Photovoltaic panel dust detection method and system and computer readable storage medium

Abstract

The invention belongs to the technical field of image processing, and particularly relates to a photovoltaic panel dust detection method and system and a computer readable storage medium. The method comprises the steps that an unmanned aerial vehicle is used for shooting a thermal imaging picture of a normal string, and a black-and-white image is obtained after opencv binarization processing and serves as a comparison sample; carrying out opencv binarization operation on thermal imaging photos shot in other photovoltaic panel detection areas; setting a pixel threshold value, and determining whether the photovoltaic string in the detection area needs to be cleaned or not; positioning the string of the image in the detection area after converting and correcting the GPS information coordinates of the image, and then correcting the data of the coordinate system; after string position information is obtained, whether cleaning is needed or not is judged by combining generating capacity data of a power station; and the photovoltaic cleaning robot autonomously plans a cleaning path. The problems that dust on the surface of a photovoltaic panel is inconvenient to see and positioning is inaccurate due to influences of the flight height, wind power and the like of an aircraft are solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of image processing, and more specifically, relates to a method and system for detecting dust on a photovoltaic panel and a computer-readable storage medium. Background Art

[0002] In today's energy field, distributed photovoltaic power stations, as a key component of clean energy, are gradually playing a crucial role. Photovoltaic panels, as the core components of solar power generation, directly determine the efficiency of solar power generation. However, in actual use, photovoltaic panels are often affected by various environmental factors, and dust is one factor that cannot be ignored. Dust adhering to the surface of photovoltaic panels will block the irradiation of sunlight, resulting in the inability of photovoltaic panels to fully absorb solar energy. Over time, the power generation efficiency of photovoltaic panels will be significantly reduced, thus affecting the economic benefits of the entire solar power generation system.

[0003] Due to the adhesion of dust, photovoltaic panels need to be cleaned regularly to maintain their performance. This not only increases the maintenance cost but also may damage the photovoltaic panels due to improper cleaning.

[0004] Chinese Patent CN116809578A discloses a method, device, and drone for cleaning a photovoltaic panel based on a drone, including determining the installation position information of a first photovoltaic panel in response to the drone entering the cleaning operation of the first photovoltaic panel, and determining the center point position information where the drone is to hover according to the installation position information; controlling the drone to dock at the center point to be hovered according to the center point position information; controlling a camera on the drone to collect an image of the first photovoltaic panel to obtain an image of the first photovoltaic panel; determining the position and type of pollutants on the first photovoltaic panel according to the image; and cleaning the first photovoltaic panel according to the position and type of pollutants in combination with a variety of cleaning structures, where the variety of cleaning structures are different cleaning structures provided on the drone.

[0005] In current technologies for cleaning and maintaining photovoltaic panels, a flying monitor is used to monitor each photovoltaic panel in a photovoltaic power station according to a preset movement trajectory to take photos of the photovoltaic panels to complete the cleaning, maintenance, and monitoring of the photovoltaic power station. However, due to the influence of factors such as the flight height of the aircraft and wind, it is not convenient to see the dust on the surface of the photovoltaic panels from the taken photos, and the positioning is inaccurate. Summary of the Invention

[0006] The present invention aims to overcome at least one defect of the above-mentioned prior art and provides a method for detecting dust on a photovoltaic panel to solve the problem that it is not convenient to see the dust on the surface of the photovoltaic panel from the taken photos due to the influence of factors such as the flight height of the aircraft and wind.

[0007] Based on the drone taking pictures of the photovoltaic panels according to a preset motion trajectory, this application adds the analysis of the power generation of the photovoltaic power station to comprehensively judge whether the photovoltaic string needs to be cleaned. It solves the problems that it is inconvenient to see the dust on the surface of the photovoltaic panels and the long time for large-area cleaning due to factors such as the flight altitude and wind force of the aircraft. This method can more accurately judge the position of the photovoltaic modules that need to be cleaned, greatly reduce the operation and maintenance costs of the power station, improve the power generation efficiency of the photovoltaic modules, and avoid potential hazards of the photovoltaic modules in the power station. After determining the string that needs to be cleaned, the system will select an appropriate time in combination with the weather information to command the photovoltaic cleaning robot to clean the detected string.

[0008] The detailed technical solution of the present invention is as follows:

[0009] S1. For a photovoltaic power station, under the same conditions, several groups of photovoltaic panel strings with clean and dust-free surfaces are divided, and a drone is used to take thermal imaging pictures of the normal strings. After binaryzation processing by opencv, a black-and-white image is obtained as a comparison sample.

[0010] S2. The thermal imaging pictures taken in other photovoltaic panel detection areas of the photovoltaic power station are subjected to opencv binaryzation operation to convert them into black-and-white images as a data set.

[0011] S3. Set a pixel threshold, compare the pixel values of the images in the data set with the pixel threshold, set the pixel values greater than the pixel threshold to 1, and set the pixel values less than the threshold to 0.

[0012] If the number of black pixels in the image of the detection area is less than the pixel threshold, the photovoltaic string in the detection area does not need to be cleaned;

[0013] If the number of black pixels in the image of the detection area is greater than the pixel threshold, then proceed to the next step;

[0014] S4. After correcting the coordinates through the GPS information of the picture, locate the string where the image of the detection area is located, and then correct the coordinate data of the coordinate system.

