photovoltaic panel anomaly detection system may specifically include a first drone 1, a second drone 2, multiple controllers 3, and electronic devices 4. The first drone 1 is equipped with a color camera 11 and an infrared camera 12 on its bottom to capture a first color image and a first infrared image of each photovoltaic panel. The first drone 1 connects to the electronic devices 4 via WiFi or Bluetooth, enabling the electronic devices 4 to receive the first color image and the first infrared image captured by the first drone 1. The acquisition path of the first drone 1 can be sent to the first drone 1 by the electronic devices 4; the acquisition path can be manually planned by the operator or automatically planned by the electronic devices 4 based on the photovoltaic panel array.

[0093] The second drone 2 is equipped with a first nozzle and a second nozzle at its bottom. The first nozzle is used for comprehensive cleaning of the entire photovoltaic panel, while the second nozzle is used for targeted cleaning of foreign objects and hot spots. The first nozzle has a larger spray range, spraying water onto the photovoltaic panel in a straight line, meaning the spray range is consistent with the width of the photovoltaic panel. As the second drone 2 moves from the top to the bottom of the photovoltaic panel, the first nozzle can complete a comprehensive cleaning of the panel. The second nozzle sprays water in a columnar shape, allowing for focused cleaning of each foreign object and hot spot. When cleaning hot spots, the second drone 2 can move in a zigzag pattern along the contour of the hot spot, enabling the second nozzle to scan and thoroughly clean the hot spot. The second drone 2 is connected to the electronic device 4 via WiFi or Bluetooth, allowing it to clean the foreign objects and hot spots on the photovoltaic panel according to the spray duration and distance. The flight path of the second drone 2 is the same as that of the first drone 1.

[0094] Each photovoltaic panel's power supply circuit is connected to a controller 3, which controls the on / off state of the photovoltaic panel's power supply circuit. The controller 3 can be a relay, a PLC controller, or other types of controllers capable of controlling circuit on / off states. An electronic device 4 is connected to each photovoltaic panel's controller 3 via wires, controlling the on / off state of the photovoltaic panel's power supply circuit by sending control signals to the controller 3. The photovoltaic panel is cleaned after its power supply circuit is disconnected, thus stopping power generation. After prolonged operation, photovoltaic panels may develop cracks, aged and damaged wiring, and loose connections. Disconnecting the power supply circuit prevents water from entering the photovoltaic panel through cracks or contacting the wiring during cleaning, which could cause short circuits and damage.

[0095] Electronic device 4 executes the above method embodiment, thereby enabling only a few photovoltaic panels to stop operating while other photovoltaic panels generate electricity normally during the cleaning of a photovoltaic panel. This ensures the normal power generation of a large number of photovoltaic panels during the photovoltaic panel cleaning operation, reduces the possibility of water erosion and damage to the photovoltaic panels during cleaning, and improves the power generation efficiency during photovoltaic panel cleaning.

[0096] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process of the photovoltaic panel anomaly detection system described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0097] This application provides an electronic device, such as... Figure 4 As shown, Figure 4The illustrated electronic device 4 includes a processor 41 and a memory 43. The processor 41 and the memory 43 are connected, for example, via a bus 42. Optionally, the electronic device 4 may also include a transceiver 44. It should be noted that in practical applications, the transceiver 304 is not limited to one unit, and the structure of this electronic device 30 does not constitute a limitation on the embodiments of this application.

[0098] Processor 41 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It can implement or execute the various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this application. Processor 41 may also be a combination that implements computational functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, etc.

[0099] Bus 42 may include a pathway for transmitting information between the aforementioned components. Bus 42 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. Bus 42 can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 4 The symbol is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0100] The memory 43 may be a ROM (Read Only Memory) or other type of static storage device capable of storing static information and instructions, RAM (Random Access Memory) or other type of dynamic storage device capable of storing information and instructions, or EEPROM (Electrically Erasable Programmable Read Only Memory), CD-ROM (Compact Disc Read Only Memory) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto.

[0101] The memory 43 is used to store application code that executes the solution of this application, and its execution is controlled by the processor 41. The processor 41 is used to execute the application code stored in the memory 43 to implement the content shown in the foregoing method embodiments.

