Photovoltaic panel dust deposition detection and power determination method and device, equipment and storage medium

Abstract

The embodiment of the invention provides a photovoltaic panel dust deposition detection and power determination method and device, equipment and a storage medium, and the method comprises the steps: carrying out the preprocessing of pre-obtained original data, obtaining sample data, and carrying out the detection of the sample data of each photovoltaic panel based on a pre-trained photovoltaic dust deposition detection model, the method comprises the steps of obtaining the dust deposition information of a photovoltaic panel and the area of the photovoltaic panel, establishing a relation model of the dust deposition information of the photovoltaic panel and the output power of the photovoltaic system, and determining the output power of the photovoltaic system through the relation model based on the dust deposition information of at least one photovoltaic panel and the area of the photovoltaic panel. And evaluating the power generation efficiency according to the output power and a preset rated power to obtain an evaluation result, and determining a photovoltaic panel cleaning scheme based on the evaluation result. According to the method, through automatic and precise dust deposition detection, the power generation efficiency evaluation accuracy and the operation and maintenance efficiency of the photovoltaic power station are improved, and the photovoltaic panel cleaning scheme is optimized.

Description

Technical Field

[0001] This application relates to the technical field, and in particular, to a method, device, equipment and storage medium for detecting dust accumulation on a photovoltaic panel and determining power. Background Art

[0002] Photovoltaic power generation, as a clean energy source, has been widely promoted and applied. However, the dust on the surface of the photovoltaic panel will directly block sunlight and reduce the effective light intensity reaching the photovoltaic cell. This blockage will cause the photoelectric conversion efficiency of the photovoltaic panel to decrease, thereby reducing the generation of electric energy. In addition, the dust accumulation results in an uneven light distribution on the surface of the photovoltaic panel, which is likely to generate a "hot spot effect" locally. This phenomenon will cause some areas to overheat, which will not only further reduce the overall conversion efficiency, but may also damage the local area of the photovoltaic cell and shorten the equipment life.

[0003] Currently, the detection of dust accumulation on photovoltaic panels usually relies on convolutional neural networks, which have a large amount of computation, low accuracy and precision. The determination of power usually relies on complex formulas and a large amount of calculation, with low efficiency.

[0004] In summary, how to flexibly, accurately and efficiently implement the detection of dust accumulation on photovoltaic panels and the determination of power is a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0005] Embodiments of this application provide a method, device, equipment and storage medium for detecting dust accumulation on a photovoltaic panel and determining power, so as to solve the problem of how to flexibly, accurately and efficiently implement the detection of dust accumulation on a photovoltaic panel and the determination of power.

[0006] In a first aspect, an embodiment of this application provides a method for detecting dust accumulation on a photovoltaic panel, including:

[0007] Preprocess the pre-acquired original data to obtain sample data, where the original data includes image data and video data of at least one photovoltaic panel collected by an airborne infrared dual-light camera on a drone;

[0008] For each photovoltaic panel, based on a pre-trained photovoltaic dust accumulation detection model, detect the sample data of the photovoltaic panel to obtain the dust accumulation information and the area of the photovoltaic panel.

[0009] In a possible implementation manner, the method further includes:

[0010] Based on a pre-set data acquisition mode, collect data from the at least one photovoltaic panel to obtain the original data.

[0011] In a possible implementation manner, the preprocessing of the pre-acquired original data to obtain sample data includes:

[0012] Crop the image data in the original data to obtain the cropped image data;

[0013] Perform enhancement processing on the cropped image data to obtain the enhanced image data;

[0014] Perform shadow detection and shadow enhancement processing on the image data in the original data to obtain the processed image data;

[0015] Perform coordinate calculation and duplicate removal and stitching on the processed image data to obtain the image data corresponding to the image data;

[0016] Construct the sample data according to the enhanced image data and the image data corresponding to the image data.

[0017] In a possible implementation manner, the method further includes:

[0018] Collect the original data of multiple photovoltaic systems;

[0019] Preprocess the original data of the multiple photovoltaic systems to obtain an initial data set;

[0020] Perform label annotation on the initial data set to obtain the label data corresponding to the initial data set;

[0021] Process the initial data set and the label data to obtain the homonymous mask image corresponding to the initial data set;

[0022] Obtain a training set according to the initial data set and the homonymous mask image corresponding to the initial data set;

[0023] Train a preset initial model based on the training set to obtain the photovoltaic dust accumulation detection model.

[0024] In a possible implementation manner, the initial model is an improved U-Net network model, and the improved U-Net network model includes multiple block blocks, and each block block is composed of two groups of depthwise separable convolutions and an ECA module.

[0025] In a second aspect, an embodiment of the present application provides a method for determining the power of a photovoltaic system, including:

[0026] Establish a relationship model between the dust accumulation information of the photovoltaic panel and the output power of the photovoltaic system;

[0027] Based on the dust accumulation information and the photovoltaic panel area of at least one photovoltaic panel, determine the output power of the photovoltaic system through the relationship model, where the dust accumulation information and the photovoltaic panel area are determined according to any one of the methods in the first aspect.

[0028] In a possible implementation, the method further includes:

[0029] Evaluating the power generation efficiency according to the output power and a preset rated power to obtain an evaluation result;

[0030] Based on the evaluation result, determining a cleaning scheme for the photovoltaic panels.

[0031] In a third aspect, an embodiment of the present application provides a photovoltaic panel dust accumulation detection device, including:

[0032] A first processing module, configured to preprocess pre-acquired original data to obtain sample data, where the original data includes image data and video data of at least one photovoltaic panel collected based on an airborne infrared dual-light camera of a drone;

[0033] A detection module, configured to detect the sample data of each photovoltaic panel based on a pre-trained photovoltaic dust accumulation detection model to obtain the dust accumulation information and the area of the photovoltaic panel of the photovoltaic panel.

