Photovoltaic panel cleaning method and system based on artificial rainfall enhancement technology

Abstract

The invention relates to the technical field of photovoltaic components, and specifically discloses a photovoltaic panel cleaning method and system based on artificial rainfall enhancement technology, the method comprising: based on a preset SAM+SVM embedded large model, identifying all multispectral remote sensing images of photovoltaic panels in all full-range multispectral remote sensing images of a photovoltaic power station area; performing spectral analysis on all multispectral remote sensing images of photovoltaic panels to obtain cleanliness data of photovoltaic panels; analyzing optimal rainfall parameters based on rainfall data of the photovoltaic power station area in a historical analysis period and cleanliness data of all photovoltaic panels in a historical analysis period; determining artificial rainfall enhancement data of the photovoltaic power station area in a future period based on the current optimal rainfall parameters of the photovoltaic power station area and future rainfall data of the photovoltaic power station area and performing artificial rainfall enhancement control to obtain artificial rainfall enhancement control results; and being used for greatly reducing water resource consumption and labor cost input in the photovoltaic panel cleaning process by accurately controlling artificial rainfall enhancement.

Description

Technical Field

[0001] The present invention relates to the technical field of photovoltaic components, and in particular to a photovoltaic panel cleaning method and system based on artificial rainfall enhancement technology. Background Art

[0002] At present, photovoltaic modules are a vital component in photovoltaic power stations. The cleanliness of their surface and the integrity of the panels are directly related to the efficiency of photoelectric conversion, which in turn affects the power generation. If there is a lot of dirt and dust on the surface of the photovoltaic module, it will hinder the transmission of light, reduce the efficiency of photoelectric conversion, and cause power loss. Therefore, daily cleaning of the surface of the panels in photovoltaic power stations is of great significance. In order to ensure the normal operation and power generation efficiency of photovoltaic power stations, we must attach great importance to the cleanliness of the surface of photovoltaic modules.

[0003] Artificial rainfall technology has been used for many years, especially in arid areas and areas with water shortages. Relevant research institutions have conducted a series of studies on artificial rainfall technology, including the use of advanced meteorological radars, satellite remote sensing and other technologies to improve the early warning capabilities and defense levels for meteorological disasters. However, the existing artificial rainfall technology is mostly used to solve regional water resource shortages and drought problems, and has not been applied to the field of photovoltaic module cleaning. Compared with the traditional photovoltaic panel cleaning, which requires a lot of labor costs and water resource costs, the application of artificial rainfall technology in the field of photovoltaic panel cleaning can save labor costs and water resource costs to a certain extent, but there is no technology in the existing technology for precise control of artificial rainfall to achieve photovoltaic panel cleaning requirements.

[0004] Therefore, the present invention proposes a photovoltaic panel cleaning method and system based on artificial rainfall enhancement technology. Summary of the invention

[0005] The present invention provides a photovoltaic panel cleaning method and system based on artificial rainfall enhancement technology, which are used to greatly reduce the water resource consumption and labor cost input in the photovoltaic panel cleaning process by accurately controlling artificial rainfall enhancement.

[0006] The present invention provides a photovoltaic panel cleaning method based on artificial rainfall enhancement technology, comprising:

[0007] S1: Based on the preset SAM+SVM embedded large model, all multispectral remote sensing images of all photovoltaic panels in the photovoltaic power station area within the historical analysis period are identified from all full-range multispectral remote sensing images of the photovoltaic power station area within the historical analysis period;

[0008] S2: Perform spectral analysis on all multispectral remote sensing images of each photovoltaic panel in the photovoltaic power station area during the historical analysis period to obtain cleanliness data of each photovoltaic panel in the photovoltaic power station area during the historical analysis period;

[0009] S3: Analyze the current optimal rainfall parameters in the photovoltaic power station area based on the rainfall data in the photovoltaic power station area during the historical analysis period and the cleanliness data of all photovoltaic panels in the photovoltaic power station area during the historical analysis period;

[0010] S4: Based on the current optimal rainfall parameters of the photovoltaic power station area and the future rainfall data of the photovoltaic power station area, determine the artificial rainfall increase data of the photovoltaic power station area in the future period;

[0011] S5: Perform artificial rainfall control based on the artificial rainfall data of the photovoltaic power station area in the future period to obtain artificial rainfall control results.

[0012] Preferably, S1: based on a preset SAM+SVM embedded large model, in all full-range multispectral remote sensing images of the photovoltaic power station area within the historical analysis period, all multispectral remote sensing images of all photovoltaic panels in the photovoltaic power station area within the historical analysis period are identified, including:

[0013] Based on the SAM model in the preset SAM+SVM embedded large model, all full-range multispectral remote sensing images of the photovoltaic power station area within the historical analysis period are segmented to obtain all segmented images in each full-range multispectral remote sensing image;

[0014] Based on the trained SVM model in the preset SAM+SVM embedded large model, all segmented images in each full-range multispectral remote sensing image are classified to obtain multispectral remote sensing images of all photovoltaic panels in the photovoltaic power station area at the corresponding analysis time within the historical analysis period;

[0015] The multispectral remote sensing images of all photovoltaic panels in the photovoltaic power station area at all analysis times in the historical analysis period are regarded as all multispectral remote sensing images of all photovoltaic panels in the photovoltaic power station area in the historical analysis period.

[0016] Preferably, S2: performing spectral analysis on all multispectral remote sensing images of each photovoltaic panel in the photovoltaic power station area within the historical analysis period to obtain cleanliness data of each photovoltaic panel in the photovoltaic power station area within the historical analysis period, including:

[0017] Perform spectral analysis on each multispectral remote sensing image of each photovoltaic panel in the photovoltaic power station area within the historical analysis period to determine the remote sensing image reflectance distribution data of each multispectral remote sensing image in each single band;

[0018] Based on the remote sensing image reflectance distribution data of each photovoltaic panel in each single band at all analysis moments in the historical analysis period, the cleanliness of each photovoltaic panel at each analysis moment in the historical analysis period is determined;

[0019] The cleanliness of each photovoltaic panel in the photovoltaic power station area at all analysis moments in the historical analysis period is summarized to obtain the cleanliness data of each photovoltaic panel in the photovoltaic power station area in the historical analysis period.

[0020] Preferably, spectral analysis is performed on each multispectral remote sensing image of each photovoltaic panel in the photovoltaic power station area within the historical analysis period to determine the remote sensing image reflectance distribution data of each multispectral remote sensing image in each single band, including:

[0021] Determine the spectral information of each single band in each multispectral remote sensing image of each photovoltaic panel in the photovoltaic power station area during the historical analysis period;

[0022] In the spectral information of each single band in each multispectral remote sensing image, the remote sensing image reflectance distribution data of each single band of each multispectral remote sensing image is determined.