[0102] Electronic devices include, but are not limited to: mobile terminals such as mobile phones, laptops, digital radio receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), and in-vehicle terminals (such as in-vehicle navigation terminals), as well as fixed terminals such as digital TVs and desktop computers. Servers can also be included. Figure 4 The electronic device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.

[0103] This application provides a computer-readable storage medium storing a computer program, which, when run on a computer, enables the computer to execute the corresponding content in the aforementioned method embodiments. Compared with related technologies, in this application embodiment, a first drone sequentially acquires a first color image and an infrared image of each photovoltaic panel. Acquiring the first color image and first infrared image of each photovoltaic panel in sequence facilitates the identification of foreign object features on each photovoltaic panel from the first color image and the identification of hot spot features from the first infrared image. Then, a second drone sequentially cleans each photovoltaic panel. Foreign object features and hot spot features are features that require focused cleaning. Therefore, the spraying duration and spraying distance corresponding to each foreign object feature and hot spot feature on the current photovoltaic panel are determined. More severe foreign object features or hot spot features correspond to longer spraying durations and shorter spraying distances. This process determines the characteristics of each foreign object feature. After the appropriate spraying time and distance are set for each foreign object feature and hot spot feature, and the previous photovoltaic panel is cleaned, the current photovoltaic panel is stopped. Then, the second drone is controlled to clean the current photovoltaic panel according to the spraying time and distance for each foreign object feature and hot spot feature. After cleaning is completed and the preset time for the current photovoltaic panel is reached, the current photovoltaic panel can be controlled to continue operating. This cycle is repeated, so that only a few photovoltaic panels stop operating while the others generate electricity normally when cleaning a photovoltaic panel. This ensures the normal power generation of a large number of photovoltaic panels during the photovoltaic panel cleaning operation, reduces the possibility of water erosion and damage to the photovoltaic panels during cleaning, and improves the power generation efficiency of photovoltaic panel cleaning.

[0104] It should be understood that although the steps in the flowcharts of the accompanying figures are shown sequentially as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the accompanying figures may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the sub-steps or stages of other steps.

[0105] The above description is only a partial embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A photovoltaic panel anomaly point detection method, characterized by, The application relates to a method for cleaning photovoltaic panels. When the first color image and the first infrared image of the current photovoltaic panel collected by the first unmanned aerial vehicle are acquired, a preset step is executed; When the detection of all foreign matter features and hot spot features is completed, the first color image and the first infrared image of the new photovoltaic panel collected by the first unmanned aerial vehicle are acquired, and the preset step is executed on the new photovoltaic panel until the new photovoltaic panel is the last one; The preset step comprises the following steps: Based on the first color image, the foreign matter features of the current photovoltaic panel are determined, and based on the first infrared image, the hot spot features of the current photovoltaic panel are determined. The spraying time and the spraying distance of the second spraying head on the second unmanned aerial vehicle to each foreign matter feature and hot spot feature are determined, the second unmanned aerial vehicle is provided with the first spraying head for overall cleaning of the photovoltaic panel and the second spraying head for cleaning of the foreign matter features or the hot spot features. The current photovoltaic panel is controlled to stop running, and the second unmanned aerial vehicle is controlled to clean the foreign matter features and the hot spot features on the current photovoltaic panel according to the spraying time and the spraying distance. When the cleaning of all foreign matter features and hot spot features is completed, the current photovoltaic panel is controlled to run after a preset time corresponding to the current photovoltaic panel is reached.

2. The method of claim 1, wherein, The spraying time and the spraying distance of the second spraying head on the second unmanned aerial vehicle to each foreign matter feature and hot spot feature are determined, the second unmanned aerial vehicle is provided with the first spraying head for overall cleaning of the photovoltaic panel and the second spraying head for cleaning of the foreign matter features or the hot spot features. The first profile area of each foreign matter feature and the corresponding foreign matter type are determined. Based on the first profile area and the preset score corresponding to the foreign matter type, the first abnormal score of each foreign matter feature is determined. Based on the first abnormal score, the spraying time and the spraying distance of each foreign matter feature are determined. The second profile area and the temperature of each hot spot feature are determined. Based on the second profile area and the temperature, the second abnormal score of each hot spot feature is determined. If there is a target hot spot feature, the first number of all foreign matter features in the range of the target hot spot feature and the first sum of the first profile areas of all foreign matter features in the range of the target hot spot feature are determined, the target hot spot feature being a hot spot feature with foreign matter features. Based on the first number and the first sum, the second abnormal score of the target hot spot feature is corrected to obtain a corrected second abnormal score. Based on the second abnormal score and the corrected second abnormal score, the spraying time and the spraying distance of each hot spot feature are determined.

