Photovoltaic panel dust detection method, electronic device and storage medium

Abstract

The application discloses a kind of photovoltaic panel dust detection method, device, electronic equipment and storage medium, the method comprises: obtaining the first image of target area photovoltaic panel, and determining the photovoltaic component image of each photovoltaic component in first image;Dust identification is carried out to photovoltaic component image based on image processing model, and the first detection result of each photovoltaic component is determined;Image processing model is the model based on artificial neural network;Dust identification is carried out to photovoltaic component image based on preset image processing method, and the second detection result of each photovoltaic component is determined;Whether photovoltaic component is dust component is determined based on first detection result and second detection result.The technical scheme of the present application, by combining the detection result of photovoltaic component image using neural network model and the detection result of photovoltaic component image using traditional image processing method, whether photovoltaic component is dust component is determined, and the accuracy of photovoltaic panel dust detection is improved.

Description

[0001] This application is a divisional application of Chinese Patent Application No. 202510774246.2, filed on June 11, 2025, entitled "Method, Apparatus, Electronic Equipment and Storage Medium for Detecting Dust Accumulation on Photovoltaic Panels". Technical Field

[0002] This invention relates to the field of image processing technology, and in particular to a method, apparatus, electronic device, and storage medium for detecting dust accumulation on photovoltaic panels. Background Technology

[0003] With the rapid development of solar power generation technology, photovoltaic (PV) panels, as the core component of solar power systems, directly affect the power generation performance of the entire system through their operating efficiency. However, during actual operation, dust, dirt, and other contaminants easily accumulate on the surface of PV panels. These contaminants can block solar radiation, reduce the photoelectric conversion efficiency of the modules, and consequently lead to a significant decrease in power generation. Therefore, timely, accurate, and efficient dust accumulation detection on the surface of PV panels is particularly important.

[0004] Currently, deep learning models are commonly used to identify images of photovoltaic (PV) panels and determine the presence of dust deposits based on the identification results. However, the performance of deep learning models is highly dependent on the richness and quality of labeled data, while in real-world business scenarios, it is difficult to obtain sufficiently rich and diverse PV panel dust image data. Furthermore, the uneven distribution of dust on the surface of PV modules and the complex and variable lighting conditions further increase the difficulty of dust detection, potentially leading to missed detections and false positives in practical applications. Summary of the Invention

[0005] This invention provides a method, apparatus, electronic device, and storage medium for detecting dust accumulation on photovoltaic panels, thereby improving the accuracy of dust accumulation detection on photovoltaic panels.

[0006] According to one aspect of the present invention, a method for detecting dust accumulation on photovoltaic panels is provided, the method comprising:

[0007] A first image of a photovoltaic panel in a target area is acquired, and the photovoltaic module images of each photovoltaic component in the first image are determined; the photovoltaic panel includes multiple photovoltaic modules; the first image is a visible light image;

[0008] Dust recognition is performed on the photovoltaic module images based on an image processing model to determine the first detection result for each photovoltaic module.

[0009] Dust is identified in the photovoltaic module images based on a preset image processing method to determine a second detection result for each photovoltaic module; the preset image processing method is a method for processing images based on statistical principles and mathematical analysis principles;

[0010] Based on the first detection result and the second detection result, it is determined whether the photovoltaic module is a dust module; the dust module is the photovoltaic module with accumulated dust.

[0011] According to another aspect of the present invention, a photovoltaic panel dust accumulation detection device is provided, the device comprising:

[0012] An image determination module is used to acquire a first image of a photovoltaic panel in a target area and determine the photovoltaic module images of each photovoltaic component in the first image; the photovoltaic panel includes multiple photovoltaic modules; the first image is a visible light image;

[0013] The first detection module is used to perform dust recognition on the photovoltaic module image based on an image processing model, and determine the first detection result for each photovoltaic module.

[0014] The second detection module is used to identify dust in the photovoltaic module image based on a preset image processing method, and determine the second detection result for each photovoltaic module; the preset image processing method is a method for processing images based on statistical principles and mathematical analysis principles;

[0015] The third detection module is used to determine whether the photovoltaic module is a dust module based on the first detection result and the second detection result; the dust module is the photovoltaic module with accumulated dust.

[0016] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0017] At least one processor; and,

[0018] A memory communicatively connected to the at least one processor; wherein,

[0019] The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to perform the photovoltaic panel dust detection method according to any embodiment of the present invention.

[0020] According to another aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions for causing a processor to execute and implement the photovoltaic panel dust accumulation detection method according to any embodiment of the present invention.

[0021] The technical solution of this invention involves acquiring a first image of a photovoltaic panel in a target area. The photovoltaic panel includes multiple photovoltaic modules. The first image is used to determine the photovoltaic module images of each module, facilitating accurate location of each module for subsequent dust detection. On one hand, dust recognition is performed on the photovoltaic module images based on an image processing model to determine a first detection result for each module. The image processing model is based on an artificial neural network. On the other hand, a preset image processing method is used to perform dust recognition on the photovoltaic module images to determine a second detection result for each module. The preset image processing method is based on statistical and mathematical analysis principles. Furthermore, the first and second detection results are combined to determine the photovoltaic module. Is it a dusty component? A dusty component is a photovoltaic module with accumulated dust. Due to the complexity of photovoltaic power plant scenarios and limitations in data volume and diversity, dust detection using artificial neural network models cannot achieve high generalization ability. In practical applications to other power plants, the algorithm still suffers from missed detections or false detections. While traditional image processing methods can distinguish between dusty and dust-free areas to some extent, their adaptability is limited by lighting and weather conditions. Therefore, combining the detection results of both methods to analyze whether dust accumulates on photovoltaic modules can improve the algorithm's accuracy when the model has insufficient data or weak generalization ability. At the same time, it can effectively overcome the dependence of traditional methods on lighting and weather conditions, effectively reduce missed detections and false detections, and improve the accuracy of dust detection on photovoltaic panels.

[0022] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This is a flowchart of a method for detecting dust accumulation on photovoltaic panels according to an embodiment of the present invention;

[0025] Figure 2 This is a schematic diagram of a photovoltaic panel and photovoltaic module applicable to embodiments of the present invention;

[0026] Figure 3 This is a flowchart of another method for detecting dust accumulation on photovoltaic panels according to an embodiment of the present invention;

[0027] Figure 4 This is a flowchart of another method for detecting dust accumulation on photovoltaic panels according to an embodiment of the present invention;

[0028] Figure 5 This is a schematic diagram of a binarized image applicable to embodiments of the present invention;

[0029] Figure 6 This is a schematic diagram of a photovoltaic panel dust accumulation detection device according to an embodiment of the present invention;

[0030] Figure 7 This is a schematic diagram of the structure of an electronic device that implements the photovoltaic panel dust accumulation detection method according to an embodiment of the present invention. Detailed Implementation

[0031] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. 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 should fall within the scope of protection of the present invention.