[0023] Preferably, the cleanliness of each photovoltaic panel at each analysis moment in the historical analysis period is determined based on the remote sensing image reflectance distribution data of each photovoltaic panel in each single band at all analysis moments in the historical analysis period, including:

[0024] Sampling local position points of each multispectral remote sensing image to obtain multiple local position points of each multispectral remote sensing image;

[0025] In the remote sensing image reflectance distribution data of each multispectral remote sensing image in each single band, the remote sensing image reflectance of each local position point of each multispectral remote sensing image in the single band is determined;

[0026] According to the order from large to small single bands, the reflectance of each local position point of each multispectral remote sensing image in all single bands is connected and fitted to obtain the wavelength-local reflectance curve of each local position point of each multispectral remote sensing image;

[0027] Calculate the similarity between the wavelength-local reflectance curve of each local position point of each multispectral remote sensing image and the wavelength-local reflectance curve of the extreme dust pixel point as the dust discrimination value of the corresponding local position point;

[0028] Calculate the similarity between the wavelength-local reflectance curve of each local position point of each multispectral remote sensing image and the wavelength-local reflectance curve of the extreme photovoltaic panel pixel point as the photovoltaic panel discrimination value of the corresponding local position point;

[0029] The dust discrimination values ​​and photovoltaic panel discrimination values ​​of all local position points of the multispectral remote sensing images of each photovoltaic panel at all analysis moments in the historical analysis period are subjected to spatiotemporal gradient analysis to determine the cleanliness of each photovoltaic panel at each analysis moment in the historical analysis period.

[0030] Preferably, the dust discrimination values ​​and photovoltaic panel discrimination values ​​of all local position points of the multispectral remote sensing image of each photovoltaic panel at all analysis moments in the historical analysis period are subjected to spatiotemporal gradient analysis to determine the cleanliness of each photovoltaic panel at each analysis moment in the historical analysis period, including:

[0031] The corresponding analysis time, the position data of the corresponding local position point, and the corresponding dust discrimination value are used as three coordinate dimensions of the three-dimensional space coordinates in sequence, and the dust discrimination values ​​of all local position points of the multispectral remote sensing images of each photovoltaic panel at all analysis times in the historical analysis period are fitted in three-dimensional space to obtain the spatiotemporal gradient representation of the dust discrimination value;

[0032] The corresponding analysis time, the position data of the corresponding local position point, and the corresponding photovoltaic panel discrimination value are sequentially used as the three coordinate dimensions of the three-dimensional space coordinates, and the photovoltaic panel discrimination values ​​of all local position points of the multispectral remote sensing images of each photovoltaic panel at all analysis times in the historical analysis period are fitted in three-dimensional space to obtain the spatiotemporal gradient representation of the photovoltaic panel discrimination value;

[0033] Based on the spatiotemporal gradient representation of dust discrimination value and photovoltaic panel discrimination value, the dust discrimination value and photovoltaic panel discrimination value of all local positions of each multispectral remote sensing image are optimized to obtain new dust discrimination value and photovoltaic panel discrimination value of each local position of each multispectral remote sensing image;

[0034] Based on the current dust discrimination value and the current photovoltaic panel discrimination value of all local position points of the multispectral remote sensing image, the photovoltaic panel cleanliness of each photovoltaic panel at each analysis moment in the historical analysis period is determined.

[0035] Preferably, S3: based on the rainfall data of the photovoltaic power station area in the historical analysis period and the cleanliness data of all photovoltaic panels in the photovoltaic power station area in the historical analysis period, analyzing the current optimal rainfall parameters of the photovoltaic power station area, including:

[0036] From the rainfall data of the photovoltaic power station area in the historical analysis period, the rainfall parameters of the photovoltaic power station area at each analysis time in the historical analysis period are determined;

[0037] From the cleanliness data of all photovoltaic panels in the photovoltaic power station area during the historical analysis period, the cleanliness of the photovoltaic panels in the photovoltaic power station area at each analysis moment during the historical analysis period is determined;

[0038] The rainfall parameters and cleanliness of photovoltaic panels in the photovoltaic power station area at each analysis moment in the historical analysis period are aligned and the relationship is quantified to obtain the relationship model between rainfall parameters and cleanliness of photovoltaic panels;

[0039] The target cleanliness of photovoltaic panels in the photovoltaic power station area is input into the relationship model to determine the current optimal rainfall parameters in the photovoltaic power station area.

[0040] Preferably, it also includes:

[0041] When the updated photovoltaic panel cleanliness of all photovoltaic panels in the photovoltaic power station area and the updated total rainfall data in the photovoltaic power station area are obtained, the relationship model is optimized based on the updated photovoltaic panel cleanliness of all photovoltaic panels in the photovoltaic power station area and the updated total rainfall data in the photovoltaic power station area to obtain a new relationship model;

[0042] The target PV panel cleanliness of the PV power station area is input into the new relationship model to obtain the updated optimal rainfall parameters of the PV power station area.

[0043] Preferably, S4: based on the current optimal rainfall parameters of the photovoltaic power station area and the future rainfall data of the photovoltaic power station area, determining the artificial rainfall increase data of the photovoltaic power station area in the future period, including:

[0044] Based on the historical meteorological data and geographic spatial data of the photovoltaic power station area and the rainfall data prediction model, the future rainfall data of the photovoltaic power station area is predicted;

[0045] From the future rainfall data of the photovoltaic power station area, the future rainfall parameters of the photovoltaic power station area at each analysis time in the future analysis period are determined;

[0046] Based on the future rainfall parameters of the photovoltaic power station area at each analysis moment in the future analysis period and the current optimal rainfall parameters of the photovoltaic power station area, the artificial rainfall parameters of the photovoltaic power station area at each analysis moment in the future period are determined;

[0047] The artificial rainfall parameters of the photovoltaic power station area at all times in the future period are regarded as the artificial rainfall data of the photovoltaic power station area in the future period.

[0048] The present invention provides a photovoltaic panel cleaning system based on artificial rainfall enhancement technology, which is used to perform any one of the photovoltaic panel cleaning methods based on artificial rainfall enhancement technology described in embodiments 1 to 9, including:

[0049] The segmentation and classification recognition module is used to identify all multispectral remote sensing images of all photovoltaic panels in the photovoltaic power station area within the historical analysis period from all full-range multispectral remote sensing images of the photovoltaic power station area within the historical analysis period based on the preset SAM+SVM embedded large model;

[0050] A cleanliness analysis module is used to perform spectral analysis on all multispectral remote sensing images of each photovoltaic panel in the photovoltaic power station area during the historical analysis period, and obtain cleanliness data of each photovoltaic panel in the photovoltaic power station area during the historical analysis period;

[0051] The optimal rainfall data analysis module is used to analyze the current optimal rainfall parameters in the photovoltaic power station area based on the rainfall data in the photovoltaic power station area during the historical analysis period and the cleanliness data of all photovoltaic panels in the photovoltaic power station area during the historical analysis period;

[0052] The artificial rainfall data determination module is used to determine the artificial rainfall data of the photovoltaic power station area in the future period based on the current optimal rainfall parameters of the photovoltaic power station area and the future rainfall data of the photovoltaic power station area;

[0053] The artificial rainfall control module is used to perform artificial rainfall control based on the artificial rainfall data of the photovoltaic power station area in the future period to obtain the artificial rainfall control result.

