Plateau photovoltaic panel smudginess identification method based on remote sensing and AI

Abstract

The invention discloses a plateau photovoltaic panel smudginess identification method based on remote sensing and AI, and relates to the technical field of photovoltaic panel smudginess identification, and the method comprises the following steps: obtaining a first gray scale function and a second gray scale function based on first type data, a first gray scale threshold value and a second gray scale threshold value; obtaining an initial dirty area based on the real-time illumination intensity, the first gray scale function and the second gray scale function; acquiring a historical power generation threshold based on the second type of data; acquiring a historical power generation function based on the second type of data and a historical power generation threshold; obtaining predicted fluctuation power generation power based on the real-time illumination intensity and a second type function; based on the real-time generated power and the predicted fluctuation generated power, judging whether the initial dirty area is dirty which affects the generating capacity or not; the method is used for solving the problem that in an existing photovoltaic panel smudginess identification technology, whether normal power generation of the photovoltaic panel is affected or not due to the fact that the identified smudginess cannot be further analyzed, and consequently the smudginess degree cannot be accurately divided.

Description

Technical Field

[0001] This invention relates to the field of photovoltaic panel dirt identification technology, specifically a method for identifying dirt on high-altitude photovoltaic panels based on remote sensing and AI. Background Technology

[0002] The unique geographical and climatic conditions of the plateau region, such as drought, strong winds and dust storms, intense ultraviolet radiation, and a fragile ecosystem, make it easy for pollutants such as dust to accumulate on the surface of photovoltaic panels. The shading caused by these pollutants will significantly reduce the light transmittance of the photovoltaic panels, resulting in a decrease in power generation efficiency. Therefore, it is necessary to identify whether the panels are dirty and then clean them. Existing photovoltaic (PV) panel dirt identification technologies, whether manual or image-based, have drawbacks. Manual inspection is highly subjective and cannot accurately classify the degree of dirt. While image detection can identify stains more accurately, some stains do not affect the power generation of the PV panel. If such stains are identified as requiring cleaning, it can lead to over-cleaning of the PV panel and waste of cleaning resources. For example, patent application CN119559113A discloses a method and system for detecting PV panel dust based on drone inspection images. This solution fails to determine whether the dirt will affect the normal power generation of the PV panel, resulting in an inaccurate classification of the degree of dirt and subsequent over-cleaning of the PV panel. Existing PV panel dirt identification technologies also fail to further analyze whether the identified dirt will affect the normal power generation of the PV panel, thus failing to accurately classify the degree of dirt. Summary of the Invention

[0003] This invention aims to at least partially solve one of the technical problems in the prior art. It obtains a first grayscale function and a second grayscale function based on a first type of data, a first grayscale threshold, and a second grayscale threshold; obtains an initial contaminated area based on real-time light intensity, the first grayscale function, and the second grayscale function; obtains a historical power generation threshold based on a second type of data; obtains a historical power generation function based on the second type of data and the historical power generation threshold; obtains a predicted fluctuating power generation based on real-time light intensity and the second type of function; and determines whether the initial contaminated area is contaminated and affects power generation based on the real-time power generation and the predicted fluctuating power generation. This addresses the problem in existing photovoltaic panel contamination identification technologies that fail to further analyze whether identified contamination affects the normal power generation of the photovoltaic panel, resulting in an inaccurate classification of contamination levels.

[0004] To achieve the above objectives, this application provides a method for identifying contaminants on high-altitude photovoltaic panels based on remote sensing and AI, comprising the following steps: Acquire an image of the photovoltaic panel to be inspected and label it as a real-time photovoltaic panel image; Obtain the first type of data, and obtain the first grayscale threshold and the second grayscale threshold based on the first type of data; The first grayscale function and the second grayscale function are obtained based on the first type of data, the first grayscale threshold, and the second grayscale threshold; Obtain the real-time illumination intensity, and obtain the initial dirty area based on the real-time illumination intensity, the first grayscale function, and the second grayscale function; Obtain the second type of data, and obtain historical power generation thresholds based on the second type of data; Historical power generation function is obtained based on the second type of data and historical power generation threshold; Predicted fluctuating power generation is obtained based on real-time illumination intensity and a second type function; Obtain real-time power generation and determine whether the initial contaminated area is contaminated and affects power generation based on the real-time power generation and the predicted fluctuating power generation.

[0005] Further, obtaining the first type of data, and obtaining the first grayscale threshold and the second grayscale threshold based on the first type of data, includes the following sub-steps: The first type of data includes clean photovoltaic panel images under different light intensities and different light intensities, which are labeled as the first historical light intensity and historical photovoltaic panel images, respectively. Historical photovoltaic panel images are obtained by converting them to grayscale. Obtain the grayscale values ​​of the power generation area in the historical grayscale image of photovoltaic panels and mark them as historical photovoltaic grayscale values.