[0032] It should be noted that the terms "first," "second," "refer to," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0033] Example 1

[0034] Figure 1 This is a flowchart illustrating a method for detecting dust accumulation on photovoltaic panels according to an embodiment of the present invention. This embodiment is applicable to detecting the presence of dust accumulation on photovoltaic panels. The method can be executed by a photovoltaic panel dust detection device, which can be implemented in hardware and / or software. This device can be configured in any electronic device with network communication capabilities. Figure 1 As shown, the photovoltaic panel dust accumulation detection method of the present invention includes the following process:

[0035] S110. Obtain a first image of the photovoltaic panel in the target area, and determine the photovoltaic module images of each photovoltaic module in the first image; the photovoltaic panel includes multiple photovoltaic modules; the first image is a visible light image.

[0036] The first image can be acquired by a drone photographing the photovoltaic panels in the target area. The drone is equipped with imaging equipment capable of capturing visible light images. It also carries a positioning module to accurately guide it to the target area for the photographing task. The positioning module can be configured with RTK positioning technology.

[0037] like Figure 2 As shown, the photovoltaic panel 1 is composed of multiple photovoltaic modules 2. The dashed box selects the photovoltaic panel, and the solid box selects one photovoltaic module on the photovoltaic panel. Figure 2 The photovoltaic panel comprises 22 photovoltaic modules. To more accurately locate areas of dust accumulation on the photovoltaic panel, feature extraction is performed on each photovoltaic module in the first image to obtain individual photovoltaic module images. This allows for subsequent dust detection on each photovoltaic module image, accurately pinpointing dust-accumulated areas and determining the presence of dust on each module for targeted cleaning.

[0038] Optionally, determining the photovoltaic module images of each photovoltaic module in the first image includes: processing the first image based on an image processing model to obtain the photovoltaic module images of each photovoltaic module in the first image. The image processing model is an artificial neural network-based model. The artificial neural network model can efficiently and quickly extract the feature information of each photovoltaic module in the first image, thereby accurately obtaining the photovoltaic module images of each photovoltaic module.

[0039] S120. Based on the image processing model, dust is identified in the photovoltaic module image to determine the first detection result for each photovoltaic module; the image processing model is a model based on artificial neural networks.

[0040] The first test result can be understood as the test result of whether there is dust accumulation on the photovoltaic module, as well as the index information of the dust accumulation status. The index information includes, but is not limited to, the level information of dust accumulation status, with more dust accumulation resulting in a higher level.

[0041] Specifically, the image processing model can identify dust features in photovoltaic module images, thereby determining the first detection result for each photovoltaic module based on the dust features.

[0042] S130. Dust recognition is performed on the photovoltaic module image based on a preset image processing method to determine the second detection result for each photovoltaic module; the preset image processing method is a method for processing images based on statistical principles and mathematical analysis principles.

[0043] The second detection result can be understood as the detection result of whether dust accumulates on the photovoltaic module, and the index information of the dust accumulation status. The index information includes, but is not limited to, the level information of dust accumulation status, with more dust and a higher level. The preset image processing method can be a method that processes images using various algorithms and formulas based on mathematical analysis principles and statistical principles.

[0044] Specifically, the preset image processing method can identify dust areas in the photovoltaic module image, and then determine the second detection result for each photovoltaic module based on the size information of the dust areas.

[0045] S140. Based on the first detection result and the second detection result, determine whether the photovoltaic module is a dust module; a dust module is a photovoltaic module with accumulated dust.

[0046] The present invention includes a first preset condition, a second preset condition, and a third preset condition. The first preset condition indicates that the photovoltaic module indicated by the first detection result obtained by the image processing model from the first image definitely has dust accumulation. The second preset condition indicates that the photovoltaic module indicated by the first detection result obtained by the image processing model from the first image may not have dust accumulation and requires further detection. The third preset condition indicates that the photovoltaic module indicated by the second detection result obtained by the preset image processing method from the first image definitely has dust accumulation.

[0047] Specifically, if the first test result meets the first preset condition, the photovoltaic module corresponding to the first test result that meets the first preset condition is regarded as a dust module; if the first test result meets the second preset condition, the photovoltaic module corresponding to the first test result that meets the second preset condition is further judged using the second test result, and the photovoltaic module corresponding to the first test result that meets the second preset condition is regarded as a candidate photovoltaic module; if the candidate photovoltaic module meets the third preset condition, the candidate photovoltaic module is a dust module, otherwise it is not a dust module.

[0048] As an optional embodiment, the first detection result includes first location information of the photovoltaic module. After determining whether the photovoltaic module is a dust module based on the first and second detection results, the method further includes steps A1-A3:

[0049] Step A1: Obtain a second image of the photovoltaic panel in the target area and determine the hot spot defects in the second image; the second image is a thermal imaging image; the hot spot defects correspond to hot spot information, which includes temperature information and second location information.

[0050] Specifically, an image processing model can be used to identify hot spots in the second image in order to accurately determine the hot spot defects in the second image.

[0051] Step A2: Based on the second location information corresponding to the hot spot defect in the second image and the first location information of each dust component, determine the reference hot spot defect that matches each dust component.

[0052] Specifically, the first position information is matched with the second position information to find the second position information corresponding to the first position information of each dust component, thereby accurately determining the reference hot spot defect matched for each dust component.

[0053] Step A3: Based on the temperature information corresponding to the reference hot spot defects matched with the dust component, determine the dust accumulation level of the dust component; the dust accumulation level is used to reflect the severity of dust accumulation.

[0054] Specifically, higher temperatures indicate more severe dust accumulation on the dust assembly. Therefore, the dust accumulation level of the dust assembly can be determined based on the preset temperature range and the temperature information corresponding to the reference hot spot defects matched to the dust assembly. The preset temperature range can be understood as the temperature range corresponding to different dust accumulation levels.