[0054] Compared with the prior art, the beneficial effects of the present invention are as follows: by classifying and segmenting all full-range multispectral remote sensing images of the photovoltaic power station area within the historical analysis period, all multispectral remote sensing images of the photovoltaic panels within the historical analysis period are identified, and by performing spectral analysis on all multispectral remote sensing images of the photovoltaic panels, the cleanliness data of the photovoltaic panels within the historical analysis period are determined, based on the correspondence between the cleanliness data of the photovoltaic panels within the historical analysis period and the rainfall data that has undergone artificial rainfall intervention, the optimal rainfall parameters that can enable the photovoltaic panels in the power supply station area to reach the required cleanliness are analyzed, and the optimal rainfall parameters and future rainfall data are used to determine detailed artificial rainfall data, and then the rainfall data is intervened through precise artificial rainfall control, so that the photovoltaic panels are cleaned by rainfall, the cleanliness of the photovoltaic panels is ensured, and the normal operation and power generation efficiency of the photovoltaic power station are ensured, that is, the water resource consumption and labor cost input in the photovoltaic panel cleaning process are greatly reduced through precise control of artificial rainfall.

[0055] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures specifically pointed out in this application document.

[0056] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] 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:

[0058] Figure 1 This is a flow chart of a photovoltaic panel cleaning method based on artificial rainfall enhancement technology in an embodiment of the present invention;

[0059] Figure 2 Schematic diagram of the photovoltaic panel identification model based on satellite remote sensing technology in an embodiment of the present invention;

[0060] Figure 3 A distribution diagram of cleanliness classification markings of all photovoltaic panels in a photovoltaic power station area in an embodiment of the present invention;

[0061] Figure 4 Schematic diagram of the internal functional modules of the photovoltaic panel cleaning system based on artificial rainfall enhancement technology in an embodiment of the present invention. DETAILED DESCRIPTION

[0062] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0063] Embodiment 1:

[0064] The present invention provides a photovoltaic panel cleaning method based on artificial rainfall enhancement technology, referring to Figure 1 ,include:

[0065] S1: Based on the preset SAM+SVM embedded large model, all multispectral remote sensing images of all photovoltaic panels in the photovoltaic power station area within the historical analysis period are identified from all full-range multispectral remote sensing images of the photovoltaic power station area within the historical analysis period;

[0066] S2: Perform spectral analysis on all multispectral remote sensing images of each photovoltaic panel in the photovoltaic power station area during the historical analysis period to obtain cleanliness data of each photovoltaic panel in the photovoltaic power station area during the historical analysis period;

[0067] S3: Analyze the current optimal rainfall parameters in the photovoltaic power station area based on the rainfall data in the photovoltaic power station area during the historical analysis period and the cleanliness data of all photovoltaic panels in the photovoltaic power station area during the historical analysis period;

[0068] S4: Based on the current optimal rainfall parameters of the photovoltaic power station area and the future rainfall data of the photovoltaic power station area, determine the artificial rainfall increase data of the photovoltaic power station area in the future period;

[0069] S5: Perform artificial rainfall control based on the artificial rainfall data of the photovoltaic power station area in the future period to obtain artificial rainfall control results.

[0070] In this embodiment, the preset SAM+SVM embedded large model is a model that combines the segmentation capability of SAM (Segment Anything Model) and the classification capability of a trained SVM (Support Vector Machines), wherein the trained SVM is trained through a large number of multispectral remote sensing images of photovoltaic panels and multispectral remote sensing images of non-photovoltaic panels.

[0071] In this embodiment, the photovoltaic power station area is a geographical area covered by the photovoltaic power station.

[0072] In this embodiment, the historical analysis period is a period of time before the current moment.

[0073] In this embodiment, the full-range multispectral remote sensing image is a multispectral remote sensing image covering the photovoltaic power station area.

[0074] In this embodiment, the multispectral remote sensing image of the photovoltaic panel is a partial full-range multispectral remote sensing image that only covers a single photovoltaic panel in the full-range multispectral remote sensing image.

[0075] In this embodiment, the rainfall data of the photovoltaic power station area in the historical analysis period includes rainfall parameters of the photovoltaic power station area at each analysis moment (eg, each single day) in the historical analysis period, wherein the rainfall parameters include rainfall duration, total rainfall, etc.

[0076] In this embodiment, the current optimal rainfall parameter in the photovoltaic power station area is the rainfall parameter required when the cleanliness of all photovoltaic panels in the photovoltaic power station area reaches the target cleanliness.

[0077] In this embodiment, the future rainfall data of the photovoltaic power station area includes rainfall parameters of the photovoltaic power station area at each analysis moment in a future analysis period, wherein the future analysis period is a period of time after the current moment.

[0078] In this embodiment, the artificial rainfall data includes artificial rainfall parameters of the photovoltaic power station area at each analysis time in the historical analysis period, wherein the artificial rainfall parameters include, for example, the artificial rainfall amount and the artificial rainfall duration.

[0079] In this embodiment, artificial rainfall control is the process of performing artificial rainfall operation control based on artificial rainfall data.

[0080] In this embodiment, the artificial rainfall control result is the result that the cleanliness of the photovoltaic panels in the photovoltaic power station area meets the cleanliness requirement after the artificial rainfall control.

[0081] The beneficial effects of the above technology are as follows: by classifying and segmenting all full-range multispectral remote sensing images of the photovoltaic power station area within the historical analysis period, all multispectral remote sensing images of the photovoltaic panels within the historical analysis period are identified, and by performing spectral analysis on all multispectral remote sensing images of the photovoltaic panels, the cleanliness data of the photovoltaic panels within the historical analysis period are determined, based on the correspondence between the cleanliness data of the photovoltaic panels within the historical analysis period and the rainfall data that has undergone artificial rainfall intervention, the optimal rainfall parameters that can enable the photovoltaic panels in the power supply station area to reach the required cleanliness are analyzed, and the optimal rainfall parameters and future rainfall data are used to determine detailed artificial rainfall data, and then the rainfall data is intervened through precise artificial rainfall control, so that the photovoltaic panels are cleaned by rainfall, the cleanliness of the photovoltaic panels is ensured, and the normal operation and power generation efficiency of the photovoltaic power station are ensured, that is, the water resource consumption and labor cost input in the photovoltaic panel cleaning process are greatly reduced through precise control of artificial rainfall.

[0082] Embodiment 2:

[0083] On the basis of Example 1, S1: Based on the preset SAM+SVM embedded large model, in all the full-range multispectral remote sensing images of the photovoltaic power station area within the historical analysis period, identify all the multispectral remote sensing images of all photovoltaic panels in the photovoltaic power station area within the historical analysis period, refer to Figure 2 ,include:

[0084] Based on the SAM model in the preset SAM+SVM embedded large model, all full-range multispectral remote sensing images of the photovoltaic power station area within the historical analysis period are segmented to obtain all segmented images in each full-range multispectral remote sensing image;

[0085] Based on the trained SVM model in the preset SAM+SVM embedded large model, all segmented images in each full-range multispectral remote sensing image are classified to obtain multispectral remote sensing images of all photovoltaic panels in the photovoltaic power station area at the corresponding analysis time within the historical analysis period;

[0086] The multispectral remote sensing images of all photovoltaic panels in the photovoltaic power station area at all analysis times in the historical analysis period are regarded as all multispectral remote sensing images of all photovoltaic panels in the photovoltaic power station area in the historical analysis period.

[0087] In this embodiment, image segmentation is performed on all full-range multispectral remote sensing images of the photovoltaic power station area within the historical analysis period based on the SAM model in the preset SAM+SVM embedded large model, including:

[0088] First, the image feature maps of all full-range multispectral remote sensing images are obtained through the image encoder in the SAM model. Secondly, the automatically generated dense points are converted into prompt maps using the prompt encoder in the SAM model. Finally, the image feature map and the prompt map are input into the mask decoder in the SAM model to generate the image segmentation result (including multiple segmented images).