[0006] Furthermore, obtaining the first grayscale threshold and the second grayscale threshold based on the first type of data also includes the following sub-steps: Under the same first historical illumination intensity, the range of historical photovoltaic grayscale values ​​is obtained, and a number axis is established with the range of historical photovoltaic grayscale values ​​as the number axis range, which is marked as the first number axis; Plot the historical photovoltaic grayscale values ​​on the first number line to obtain coordinate points, and mark them as the first coordinate points; Obtain the distribution length of the first coordinate point on the first number line and mark it as the first overall length; Get the number of the first coordinate points and mark it as the first overall count; Mark the line segment of length Z1 on the first number line as the first line segment; The threshold for the first number is calculated as: C1 = v1 × [(Z1 ÷ B1) × N1]; where C1 is the threshold for the first number, v1 is the percentage of the first number, B1 is the length of the first whole, and N1 is the number of the first whole. Mark the number of the first coordinate points on the first line segment as the first search count; Move the leftmost end of the first line segment to the leftmost first coordinate point. If the number of searches is less than the first number threshold, delete the leftmost first coordinate point and move the first line segment to the right so that the leftmost end of the first line segment coincides with the next first coordinate point. Stop when the number of searches is greater than or equal to the first number threshold. If not, continue deleting and moving. Obtain the minimum value of the first coordinate point after deletion and mark it as the first grayscale threshold. Move the rightmost end of the first line segment to the rightmost first coordinate point. If the number of searches is less than the first number threshold, delete the rightmost first coordinate point. Move the first line segment to the left so that the rightmost end of the first line segment coincides with the next first coordinate point. Stop when the number of searches is greater than or equal to the first number threshold. If not, continue deleting and moving. Obtain the maximum value of the first coordinate point after deletion and mark it as the second grayscale threshold.

[0007] Furthermore, obtaining the first grayscale function and the second grayscale function based on the first type of data, the first grayscale threshold, and the second grayscale threshold includes the following sub-steps: Obtain the first grayscale threshold and the second grayscale threshold for all first historical illumination intensities; A Cartesian coordinate system is established with the first historical illumination intensity as the horizontal axis value and the first grayscale threshold as the vertical axis value, and it is marked as the first grayscale coordinate system. The coordinate point with the first historical illumination intensity x-axis value and the first grayscale threshold y-axis value is marked as the first grayscale coordinate point; Plot all the first grayscale coordinate points in the first grayscale coordinate system and perform function fitting on all the first grayscale coordinate points to obtain a function, which is then labeled as the first grayscale function.

[0008] Furthermore, obtaining the first grayscale function and the second grayscale function based on the first type of data, the first grayscale threshold, and the second grayscale threshold also includes the following sub-steps: A Cartesian coordinate system is established with the first historical illumination intensity as the horizontal axis value and the second grayscale threshold as the vertical axis value, and it is marked as the second grayscale coordinate system. Mark the coordinate point with the first historical illumination intensity x-axis value and the second grayscale threshold y-axis value as the second grayscale coordinate point; Plot all the second grayscale coordinate points in the second grayscale coordinate system and perform function fitting on all the second grayscale coordinate points to obtain the function, which is then labeled as the second grayscale function.

[0009] Furthermore, obtaining the initial dirty area based on real-time illumination intensity, the first grayscale function, and the second grayscale function includes the following sub-steps: The real-time light intensity is used as the horizontal axis value and substituted into the first grayscale function and the second grayscale function respectively to obtain the vertical axis value, which is marked as the first real-time grayscale value and the second real-time grayscale value respectively. The real-time photovoltaic panel image is converted to grayscale to obtain a real-time photovoltaic panel grayscale image. Pixels whose grayscale values ​​in the real-time photovoltaic panel grayscale image are less than the first real-time grayscale value or greater than the second real-time grayscale value are marked as dirty pixels. Obtain the region consisting of all dirty pixels and mark it as the initial dirty region.

[0010] Further, obtaining the second type of data, and obtaining the historical power generation threshold based on the second type of data, includes the following sub-steps: The second type of data includes the power generation of photovoltaic panels under different light intensities and different light intensities, which are respectively labeled as the second historical light intensity and historical power generation. Under the same second historical illumination intensity, the range of historical power generation is obtained, and a number axis is established with the range of historical power generation as the number axis, which is marked as the second number axis. Plot the historical power generation on the second number line to obtain coordinate points, and mark them as the second coordinate points; Obtain the distribution length of the second coordinate point on the second number line and mark it as the second overall length; Get the number of second coordinate points and mark it as the second overall count; Mark the line segment of length Z2 on the second number line as the second line segment; The threshold for the second number is calculated as: C2 = v2 × [(Z2 ÷ B2) × N2]; where C2 is the threshold for the second number, v1 is the percentage of the second number, B2 is the length of the second whole, and N2 is the number of the second whole. Mark the number of second coordinate points on the second line segment as the second search count; Move the leftmost end of the second line segment to the leftmost second coordinate point. If the number of searches is less than the second threshold, delete the leftmost second coordinate point and move the second line segment to the right so that the leftmost end of the second line segment coincides with the next second coordinate point. Stop when the number of searches is greater than or equal to the second threshold. If not, continue deleting and moving. Obtain the minimum value of the second coordinate point after deletion and mark it as the historical power generation threshold.