[0055] Optionally, the dust accumulation level of the dust component is determined based on the temperature information corresponding to the reference hot spot defect matched with the dust component, including: if the temperature information corresponding to the reference hot spot defect matched with the dust component is less than a first preset temperature, then the dust accumulation level of the dust component is level one; if the temperature information corresponding to the reference hot spot defect matched with the dust component is greater than the first preset temperature and less than a second preset temperature, then the dust accumulation level of the dust component is level two; if the temperature information corresponding to the reference hot spot defect matched with the dust component is greater than the second preset temperature, then the dust accumulation level of the dust component is level three; the dust accumulation severity of level one is less than the dust accumulation severity of level two, and the dust accumulation severity of level two is less than the dust accumulation severity of level three.

[0056] For example, the first preset temperature can be 40 degrees Celsius, and the second preset temperature can be 60 degrees Celsius. If the temperature information corresponding to the reference hot spot defect matched by the dust component is less than 40 degrees Celsius, then the dust accumulation level of the dust component is level one. If the temperature information corresponding to the reference hot spot defect matched by the dust component is greater than 40 degrees Celsius and less than 60 degrees Celsius, then the dust accumulation level of the dust component is level two. If the temperature information corresponding to the reference hot spot defect matched by the dust component is greater than 60 degrees Celsius, then the dust accumulation level of the dust component is level three.

[0057] This embodiment's technical solution involves acquiring a second image of the photovoltaic panel in the target area and identifying hot spot defects within that image. The second image is a thermal imaging image; the hot spot defects correspond to hot spot information, which includes temperature information and second location information. This allows for the determination of a matching reference hot spot defect for each dust component based on the second location information corresponding to the hot spot defects in the second image and the first location information of each dust component. Furthermore, based on the temperature information corresponding to the matching reference hot spot defects of the dust components, the dust accumulation level of the dust components is determined. The severity of dust accumulation is graded by hot spot temperature, thereby providing operational guidance for subsequent photovoltaic panel cleaning plans.

[0058] The technical solution of this invention involves acquiring a first image of a photovoltaic panel in a target area. The photovoltaic panel includes multiple photovoltaic modules. The first image is used to determine the photovoltaic module images of each module, facilitating accurate location of each module for subsequent dust detection. On one hand, dust recognition is performed on the photovoltaic module images based on an image processing model to determine a first detection result for each module. The image processing model is based on an artificial neural network. On the other hand, a preset image processing method is used to perform dust recognition on the photovoltaic module images to determine a second detection result for each module. The preset image processing method is based on statistical and mathematical analysis principles. Furthermore, the first and second detection results are combined to determine the photovoltaic module. Is it a dusty component? A dusty component is a photovoltaic module with accumulated dust. Due to the complexity of photovoltaic power plant scenarios and limitations in data volume and diversity, dust detection using artificial neural network models cannot achieve high generalization ability. In practical applications to other power plants, the algorithm still suffers from missed detections or false detections. While traditional image processing methods can distinguish between dusty and dust-free areas to some extent, their adaptability is limited by lighting and weather conditions. Therefore, combining the detection results of both methods to analyze whether dust accumulates on photovoltaic modules can improve the algorithm's accuracy when the model has insufficient data or weak generalization ability. At the same time, it can effectively overcome the dependence of traditional methods on lighting and weather conditions, effectively reduce missed detections and false detections, and improve the accuracy of dust detection on photovoltaic panels.

[0059] Example 2

[0060] Figure 3 This is a flowchart of another photovoltaic panel dust accumulation detection method provided by an embodiment of the present invention. The technical solution of this embodiment further optimizes the process of S140 in the aforementioned embodiments based on the above embodiments. This embodiment can be combined with various optional solutions in one or more of the above embodiments. Figure 3 As shown, the methods for detecting dust accumulation on photovoltaic panels include:

[0061] S210. Obtain a first image of the photovoltaic panel in the target area, and determine the photovoltaic module image of each photovoltaic module in the first image; the photovoltaic panel includes multiple photovoltaic modules; the first image is a visible light image.

[0062] S220. Based on the image processing model, dust is identified in the image of the photovoltaic module to determine the first detection result of each photovoltaic module; the image processing model is a model based on artificial neural network; the first detection result includes the first evaluation index of the photovoltaic module, which is used to describe the probability value of dust accumulation on the photovoltaic module.

[0063] The first evaluation indicator can be the confidence level value.

[0064] S230. Dust recognition is performed on the photovoltaic module image based on a preset image processing method to determine the second detection result for each photovoltaic module; the preset image processing method is a method for processing images based on statistical principles and mathematical analysis principles.

[0065] S240. If the first evaluation index of the photovoltaic module is greater than the first preset threshold, the photovoltaic module is determined to be a dust module.

[0066] The first preset threshold can be obtained from the analysis of experimental data. For example, the first preset threshold can be 0.65.

[0067] S250. If the first evaluation index of the photovoltaic module is less than the first preset threshold, then the photovoltaic module whose first evaluation index is less than the preset threshold is used as the reference photovoltaic module. Based on the second detection result, it is determined whether the reference photovoltaic module is a dust module.

[0068] If the first evaluation index of the photovoltaic module is less than the first preset threshold, it means that the photovoltaic module may not necessarily have dust. Further determination based on the second test results is needed to determine whether the photovoltaic module with the first evaluation index less than the preset threshold is a dust module.

[0069] Specifically, the present invention includes a third preset condition, which is used to indicate that the photovoltaic module indicated by the second detection result obtained by the preset image processing method from the first image must have dust accumulation. Therefore, determining whether a reference photovoltaic module is a dusty module based on the second detection result may include: if the reference photovoltaic module meets the third preset condition, then the reference photovoltaic module is determined to be a dusty module; otherwise, the reference photovoltaic module is not a dusty module.

[0070] In this embodiment, optionally, the second detection result includes a second evaluation index of the photovoltaic module. The second evaluation index is a score for evaluating the dust accumulation on the photovoltaic module. Based on the second detection result, determining whether a reference photovoltaic module is a dust module includes: using a reference photovoltaic module with a first evaluation index of zero as a first reference photovoltaic module, and using a reference photovoltaic module with a first evaluation index that is not zero as a second reference photovoltaic module; if the second evaluation index of the first reference photovoltaic module is greater than a second preset threshold, then the first reference photovoltaic module is a dust module; if the second evaluation index of the second reference photovoltaic module is greater than a third preset threshold, then the second reference photovoltaic module is a dust module; the second preset threshold is greater than the third preset threshold.

[0071] The second evaluation index can be understood as a score for the pollution level of the photovoltaic module. The second and third preset thresholds can be obtained based on experimental data analysis.

[0072] In this embodiment, the photovoltaic modules whose first evaluation index is less than the first preset threshold are further tested by combining the second evaluation index, thereby achieving more precise testing and effectively reducing the problems of missed detection and false detection.