[0089] In this embodiment, the segmented image is an image obtained by segmenting the full-range multispectral remote sensing image using the SAM model.

[0090] In this embodiment, based on the trained SVM model in the preset SAM+SVM embedded large model, all segmented images in each full-range multispectral remote sensing image are classified, including:

[0091] The band features of each mask obtained during the image segmentation process are extracted and input into the trained SVM model. The segmentation results are classified and all multispectral remote sensing images of photovoltaic panels are screened out from all segmented images.

[0092] The beneficial effects of the above technology are: based on the preset SAM+SVM embedded large model, all full-range multispectral remote sensing images of the photovoltaic power station area within the historical analysis period are segmented and reclassified, and all multispectral remote sensing images of all photovoltaic panels in the photovoltaic power station area within the historical analysis period are accurately identified.

[0093] Embodiment 3:

[0094] Based on Example 1, S2: spectral analysis is performed on all multispectral remote sensing images of each photovoltaic panel in the photovoltaic power station area within the historical analysis period to obtain cleanliness data of each photovoltaic panel in the photovoltaic power station area within the historical analysis period, referring to Figure 3 ,include:

[0095] Perform spectral analysis on each multispectral remote sensing image of each photovoltaic panel in the photovoltaic power station area within the historical analysis period to determine the remote sensing image reflectance distribution data of each multispectral remote sensing image in each single band;

[0096] Based on the remote sensing image reflectance distribution data of each photovoltaic panel in each single band at all analysis moments in the historical analysis period, the cleanliness of each photovoltaic panel at each analysis moment in the historical analysis period is determined;

[0097] The cleanliness of each photovoltaic panel in the photovoltaic power station area at all analysis moments in the historical analysis period is summarized to obtain the cleanliness data of each photovoltaic panel in the photovoltaic power station area in the historical analysis period.

[0098] In this embodiment, Figure 3Cleanliness is divided into three categories: good cleanliness, average cleanliness, and poor cleanliness.

[0099] In this embodiment, a single wavelength band is a continuous wavelength value range.

[0100] In this embodiment, the remote sensing image reflectance distribution data of the multispectral remote sensing image in a single band includes the local reflectance of each point in the multispectral remote sensing image in the single band.

[0101] In this embodiment, the multispectral remote sensing image is an image composed of many channels. The specific number of channels depends on the wavelength resolution of the sensor. Each channel captures light of a specified wavelength. Imagine the spectrum as a straight line. Due to the existence of wavelength resolution, only one wavelength can be seen at a certain distance. When this wavelength is seen, the information corresponding to this wavelength and a small range of bands nearby can be collected to form a channel, that is, one band corresponds to one channel. If a pixel in a multispectral remote sensing image is sampled, a spectral feature can be generated for this pixel.

[0102] The beneficial effect of the above technology is: based on the analyzed multi-spectral remote sensing image reflectance distribution data in each single band of the remote sensing image, the cleanliness of the photovoltaic panels at each analysis moment in the historical analysis period is accurately analyzed.

[0103] Embodiment 4:

[0104] On the basis of Example 3, spectral analysis is performed on each multispectral remote sensing image of each photovoltaic panel in the photovoltaic power station area during the historical analysis period to determine the remote sensing image reflectance distribution data of each multispectral remote sensing image in each single band, including:

[0105] Determine the spectral information of each single band in each multispectral remote sensing image of each photovoltaic panel in the photovoltaic power station area during the historical analysis period;

[0106] In the spectral information of each single band in each multispectral remote sensing image, the remote sensing image reflectance distribution data of each single band of each multispectral remote sensing image is determined.

[0107] In this embodiment, the spectral information of a single band includes information of the remote sensing image reflectance distribution data of the multi-spectral remote sensing image in a single band, that is, includes all spectral features of the multi-spectral remote sensing image in a single band.

[0108] The beneficial effect of the above technology is: by spectral analysis of the multispectral remote sensing image, the remote sensing image reflectance distribution data of the multispectral remote sensing image in each single band is read out.

[0109] Embodiment 5:

[0110] On the basis of Example 3, based on the remote sensing image reflectance distribution data of each photovoltaic panel in each single band at all analysis moments in the historical analysis period, the cleanliness of each photovoltaic panel at each analysis moment in the historical analysis period is determined, including:

[0111] Sampling local position points of each multispectral remote sensing image to obtain multiple local position points of each multispectral remote sensing image;

[0112] In the remote sensing image reflectance distribution data of each multispectral remote sensing image in each single band, the remote sensing image reflectance of each local position point of each multispectral remote sensing image in the single band is determined;

[0113] According to the order from large to small single bands, the reflectance of each local position point of each multispectral remote sensing image in all single bands is connected and fitted to obtain the wavelength-local reflectance curve of each local position point of each multispectral remote sensing image;

[0114] Calculate the similarity between the wavelength-local reflectance curve of each local position point of each multispectral remote sensing image and the wavelength-local reflectance curve of the extreme dust pixel point as the dust discrimination value of the corresponding local position point;

[0115] Calculate the similarity between the wavelength-local reflectance curve of each local position point of each multispectral remote sensing image and the wavelength-local reflectance curve of the extreme photovoltaic panel pixel point as the photovoltaic panel discrimination value of the corresponding local position point;

[0116] The dust discrimination values ​​and photovoltaic panel discrimination values ​​of all local position points of the multispectral remote sensing images of each photovoltaic panel at all analysis moments in the historical analysis period are subjected to spatiotemporal gradient analysis to determine the cleanliness of each photovoltaic panel at each analysis moment in the historical analysis period.

[0117] In this embodiment, the local position point sampling for each multispectral remote sensing image is to uniformly sample a plurality of local position points in each multispectral remote sensing image.

[0118] In this embodiment, the wavelength-local reflectivity curve is a curve in a two-dimensional coordinate system with the wavelength as the abscissa value and the local reflectivity as the ordinate value.

[0119] In this embodiment, calculating the similarity between the wavelength-local reflectance curve of each local position point of each multispectral remote sensing image and the wavelength-local reflectance curve of the extreme dust pixel point includes:

[0120] The average of the ratios of the local reflectances corresponding to all points in the wavelength-local reflectance curve of each local position point of each multispectral remote sensing image and the local reflectances corresponding to the points with the same abscissa value in the wavelength-local reflectance curve of the extreme dust pixel point is taken as the similarity between the wavelength-local reflectance curve of each local position point of each multispectral remote sensing image and the wavelength-local reflectance curve of the extreme dust pixel point.

[0121] In this embodiment, the wavelength-local reflectivity curve of the extreme dust pixel point is a curve obtained by aligning and averaging the wavelength-local reflectivity curves of the pixel points in multiple multispectral remote sensing images that only contain dust areas of the photovoltaic panel.

[0122] In this embodiment, the wavelength-local reflectivity curve of the extreme photovoltaic panel pixel point is a curve obtained by aligning and averaging the wavelength-local reflectivity curves of the pixel points in multiple multispectral remote sensing images that only contain photovoltaic panels (excluding dust areas).