[0034] In a possible implementation, the device further includes:

[0035] A first acquisition module, configured to collect data of the at least one photovoltaic panel based on a preset data acquisition mode to obtain the original data.

[0036] In a possible implementation, the first processing module is specifically configured to:

[0037] Crop the image data in the original data to obtain cropped image data;

[0038] Perform enhancement processing on the cropped image data to obtain enhanced image data;

[0039] Perform shadow detection and shadow enhancement processing on the video data in the original data to obtain processed video data;

[0040] Perform coordinate calculation and duplicate removal and stitching on the processed video data to obtain the image data corresponding to the video data;

[0041] Construct the sample data according to the enhanced image data and the image data corresponding to the video data.

[0042] In a possible implementation, the device further includes:

[0043] A second acquisition module, configured to collect original data of multiple photovoltaic systems;

[0044] A second processing module, configured to preprocess the original data of the multiple photovoltaic systems to obtain an initial data set;

[0045] An annotation module, configured to perform label annotation on the initial data set to obtain label data corresponding to the initial data set;

[0046] A third processing module, configured to process the initial data set and the label data to obtain a homonymous mask image corresponding to the initial data set;

[0047] A generation module, configured to obtain a training set according to the initial data set and the homonymous mask image corresponding to the initial data set;

[0048] A training module, configured to train a preset initial model based on the training set to obtain the photovoltaic dust accumulation detection model.

[0049] In a possible implementation manner, the initial model is an improved U-Net network model, and the improved U-Net network model includes multiple block blocks, and each block block is composed of two groups of depthwise separable convolutions and an ECA module.

[0050] In a fourth aspect, an embodiment of the present application provides a photovoltaic system power determination device, including:

[0051] A building module, configured to establish a relationship model between the dust accumulation information of a photovoltaic panel and the output power of a photovoltaic system;

[0052] A first determination module, configured to determine the output power of the photovoltaic system based on the dust accumulation information and the photovoltaic panel area of at least one photovoltaic panel through the relationship model, where the dust accumulation information and the photovoltaic panel area are determined according to any method in the first aspect.

[0053] In a possible implementation manner, the device further includes:

[0054] An evaluation module, configured to evaluate the power generation efficiency according to the output power and a preset rated power to obtain an evaluation result;

[0055] A second determination module, configured to determine a photovoltaic panel cleaning scheme based on the evaluation result.

[0056] In a fifth aspect, an embodiment of the present application provides an electronic device, including: a memory, a processor;

[0057] The memory stores computer execution instructions;

[0058] The processor executes the computer execution instructions stored in the memory, so that the processor executes the various possible implementation manners in the first aspect and the second aspect as above.

[0059] In a sixth aspect, an embodiment of the present application provides a computer-readable storage medium storing computer-executable instructions, which are used to implement various possible implementation manners of the first aspect and the second aspect as described above when executed by a processor.

[0060] In a seventh aspect, an embodiment of the present application provides a computer program product including a computer program, which implements various possible implementation manners of the first aspect and the second aspect as described above when executed by a processor.

[0061] The method, device, equipment, and storage medium for detecting dust accumulation and determining power of a photovoltaic panel provided by the embodiments of the present application preprocess the pre-acquired original data to obtain sample data. For each photovoltaic panel, based on a pre-trained photovoltaic dust accumulation detection model, the sample data of the photovoltaic panel is detected to obtain the dust accumulation information and the area of the photovoltaic panel. A relationship model between the dust accumulation information of the photovoltaic panel and the output power of the photovoltaic system is established. Based on the dust accumulation information and the area of at least one photovoltaic panel, through the relationship model, the output power of the photovoltaic system is determined. The power generation efficiency is evaluated according to the output power and a preset rated power to obtain an evaluation result. Based on the evaluation result, a cleaning plan for the photovoltaic panel is determined. The above method improves the accuracy of power generation efficiency evaluation and the operation and maintenance efficiency of the photovoltaic power station, and optimizes the cleaning plan of the photovoltaic panel through automated and precise dust accumulation detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] The drawings herein are incorporated into the specification and constitute a part of the specification, showing embodiments consistent with the present application, and are used together with the specification to explain the principles of the present application.

[0063] Figure 1 Schematic flow chart of the method for detecting dust accumulation on a photovoltaic panel provided by the present application Figure 1 ;

[0064] Figure 2 Schematic flow chart of the method for detecting dust accumulation on a photovoltaic panel provided by the present application Figure 2 ;

[0065] Figure 3 Schematic flow chart of the method for detecting dust accumulation on a photovoltaic panel provided by the present application Figure 3 ;

[0066] Figure 4 Schematic flow chart of determining the power of a photovoltaic system provided by the present application Figure 1 ;

[0067] Figure 5 Schematic structural diagram of the device for detecting dust accumulation on a photovoltaic panel provided by the present application;

[0068] Figure 6Structural schematic diagram of the photovoltaic system power determination device provided by the present application;

[0069] Figure 7 Structural schematic diagram of the electronic device provided by the present application.