[0073] The technical solution of this invention involves acquiring a first image of a photovoltaic panel in a target area and determining the photovoltaic module images of each photovoltaic component in the first image; the photovoltaic panel includes multiple photovoltaic modules; the first image is a visible light image. Dust is identified in the photovoltaic module images based on an image processing model to determine a first detection result for each photovoltaic module; the image processing model is based on an artificial neural network; the first detection result includes a first evaluation index for the photovoltaic module, which describes the probability of dust accumulation on the photovoltaic module. This invention displays the degree of dust accumulation on the photovoltaic module numerically, achieving data quantification and comparability, and providing greater objectivity. A second detection result for each photovoltaic module is determined based on a preset image processing method for dust identification in the photovoltaic module images; the preset image processing method is a method for processing images based on statistical principles and mathematical analysis principles. If the first evaluation index of the photovoltaic module is greater than a first preset threshold, the photovoltaic module is determined to be a dusty module. If the first evaluation index of a photovoltaic module is less than the first preset threshold, then the photovoltaic module whose first evaluation index is less than the preset threshold is used as a reference photovoltaic module. Based on the second detection result, it is determined whether the reference photovoltaic module is a dust module. This invention combines the second detection result to further detect the photovoltaic modules whose first evaluation index is less than the first preset threshold, thereby achieving more refined detection, effectively reducing the problems of missed detection and false detection, and improving the accuracy of photovoltaic panel dust detection.

[0074] Example 3

[0075] Figure 4 This is a flowchart of another photovoltaic panel dust accumulation detection method provided by an embodiment of the present invention. The technical solution of this embodiment further optimizes the process of identifying dust in photovoltaic module images based on a preset image processing method to determine the second detection result for each photovoltaic module, based on the aforementioned embodiments. This embodiment can be combined with various optional solutions in one or more of the above embodiments. Figure 4 As shown, the methods for detecting dust accumulation on photovoltaic panels include:

[0076] S310. Obtain a first image of the photovoltaic panel in the target area, and determine the photovoltaic module images of each photovoltaic module in the first image; the photovoltaic panel includes multiple photovoltaic modules; the first image is a visible light image.

[0077] S320. Based on the image processing model, dust is identified in the image of the photovoltaic module to determine the first detection result of each photovoltaic module; the image processing model is a model based on artificial neural networks.

[0078] S330. Preprocess the photovoltaic module image to obtain a preprocessed photovoltaic module image; the preprocessing includes at least one of grayscale conversion, Gaussian filtering and morphological operations.

[0079] Grayscale conversion is the process of converting a color image into a single-channel grayscale image, typically achieved by calculating the grayscale value of each pixel using a weighted average method. The formula for calculating the grayscale value is shown below:

[0080] I gray =0.299*R + 0.587*G + 0.114*B;

[0081] Where R, G, and B represent the red, green, and blue channel values ​​of the color image, respectively, and I gray These are the pixel values ​​after grayscale conversion.

[0082] Gaussian filtering is a linear smoothing filter that effectively suppresses high-frequency noise in images while preserving key edge information. The core of Gaussian filtering is to use a Gaussian kernel function to perform a weighted average on the image through convolution. The formula for the Gaussian kernel function is shown below:

[0083]

[0084] Where σ is the standard deviation of the Gaussian kernel, controlling the smoothness of the filter. Gaussian filtering can reduce noise interference in subsequent steps.

[0085] Morphological operations can include erosion and dilation. After noise removal, morphological operations are used to further process the image. The specific process is as follows: First, erosion is used to remove small disturbances such as white streaks on the photovoltaic panel surface. Erosion can reduce bright areas in the image and eliminate small white streaks or noise points. Next, dilation is used to enlarge small dust areas in the image, making them more visible. Dilation can expand bright areas in the image and enhance the connectivity of dust areas. The mathematical expressions for erosion and dilation are shown below:

[0086]

[0087] Where A is the input image, and B is the structuring element. Reflection of a structuring element.

[0088] S340. The preprocessed photovoltaic module image is analyzed based on the histogram analysis method to obtain a grayscale histogram.

[0089] One method is histogram analysis, which involves calculating the histogram of the preprocessed photovoltaic module image, and then performing Gaussian smoothing on the histogram to obtain a grayscale histogram.

[0090] S350. Binarize the grayscale histogram based on a preset threshold to obtain a binarized image of the photovoltaic module.

[0091] The preset threshold can be pre-set or calculated based on the image processing requirements. Specifically, by binarizing the pixels in the grayscale histogram using the preset threshold, the pixels in the grayscale histogram are divided into two regions corresponding to two values, as seen in the binarized image of a photovoltaic module. For example,... Figure 5 A schematic diagram of the binarized image shown.

[0092] In this embodiment of the invention, optionally, the preset threshold includes a first grayscale threshold and a second grayscale threshold. The grayscale histogram is binarized based on the preset threshold to obtain a binarized image of the photovoltaic module, including steps B1-B3:

[0093] Step B1: Determine the minimum gray value between the two largest peaks in the gray-level histogram based on the gray-level histogram, and use the minimum gray value as the first gray-level threshold; the first gray-level threshold is used to distinguish between normal areas and dust areas in the photovoltaic module image.

[0094] Specifically, determining the minimum gray value between the two largest peaks in a gray-level histogram based on the gray-level histogram can include: finding two peak points within a preset gray-level range; if two peaks exist within the preset gray-level range, calculating the minimum gray value between these two peaks; if not, searching for the minimum gray value between the two largest peaks globally. The preset gray-level range can be a gray-level value between 100 and 150.

[0095] Step B2: Set the pixels in the grayscale histogram that are greater than the second grayscale threshold to the first grayscale value; the second grayscale threshold is greater than the first grayscale threshold.

[0096] The second grayscale threshold can be a minimum pixel value corresponding to a normal area, set according to actual needs. For example, the second grayscale threshold can be 235.

[0097] Step B3: Set the pixels in the grayscale histogram that are less than the first grayscale threshold to the first grayscale value; set the pixels in the grayscale histogram that are greater than the first grayscale threshold to the second grayscale value; the area corresponding to the first grayscale value is the normal area, and the area corresponding to the second grayscale value is the dust area.

[0098] The first grayscale value and the second grayscale value can be two grayscale values ​​that clearly distinguish color regions. For example, the first grayscale value is 100 and the second grayscale value is 200.

[0099] In addition, it should be noted that in order to avoid the impact of the expansion of the white edge of the photovoltaic panel on the detection results, the reference dust area in the binarized image is removed; the reference dust area is the dust area in which the number of pixel rows and / or columns is less than 8% of the total area.