[0123] In this embodiment, the dust discrimination value of the local position point is a discrimination value for judging whether the local position point belongs to a pixel point in a dust area. A larger dust discrimination value indicates that the corresponding local position point is more likely to belong to a dust area, and vice versa.

[0124] In this embodiment, the photovoltaic panel discrimination value of the local position point is a discrimination value used to determine whether the local position point belongs to a pixel point in the photovoltaic panel area. The larger the photovoltaic panel discrimination value, the more likely the corresponding local pixel point is to belong to the non-dust area in the photovoltaic panel area, and vice versa.

[0125] The beneficial effect of the above technology is that by sampling the position points of the multispectral remote sensing image and performing spatiotemporal gradient analysis of the dust discrimination values ​​and photovoltaic panel discrimination values ​​of all local position points obtained by the calculation, the cleanliness of each photovoltaic panel at each analysis moment during the historical analysis period can be accurately analyzed.

[0126] Embodiment 6:

[0127] On the basis of Example 5, the dust discrimination values ​​and photovoltaic panel discrimination values ​​of all local position points of the multispectral remote sensing image of each photovoltaic panel at all analysis moments in the historical analysis period are subjected to spatiotemporal gradient analysis to determine the cleanliness of each photovoltaic panel at each analysis moment in the historical analysis period, including:

[0128] The corresponding analysis time, the position data of the corresponding local position point, and the corresponding dust discrimination value are used as three coordinate dimensions of the three-dimensional space coordinates in sequence, and the dust discrimination values ​​of all local position points of the multispectral remote sensing images of each photovoltaic panel at all analysis times in the historical analysis period are fitted in three-dimensional space to obtain the spatiotemporal gradient representation of the dust discrimination value;

[0129] The corresponding analysis time, the position data of the corresponding local position point, and the corresponding photovoltaic panel discrimination value are sequentially used as the three coordinate dimensions of the three-dimensional space coordinates, and the photovoltaic panel discrimination values ​​of all local position points of the multispectral remote sensing images of each photovoltaic panel at all analysis times in the historical analysis period are fitted in three-dimensional space to obtain the spatiotemporal gradient representation of the photovoltaic panel discrimination value;

[0130] Based on the spatiotemporal gradient representation of dust discrimination value and photovoltaic panel discrimination value, the dust discrimination value and photovoltaic panel discrimination value of all local positions of each multispectral remote sensing image are optimized to obtain new dust discrimination value and photovoltaic panel discrimination value of each local position of each multispectral remote sensing image;

[0131] Based on the current dust discrimination value and the current photovoltaic panel discrimination value of all local position points of the multispectral remote sensing image, the photovoltaic panel cleanliness of each photovoltaic panel at each analysis moment in the historical analysis period is determined.

[0132] In this embodiment, the spatiotemporal gradient representation of the dust discrimination value is the dust discrimination value of all local position points in the multispectral remote sensing image of each photovoltaic panel at all analysis moments in the historical analysis period, which is a spatial stereo function representation in three-dimensional space when the corresponding analysis moment, the corresponding position data of the local position point, and the corresponding dust discrimination value are sequentially used as the three coordinate dimensions of the three-dimensional space coordinate.

[0133] In this embodiment, the spatiotemporal gradient representation of the photovoltaic panel discrimination value is the photovoltaic panel discrimination value of all local position points of the multispectral remote sensing image of each photovoltaic panel at all analysis moments in the historical analysis period, which is a spatial stereo function representation in three-dimensional space when the corresponding analysis moment, the corresponding position data of the local position point, and the corresponding photovoltaic panel discrimination value are sequentially used as the three coordinate dimensions of the three-dimensional space coordinate.

[0134] In this embodiment, based on the spatiotemporal gradient representation of the dust discrimination value and the spatiotemporal gradient representation of the photovoltaic panel discrimination value, the dust discrimination value and the photovoltaic panel discrimination value of all local positions of each multispectral remote sensing image are optimized to obtain new dust discrimination value and photovoltaic panel discrimination value of each local position of each multispectral remote sensing image, which is achieved through a preset optimization model, specifically:

[0135] The spatiotemporal gradient representation of the dust discrimination value and the spatiotemporal gradient representation of the photovoltaic panel discrimination value and the corresponding dust discrimination value and photovoltaic panel discrimination value of all local positions of the multispectral remote sensing image are input into the optimization model to obtain a new dust discrimination value and photovoltaic panel discrimination value for each local position of each multispectral remote sensing image;

[0136] The optimization model is trained by using the dust discrimination values ​​and photovoltaic panel discrimination values ​​of all local position points of the multispectral remote sensing images before and after optimization and the corresponding spatiotemporal gradient representations of the dust discrimination values ​​and the spatiotemporal gradient representations of the photovoltaic panel discrimination values ​​as training samples;

[0137] During the training process, the dust discrimination values ​​and photovoltaic panel discrimination values ​​of all local positions of the multispectral remote sensing image before optimization and the corresponding spatiotemporal gradient representation of the dust discrimination value and the spatiotemporal gradient representation of the photovoltaic panel discrimination value are used as model inputs, and the dust discrimination values ​​and photovoltaic panel discrimination values ​​of all local positions of the multispectral remote sensing image after optimization are used as model outputs for training.

[0138] In this embodiment, based on the current dust discrimination value and the current photovoltaic panel discrimination value of all local position points of the multispectral remote sensing image, the photovoltaic panel cleanliness of each photovoltaic panel at each analysis time in the historical analysis period is determined, including:

[0139] The local position points in the multispectral remote sensing image where the dust discrimination value is greater than the photovoltaic panel discrimination value are regarded as dust position points;

[0140] The local pixel points in the multispectral remote sensing image whose photovoltaic panel discrimination value is greater than the dust discrimination value are regarded as photovoltaic panel position points;

[0141] Determine the sum of the dust position points and the photovoltaic panel position points as the total number of pixels;

[0142] The ratio of the total number of photovoltaic panel position points to the total number of pixels in the multispectral remote sensing image is regarded as the cleanliness of the corresponding photovoltaic panel at the corresponding analysis moment in the historical analysis period.

[0143] The beneficial effect of the above technology is: by performing three-dimensional spatial fitting on the dust discrimination values ​​and photovoltaic panel discrimination values ​​of all local position points of the multispectral remote sensing images of each photovoltaic panel at all analysis moments in the historical analysis period, and performing local optimization based on the results of the three-dimensional spatial fitting, a global correction of the dust discrimination values ​​and photovoltaic panel discrimination values ​​of all local position points of the multispectral remote sensing images of each photovoltaic panel at all analysis moments in the historical analysis period is achieved, so as to ensure the accuracy of the cleanliness of the photovoltaic panel determined based on the dust discrimination values ​​and photovoltaic panel discrimination values ​​of all local position points of the multispectral remote sensing images.