[0070] Through the above-mentioned drawings, specific embodiments of the present application have been shown, and more detailed descriptions will be given in the following text. These drawings and textual descriptions are not intended to limit the scope of the concept of the present application in any way, but to illustrate the concept of the present application to those skilled in the art by referring to specific embodiments. Detailed implementation manners

[0071] Here, the exemplary embodiments will be described in detail, and the examples are shown in the drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementation manners described in the following exemplary embodiments do not represent all implementation manners consistent with the present application. On the contrary, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0072] With the increasing global demand for clean energy, photovoltaic power generation, as an efficient and environmentally friendly way of energy utilization, has been widely promoted and applied. A photovoltaic power station usually consists of a large number of photovoltaic panels, which convert sunlight into electrical energy and provide clean energy for the power grid. However, during the long-term operation of photovoltaic panels, dust is likely to accumulate on their surfaces, which not only affects the light transmittance of the photovoltaic panels but also reduces the photoelectric conversion efficiency, thus affecting the overall power generation performance of the photovoltaic power station. Therefore, regular inspection and cleaning and maintenance of photovoltaic panels to ensure their optimal working state are key links to ensure the efficient operation of photovoltaic power stations. At present, for the inspection and maintenance of photovoltaic panels, mainly two methods of manual inspection and drone inspection are adopted. Although manual inspection can visually check the dust accumulation on photovoltaic panels, there are problems such as low inspection efficiency and high costs. While drone inspection can improve the inspection efficiency, there are still some challenges in image processing and analysis. The existing image processing technologies mainly rely on traditional image segmentation and recognition algorithms, and these algorithms often have poor effects when dealing with complex backgrounds, light changes, etc., resulting in inaccurate segmentation of photovoltaic panels and low recognition accuracy of the dust accumulation state. In addition, when the existing technologies calculate the area and output power of photovoltaic panels, they usually need to rely on complex mathematical models and a large amount of computing resources, which not only increases the processing cost but also limits the wide application of the technology.

[0073] In view of the above problems, the present application provides a method, device, equipment and storage medium for detecting dust accumulation on photovoltaic panels and determining power, which realizes flexible, accurate and high-efficiency detection of dust accumulation on photovoltaic panels and determination of the output power of a photovoltaic system. Specifically, with the development of technology, the application of unmanned aerial vehicle (UAV) technology is becoming more and more extensive, especially in data collection and monitoring. At the same time, the application of image processing technology in various fields has also become increasingly important, especially in image segmentation and recognition. However, there are some problems in the existing technologies in this field. First, in terms of data collection and preprocessing, the existing technologies usually require a large amount of manual operations and have low efficiency. Second, in terms of automatic segmentation of photovoltaic panels and recognition of dust accumulation status, the existing technologies usually rely on traditional convolutional neural networks, which have a large amount of computation, long training time and low recognition accuracy, and cannot meet the actual needs. In addition, in terms of calculating the actual area and output power of photovoltaic panels, the existing technologies usually rely on complex formulas and a large amount of computation, and have low efficiency. Considering these problems, the inventors studied whether it is possible to intelligently and efficiently extract the contours of photovoltaic panels through algorithms, automatically calculate the area of the photovoltaic region and the severity of dust accumulation, so as to evaluate the power generation efficiency of a photovoltaic power station. Based on this, the solution of the present application is proposed.

[0074] The technical solution of the present application and how the technical solution of the present application solves the above technical problems will be described in detail below with specific embodiments. These specific embodiments below can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below with reference to the accompanying drawings.

[0075] Figure 1 Schematic flow of the method for detecting dust accumulation on photovoltaic panels provided by the present application Figure 1 , as Figure 1 shown, the method includes:

[0076] S101: Preprocess the pre-acquired raw data to obtain sample data.

[0077] In this step, in order to accurately detect the degree of dust accumulation on photovoltaic panels, after the raw data of the photovoltaic panels of a photovoltaic power station is collected, the raw data needs to be preprocessed to obtain sample data, where the raw data includes image data and video data of at least one photovoltaic panel collected by an infrared dual-light camera on an unmanned aerial vehicle.

[0078] Specifically, crop the image data in the original data to obtain the cropped image data, perform enhancement processing on the cropped image data to obtain the enhanced image data, perform shadow detection and shadow enhancement processing on the image data in the original data to obtain the processed image data, perform coordinate calculation and duplicate removal and stitching on the processed image data to obtain the image data corresponding to the image data, and construct sample data according to the enhanced image data and the image data corresponding to the image data.

[0079] S102: For each photovoltaic panel, based on a pre-trained photovoltaic panel dust accumulation detection model, detect the sample data of the photovoltaic panel to obtain the dust accumulation information and the area of the photovoltaic panel.

[0080] In this step, after completing the preprocessing of the original data, the obtained sample data is input into the trained photovoltaic panel dust accumulation detection model for dust accumulation detection, so as to obtain the dust accumulation information and the area of each photovoltaic panel.

[0081] Optionally, before step S101, it is also necessary to collect data from at least one photovoltaic panel based on a pre-set data collection mode to obtain the original data.

[0082] Specifically, control the drone to preset the data collection mode. In order to accurately collect the original data, use the drone oblique photography modeling method to collect data from the photovoltaic power station to generate a three-dimensional point cloud model of the photovoltaic power station, then plan the drone inspection route on the three-dimensional point cloud model, set the drone photo-taking points according to the coordinates of the photovoltaic panels, cover all the photovoltaic panels in the field area with global waypoints, and finally use the drone equipped with a visible light and infrared dual-light sensor to autonomously inspect and photograph the photovoltaic panels according to the preset task points, and collect high-resolution images and images containing different photovoltaic panel installation scenarios, including normal photovoltaic panels and dust-accumulated photovoltaic panels, so as to form the original data.