3. The method of claim 2, wherein the step of determining the abnormal point of the photovoltaic panel comprises: Based on the first abnormal score and the first preset coefficient, the spraying time of each foreign matter feature is determined, and based on the first abnormal score and the second preset coefficient, the spraying distance of each foreign matter feature is determined. Based on the second abnormal score and the corrected second abnormal score and the first preset coefficient, the spraying time of each hot spot feature is determined, and based on the second abnormal score and the corrected second abnormal score and the second preset coefficient, the spraying distance of each hot spot feature is determined. ​ 4. The method of claim 2, wherein the step of determining the abnormal point of the photovoltaic panel comprises: The second anomaly score of the target hot spot feature is corrected based on the first quantity, the first sum, and the second anomaly score to obtain a corrected second anomaly score. A correction value is determined based on the first quantity, the first sum, and the respective corresponding weight. The correction value and the second anomaly score are summed to obtain a corrected second anomaly score of each target hot spot feature.

5. The method of claim 1, further comprising: determining a preset time period corresponding to the current photovoltaic panel, wherein the method further comprises: determining a second sum of the water spraying time period of each foreign object feature and hot spot feature of the current photovoltaic panel, and determining a total quantity of all foreign object features and hot spot features of the current photovoltaic panel; determining a color saturation of the first color image of the current photovoltaic panel and a color saturation of the reference photovoltaic panel image, and determining a saturation difference value; determining the preset time period corresponding to the current photovoltaic panel based on the second sum, the total quantity, and the saturation difference value.

6. The method of claim 1, wherein, The method further comprises: controlling the first unmanned aerial vehicle to sequentially collect a second color image and a second infrared image of each photovoltaic panel; determining a second quantity of foreign object features on each photovoltaic panel based on the second color image, and determining a third quantity of hot spot features on each photovoltaic panel based on the second infrared image; determining a similarity between each hot spot feature on the second infrared image and a hot spot feature at a corresponding position on the first infrared image; determining permanent hot spots based on the similarity, and determining a difference value by subtracting a fourth quantity of all permanent hot spots from the third quantity; determining a cleaning quality of each photovoltaic panel based on the second quantity and the difference value.

7. The method of claim 1, wherein, The method further comprises: determining a temperature corresponding to each permanent hot spot and a third contour area of each permanent hot spot from the first infrared image; determining a severity score of each permanent hot spot based on the temperature and the third contour area; determining a third sum of the third contour areas of all permanent hot spots on each photovoltaic panel; determining a damage degree of each photovoltaic panel based on the severity score of each hot spot feature, a fourth quantity of all hot spot features, and the third sum.

8. A photovoltaic panel anomaly point detection system characterized by, The method further comprises: a first unmanned aerial vehicle configured to collect a first color image and a first infrared image of each photovoltaic panel; a second unmanned aerial vehicle provided with a first spray head configured to comprehensively clean the photovoltaic panel and a second spray head configured to clean foreign object features or hot spot features, and configured to clean the foreign object features and hot spot features on the current photovoltaic panel according to the water spraying time period and the water spraying distance; a plurality of controllers configured to control the operation or stop of the operation of each photovoltaic panel; an electronic device communicatively connected to the first unmanned aerial vehicle, the second unmanned aerial vehicle, and the plurality of controllers, and configured to execute the method of any one of claims 1-7.

9. An electronic device, comprising: The electronic device comprises: at least one processor; a memory; at least one application program stored in the memory and configured to be executed by the at least one processor, wherein the at least one application program is configured to execute the method of any one of claims 1-7. The electronic device comprises: at least one processor; a memory; at least one application program stored in the memory and configured to be executed by the at least one processor, wherein the at least one application program is configured to execute the method of any one of claims 1-7.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program product comprises computer executable instructions for causing a computer to perform the method of any one of claims 1 to 7 when the computer program is executed in the computer.

Images