[0144] Embodiment 7:

[0145] On the basis of Example 1, S3: based on the rainfall data of the photovoltaic power station area during the historical analysis period and the cleanliness data of all photovoltaic panels in the photovoltaic power station area during the historical analysis period, the optimal rainfall parameters of the photovoltaic power station area are analyzed, including:

[0146] From the rainfall data of the photovoltaic power station area in the historical analysis period, the rainfall parameters of the photovoltaic power station area at each analysis time in the historical analysis period are determined;

[0147] From the cleanliness data of all photovoltaic panels in the photovoltaic power station area during the historical analysis period, the cleanliness of the photovoltaic panels in the photovoltaic power station area at each analysis moment during the historical analysis period is determined;

[0148] The rainfall parameters and cleanliness of photovoltaic panels in the photovoltaic power station area at each analysis moment in the historical analysis period are aligned and the relationship is quantified to obtain the relationship model between rainfall parameters and cleanliness of photovoltaic panels;

[0149] The target cleanliness of photovoltaic panels in the photovoltaic power station area is input into the relationship model to determine the current optimal rainfall parameters in the photovoltaic power station area.

[0150] In this embodiment, the rainfall parameters and the cleanliness of the photovoltaic panels in the photovoltaic power station area at each analysis moment in the historical analysis period are aligned and the relationship is quantified to obtain a relationship model between the rainfall parameters and the cleanliness of the photovoltaic panels, including:

[0151] Determine the quantitative relationship between the rainfall parameters and the cleanliness of the photovoltaic panels in the photovoltaic power station area at each analysis moment during the historical analysis period;

[0152] The quantitative relationship between rainfall parameters and PV panel cleanliness in the PV power station area at all analysis moments in the historical analysis period was averaged to obtain a relationship model between rainfall parameters and PV panel cleanliness.

[0153] In this embodiment, the relationship model between rainfall parameters and photovoltaic panel cleanliness is a model that represents the quantitative relationship between rainfall parameters and photovoltaic panel cleanliness at a single analysis moment in the photovoltaic power station region within the historical analysis moments.

[0154] In this embodiment, the target photovoltaic panel cleanliness of the photovoltaic power station area is a preset minimum photovoltaic panel cleanliness that the photovoltaic panels in the photovoltaic power station area need to achieve.

[0155] The beneficial effect of the above technology is: by quantifying the relationship between the rainfall parameters and the cleanliness of the photovoltaic panels at each analysis moment in the historical analysis moment, and substituting the target photovoltaic panel cleanliness of the photovoltaic power station area into the quantitative relationship determined in the previous step, the optimal rainfall parameters required when the photovoltaic panel cleanliness of all photovoltaic panels in the photovoltaic power station area is the target photovoltaic panel cleanliness are accurately determined.

[0156] Embodiment 8:

[0157] On the basis of Example 7, it also includes:

[0158] When the updated photovoltaic panel cleanliness of all photovoltaic panels in the photovoltaic power station area and the updated total rainfall data in the photovoltaic power station area are obtained, the relationship model is optimized based on the updated photovoltaic panel cleanliness of all photovoltaic panels in the photovoltaic power station area and the updated total rainfall data in the photovoltaic power station area to obtain a new relationship model;

[0159] The target PV panel cleanliness of the PV power station area is input into the new relationship model to obtain the updated optimal rainfall parameters of the PV power station area.

[0160] In this embodiment, the updated photovoltaic panel cleanliness is a photovoltaic panel cleanliness obtained at a later time than the previously obtained photovoltaic panel cleanliness.

[0161] In this embodiment, the updated total rainfall data is total rainfall data acquired at a later time than the acquisition time corresponding to the previously acquired total rainfall data.

[0162] In this embodiment, based on the updated photovoltaic panel cleanliness of all photovoltaic panels in the photovoltaic power station area and the updated total rainfall data in the photovoltaic power station area, the relationship model is optimized to obtain a new relationship model, including:

[0163] The updated photovoltaic panel cleanliness of all photovoltaic panels in the photovoltaic power station area and the corresponding updated total rainfall data are used as training samples, and the relationship model is continuously trained to obtain a new relationship model.

[0164] In this embodiment, the optimal rainfall parameters updated in the photovoltaic power station area are determined based on the updated photovoltaic panel cleanliness of all photovoltaic panels in the photovoltaic power station area and the total rainfall data updated in the photovoltaic power station area, and the optimal rainfall parameters required when the photovoltaic panel cleanliness of all photovoltaic panels in the photovoltaic power station area is the target photovoltaic panel cleanliness.

[0165] In this embodiment, the total rainfall data is the sum of rainfall parameters of natural rainfall and artificial rainfall, for example, the sum of rainfall amount of natural rainfall and rainfall amount of artificial rainfall.

[0166] The beneficial effect of the above technology is that the relationship model can be continuously optimized using the newly obtained photovoltaic panel cleanliness and total rainfall data of all photovoltaic panels in the photovoltaic power station area, so that the relationship model has better performance and gradually improves the accuracy of the determined optimal rainfall parameters for the photovoltaic power station area update.

[0167] Embodiment 9:

[0168] On the basis of Example 1, S4: based on the current optimal rainfall parameters of the photovoltaic power station area and the future rainfall data of the photovoltaic power station area, determining the artificial rainfall data of the photovoltaic power station area in the future period, including:

[0169] Based on the historical meteorological data and geographic spatial data of the photovoltaic power station area and the rainfall data prediction model, the future rainfall data of the photovoltaic power station area is predicted;

[0170] From the future rainfall data of the photovoltaic power station area, the future rainfall parameters of the photovoltaic power station area at each analysis time in the future analysis period are determined;

[0171] Based on the future rainfall parameters of the photovoltaic power station area at each analysis moment in the future analysis period and the current optimal rainfall parameters of the photovoltaic power station area, the artificial rainfall parameters of the photovoltaic power station area at each analysis moment in the future period are determined;

[0172] The artificial rainfall parameters of the photovoltaic power station area at all times in the future period are regarded as the artificial rainfall data of the photovoltaic power station area in the future period.

[0173] In this embodiment, the historical meteorological data of the photovoltaic power station area includes meteorological data of the photovoltaic power station area in the past period of time (such as temperature, air pressure, humidity, wind direction, wind speed, aerosol particle concentration, etc.).

[0174] In this embodiment, the geographic spatial data of the photovoltaic power station area includes topographic data (such as vegetation coverage data, hydrological data, etc.) of the photovoltaic power station area in the past period of time.

[0175] In this embodiment, the rainfall data prediction model is a pre-trained model that can predict the future rainfall data of a region based on the historical meteorological data and geospatial data of a certain region that are input. The rainfall data prediction model is trained using the historical meteorological data and geospatial data of multiple regions and the corresponding future rainfall data as training samples. During the training process, the historical meteorological data and geospatial data of the region are the model input, and the future rainfall data of the region are the model output.

[0176] In this embodiment, the future rainfall parameter is a rainfall parameter of the photovoltaic power station area at a certain analysis time in the future, and the rainfall parameter includes, for example, the total duration of rainfall, the total rainfall amount, and the like.

[0177] In this embodiment, based on the future rainfall parameters of the photovoltaic power station area at each analysis moment in the future analysis period and the current optimal rainfall parameters of the photovoltaic power station area, the artificial rainfall parameters of the photovoltaic power station area at each analysis moment in the future period are determined, including:

[0178] When the future rainfall parameter of the photovoltaic power station area at a single analysis moment in the future analysis period is not less than the current optimal rainfall parameter of the photovoltaic power station area, the artificial rainfall parameter of the photovoltaic power station area at the corresponding analysis moment in the future period is set to 0;

[0179] When the future rainfall parameter of the photovoltaic power station area at a single analysis moment in the future analysis period is less than the current optimal rainfall parameter of the photovoltaic power station area, the difference between the current optimal rainfall parameter of the photovoltaic power station area and the future rainfall parameter of the photovoltaic power station area at the corresponding analysis moment in the future analysis period is used as the artificial rainfall parameter of the photovoltaic power station area at the corresponding analysis moment in the future period.