[0083] The photovoltaic panel dust accumulation detection method provided by the embodiment of the present application preprocesses the pre-obtained original data to obtain sample data, and for each photovoltaic panel, based on a pre-trained photovoltaic panel dust accumulation detection model, detects the sample data of the photovoltaic panel to obtain the dust accumulation information and the area of the photovoltaic panel. The above method can effectively improve the operation and maintenance efficiency and power generation efficiency of the photovoltaic power station and reduce the maintenance cost through automated and precise dust accumulation detection.

[0084] Figure 2 It is a flowchart of the photovoltaic panel dust accumulation detection method provided by the present application Figure 2 , such as Figure 2 shown. On the basis of the above embodiment, step S101 specifically includes:

[0085] S201: Crop the image data in the original data to obtain the cropped image data.

[0086] In this step, the image data in the original data contains a lot of irrelevant backgrounds (such as sky, ground, brackets, etc.). In order to focus on the photovoltaic panel area, reduce interference, and improve the efficiency of subsequent processing, the image data can be cropped.

[0087] Specifically, the position of the photovoltaic panel can be manually marked to determine the cropping area; or a target detection model (such as YOLO, Faster R-CNN) or an edge detection algorithm (such as Canny) can be used to automatically identify the photovoltaic panel area. If the photovoltaic panels are arranged regularly, batch cropping can be performed according to fixed dimensions and positions.

[0088] S202: Perform enhancement processing on the cropped image data to obtain the enhanced image data.

[0089] In this step, in order to make the cropped image data adapt to different changes such as illumination, angle, and dust accumulation degree, and improve the robustness of the model, the cropped image data can be enhanced.

[0090] Specifically, the enhancement processing includes image flipping, image rotation, and image color transformation. Among them, image flipping and image rotation can simulate the photovoltaic panel images under different perspectives, and image color transformation can adjust brightness, contrast, and saturation to simulate the images under different illumination conditions.

[0091] Optionally, for the enhancement processing, operations such as noise addition, blurring, and random occlusion can also be performed to improve the diversity of the data.

[0092] S203: Perform shadow detection and shadow enhancement processing on the image data in the original data to obtain the processed image data.

[0093] In this step, shadow is one of the main interference factors in the dust accumulation detection of photovoltaic panels. Through shadow detection and enhancement processing, the influence of shadow on dust accumulation detection can be reduced.

[0094] Specifically, threshold segmentation method (such as Otsu algorithm) or color space-based method (such as HSV space) can be used for shadow detection to detect the shadow area. For shadow enhancement, the brightness of the shadow area can be adjusted to make the contrast with the non-shadow area more obvious. Histogram equalization or adaptive contrast enhancement (CLAHE) method can also be used to enhance the details of the shadow area.

[0095] S204: Perform coordinate calculation and duplicate removal and stitching on the processed image data to obtain the image data corresponding to the image data.

[0096] In this step, the image data usually comes from drones or satellites and contains multiple images. Through coordinate calculation and stitching, multiple images can be integrated into a complete panoramic view of the power station.

[0097] Specifically, for coordinate calculation, professional tools or image processing software can be used to calculate the geographical coordinates of the image data to determine the position information of each image. For duplicate removal and stitching, feature point matching algorithms (such as SIFT, SURF) can be used to match the overlapping areas. Image stitching algorithms (such as multi-band fusion, Poisson fusion) can also be used to stitch multiple images into a complete panoramic view, removing duplicate areas to ensure seamless connection of the stitched images.

[0098] S205: Construct sample data based on the enhanced image data and the image data corresponding to the image data.

[0099] In this step, the enhanced image data is combined with the image data corresponding to the image data to construct sample data for actual detection. These sample data will be directly input into a pre-trained photovoltaic dust detection model for dust detection and photovoltaic panel area calculation.

[0100] Align the enhanced image data with the image data corresponding to the panoramic image data to ensure the same position and size of each photovoltaic panel. Alignment can be achieved in the following ways:

[0101] Feature point matching: Use feature point detection algorithms (such as SIFT, SURF) to match the photovoltaic panel areas in the enhanced image and the panoramic image.

[0102] Coordinate mapping: According to the coordinate information of the image data, map the enhanced image data to the corresponding position in the panoramic image.

[0103] Integrate the enhanced image data with the panoramic image data to form complete sample data. The integrated data can include the following information:

[0104] High-resolution images of each photovoltaic panel (from the enhanced image data).

[0105] The position and boundary information of each photovoltaic panel in the panoramic image (from the image data corresponding to the image data).

[0106] Format the sample data into the input format required by the model. For example: The image data can be converted into an RGB image of a fixed size. The position information can be converted into bounding box coordinates or mask data.

[0107] The method for detecting dust accumulation on photovoltaic panels provided by the embodiments of the present application crops the image data in the original data to obtain the cropped image data, performs enhancement processing on the cropped image data to obtain the enhanced image data, performs shadow detection and shadow enhancement processing on the image data in the original data to obtain the processed image data, performs coordinate calculation and duplicate removal and stitching on the processed image data to obtain the image data corresponding to the image data, and constructs sample data according to the enhanced image data and the image data corresponding to the image data. Through steps such as image cropping, enhancement, shadow detection, coordinate calculation, and duplicate removal and stitching, the above method constructs high-quality sample data for actual detection. These processing steps act together in the actual detection process, improving the accuracy and efficiency of detection and providing strong support for the operation and maintenance management of photovoltaic power stations.

[0108] Figure 3 Schematic flow of the method for detecting dust accumulation on photovoltaic panels provided by the present application Figure 3 , such as Figure 3 shown. On the basis of the above embodiments, the method further includes:

[0109] S301: Collect the original data of multiple photovoltaic systems.