[0011] Furthermore, obtaining the historical power generation function based on the second type of data and the historical power generation threshold includes the following sub-steps: Obtain all historical power generation thresholds for the second historical light intensity; A Cartesian coordinate system is established with the second historical light intensity as the horizontal axis value and the historical power generation threshold as the vertical axis value, and it is marked as the historical power generation coordinate system. Mark the coordinate point with the second historical light intensity horizontal axis value and the historical power generation threshold vertical axis value as the historical power generation coordinate point; All historical power generation coordinates are plotted in the historical power generation coordinate system, and a function is obtained by fitting all historical power generation coordinates to obtain a function, which is then marked as the historical power generation function.

[0012] Furthermore, obtaining the predicted fluctuating power generation based on real-time illumination intensity and the second type function includes the following sub-steps: The real-time light intensity is substituted into the historical power generation function as the horizontal axis value to obtain the vertical axis value, which is then marked as the predicted fluctuating power generation.

[0013] Furthermore, obtaining real-time power generation and determining whether the initial contaminated area is contaminated and affecting power generation based on the real-time power generation and the predicted fluctuating power generation includes the following sub-steps: If the real-time power generation is less than the predicted fluctuating power generation, the initial contaminated area is the contamination that affects power generation; if the real-time power generation is greater than or equal to the predicted fluctuating power generation, the initial contaminated area is the contamination that does not affect power generation.

[0014] The beneficial effects of this invention are as follows: This invention obtains a first grayscale function and a second grayscale function based on a first type of data, a first grayscale threshold, and a second grayscale threshold; obtains an initial dirty area based on real-time light intensity, the first grayscale function, and the second grayscale function; obtains a historical power generation threshold based on a second type of data; obtains a historical power generation function based on the second type of data and the historical power generation threshold; obtains a predicted fluctuating power generation based on real-time light intensity and the second type of function; and determines whether the initial dirty area is dirty and affects power generation based on real-time power generation and predicted fluctuating power generation. The advantage is that it can further determine whether the existing dirt will affect the normal power generation of the photovoltaic panel, improve the accuracy of dirt classification, and prevent excessive cleaning of the photovoltaic panel in the future. This invention obtains the initial dirty area based on real-time light intensity, a first grayscale function, and a second grayscale function. Its advantage lies in its ability to identify dirt under different light intensities, thereby improving the accuracy of dirt identification. Attached Figure Description

[0015] Figure 1 This is a flowchart illustrating the steps of the method of the present invention; Figure 2 This is a schematic diagram of the first grayscale function of the present invention; Figure 3 This is a schematic diagram of the second grayscale function of the present invention; Figure 4 This is a schematic diagram of the historical power generation function of the present invention. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] Example 1, please refer to Figure 1 As shown, this application provides a method for identifying contaminants on high-altitude photovoltaic panels based on remote sensing and AI, including the following steps: Step S1: Obtain an image of the photovoltaic panel to be detected and mark it as a real-time photovoltaic panel image; the real-time photovoltaic panel image is a surface image of the photovoltaic panel.