[0100] The technical solution of this embodiment performs binarization processing on the grayscale histogram through a first grayscale threshold and a second grayscale threshold, resulting in a more accurate binarized image of the photovoltaic module.

[0101] S360. The ratio of the corresponding dust area in the binarized image to the entire area of ​​the binarized image is used as the second detection result of the photovoltaic module.

[0102] S370. Based on the first and second detection results, determine whether the photovoltaic module is a dust module; a dust module is a photovoltaic module with accumulated dust.

[0103] The technical solution of this invention involves acquiring a first image of a photovoltaic panel in a target area and determining the photovoltaic module images of each photovoltaic component in the first image; the photovoltaic panel includes multiple photovoltaic modules; the first image is a visible light image. Dust recognition is performed on the photovoltaic module images based on an image processing model to determine a first detection result for each photovoltaic module; the image processing model is based on an artificial neural network. The photovoltaic module images are preprocessed to obtain preprocessed photovoltaic module images; the preprocessing includes grayscale conversion, Gaussian filtering, and morphological operations to effectively simplify the image while preserving its main structural information. The preprocessed photovoltaic module images are analyzed using a histogram analysis method to obtain a grayscale histogram, facilitating the subsequent determination of preset thresholds for dust and normal regions. The grayscale histogram is binarized based on the preset thresholds to obtain a binarized image of the photovoltaic module, effectively distinguishing between dust and normal regions. The ratio of the corresponding dust region in the binarized image to the entire region of the binarized image is used as the second detection result of the photovoltaic module, quantifying the second detection result. Finally, based on the first and second test results, it is determined whether the photovoltaic module is a dust module. This effectively overcomes the dependence of traditional methods on light and weather conditions, effectively reduces the problems of missed detection and false detection, and improves the accuracy of photovoltaic panel dust detection.

[0104] Example 4

[0105] In this embodiment, the image processing model is based on the YOLO11S-OBB algorithm. The YOLO11S-OBB algorithm is an algorithm formed by combining the YOLO11S algorithm with the directed bounding box algorithm. The image obtained by image processing using the YOLO11S-OBB algorithm has directional information.

[0106] Specifically, the orientation information can be described using oriented rotating boxes. The YOLO11S-OBB algorithm can robustly predict angles and output oriented rotating boxes. Because the photovoltaic panels are tilted to a certain extent in the image, the YOLO11S-OBB algorithm can detect the first image and effectively distinguish the boundary information of each photovoltaic module using oriented rotating boxes, accurately locating the position information of each photovoltaic module. Thus, the photovoltaic module images of each module in the first image can be accurately obtained.

[0107] In this embodiment, optionally, the YOLO11S-OBB algorithm also includes a C3K2-T module and a C2PSA-S module; the C3K2-T module has the ability to capture the detailed features of the photovoltaic panel in the image; the C2PSA-S module has the ability to segment the background information of the image; the C3K2-T module is formed by integrating a triple attention mechanism into the C3K2 module; the C2PSA-S module is formed by integrating the SEAM module into the PSA attention layer of the C2PSA module, and the repulsion loss of the SEAM module includes classification loss, equivariant regularization loss and equivariant cross regularization loss.

[0108] The triple attention mechanism includes channel attention, spatial attention, and contextual attention. The SEAM (Semantic Encoding with Attention Modules) module can effectively segment the background of an image.

[0109] Specifically, both the C3K2 module and the C2PSA module are part of the YOLO11S algorithm architecture.

[0110] While the C3K2 module demonstrates certain advantages in photovoltaic panel detection, it exhibits limitations when handling complex weather conditions, particularly in its potential inaccuracy in capturing detailed features of the photovoltaic panels. To further improve detection performance, this invention integrates a triple attention mechanism into the C3K2 module, forming the C3K2-T module. The C3K2-T module enhances the sensitivity of the image processing model to complex features by fusing channel attention, spatial attention, and contextual attention mechanisms, thereby improving the accuracy and robustness of photovoltaic panel detection.

[0111] Furthermore, the channel attention mechanism primarily enhances the influence of key feature channels by adjusting the weights of each channel, as shown in the following formula:

[0112] C=σ1(W c ·X);

[0113] Where C is the parameter corresponding to the channel attention mechanism, X is the input image, and W... c σ1 represents the weights of the channel attention mechanism, and σ1 is the first activation function.

[0114] Spatial attention mechanisms focus on key regions in an image by weighting the location of each pixel, as shown in the following formula:

[0115] S = σ²(W·X);

[0116] Where S is the parameter corresponding to the spatial attention mechanism, X is the input image, W is the weight of the spatial attention mechanism, and σ2 is the second activation function;

[0117] Context attention mechanisms capture more long-range dependencies by integrating global information. They are typically calculated using the following formula:

[0118] S context =Soft Max(W context ·X);

[0119] Among them, S context These are the parameters corresponding to the context attention mechanism, where X is the input image and W is the input image. context These are the weights of the context attention mechanism.

[0120] Finally, the weighted feature formula for the triple attention mechanism can be expressed as follows:

[0121] F out =X·C·S·S context .

[0122] Furthermore, it should be noted that while the YOLO11S algorithm performs excellently in feature extraction and object detection, it is prone to overlapping predicted bounding boxes or misclassification when faced with lighting, shadows, and background interference, leading to missed detections and false positives. The SEAM module, however, enhances the correlation between different channels by optimizing the convolutional structure. Combined with spatial attention and feature enhancement mechanisms, it improves detection performance under occluded conditions by focusing on the importance of unoccluded regions and improving overall feature representation. This method not only improves recognition accuracy under lighting and shadow conditions but also enhances the understanding of features in complex scenes, contributing to the model's accurate localization and recognition of photovoltaic modules. Therefore, the PSA attention layer in the C2PSA module of the YOLO11S algorithm is integrated with the SEAM module to form the C2PSA-S module of this invention.

[0123] The SEAM module's exclusion loss includes classification loss, isovariant regularization loss, and isovariant cross regularization loss. The classification loss is used for coarse object localization, the ER loss is used to narrow the gap between pixel-level and image-level monitoring, and the ECR loss is used to integrate PCM with the network to make consistent predictions across various affine transformations.