[0180] The beneficial effects of the above technology are: using the historical meteorological data and geospatial data of the photovoltaic power station area and the rainfall data prediction model to achieve accurate prediction of future rainfall data in the photovoltaic power station area, and based on the future rainfall parameters of the photovoltaic power station area at each analysis moment in the future analysis period and the current optimal rainfall parameters of the photovoltaic power station area, accurate determination of artificial rainfall data in the photovoltaic power station area in the future period is achieved.

[0181] Embodiment 10:

[0182] The present invention provides a photovoltaic panel cleaning system based on artificial rainfall enhancement technology, which is used in any one of the photovoltaic panel cleaning methods based on artificial rainfall enhancement technology described in embodiments 1 to 9, referring to Figure 4 ,include:

[0183] The segmentation and classification recognition module is used to identify all multispectral remote sensing images of all photovoltaic panels in the photovoltaic power station area within the historical analysis period from all full-range multispectral remote sensing images of the photovoltaic power station area within the historical analysis period based on the preset SAM+SVM embedded large model;

[0184] A cleanliness analysis module is used to perform spectral analysis on all multispectral remote sensing images of each photovoltaic panel in the photovoltaic power station area during the historical analysis period, and obtain cleanliness data of each photovoltaic panel in the photovoltaic power station area during the historical analysis period;

[0185] The optimal rainfall data analysis module is used to analyze the current optimal rainfall parameters in the photovoltaic power station area based on the rainfall data in the photovoltaic power station area during the historical analysis period and the cleanliness data of all photovoltaic panels in the photovoltaic power station area during the historical analysis period;

[0186] The artificial rainfall data determination module is used to determine the artificial rainfall data of the photovoltaic power station area in the future period based on the current optimal rainfall parameters of the photovoltaic power station area and the future rainfall data of the photovoltaic power station area;

[0187] The artificial rainfall control module is used to perform artificial rainfall control based on the artificial rainfall data of the photovoltaic power station area in the future period to obtain the artificial rainfall control result.

[0188] The beneficial effects of the above technology are as follows: by classifying and segmenting all full-range multispectral remote sensing images of the photovoltaic power station area within the historical analysis period, all multispectral remote sensing images of the photovoltaic panels within the historical analysis period are identified, and by performing spectral analysis on all multispectral remote sensing images of the photovoltaic panels, the cleanliness data of the photovoltaic panels within the historical analysis period are determined, based on the correspondence between the cleanliness data of the photovoltaic panels within the historical analysis period and the rainfall data that has undergone artificial rainfall intervention, the optimal rainfall parameters that can enable the photovoltaic panels in the power supply station area to reach the required cleanliness are analyzed, and the optimal rainfall parameters and future rainfall data are used to determine detailed artificial rainfall data, and then the rainfall data is intervened through precise artificial rainfall control, so that the photovoltaic panels are cleaned by rainfall, the cleanliness of the photovoltaic panels is ensured, and the normal operation and power generation efficiency of the photovoltaic power station are ensured, that is, the water resource consumption and labor cost input in the photovoltaic panel cleaning process are greatly reduced through precise control of artificial rainfall.

[0189] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A photovoltaic panel cleaning method based on artificial rainfall enhancement technology, characterized in that: include: S1: Based on the preset SAM+SVM embedded large model, all multispectral remote sensing images of all photovoltaic panels in the photovoltaic power station area within the historical analysis period are identified from all full-range multispectral remote sensing images of the photovoltaic power station area within the historical analysis period; S2: Perform spectral analysis on all multispectral remote sensing images of each photovoltaic panel in the photovoltaic power station area during the historical analysis period to obtain cleanliness data of each photovoltaic panel in the photovoltaic power station area during the historical analysis period; S3: Analyze the current optimal rainfall parameters in the photovoltaic power station area based on the rainfall data in the photovoltaic power station area during the historical analysis period and the cleanliness data of all photovoltaic panels in the photovoltaic power station area during the historical analysis period; S4: Based on the current optimal rainfall parameters of the photovoltaic power station area and the future rainfall data of the photovoltaic power station area, determine the artificial rainfall increase data of the photovoltaic power station area in the future period; S5: Perform artificial rainfall control based on the artificial rainfall data of the photovoltaic power station area in the future period to obtain artificial rainfall control results.

2. The photovoltaic panel cleaning method based on artificial rainfall enhancement technology according to claim 1 is characterized in that: S1: Based on the preset SAM+SVM embedded large model, all multispectral remote sensing images of all photovoltaic panels in the photovoltaic power station area within the historical analysis period are identified, including: Based on the SAM model in the preset SAM+SVM embedded large model, all full-range multispectral remote sensing images of the photovoltaic power station area within the historical analysis period are segmented to obtain all segmented images in each full-range multispectral remote sensing image; Based on the trained SVM model in the preset SAM+SVM embedded large model, all segmented images in each full-range multispectral remote sensing image are classified to obtain multispectral remote sensing images of all photovoltaic panels in the photovoltaic power station area at the corresponding analysis time within the historical analysis period; The multispectral remote sensing images of all photovoltaic panels in the photovoltaic power station area at all analysis times in the historical analysis period are regarded as all multispectral remote sensing images of all photovoltaic panels in the photovoltaic power station area in the historical analysis period.

3. The photovoltaic panel cleaning method based on artificial rainfall enhancement technology according to claim 1 is characterized in that: S2: Perform spectral analysis on all multispectral remote sensing images of each photovoltaic panel in the photovoltaic power station area during the historical analysis period to obtain the cleanliness data of each photovoltaic panel in the photovoltaic power station area during the historical analysis period, including: Perform spectral analysis on each multispectral remote sensing image of each photovoltaic panel in the photovoltaic power station area within the historical analysis period to determine the remote sensing image reflectance distribution data of each multispectral remote sensing image in each single band; Based on the remote sensing image reflectance distribution data of each photovoltaic panel in each single band at all analysis moments in the historical analysis period, the cleanliness of each photovoltaic panel at each analysis moment in the historical analysis period is determined; The cleanliness of each photovoltaic panel in the photovoltaic power station area at all analysis moments in the historical analysis period is summarized to obtain the cleanliness data of each photovoltaic panel in the photovoltaic power station area in the historical analysis period.

4. The photovoltaic panel cleaning method based on artificial rainfall enhancement technology according to claim 3 is characterized in that: Perform spectral analysis on each multispectral remote sensing image of each photovoltaic panel in the photovoltaic power station area during the historical analysis period to determine the remote sensing image reflectance distribution data of each multispectral remote sensing image in each single band, including: Determine the spectral information of each single band in each multispectral remote sensing image of each photovoltaic panel in the photovoltaic power station area during the historical analysis period; In the spectral information of each single band in each multispectral remote sensing image, the remote sensing image reflectance distribution data of each single band of each multispectral remote sensing image is determined.