[0018] Step S2: Obtain the first type of data, and obtain the first grayscale threshold and the second grayscale threshold based on the first type of data; Step S2 includes the following sub-steps: Step S201, wherein the first type of data includes images of clean photovoltaic panels under different light intensities and images of clean photovoltaic panels under different light intensities, which are respectively labeled as the first historical light intensity and historical photovoltaic panel images; in order to obtain the gray value distribution range of the normal photovoltaic power generation area under different light intensities and improve the accuracy of dirt identification; Step S202: Perform grayscale processing on the historical photovoltaic panel image to obtain a historical photovoltaic panel grayscale image; Step S203: Obtain the grayscale value of the power generation area in the historical photovoltaic panel grayscale image and mark it as the historical photovoltaic grayscale value; Step S204: Under the same first historical illumination intensity, obtain the range of historical photovoltaic grayscale values, and establish a number axis with the range of historical photovoltaic grayscale values ​​as the number axis, which is marked as the first number axis; In order to obtain the range of historical photovoltaic grayscale values ​​under the same first historical illumination intensity, the first number axis is established to observe the distribution of historical photovoltaic grayscale values, facilitate the filtering out of historical photovoltaic grayscale values ​​that are too small or too large, and thus obtain the accurate distribution range of historical photovoltaic grayscale values; Step S205: Plot the historical photovoltaic grayscale values ​​on the first number line to obtain coordinate points, and mark them as the first coordinate points; Step S206: Obtain the distribution length of the first coordinate point on the first number line and mark it as the first overall length; Step S207: Obtain the number of first coordinate points and mark them as the first overall count; Step S208: Mark the line segment of length Z1 on the first number line as the first line segment; the first line segment is set in order to obtain the first coordinate points with sparse distribution, so the length of the first line segment should not be too long, for example, Z1 is 1cm; Step S209, calculate the first number threshold as: C1=v1×[(Z1÷B1)×N1]; where C1 is the first number threshold, v1 is the percentage of the first number, B1 is the length of the first whole, and N1 is the number of the first whole; the first number threshold is set to obtain the sparsely distributed first coordinate points, so the first number threshold should not be set too large. At the same time, since (Z1÷B1)×N1 represents the average number of the first coordinate points on the first line segment, v1 should not be set too large, for example, v1 is 0.1; In practical applications, for example, if the number of the first whole is 100 and the length of the first whole is 10cm, the threshold for the first number is calculated as: C1 = 0.1 × [(1 ÷ 20) × 100] = 2.

[0019] Step S210: Mark the number of first coordinate points on the first line segment as the first search count; Step S211: Move the leftmost end of the first line segment to the leftmost first coordinate point. If the number of searches is less than the first number threshold, delete the leftmost first coordinate point. Move the first line segment to the right so that the leftmost end of the first line segment coincides with the next first coordinate point. Stop when the number of searches is greater than or equal to the first number threshold. If not, continue deleting and moving. Obtain the minimum value of the first coordinate point after deletion and mark it as the first grayscale threshold. Delete excessively small and sparsely distributed first coordinate points by moving the first line segment, thereby filtering out excessively small historical photovoltaic grayscale values. Step S212: Move the rightmost end of the first line segment to the rightmost first coordinate point. If the number of searches is less than the first number threshold, delete the rightmost first coordinate point. Move the first line segment to the left so that the rightmost end of the first line segment coincides with the next first coordinate point. Stop if the number of searches is greater than or equal to the first number threshold. If not, continue deleting and moving. Obtain the maximum value of the first coordinate point after deletion and mark it as the second grayscale threshold. Delete excessively large and sparsely distributed first coordinate points by moving the first line segment, thereby filtering out excessively large historical photovoltaic grayscale values. In practical applications, for example, when the first historical illuminance is 500W / m 2 At that time, the first grayscale threshold obtained was 90, and the second grayscale threshold was 110.

[0020] Step S3 involves obtaining a first grayscale function and a second grayscale function based on the first type of data, the first grayscale threshold, and the second grayscale threshold. Step S3 includes the following sub-steps: Step S301: Obtain the first grayscale threshold and the second grayscale threshold for all first historical illumination intensities; Step S302: Establish a Cartesian coordinate system with the first historical illumination intensity as the horizontal axis value and the first grayscale threshold as the vertical axis value, and mark it as the first grayscale coordinate system; Step S303: Mark the coordinate point with the first historical illumination intensity horizontal axis value and the first grayscale threshold vertical axis value as the first grayscale coordinate point; Step S304: Plot all first grayscale coordinate points in the first grayscale coordinate system and perform function fitting on all first grayscale coordinate points to obtain a function, which is marked as the first grayscale function; in order to obtain the minimum grayscale value of the photovoltaic panel based on the light intensity; For practical applications, please refer to Figure 2 As shown, the first grayscale function is obtained.

[0021] Step S305: Establish a Cartesian coordinate system with the first historical illumination intensity as the horizontal axis value and the second grayscale threshold as the vertical axis value, and mark it as the second grayscale coordinate system; Step S306: Mark the coordinate point with the first historical illumination intensity horizontal axis value and the second grayscale threshold vertical axis value as the second grayscale coordinate point; Step S307: Plot all the second grayscale coordinate points in the second grayscale coordinate system and perform function fitting on all the second grayscale coordinate points to obtain a function, which is marked as the second grayscale function; in order to obtain the maximum grayscale value of the photovoltaic panel based on the light intensity; For practical applications, please refer to Figure 3 As shown, the second grayscale function is obtained.