[0124] The feature map used for classification loss is then subjected to global average pooling and used to calculate the loss Z with the classification label. 0 and Z t These are two different sets of prediction results, where l is the true classification label, and l cls The basic classification loss function (such as cross-entropy loss, FocalLoss, etc.) outputs the classification error for a single sample or view, and the classification loss L... cls The formula is expressed as follows:

[0125]

[0126] Isovariant regularization loss is an index of the similarity between the CAM of the original image and the CAM of the affine transformed image. The predicted values... Mapping to a certain space using matrix A, and then comparing it with the target value The absolute value of the difference (L1 norm) is used as the error measure. The isovariant regularization loss L... ER The formula is expressed as follows:

[0127]

[0128] The equal-variable cross regularization loss will affect the observed value y 0 After mapping through matrix A, and with the target value Calculate the L1 norm of the difference, and then predict the value. Compared with another set of target values ​​y t The L1 norm is calculated from the differences, and finally, the absolute values ​​of the differences are taken and summed to measure the absolute error between the two vectors, i.e., the equivariant cross regularization loss. The formula for the absolute error between two vectors, i.e., the equivariant cross regularization loss L... ECR It is expressed as follows:

[0129]

[0130] Finally, the total loss function formula for the SEAM module is expressed as follows:

[0131] L = L cls +L ER +L ECR .

[0132] Optionally, the superiority of the YOLO11S-OBB algorithm of this invention can be demonstrated through comparative experiments on photovoltaic panel detection. Table 1 shows a comparative experiment between the YOLO11S-OBB algorithm of this invention and popular algorithms in recent years on the same dataset. To demonstrate the model's generalization ability, data from different photovoltaic power plants under different weather conditions were selected for testing to reflect the model's generalization performance. The method of this invention shows significant advantages in detection accuracy, robustness, and adaptability. Specifically, the YOLO11S-OBB algorithm can efficiently and accurately locate photovoltaic panel modules.

[0133] Table 1 Comparison Experiment with Mainstream Algorithms

[0134] algorithm Recall Map@0.5 mAP@0.5:0.95 Param / M GFLOPs Faster-RCNN 0.76 0.77 0.78 370.0 220 SSD 0.73 0.76 0.72 12.3 62 YOLOv5s 0.73 0.75 0.75 7.2 17 YOLOv8s 0.76 0.78 0.77 11.2 28 YOL011s-0BB 0.80 0.82 0.79 13.5 20 0urs 0.85 0.88 0.83 7.8 20

[0135] Table 2 shows the ablation experiment of the algorithm of this invention. Analysis of Table 2 shows that (1) YOLO11s-OBB+C2PSA-S, (2) YOLO11s-OBB+C3K2-T, and (3) YOLO11s-OBB+C2PSA-s+C3K2-T, compared with YOLO11s-OBB, respectively, improved the average accuracy by 1.2%, 2.4%, and 7.3%. Through comparison, the improved model in this paper achieved the highest accuracy improvement of 0.88, improving accuracy without significantly increasing computational load, demonstrating that the YOLO11s-OBB algorithm of this invention is superior in photovoltaic module detection tasks.

[0136] Table 2 Comparison of ablation experiments

[0137]

[0138] Optionally, the superiority of the YOLO11S-OBB algorithm of this invention can be demonstrated through comparative experiments on dust detection of photovoltaic panels. Table 3 below shows a comparative experiment on dust detection between the method of this invention and mainstream algorithms. The results show that the algorithm proposed in this paper has significant advantages in terms of detection effect and accuracy.

[0139] Table 3. Comparison of Dust Detection Experiments

[0140] algorithm Recall Map@0.5 mAP@0.5:0.95 Param / M GFLOPs Faster-RCNN 0.53 0.55 0.55 370.0 220 SSD 0.51 0.52 0.52 12.3 62 YOLOv5s 0.55 0.56 0.58 7.2 17 YOLOv8s 0.56 0.58 0.58 11.2 28 YOL011s-0BB 0.56 0.59 0.61 13.5 20 Ours 0.62 0.65 0.63 7.8 20

[0141] Due to the scarcity of dust data for photovoltaic (PV) panels, traditional single-module dust detection methods struggle to improve accuracy. To address this issue, this invention employs the YOLO11s-OBB algorithm to first detect the PV strings of the PV panel, then divides the PV strings into PV modules for dust detection. This strategy reduces background noise interference and provides high-quality data for dust detection. This method not only improves detection accuracy but also further optimizes the detection effect, providing more reliable data support for PV system maintenance.

[0142] This embodiment utilizes the YOLO11S-OBB algorithm, formed by combining the YOLO11S algorithm with the OBB algorithm, to detect the first image. The directional rotating bounding box effectively distinguishes the boundary information of each photovoltaic module, accurately locating their positions and thus obtaining accurate images of each module in the first image. Furthermore, improvements to the C3K2 and C2PSA modules in YOLO11S enhance the image processing model's sensitivity to complex features, thereby improving the accuracy and robustness of photovoltaic panel detection. It also eliminates the impact of complex weather conditions on dust detection, effectively improving recognition accuracy under varying lighting and shadow conditions, and enhancing the understanding of features in complex scenes.

[0143] Example 5

[0144] Figure 6 This is a schematic diagram of a photovoltaic panel dust accumulation detection device provided in an embodiment of the present invention. This embodiment is applicable to detecting the presence of dust accumulation on photovoltaic panels. The photovoltaic panel dust accumulation detection device can be implemented in hardware and / or software, and can be configured in any electronic device with network communication capabilities. Figure 3 As shown, the photovoltaic panel dust detection device includes:

[0145] Image determination module 410 is used to acquire a first image of a photovoltaic panel in a target area and determine the photovoltaic module images of each photovoltaic component in the first image; the photovoltaic panel includes multiple photovoltaic modules; the first image is a visible light image;

[0146] The first detection module 420 is used to perform dust recognition on the photovoltaic module image based on an image processing model, and determine the first detection result for each photovoltaic module; the image processing model is a model based on an artificial neural network.

[0147] The second detection module 430 is used to perform dust identification on the photovoltaic module image based on a preset image processing method, and determine the second detection result for each photovoltaic module; the preset image processing method is a method for processing images based on statistical principles and mathematical analysis principles;

[0148] The third detection module 440 is used to determine whether the photovoltaic module is a dust module based on the first detection result and the second detection result; the dust module is the photovoltaic module with accumulated dust.

[0149] Based on the above embodiments, optionally, the image determination module is used to: process the first image based on the image processing model to obtain photovoltaic module images of each photovoltaic module in the first image.

[0150] Based on the above embodiments, optionally, the image processing model is a model for image processing based on the YOLO11S-OBB algorithm; wherein, the YOLO11S-OBB algorithm is an algorithm formed by combining the YOLO11S algorithm with the directed bounding box algorithm. The image obtained by image processing using the YOLO11S-OBB algorithm has directional information.