[0110] In this step, diverse original data is collected to ensure a wide coverage of the training set and to be able to reflect the dust accumulation conditions under different photovoltaic systems and different environmental conditions.

[0111] Optionally, the data source can be one or more of the following methods:

[0112] UAV aerial photography: Obtain high-resolution images of photovoltaic panels.

[0113] Satellite imagery: Cover large areas of photovoltaic power stations.

[0114] Ground cameras: Monitor the dust accumulation situation of photovoltaic panels in real time.

[0115] Historical data: Obtain historical image data from the operation and maintenance records of photovoltaic power stations.

[0116] S302: Preprocess the original data of multiple photovoltaic systems to obtain an initial data set.

[0117] In this step, the original data is cleaned and standardized to remove noise and invalid data, providing high-quality input for subsequent annotation and training.

[0118] Optionally, the preprocessing can include the following methods:

[0119] Image denoising: Use filtering algorithms (such as Gaussian filtering, median filtering) to remove noise in the image.

[0120] Image enhancement: Adjust brightness, contrast, and sharpness to improve image quality.

[0121] Image cropping: Focus on the photovoltaic panel area and remove irrelevant backgrounds.

[0122] Data format standardization: Convert all data to a unified format (such as JPEG, PNG) and resolution.

[0123] S303: Perform label annotation on the initial dataset to obtain the label data corresponding to the initial dataset.

[0124] In this step, in order to achieve data diversification, the images in the initial dataset can be labeled to generate a JSON file containing the coordinates of the target area, providing label data for subsequent mask image generation.

[0125] Specifically, use a labeling tool to load the image data in the initial dataset, annotate the dust accumulation area through polygons, set corresponding labels for each area, save the annotation results, and generate a JSON file with the same name as the original image. The JSON file contains the coordinate information and label information of the target area.

[0126] S304: Process the initial dataset and the label data to obtain the mask image with the same name as the initial dataset.

[0127] In this step, convert the annotation information in the JSON file into a mask image. Different pixel values in the mask image represent areas with different dust accumulation levels, providing label data for subsequent model training.

[0128] Specifically, read the JSON file, extract the polygon coordinate information and corresponding labels stored in the file (such as "mild dust accumulation", "moderate dust accumulation", "severe dust accumulation"). According to the size of the original image, create a blank image of the same size (single-channel grayscale image), and set all initial pixel values to 0, indicating the area without dust accumulation. Traverse each annotation area in the JSON file, and assign different pixel values to areas with different dust accumulation levels according to the label information (such as 1 for "mild dust accumulation", 2 for "moderate dust accumulation", and 3 for "severe dust accumulation"). Use a polygon filling algorithm (such as scan-line filling) to draw the annotation area on the blank image and fill in the corresponding pixel values. Save the generated mask image as a file with the same name as the original image to ensure that the image and the mask correspond one by one.

[0129] S305: Obtain the training set based on the initial dataset and the mask image with the same name as the initial dataset.

[0130] In this step, construct a training set for model training, including input images and corresponding mask images. To further increase the quantity and diversity of the dataset, perform data augmentation.

[0131] S306: Train the preset initial model based on the training set to obtain a photovoltaic dust accumulation detection model.

[0132] In this step, after obtaining the training set, the initial model is trained based on the training set to obtain a photovoltaic dust accumulation detection model. Among them, the initial model is an improved Convolutional Networks for Biomedical Image Segmentation (U-Net) network model. The improved U-Net network model includes multiple block blocks, and each block block is composed of two groups of depthwise separable convolutions and an Efficient Channel Attention (ECA) module.

[0133] Optionally, depthwise separable convolutions are used to reduce the network computation amount and shrink the model size. The efficient ECA attention module is added between the two groups of depthwise separable convolutions, and a block block is composed of two groups of depthwise separable convolutions and an ECA attention module. Multiple block blocks are used to improve the segmentation performance of the multi-layer network.

[0134] Since the edges of the photovoltaic panel are relatively clear polygons, if the traditional cross-entropy is used, the weights of the edge region and the central region will be the same, which easily leads to average edge recognition effect. Therefore, the loss function needs to be modified. Optionally, the edge part can be extracted by contour detection, and then the pixels in the edge part are weighted when calculating the loss function, so that the model can pay more attention to the pixels in the edge part, thereby improving the model effect. The improved U-Net network meets the actual requirements of accurate segmentation of photovoltaic panel image data, rapid model training, and lightweight deployment.

[0135] Input the training set into the model for continuous training to realize the automatic segmentation of the photovoltaic panel and the recognition of the dust accumulation state, establish a photovoltaic panel dust accumulation severity model f(d), where d is the dust accumulation thickness, and calculate the pixel area of the photovoltaic panel. Calculate the actual area S of the photovoltaic panel through the field of view angle, flight height, and image resolution of the camera.

[0136] Exemplarily, to improve the segmentation effect of the edge region, the pixels in the edge region can be weighted. This can extract the edge part through a contour detection algorithm (such as Canny edge detection) and assign higher weights to the edge pixels when calculating the loss. The form of the weighted cross-entropy loss function can be expressed as:

[0137]

[0138] Among them, is the weight of pixel i, is the true label, is the predicted probability.

[0139] The method for detecting dust accumulation on photovoltaic panels provided by the embodiments of the present application collects the original data of multiple photovoltaic systems, preprocesses the original data of multiple photovoltaic systems to obtain an initial data set, performs label annotation on the initial data set to obtain label data corresponding to the initial data set, processes the initial data set and the label data to obtain a homonymous mask image corresponding to the initial data set, and based on the initial data set and the homonymous mask image corresponding to the initial data set, obtains a training set, and trains a preset initial model based on the training set to obtain a photovoltaic dust accumulation detection model. The above method realizes the construction and optimization of the photovoltaic dust accumulation detection model, and improves the model generalization ability, detection accuracy and practical application value.