[0022] Step S4: Obtain real-time illumination intensity, and obtain the initial dirty area based on the real-time illumination intensity, the first grayscale function, and the second grayscale function; Step S4 includes the following sub-steps: Step S401: Substitute the real-time light intensity as the horizontal axis value into the first grayscale function and the second grayscale function respectively to obtain the vertical axis value, and mark them as the first real-time grayscale value and the second real-time grayscale value respectively; obtain different first real-time grayscale values ​​and second real-time grayscale values ​​under different real-time light intensities to improve the recognition of the photovoltaic panel itself, obtain the distribution of grayscale values ​​of the clean photovoltaic panel, and thus more accurately obtain the dirty area; Step S402: Perform grayscale processing on the real-time photovoltaic panel image to obtain a real-time photovoltaic panel grayscale image; Step S403: Mark the pixels in the real-time photovoltaic panel grayscale image whose grayscale value is less than the first real-time grayscale value or greater than the second real-time grayscale value as dirty pixels. Step S404: Obtain the region composed of all dirty pixels and mark it as the initial dirty region; if dirt occurs, it will change the gray value, and when dirt occurs, the gray value will not be between the first real-time gray value and the second real-time gray value. For practical applications, please refer to Figure 2 and Figure 3 As shown, for example, the real-time illumination intensity obtained is 500W / m².2 Real-time light intensity 500W / m 2 Substituting the horizontal axis values ​​into the first and second grayscale functions respectively, we obtain vertical axis values ​​of 90 and 110. Thus, the first real-time grayscale value is 90 and the second real-time grayscale value is 110. Pixels in the real-time photovoltaic panel grayscale image whose grayscale value is less than the first real-time grayscale value of 90 or greater than the second real-time grayscale value of 110 are marked as dirty pixels.

[0023] Step S5: Obtain the second type of data, and obtain the historical power generation threshold based on the second type of data; Step S5 includes the following sub-steps: Step S501, the second type of data includes different light intensities and the power generation of the photovoltaic panel under different light intensities, which are respectively labeled as the second historical light intensity and historical power generation. Step S502: Under the same second historical illumination intensity, obtain the range of historical power generation, establish a number axis with the range of historical power generation as the range of the number axis, and mark it as the second number axis; establish the second number axis to observe the distribution of historical power generation and filter out power generation that is too small; Step S503: Plot the historical power generation on the second number line to obtain coordinate points, and mark them as the second coordinate points; Step S504: Obtain the distribution length of the second coordinate point on the second number line and mark it as the second overall length; Step S505: Obtain the number of second coordinate points and mark them as the second overall count; Step S506: Mark the line segment of length Z2 on the second number line as the second line segment; in order to obtain sparsely distributed second coordinate points, Z2 is set to be small, for example, Z2 is 1cm; Step S507, calculate the threshold of the second number as: C2=v2×[(Z2÷B2)×N2]; where C2 is the threshold of the second number, v1 is the percentage of the second number, B2 is the length of the second whole, and N2 is the number of the second whole; In practical applications, for example, if the number of the second whole is 100 and the length of the second whole is 10cm, the threshold for the second number is calculated as: C2 = 0.1 × [(1 ÷ 20) × 100] = 2.

[0024] Step S508: Mark the number of second coordinate points on the second line segment as the second search count; Step S509: Move the leftmost end of the second line segment to the leftmost second coordinate point. If the number of searches is less than the second threshold, delete the leftmost second coordinate point and move the second line segment to the right so that the leftmost end of the second line segment coincides with the next second coordinate point. Stop when the number of searches is greater than or equal to the second threshold. If not, continue deleting and moving. Obtain the minimum value of the second coordinate point after deletion and mark it as the historical power generation threshold. By moving the second line segment, search for historical power generation that is too small and too small in distribution. Obtain the minimum value of the accurate historical power generation under this second historical light intensity. For practical applications, please refer to Figure 4 As shown, for example, the second historical illuminance is 500W / m². 2 The historical power generation threshold obtained was 370W.

[0025] Step S6: Obtain the historical power generation function based on the second type of data and the historical power generation threshold; Step S6 includes the following sub-steps: Step S601: Obtain all historical power generation thresholds for the second historical light intensity; Step S602: Establish a Cartesian coordinate system with the second historical light intensity as the horizontal axis value and the historical power generation threshold as the vertical axis value, and mark it as the historical power generation coordinate system; Step S603: Mark the coordinate point with the second historical light intensity horizontal axis value and the historical power generation threshold vertical axis value as the historical power generation coordinate point; Step S604: Plot all historical power generation coordinate points in the historical power generation coordinate system and perform function fitting on all historical power generation coordinate points to obtain a function, which is then marked as the historical power generation function. For practical applications, please refer to Figure 4 The diagram shown illustrates the acquisition of historical power generation functions.