5. The photovoltaic panel cleaning method based on artificial rainfall enhancement technology according to claim 3 is characterized in that: Based on the reflectance distribution data of the multispectral remote sensing images of each photovoltaic panel at all analysis moments in the historical analysis period in each single band, the cleanliness of each photovoltaic panel at each analysis moment in the historical analysis period is determined, including: Sampling local position points of each multispectral remote sensing image to obtain multiple local position points of each multispectral remote sensing image; In the remote sensing image reflectance distribution data of each multispectral remote sensing image in each single band, the remote sensing image reflectance of each local position point of each multispectral remote sensing image in the single band is determined; According to the order from large to small single bands, the reflectance of each local position point of each multispectral remote sensing image in all single bands is connected and fitted to obtain the wavelength-local reflectance curve of each local position point of each multispectral remote sensing image; Calculate the similarity between the wavelength-local reflectance curve of each local position point of each multispectral remote sensing image and the wavelength-local reflectance curve of the extreme dust pixel point as the dust discrimination value of the corresponding local position point; Calculate the similarity between the wavelength-local reflectance curve of each local position point of each multispectral remote sensing image and the wavelength-local reflectance curve of the extreme photovoltaic panel pixel point as the photovoltaic panel discrimination value of the corresponding local position point; The dust discrimination values ​​and photovoltaic panel discrimination values ​​of all local position points of the multispectral remote sensing images of each photovoltaic panel at all analysis moments in the historical analysis period are subjected to spatiotemporal gradient analysis to determine the cleanliness of each photovoltaic panel at each analysis moment in the historical analysis period.

6. The photovoltaic panel cleaning method based on artificial rainfall enhancement technology according to claim 5 is characterized in that: The dust discrimination values ​​and photovoltaic panel discrimination values ​​of all local position points of the multispectral remote sensing images of each photovoltaic panel at all analysis moments in the historical analysis period are analyzed in time and space to determine the cleanliness of each photovoltaic panel at each analysis moment in the historical analysis period, including: The corresponding analysis time, the position data of the corresponding local position point, and the corresponding dust discrimination value are used as three coordinate dimensions of the three-dimensional space coordinates in sequence, and the dust discrimination values ​​of all local position points of the multispectral remote sensing images of each photovoltaic panel at all analysis times in the historical analysis period are fitted in three-dimensional space to obtain the spatiotemporal gradient representation of the dust discrimination value; The corresponding analysis time, the position data of the corresponding local position point, and the corresponding photovoltaic panel discrimination value are sequentially used as the three coordinate dimensions of the three-dimensional space coordinates, and the photovoltaic panel discrimination values ​​of all local position points of the multispectral remote sensing images of each photovoltaic panel at all analysis times in the historical analysis period are fitted in three-dimensional space to obtain the spatiotemporal gradient representation of the photovoltaic panel discrimination value; Based on the spatiotemporal gradient representation of dust discrimination value and photovoltaic panel discrimination value, the dust discrimination value and photovoltaic panel discrimination value of all local positions of each multispectral remote sensing image are optimized to obtain new dust discrimination value and photovoltaic panel discrimination value of each local position of each multispectral remote sensing image; Based on the current dust discrimination value and the current photovoltaic panel discrimination value of all local position points of the multispectral remote sensing image, the photovoltaic panel cleanliness of each photovoltaic panel at each analysis moment in the historical analysis period is determined.

7. The photovoltaic panel cleaning method based on artificial rainfall enhancement technology according to claim 1, characterized in that: S3: Based on the rainfall data of the photovoltaic power station area during the historical analysis period and the cleanliness data of all photovoltaic panels in the photovoltaic power station area during the historical analysis period, the current optimal rainfall parameters of the photovoltaic power station area are analyzed, including: From the rainfall data of the photovoltaic power station area in the historical analysis period, the rainfall parameters of the photovoltaic power station area at each analysis time in the historical analysis period are determined; From the cleanliness data of all photovoltaic panels in the photovoltaic power station area during the historical analysis period, the cleanliness of the photovoltaic panels in the photovoltaic power station area at each analysis moment during the historical analysis period is determined; The rainfall parameters and cleanliness of photovoltaic panels in the photovoltaic power station area at each analysis moment in the historical analysis period are aligned and the relationship is quantified to obtain the relationship model between rainfall parameters and cleanliness of photovoltaic panels; The target cleanliness of photovoltaic panels in the photovoltaic power station area is input into the relationship model to determine the current optimal rainfall parameters in the photovoltaic power station area.

8. The photovoltaic panel cleaning method based on artificial rainfall enhancement technology according to claim 7 is characterized in that: Also includes: When the updated photovoltaic panel cleanliness of all photovoltaic panels in the photovoltaic power station area and the updated total rainfall data in the photovoltaic power station area are obtained, the relationship model is optimized based on the updated photovoltaic panel cleanliness of all photovoltaic panels in the photovoltaic power station area and the updated total rainfall data in the photovoltaic power station area to obtain a new relationship model; The target PV panel cleanliness of the PV power station area is input into the new relationship model to obtain the updated optimal rainfall parameters of the PV power station area.

9. The photovoltaic panel cleaning method based on artificial rainfall enhancement technology according to claim 1, characterized in that: S4: Based on the current optimal rainfall parameters in the photovoltaic power station area and the future rainfall data in the photovoltaic power station area, determine the artificial rainfall increase data in the photovoltaic power station area in the future period, including: Based on the historical meteorological data and geographic spatial data of the photovoltaic power station area and the rainfall data prediction model, the future rainfall data of the photovoltaic power station area is predicted; From the future rainfall data of the photovoltaic power station area, the future rainfall parameters of the photovoltaic power station area at each analysis time in the future analysis period are determined; Based on the future rainfall parameters of the photovoltaic power station area at each analysis moment in the future analysis period and the current optimal rainfall parameters of the photovoltaic power station area, the artificial rainfall parameters of the photovoltaic power station area at each analysis moment in the future period are determined; The artificial rainfall parameters of the photovoltaic power station area at all times in the future period are regarded as the artificial rainfall data of the photovoltaic power station area in the future period.

10. A photovoltaic panel cleaning system based on artificial rainfall enhancement technology, characterized in that: A photovoltaic panel cleaning method based on artificial rainfall enhancement technology for performing any one of claims 1 to 9, comprising: The segmentation and classification recognition module is used to identify all multispectral remote sensing images of all photovoltaic panels in the photovoltaic power station area within the historical analysis period from all full-range multispectral remote sensing images of the photovoltaic power station area within the historical analysis period based on the preset SAM+SVM embedded large model; A cleanliness analysis module is used to perform spectral analysis on all multispectral remote sensing images of each photovoltaic panel in the photovoltaic power station area during the historical analysis period, and obtain cleanliness data of each photovoltaic panel in the photovoltaic power station area during the historical analysis period; The optimal rainfall data analysis module is used to analyze the current optimal rainfall parameters in the photovoltaic power station area based on the rainfall data in the photovoltaic power station area during the historical analysis period and the cleanliness data of all photovoltaic panels in the photovoltaic power station area during the historical analysis period; The artificial rainfall data determination module is used to determine the artificial rainfall data of the photovoltaic power station area in the future period based on the current optimal rainfall parameters of the photovoltaic power station area and the future rainfall data of the photovoltaic power station area; The artificial rainfall control module is used to perform artificial rainfall control based on the artificial rainfall data of the photovoltaic power station area in the future period to obtain the artificial rainfall control result.

Images