[0151] Based on the above embodiments, optionally, the YOLO11S-OBB algorithm further includes a C3K2-T module and a C2PSA-S module; the C3K2-T module has the ability to capture detailed features of photovoltaic panels in the image; the C2PSA-S module has the ability to segment background information of the image; the C3K2-T module is formed by integrating a triple attention mechanism into the C3K2 module; the C2PSA-S module is formed by integrating a SEAM module into the PSA attention layer of the C2PSA module, and the repulsion loss of the SEAM module includes classification loss, equivariant regularization loss and equivariant cross regularization loss.

[0152] Based on the above embodiments, optionally, the first detection result includes a first evaluation index of the photovoltaic module, the first evaluation index being used to describe the probability value of dust accumulation on the photovoltaic module; the third detection module includes a first judgment unit, a second judgment unit, and a first detection unit;

[0153] The first judgment unit is used to determine that the photovoltaic module is a dust module if the first evaluation index of the photovoltaic module is greater than the first preset threshold.

[0154] The second judgment unit is used to take the photovoltaic module whose first evaluation index is less than the first preset threshold as a reference photovoltaic module if the first evaluation index of the photovoltaic module is less than the first preset threshold.

[0155] The first detection unit is used to determine whether the reference photovoltaic module is a dust module based on the second detection result.

[0156] Based on the above embodiments, optionally, the second detection result includes a second evaluation index of the photovoltaic module, the second evaluation index being a score for evaluating the dust accumulation on the photovoltaic module; the first detection unit is used to: use the reference photovoltaic module with a first evaluation index of zero as the first reference photovoltaic module, and use the reference photovoltaic module with a first evaluation index of non-zero as the second reference photovoltaic module; if the second evaluation index of the first reference photovoltaic module is greater than a second preset threshold, then the first reference photovoltaic module is a dust module; if the second evaluation index of the second reference photovoltaic module is greater than a third preset threshold, then the second reference photovoltaic module is a dust module; the second preset threshold is greater than the third preset threshold.

[0157] Based on the above embodiments, optionally, the first detection result includes the first position information of the photovoltaic module, and the photovoltaic panel dust accumulation detection device includes a fourth detection module, which includes an image determination unit, a hot spot defect determination unit, and a dust accumulation level determination unit.

[0158] An image determination unit is used to acquire a second image of a photovoltaic panel in a target area and determine hot spot defects in the second image; the hot spot defects correspond to hot spot information, and the hot spot information includes temperature information and second location information.

[0159] A hot spot defect determination unit is used to determine a reference hot spot defect matching each dust component based on the second location information corresponding to the hot spot defect in the second image and the first location information of each dust component.

[0160] The dust accumulation level determination unit is used to determine the dust accumulation level of the dust component based on the temperature information corresponding to the reference hot spot defect matched by the dust component; the dust accumulation level is used to reflect the severity of dust accumulation.

[0161] Based on the above embodiments, optionally, the dust accumulation level determination unit is configured to: if the temperature information corresponding to the reference hot spot defect matched by the dust component is less than a first preset temperature, then the dust accumulation level of the dust component is a first level; if the temperature information corresponding to the reference hot spot defect matched by the dust component is greater than the first preset temperature, and the temperature information corresponding to the reference hot spot defect matched by the dust component is less than a second preset temperature, then the dust accumulation level of the dust component is a second level; if the temperature information corresponding to the reference hot spot defect matched by the dust component is greater than the second preset temperature, then the dust accumulation level of the dust component is a third level; the dust accumulation severity of the first level is less than the dust accumulation severity of the second level, and the dust accumulation severity of the second level is less than the dust accumulation severity of the third level.

[0162] Based on the above embodiments, optionally, the preset image processing method is a histogram analysis method, and the second detection module includes a first image processing unit, an image analysis unit, a second image processing unit, and a second detection unit;

[0163] The first image processing unit is used to preprocess the photovoltaic module image to obtain a preprocessed photovoltaic module image; the preprocessing includes at least one of grayscale conversion, Gaussian filtering and morphological operations;

[0164] The image analysis unit is used to analyze the preprocessed photovoltaic module image based on the histogram analysis method to obtain a grayscale histogram.

[0165] The second image processing unit is used to perform binarization processing on the grayscale histogram based on a preset threshold to obtain a binarized image of the photovoltaic module.

[0166] The second detection unit is used to take the ratio of the corresponding dust area in the binarized image to the entire area of ​​the binarized image as the second detection result of the photovoltaic module.

[0167] Based on the above embodiments, optionally, the preset threshold includes a first grayscale threshold and a second grayscale threshold, and the second image processing unit is used for:

[0168] Based on the grayscale histogram, the minimum grayscale value between the two largest peaks in the grayscale histogram is determined, and the minimum grayscale value is used as the first grayscale threshold; the first grayscale threshold is used to distinguish between normal areas and dust areas in the photovoltaic module image;

[0169] Pixels in the grayscale histogram that are greater than a second grayscale threshold are set to a first grayscale value; the second grayscale threshold is greater than the first grayscale threshold.

[0170] Pixels in the grayscale histogram that are less than the first grayscale threshold are set to the first grayscale value;

[0171] Pixels in the grayscale histogram that are greater than the first grayscale threshold are set to the second grayscale value;

[0172] The area corresponding to the first gray value is the normal area, and the area corresponding to the second gray value is the dust area.

[0173] The photovoltaic panel dust accumulation detection device provided in this embodiment of the invention can execute the photovoltaic panel dust accumulation detection method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the method.

[0174] Example 6

[0175] According to embodiments of this disclosure, this disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0176] Figure 7 A schematic diagram of an electronic device that can be used to implement the photovoltaic panel dust accumulation detection method of embodiments of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workbenches, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0177] like Figure 7 As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded from storage unit 18 into the RAM 13. The RAM 13 may also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0178] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0179] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as the photovoltaic panel dust detection method.

[0180] In some embodiments, the photovoltaic panel dust detection method can be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the photovoltaic panel dust detection method described above can be performed. Alternatively, in other embodiments, processor 11 can be configured to perform the photovoltaic panel dust detection method by any other suitable means (e.g., by means of firmware).

[0181] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), system-on-a-chip (SoCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transferring data and instructions to the storage system, the at least one input device, and the at least one output device.