[0026] Step S7: Obtain the predicted fluctuating power generation based on real-time light intensity and the second type function; Step S7 includes the following sub-steps: Step S701: Substitute the real-time light intensity as the horizontal axis value into the historical power generation function to obtain the vertical axis value, and mark it as the predicted fluctuating power generation; the predicted fluctuating power generation represents the minimum power generation under the real-time light intensity. For practical applications, please refer to Figure 4 As shown, the real-time light intensity is 500W / m. 2 Substituting the horizontal axis value into the historical power generation function to obtain the vertical axis value of 370W, the predicted fluctuating power generation is 370W.

[0027] Step S8: Obtain real-time power generation; based on the real-time power generation and the predicted fluctuating power generation, determine whether the initial contaminated area is contaminated and affects power generation. Step S8 includes the following sub-steps: Step S801: If the real-time power generation is less than the predicted fluctuating power generation, the initial dirty area is the dirt that affects the power generation; if the real-time power generation is greater than or equal to the predicted fluctuating power generation, the initial dirty area is the dirt that does not affect the power generation. The predicted fluctuating power generation represents the minimum power generation under the real-time light intensity. If the real-time power generation is less than the predicted fluctuating power generation, it can be determined that the identified dirt affects the power generation, so the dirt is cleaned. In practical applications, the real-time power generation is 300W, which is less than the predicted fluctuating power generation of 370W. The initial dirty area is the dirt that affects the power generation and needs to be cleaned.

[0028] Example 2: This application also provides an electronic device, which may include: a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus. The memory stores computer-readable instructions, and the processor can call the instructions in the memory. When the computer-readable instructions are executed by the processor, the steps in the "High-Altitude Photovoltaic Panel Dirt Identification Method Based on Remote Sensing and AI" are performed, which can achieve the following functions: acquiring an image of the photovoltaic panel to be detected and marking it as a real-time photovoltaic panel image; acquiring first type data, and acquiring a first grayscale threshold and a second grayscale threshold based on the first type data; acquiring a first grayscale function and a second grayscale function based on the first type data, the first grayscale threshold, and the second grayscale threshold; acquiring real-time light intensity, and acquiring an initial dirty area based on the real-time light intensity, the first grayscale function, and the second grayscale function; acquiring second type data, and acquiring a historical power generation threshold based on the second type data; acquiring a historical power generation function based on the second type data and the historical power generation threshold; acquiring a predicted fluctuating power generation based on the real-time light intensity and the second type function; acquiring real-time power generation, and determining whether the initial dirty area is dirt affecting power generation based on the real-time power generation and the predicted fluctuating power generation.

[0029] Furthermore, when the logical instructions in the aforementioned memory can be implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0030] Example 3: This application also provides a computer program product, which includes a computer program stored on a computer-readable storage medium. The computer program includes program instructions. When the program instructions are executed by a computer, the computer can execute the plateau photovoltaic panel dirt identification method based on remote sensing and AI provided by the above methods. The method includes: acquiring an image of the photovoltaic panel to be detected and marking it as a real-time photovoltaic panel image; acquiring first type data and acquiring a first grayscale threshold and a second grayscale threshold based on the first type data; acquiring a first grayscale function and a second grayscale function based on the first type data, the first grayscale threshold, and the second grayscale threshold; acquiring real-time light intensity and acquiring an initial dirt area based on the real-time light intensity, the first grayscale function, and the second grayscale function; acquiring second type data and acquiring a historical power generation threshold based on the second type data; acquiring a historical power generation function based on the second type data and the historical power generation threshold; acquiring a predicted fluctuating power generation based on the real-time light intensity and the second type function; acquiring real-time power generation and determining whether the initial dirt area is dirt affecting power generation based on the real-time power generation and the predicted fluctuating power generation.

[0031] Example 4: This application also provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it performs the steps of the above-mentioned method for identifying contamination in high-altitude photovoltaic panels based on remote sensing and AI, to achieve the following functions: acquiring an image of the photovoltaic panel to be detected and marking it as a real-time photovoltaic panel image; acquiring first type data and acquiring a first grayscale threshold and a second grayscale threshold based on the first type data; acquiring a first grayscale function and a second grayscale function based on the first type data, the first grayscale threshold, and the second grayscale threshold; acquiring real-time light intensity and acquiring an initial contamination area based on the real-time light intensity, the first grayscale function, and the second grayscale function; acquiring second type data and acquiring a historical power generation threshold based on the second type data; acquiring a historical power generation function based on the second type data and the historical power generation threshold; acquiring a predicted fluctuating power generation based on the real-time light intensity and the second type function; acquiring real-time power generation and determining whether the initial contamination area is contamination affecting power generation based on the real-time power generation and the predicted fluctuating power generation.

[0032] Based on the above description of the embodiments, the embodiments of the present invention can be provided as methods, systems, or computer program products. Based on this understanding, the above technical solutions, in essence or in terms of their contribution to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or certain parts of the embodiments.