[0140] Figure 4 is a schematic flow chart of determining the power of the photovoltaic system provided by the present application Figure 1 , as Figure 4 shown, the method specifically includes:

[0141] S401: Establish a relationship model between the dust accumulation information of the photovoltaic panel and the output power of the photovoltaic system.

[0142] In this step, based on field measurement and simulation, a relationship model between the dust accumulation severity and the photovoltaic conversion efficiency is established. The calculation formula is as follows:

[0143]

[0144] Where: is the efficiency of the photovoltaic panel when the dust accumulation severity is ; is the standard efficiency under the condition of no dust accumulation; is the linear attenuation coefficient; is the exponential attenuation coefficient.

[0145] S402: Based on the dust accumulation information and the photovoltaic panel area of at least one photovoltaic panel, determine the output power of the photovoltaic system through the relationship model.

[0146] In this step, first calculate the photovoltaic module efficiency , the photovoltaic module efficiency will change with the temperature, and the calculation formula is as follows:

[0147]

[0148] Where: is the module efficiency at the current module temperature Tc; is the component efficiency under standard test conditions (STC); is the temperature coefficient of the component. Tc can be obtained through the temperature measurement technology of an infrared dual-light camera.

[0149] Then calculate the photovoltaic system efficiency , the photovoltaic system efficiency is the synthesis of the efficiencies of all parts of the system, and the calculation formula is as follows:

[0150]

[0151] Where: is the efficiency of the photovoltaic component; is the efficiency of the inverter, usually 95% to 98%.

[0152] Then calculate the solar radiation intensity on the inclined plane , the solar radiation intensity on the inclined plane is one of the core parameters determining the output of the photovoltaic system. It includes the combination of direct radiation, scattered radiation and reflected radiation, and the calculation formula is as follows:

[0153]

[0154] Where: is the direct radiation intensity; is the incident angle; is the scattered radiation intensity; is the ground reflected radiation intensity.

[0155] Finally, calculate the output power of the photovoltaic system

[0156]

[0157] Where: is the rated power of the photovoltaic component (unit: kilowatt, kW), usually based on standard test conditions (STC: radiation intensity of 1000 W / m², component temperature of 25°C, AM1.5 spectrum); is the total area of the photovoltaic panels (unit: square meter, m²), which is the sum of the areas S of all photovoltaic panels.

[0158] Optionally, the method further includes:

[0159] S403: Evaluate the power generation efficiency according to the output power and the preset rated power to obtain an evaluation result.

[0160] S404: Determine the photovoltaic panel cleaning plan based on the evaluation result.

[0161] The rated power is the maximum power that a photovoltaic system can continuously output under standard test conditions (such as irradiance of 1000 W / m² and temperature of 25°C). This value is usually given in the technical specifications of the photovoltaic system.

[0162] The power generation efficiency can be calculated by the ratio of the output power to the rated power, i.e., Power generation efficiency = Output power / Rated power × 100%.

[0163] This value reflects the relative relationship between the power generation performance of the photovoltaic system under the current conditions and its rated performance.

[0164] If the power generation efficiency is high (close to or reaching 100%), it indicates that the photovoltaic system is operating well under the current conditions and no special cleaning or maintenance is required.

[0165] If the power generation efficiency is low (far lower than 100%), it may indicate that there is dust, dirt or other obstructions on the surface of the photovoltaic panels, resulting in a decline in the power generation performance of the system. At this time, a cleaning plan needs to be formulated and implemented to improve the power generation efficiency.

[0166] Based on the evaluation results of the power generation efficiency, determine whether the photovoltaic panels need to be cleaned and the urgency of cleaning. If the power generation efficiency drops significantly and the dust accumulation is serious, cleaning needs to be arranged as soon as possible.

[0167] According to the actual situation of the photovoltaic power station (such as scale, terrain, component type, etc.), select a suitable cleaning method. Common cleaning methods include manual cleaning, mechanical cleaning, drone cleaning and self-cleaning technology, etc.

[0168] Manual cleaning is suitable for small or scattered photovoltaic power stations, but it has low efficiency and high cost; mechanical cleaning and drone cleaning are suitable for large-scale and complex terrain photovoltaic power stations, with high efficiency but large initial investment; self-cleaning technology realizes self-cleaning through special materials, but it is still in the development stage.

[0169] Determine the specific time, frequency and scope of cleaning. The cleaning time should be selected during a period with weak light and suitable temperature; the cleaning frequency should be reasonably arranged according to the actual situation of the photovoltaic power station and the local climate conditions; the cleaning scope should include all photovoltaic panels affected by dust accumulation.

[0170] After cleaning is completed, re-measure the output power of the photovoltaic system and calculate the power generation efficiency to evaluate the cleaning effect. If the power generation efficiency is significantly improved, it indicates that the cleaning plan is effective; if the effect is not obvious, the reasons need to be further analyzed and the cleaning plan needs to be adjusted.

[0171] The power determination of the photovoltaic system provided by the embodiment of the present application establishes a relationship model between the dust accumulation information of the photovoltaic panel and the output power of the photovoltaic system. Based on the dust accumulation information and the area of at least one photovoltaic panel, the output power of the photovoltaic system is determined through the relationship model. The power generation efficiency is evaluated according to the output power and the preset rated power to obtain an evaluation result. Based on the evaluation result, a cleaning plan for the photovoltaic panel is determined. The above method improves the accuracy of power generation efficiency evaluation, optimizes the cleaning plan of the photovoltaic panel, improves the operation and maintenance efficiency and economic benefits, and promotes the sustainable development of photovoltaic technology.