[0182] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0183] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0184] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0185] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or computing systems that include middleware components (e.g., application servers), or computing systems that include frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0186] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0187] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0188] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method for detecting dust accumulation on photovoltaic panels, characterized in that, The method includes: The drone is controlled to photograph the photovoltaic panel in the target area to obtain a first image; wherein the photovoltaic panel includes multiple photovoltaic modules, and the first image is a visible light image; Dust is identified in the photovoltaic modules in the first image based on an image processing model to determine the first detection result for each photovoltaic module; wherein, the image processing model is a model based on an artificial neural network; Dust is identified in the photovoltaic modules in the first image based on a preset image processing method, and a second detection result is determined for each photovoltaic module; the preset image processing method is a method for processing images based on statistical principles and mathematical analysis principles; Based on the first detection result and the second detection result, it is determined whether the photovoltaic module is a dust module; the dust module is a photovoltaic module with accumulated dust.

2. The method according to claim 1, characterized in that, The image processing model is based on the YOLO11S-OBB algorithm; the YOLO11S-OBB algorithm is an algorithm formed by combining the YOLO11S algorithm with the directed bounding box algorithm; the image obtained by image processing using the YOLO11S-OBB algorithm has directional information.

3. The method according to claim 2, characterized in that, The YOLO11S-OBB algorithm also includes a C3K2-T module and a C2PSA-S module; the C3K2-T module has the ability to capture detailed features of photovoltaic panels in the image; the C2PSA-S module has the ability to segment background information of the image; the C3K2-T module is formed by integrating a triple attention mechanism into the C3K2 module; the C2PSA-S module is formed by integrating a SEAM module into the PSA attention layer of the C2PSA module, and the repulsion loss of the SEAM module includes classification loss, equivariant regularization loss and equivariant cross regularization loss.

4. The method according to any one of claims 1 to 3, characterized in that, The first detection result includes a first evaluation index of the photovoltaic module, which describes the probability value of dust accumulation on the photovoltaic module; determining whether the photovoltaic module is a dust module based on the first detection result and the second detection result includes: If the first evaluation index of the photovoltaic module is greater than the first preset threshold, then the photovoltaic module is determined to be a dust module. If the first evaluation index of the photovoltaic module is less than the first preset threshold, then the photovoltaic module whose first evaluation index is less than the preset threshold is used as a reference photovoltaic module. Based on the second detection result, it is determined whether the reference photovoltaic module is a dust module.

5. The method according to claim 4, characterized in that, The second test result includes a second evaluation index for the photovoltaic module, which is a score for evaluating the dust accumulation on the photovoltaic module; The step of determining whether the reference photovoltaic module is a dust module based on the second detection result includes: The reference photovoltaic module with a first evaluation index of zero is used as the first reference photovoltaic module, and the reference photovoltaic module with a first evaluation index of non-zero is used as the second reference photovoltaic module. If the second evaluation index of the first reference photovoltaic module is greater than the second preset threshold, then the first reference photovoltaic module is a dust module; If the second evaluation index of the second reference photovoltaic module is greater than the third preset threshold, then the second reference photovoltaic module is a dust module; the second preset threshold is greater than the third preset threshold.

6. The method according to any one of claims 1 to 3, characterized in that, The first detection result includes the first location information of the photovoltaic module. After determining whether the photovoltaic module is a dust module based on the first detection result and the second detection result, the method further includes: A second image of the photovoltaic panel in the target area is acquired, and hot spot defects in the second image are identified; the second image is a thermal imaging image; the hot spot defects correspond to hot spot information, and the hot spot information includes temperature information and second location information. Based on the second location information corresponding to the hot spot defect in the second image and the first location information of each dust component, a reference hot spot defect matching each dust component is determined; Based on the temperature information corresponding to the reference hot spot defect matched by the dust component, the dust accumulation level of the dust component is determined; the dust accumulation level is used to reflect the severity of dust accumulation.

7. The method according to claim 6, characterized in that, The step of determining the dust accumulation level of the dust component based on the temperature information corresponding to the reference hot spot defects matched by the dust component includes: If the temperature information corresponding to the reference hot spot defect matched by the dust component is less than the first preset temperature, then the dust accumulation level of the dust component is the first level. If the temperature information corresponding to the reference hot spot defect matched by the dust component is greater than a first preset temperature, and the temperature information corresponding to the reference hot spot defect matched by the dust component is less than a second preset temperature, then the dust accumulation level of the dust component is the second level. If the temperature information corresponding to the reference hot spot defect matched by the dust component is greater than the second preset temperature, then the dust accumulation level of the dust component is the third level; the dust accumulation severity of the first level is less than the dust accumulation severity of the second level, and the dust accumulation severity of the second level is less than the dust accumulation severity of the third level.

8. The method according to any one of claims 1 to 3, characterized in that, The preset image processing method is a histogram analysis method. The step of identifying dust on the photovoltaic modules in the first image based on the preset image processing method, and determining a second detection result for each photovoltaic module, includes: The photovoltaic module in the first image is preprocessed to obtain a preprocessed photovoltaic module image; the preprocessing includes at least one of grayscale conversion, Gaussian filtering and morphological operations. The preprocessed photovoltaic module image is analyzed based on the histogram analysis method to obtain a grayscale histogram. The grayscale histogram is binarized based on a preset threshold to obtain a binarized image of the photovoltaic module. The ratio of the corresponding dust area in the binarized image to the entire area of ​​the binarized image is used as the second detection result of the photovoltaic module.

9. The method according to claim 8, characterized in that, The preset threshold includes a first grayscale threshold and a second grayscale threshold. Based on the preset threshold, the grayscale histogram is binarized to obtain a binarized image of the photovoltaic module, including: Based on the grayscale histogram, the minimum grayscale value between the two largest peaks in the grayscale histogram is determined, and the minimum grayscale value is used as the first grayscale threshold; the first grayscale threshold is used to distinguish between normal areas and dust areas in the photovoltaic module image; Pixels in the grayscale histogram that are greater than the second grayscale threshold are set to the first grayscale value; the second grayscale threshold is greater than the first grayscale threshold; Pixels in the grayscale histogram that are less than the first grayscale threshold are set to the first grayscale value; Pixels in the grayscale histogram that are greater than the first grayscale threshold are set to the second grayscale value; The area corresponding to the first gray value is the normal area, and the area corresponding to the second gray value is the dust area.

10. An electronic device, characterized in that, The electronic device includes: At least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the photovoltaic panel dust accumulation detection method according to any one of claims 1-9.

11. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that, when executed by a processor, implement the photovoltaic panel dust accumulation detection method according to any one of claims 1-9.

Images