[0033] In the embodiments provided in this application, it should be understood that the disclosed system or method can be implemented in other ways. The embodiments described above are merely illustrative. For example, the division of modules or units is only a logical functional division, and there may be other division methods in actual implementation. Furthermore, multiple modules or units may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the coupling or direct coupling or communication connection shown or discussed may be through some communication interfaces. The indirect coupling or communication connection between systems, modules, and units may be electrical, mechanical, or other forms.

[0034] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for identifying contamination on high-altitude photovoltaic panels based on remote sensing and AI, characterized in that, Includes the following steps: Acquire an image of the photovoltaic panel to be inspected and label it as a real-time photovoltaic panel image; Obtain the first type of data, and obtain the first grayscale threshold and the second grayscale threshold based on the first type of data; The first grayscale function and the second grayscale function are obtained based on the first type of data, the first grayscale threshold, and the second grayscale threshold; Obtain the real-time illumination intensity, and obtain the initial dirty area based on the real-time illumination intensity, the first grayscale function, and the second grayscale function; Obtain the second type of data, and obtain historical power generation thresholds based on the second type of data; Historical power generation function is obtained based on the second type of data and historical power generation threshold; Predicted fluctuating power generation is obtained based on real-time illumination intensity and a second type function; Obtain real-time power generation and determine whether the initial contaminated area is contaminated and affects power generation based on the real-time power generation and the predicted fluctuating power generation.

2. The method for identifying contamination on high-altitude photovoltaic panels based on remote sensing and AI according to claim 1, characterized in that, Obtaining the first type of data, and obtaining the first grayscale threshold and the second grayscale threshold based on the first type of data, includes the following sub-steps: The first type of data includes clean photovoltaic panel images under different light intensities and different light intensities, which are labeled as the first historical light intensity and historical photovoltaic panel images, respectively. Historical photovoltaic panel images are obtained by converting them to grayscale. Obtain the grayscale values ​​of the power generation area in the historical grayscale image of photovoltaic panels and mark them as historical photovoltaic grayscale values.

3. The method for identifying contamination on high-altitude photovoltaic panels based on remote sensing and AI according to claim 2, characterized in that, Obtaining the first grayscale threshold and the second grayscale threshold based on the first type of data also includes the following sub-steps: Under the same first historical illumination intensity, the range of historical photovoltaic grayscale values ​​is obtained, and a number axis is established with the range of historical photovoltaic grayscale values ​​as the number axis range, which is marked as the first number axis; Plot the historical photovoltaic grayscale values ​​on the first number line to obtain coordinate points, and mark them as the first coordinate points; Obtain the distribution length of the first coordinate point on the first number line and mark it as the first overall length; Get the number of the first coordinate points and mark it as the first overall count; Mark the line segment of length Z1 on the first number line as the first line segment; The threshold for the first number is calculated as: C1 = v1 × [(Z1 ÷ B1) × N1]; where C1 is the threshold for the first number, v1 is the percentage of the first number, B1 is the length of the first whole, and N1 is the number of the first whole. Mark the number of the first coordinate points on the first line segment as the first search count; Move the leftmost end of the first line segment to the leftmost first coordinate point. If the number of searches is less than the first number threshold, delete the leftmost first coordinate point and move the first line segment to the right so that the leftmost end of the first line segment coincides with the next first coordinate point. Stop when the number of searches is greater than or equal to the first number threshold. If not, continue deleting and moving. Obtain the minimum value of the first coordinate point after deletion and mark it as the first grayscale threshold. Move the rightmost end of the first line segment to the rightmost first coordinate point. If the number of searches is less than the first number threshold, delete the rightmost first coordinate point. Move the first line segment to the left so that the rightmost end of the first line segment coincides with the next first coordinate point. Stop when the number of searches is greater than or equal to the first number threshold. If not, continue deleting and moving. Obtain the maximum value of the first coordinate point after deletion and mark it as the second grayscale threshold.

4. The method for identifying contamination on high-altitude photovoltaic panels based on remote sensing and AI according to claim 3, characterized in that, Obtaining the first grayscale function and the second grayscale function based on the first type of data, the first grayscale threshold, and the second grayscale threshold includes the following sub-steps: Obtain the first grayscale threshold and the second grayscale threshold for all first historical illumination intensities; A Cartesian coordinate system is established with the first historical illumination intensity as the horizontal axis value and the first grayscale threshold as the vertical axis value, and it is marked as the first grayscale coordinate system. The coordinate point with the first historical illumination intensity x-axis value and the first grayscale threshold y-axis value is marked as the first grayscale coordinate point; Plot all the first grayscale coordinate points in the first grayscale coordinate system and perform function fitting on all the first grayscale coordinate points to obtain a function, which is then labeled as the first grayscale function.