[0172] Figure 5 As shown in the structure schematic diagram of the photovoltaic panel dust accumulation detection device provided by the present application, Figure 5 as shown, the photovoltaic panel dust accumulation detection device 500 provided in this embodiment includes:

[0173] A first processing module 501, configured to preprocess the pre-acquired original data to obtain sample data, where the original data includes image data and video data of at least one photovoltaic panel collected based on an airborne infrared dual-light camera of a drone.

[0174] A detection module 502, configured to detect the sample data of each photovoltaic panel based on a pre-trained photovoltaic dust accumulation detection model to obtain the dust accumulation information and the area of the photovoltaic panel.

[0175] Optionally, the photovoltaic panel dust accumulation detection device 500 further includes:

[0176] A first acquisition module 503, configured to collect data of at least one photovoltaic panel based on a preset data acquisition mode to obtain original data.

[0177] In a possible implementation manner, the first processing module 501 is specifically configured to:

[0178] Crop the image data in the original data to obtain the cropped image data;

[0179] Perform enhancement processing on the cropped image data to obtain the enhanced image data;

[0180] Perform shadow detection and shadow enhancement processing on the video data in the original data to obtain the processed video data;

[0181] Perform coordinate calculation and duplicate removal and stitching on the processed video data to obtain the image data corresponding to the video data;

[0182] Construct sample data according to the enhanced image data and the image data corresponding to the video data.

[0183] In a possible implementation manner, the photovoltaic panel dust accumulation detection device 500 further includes:

[0184] The second acquisition module 504 is configured to acquire the original data of multiple photovoltaic systems.

[0185] The second processing module 505 is configured to preprocess the original data of multiple photovoltaic systems to obtain an initial data set.

[0186] The annotation module 506 is configured to perform label annotation on the initial data set to obtain label data corresponding to the initial data set.

[0187] The third processing module 507 is configured to process the initial data set and the label data to obtain a homonymous mask image corresponding to the initial data set.

[0188] The generation module 508 is configured to obtain a training set according to the initial data set and the homonymous mask image corresponding to the initial data set.

[0189] The training module 509 is configured to train a preset initial model based on the training set to obtain a photovoltaic dust accumulation detection model.

[0190] Optionally, the initial model is an improved U-Net network model, and the improved U-Net network model includes multiple block blocks, and each block block is composed of two groups of depthwise separable convolutions and an ECA module.

[0191] The photovoltaic panel dust accumulation detection device provided in this embodiment can execute the photovoltaic panel dust accumulation detection method provided in the above method embodiment, and its implementation principle and technical effect are similar, and will not be elaborated here in this embodiment.

[0192] Figure 6 It is a schematic structural diagram of a photovoltaic system power determination device provided by the present application. As Figure 6 shown, the photovoltaic system power determination device 600 provided in this embodiment includes:

[0193] The establishment module 601 is configured to establish a relationship model between the dust accumulation information of the photovoltaic panel and the output power of the photovoltaic system.

[0194] The first determination module 602 is configured to determine the output power of the photovoltaic system based on the dust accumulation information and the photovoltaic panel area of at least one photovoltaic panel through the relationship model.

[0195] In a possible implementation manner, the photovoltaic system power determination device 600 further includes:

[0196] The evaluation module 603 is configured to evaluate the power generation efficiency according to the output power and a preset rated power to obtain an evaluation result.

[0197] The second determination module 604 is configured to determine a photovoltaic panel cleaning scheme based on the evaluation result.

[0198] The photovoltaic system power determination device provided in this embodiment can execute the photovoltaic system power determination method provided in the above method embodiment. The implementation principle and technical effect are similar, and will not be elaborated here in this embodiment.

[0199] Figure 7 It is a schematic structural diagram of the electronic device provided in this application. As Figure 7 shown, the electronic device 700 provided in this embodiment includes: at least one processor 701 and a memory 702. Optionally, the device 700 further includes a communication component 703. Among them, the processor 701, the memory 702, and the communication component 703 are connected through a bus 704.

[0200] In the specific implementation process, at least one processor 701 executes the computer execution instructions stored in the memory 702, so that at least one processor 701 executes the methods of the above various embodiments.

[0201] For the specific implementation process of the processor 701, reference can be made to the above various method embodiments. The implementation principle and technical effect are similar, and will not be elaborated here in this embodiment.

[0202] In the above embodiment, it should be understood that the processor may be a central processing unit (English: Central Processing Unit, abbreviated: CPU), or other general-purpose processors, digital signal processors (English: Digital Signal Processor, abbreviated: DSP), application specific integrated circuits (English: Application Specific Integrated Circuit, abbreviated: ASIC), etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The steps of the method disclosed in combination with the invention can be directly embodied as being executed by a hardware processor, or executed by a combination of hardware and software modules in the processor.

[0203] The memory may include a high-speed memory (Random Access Memory, RAM), and may also include a non-volatile memory (Non-volatile Memory, NVM), such as at least one disk memory.

[0204] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For the sake of convenience in representation, the buses in the attached drawings of this application are not limited to only one bus or one type of bus.

[0205] This application also provides a computer program product, including a computer program, which implements the methods of the above various embodiments when executed by a processor.

[0206] This application also provides a computer-readable storage medium, in which computer-executable instructions are stored. When the processor executes the computer-executable instructions, the methods of the above various embodiments are implemented.