5. The method for identifying contamination on high-altitude photovoltaic panels based on remote sensing and AI according to claim 4, characterized in that, Obtaining the first grayscale function and the second grayscale function based on the first type of data, the first grayscale threshold, and the second grayscale threshold also includes the following sub-steps: A Cartesian coordinate system is established with the first historical illumination intensity as the horizontal axis value and the second grayscale threshold as the vertical axis value, and it is marked as the second grayscale coordinate system. Mark the coordinate point with the first historical illumination intensity x-axis value and the second grayscale threshold y-axis value as the second grayscale coordinate point; Plot all the second grayscale coordinate points in the second grayscale coordinate system and perform function fitting on all the second grayscale coordinate points to obtain the function, which is then labeled as the second grayscale function.

6. The method for identifying contamination on high-altitude photovoltaic panels based on remote sensing and AI according to claim 5, characterized in that, Obtaining the initial dirty area based on real-time illumination intensity, a first grayscale function, and a second grayscale function includes the following sub-steps: The real-time light intensity is used as the horizontal axis value and substituted into the first grayscale function and the second grayscale function respectively to obtain the vertical axis value, which is marked as the first real-time grayscale value and the second real-time grayscale value respectively. The real-time photovoltaic panel image is converted to grayscale to obtain a real-time photovoltaic panel grayscale image. Pixels whose grayscale value is less than the first real-time grayscale value or greater than the second real-time grayscale value in the real-time photovoltaic panel grayscale image are marked as dirty pixels. Obtain the region consisting of all dirty pixels and mark it as the initial dirty region.

7. The method for identifying contamination on high-altitude photovoltaic panels based on remote sensing and AI according to claim 6, characterized in that, Obtaining the second type of data and then using it to determine historical power generation thresholds includes the following sub-steps: The second type of data includes the power generation of photovoltaic panels under different light intensities and different light intensities, which are respectively labeled as the second historical light intensity and historical power generation. Under the same second historical illumination intensity, the range of historical power generation is obtained, and a number axis is established with the range of historical power generation as the number axis, which is marked as the second number axis. Plot the historical power generation on the second number line to obtain coordinate points, and mark them as the second coordinate points; Obtain the distribution length of the second coordinate point on the second number line and mark it as the second overall length; Get the number of second coordinate points and mark it as the second overall count; Mark the line segment of length Z2 on the second number line as the second line segment; The threshold for the second number is calculated as: C2 = v2 × [(Z2 ÷ B2) × N2]; where C2 is the threshold for the second number, v1 is the percentage of the second number, B2 is the length of the second whole, and N2 is the number of the second whole. Mark the number of second coordinate points on the second line segment as the second search count; Move the leftmost end of the second line segment to the leftmost second coordinate point. If the number of searches is less than the second threshold, delete the leftmost second coordinate point and move the second line segment to the right so that the leftmost end of the second line segment coincides with the next second coordinate point. Stop when the number of searches is greater than or equal to the second threshold. If not, continue deleting and moving. Obtain the minimum value of the second coordinate point after deletion and mark it as the historical power generation threshold.

8. The method for identifying contamination on high-altitude photovoltaic panels based on remote sensing and AI according to claim 7, characterized in that, Obtaining the historical power generation function based on the second type of data and historical power generation threshold includes the following sub-steps: Obtain all historical power generation thresholds for the second historical light intensity; A Cartesian coordinate system is established with the second historical light intensity as the horizontal axis value and the historical power generation threshold as the vertical axis value, and it is marked as the historical power generation coordinate system. Mark the coordinate point with the second historical light intensity horizontal axis value and the historical power generation threshold vertical axis value as the historical power generation coordinate point; All historical power generation coordinates are plotted in the historical power generation coordinate system, and a function is obtained by fitting all historical power generation coordinates to obtain a function, which is then marked as the historical power generation function.

9. The method for identifying contamination on high-altitude photovoltaic panels based on remote sensing and AI according to claim 8, characterized in that, Obtaining the predicted fluctuating power generation based on real-time illumination intensity and the second type function includes the following sub-steps: The real-time light intensity is substituted into the historical power generation function as the horizontal axis value to obtain the vertical axis value, which is then marked as the predicted fluctuating power generation.

10. The method for identifying contamination on high-altitude photovoltaic panels based on remote sensing and AI according to claim 9, characterized in that, Obtaining real-time power generation and determining whether the initial contaminated area is contaminated and affecting power generation based on real-time power generation and predicted fluctuating power generation includes the following sub-steps: If the real-time power generation is less than the predicted fluctuating power generation, the initial contaminated area is the contamination that affects power generation; if the real-time power generation is greater than or equal to the predicted fluctuating power generation, the initial contaminated area is the contamination that does not affect power generation.

Images