[0207] The above-mentioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as a static random access memory (SRAM), an electrically erasable programmable read-only memory (EEPROM), an erasable programmable read-only memory (EPROM), a programmable read-only memory (PROM), a read-only memory (ROM), a magnetic memory, a flash memory, a magnetic disk or an optical disc. The readable storage medium can be any available medium accessible by a general-purpose or special-purpose computer.

[0208] An exemplary readable storage medium is coupled to the processor, so that the processor can read information from the readable storage medium and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can be located in an Application Specific Integrated Circuit (ASIC). Of course, the processor and the readable storage medium can also exist as discrete components in a device.

[0209] The division of units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection to each other can be through some interfaces, and the indirect coupling or communication connection of devices or units can be in an electrical, mechanical or other forms.

[0210] The unit described as a separation component may or may not be physically separated. The component displayed as a unit may or may not be a physical unit, that is, it may be located in one place or distributed across multiple network units. Some or all of these units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0211] In addition, in each embodiment of the present invention, each functional unit may be integrated into a processing unit, or each unit may exist physically alone, or two or more units may be integrated into one unit.

[0212] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this 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 for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods in each embodiment of the present invention. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs that can store program codes.

[0213] Those of ordinary skill in the art can understand that all or part of the steps to implement the above method embodiments can be completed by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When this program is executed, it executes the steps including the above method embodiments; and the aforementioned storage medium includes: various media such as ROM, RAM, magnetic disks, or optical discs that can store program codes.

[0214] Finally, it should be noted that: After considering the specification and practicing the invention disclosed herein, those skilled in the art will readily think of other implementation schemes of the present invention. The present invention aims to cover any variations, uses, or adaptive changes of the present invention. These variations, uses, or adaptive changes follow the general principles of the present invention and include the common general knowledge or conventional technical means in the technical field not disclosed in the present invention. It is not limited to the precise structures already described and shown in the drawings, and various modifications and changes can be made without departing from its scope. The scope of the present invention is only limited by the appended claims.

Claims

1. A photovoltaic panel dust accumulation detection method, characterized in that: include: Preprocessing the pre-acquired raw data to obtain sample data, wherein the raw data includes image data and video data of at least one photovoltaic panel collected by an infrared dual-light camera onboard the drone; For each photovoltaic panel, based on a pre-trained photovoltaic dust accumulation detection model, sample data of the photovoltaic panel is detected to obtain dust accumulation information and photovoltaic panel area of ​​the photovoltaic panel.

2. The method according to claim 1, characterized in that The method further comprises: Based on a preset data collection mode, data is collected on the at least one photovoltaic panel to obtain the raw data.

3. The method according to claim 1, characterized in that The preprocessing of the pre-acquired raw data to obtain sample data includes: Cropping the image data in the original data to obtain cropped image data; Performing enhancement processing on the cropped image data to obtain enhanced image data; Performing shadow detection and shadow enhancement processing on the image data in the original data to obtain processed image data; Performing coordinate calculation and de-duplication splicing on the processed image data to obtain image data corresponding to the image data; The sample data is constructed based on the enhanced image data and the image data corresponding to the image data.

4. The method according to claim 1, characterized in that The method further comprises: Collect raw data from multiple photovoltaic systems; Preprocessing the raw data of the plurality of photovoltaic systems to obtain an initial data set; Labeling the initial data set to obtain label data corresponding to the initial data set; Processing the initial data set and the label data to obtain a mask image with the same name corresponding to the initial data set; Obtaining a training set according to the initial data set and a mask image with the same name corresponding to the initial data set; The pre-set initial model is trained based on the training set to obtain the photovoltaic dust accumulation detection model.

5. The method according to claim 4, characterized in that The initial model is an improved U-Net network model, which includes multiple blocks, each of which is composed of two groups of depth-separable convolutions and high-speed channel attention ECA modules.

6. A method for determining the power of a photovoltaic system, characterized in that: include: Establish a relationship model between the dust accumulation information of photovoltaic panels and the output power of the photovoltaic system; Based on dust accumulation information and the area of ​​the photovoltaic panel of at least one photovoltaic panel, the output power of the photovoltaic system is determined through the relationship model, and the dust accumulation information and the area of ​​the photovoltaic panel are determined according to any one of the methods of claims 1 to 5.

7. The method according to claim 6, characterized in that The method further comprises: Evaluating the power generation efficiency according to the output power and the preset rated power to obtain an evaluation result; Based on the evaluation results, a photovoltaic panel cleaning plan is determined.

8. A photovoltaic panel dust accumulation detection device, characterized in that: include: A first processing module is used to pre-process the pre-acquired raw data to obtain sample data, wherein the raw data includes image data and video data of at least one photovoltaic panel collected by an infrared dual-light camera onboard the drone; The detection module is used to detect the sample data of each photovoltaic panel based on a pre-trained photovoltaic dust accumulation detection model to obtain the dust accumulation information and photovoltaic panel area of ​​the photovoltaic panel.

9. A photovoltaic system power determination device, characterized in that: include: Establish a module for establishing a relationship model between the dust accumulation information of the photovoltaic panel and the output power of the photovoltaic system; The first determination module is used to determine the output power of the photovoltaic system through the relationship model based on the dust accumulation information and the photovoltaic panel area of ​​at least one photovoltaic panel, and the dust accumulation information and the photovoltaic panel area are determined according to any one of the methods of claims 1 to 5.

10. An electronic device, characterized in that: include: Memory, processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory, so that the processor performs the method according to any one of claims 1 to 7.

11. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, which are used to implement the method according to any one of claims 1 to 7 when executed by a processor.

Images