[0015] S5. After obtaining the string position information, in combination with the power generation data of the power station, if the power generation of the target detection string is less than 80% of that of the normal photovoltaic string, then it needs to be cleaned;

[0016] Otherwise, the impact on power generation is small and it does not need to be cleaned at present;

[0017] S6. The photovoltaic cleaning robot uses sensor technology and intelligent algorithms to autonomously plan the cleaning path according to the string to be cleaned;

[0018] Specifically, when it is detected that the photovoltaic string needs to be cleaned, an appropriate time is selected in combination with the weather information to command the photovoltaic cleaning robot to clean the detected string.

[0019] Furthermore, the opencv binarization is based on adaptive threshold parameters, specifically including:

[0020] S11. Calculate the histogram of the grayscale image, and calculate the number of pixels occupied by each pixel value from 0 to 255. The grayscale image is directly generated by the drone shooting.

[0021] S12. Traverse the threshold from 0 to 255. The pixels less than or equal to the threshold are the background, and the pixels greater than the threshold are the foreground.

[0022] S13. Calculate the proportion of the number of background pixels in the total number of pixels and the average value of the background pixels.

[0023] S14. Calculate the proportion of the number of foreground pixels in the total number of pixels and the average value of the foreground pixels.

[0024] S15. Calculate the between-class variance or within-class variance. The threshold that maximizes the between-class variance or minimizes the within-class variance is used as the optimal threshold. Preferably, the optimal threshold is the pixel threshold set by S3.

[0025] When the between-class variance is the largest, the difference between the foreground and the background is the largest, and the threshold selected at this time is the best:

[0026]

[0027] In formula (1), is the between-class variance, ω0 and ω1 are the pixel ratios of the foreground and the background respectively; μ0 and μ1 are the pixel means of the foreground and the background respectively.

[0028] When the within-class variance is the smallest, the pixel values within each category are more concentrated, and the threshold selected at this time is the best;

[0029]

[0030] In formula (2), is the within-class variance, and are the variances of the foreground and the background respectively;

[0031] S16. Use the optimal threshold to perform binarization processing on the image.

[0032] Furthermore, the correction of the coordinate system data specifically includes:

[0033] S41. Determine the coordinate system, including determining the source coordinate system and the target coordinate system.

[0034] S42. Extract the GPS information in the image, which is usually related to the Exif metadata. The GPS coordinates of the image are stored in the Exif data of the image.

[0035] S43. Perform latitude and longitude offset on the GPS information, and then perform offset normalization:

[0036] Convert the format of the GPS information of the picture to the decimal DD format:

[0037]

[0038] In formula (3), minutes is minutes, seconds is seconds, and Degress is degrees;

[0039] deltaLat and deltaLat are the offsets of latitude and longitude, which define the latitude and longitude offset:

[0040] deltaLat = transformLat(x, y) (4);

[0041] deltaLon = transformLon(x, y) (5);

[0042] In formulas (4)-(5), transformLat and transformLon represent the offsets of latitude and longitude calculated through complex non-linear functions;

[0043] Convert the latitude from angular units to radian units:

[0044]

[0045] Then, correct the influence of the Earth ellipsoid, with eccentricity e 2 That is, the difference between the equatorial radius and the polar radius, and magic is the sine value at the latitude, which is used to calculate the eccentricity correction at the latitude:

[0046] magic = sin(radLat) (7);

[0047] magic = 1 - e 2 *magic 2 (8);

[0048] Finally, perform offset normalization. a is the semi-major axis of the Earth ellipsoid, and 1 - e 2 is the eccentricity correction of the Earth ellipsoid, to obtain the normalized latitude adjustedLat and longitude adjustedLon:

[0049]

[0050]

[0051] S44. Reverse Geocoding: Receive a geographical coordinate (latitude, longitude) as input, and find the matching address by comparing the input coordinate with the location information in the geographical database;

[0052] Output the address information corresponding to the input geographical coordinate, including street, city, and country information. Use the database including address information to match the address through coordinate range search to achieve reverse geocoding;

[0053] Alternatively, use the ready-made reverse geocoding function of external APIs, such as the ready-made reverse geocoding functions of Google Maps API, Bing Maps, OpenStreetMap, etc.;

[0054] S45. Check whether the rectification result meets the expected range. If it meets, the rectification is completed; if not, rectify again;

[0055] Preferably, perform statistical analysis on the rectified data, calculate the statistic of the rectified coordinate system data, and if the statistic is within the set statistical threshold, it meets the expectation.

[0056] This method can greatly increase the accuracy of the coordinate system data and avoid potential hazards of power station photovoltaic modules.

[0057] The present invention further includes a photovoltaic panel dust detection system, comprising:

[0058] At least one processor; and

[0059] A memory storing instructions, which when executed by the at least one processor, cause the at least one processor to execute a photovoltaic panel dust detection method as described above.

[0060] On the other hand, the present invention also provides a computer-readable storage medium storing executable instructions, which when executed cause the machine to execute a photovoltaic panel dust detection method as described above.

[0061] Compared with the prior art, the beneficial effects of the present invention are:

[0062] A photovoltaic panel dust detection method, system, and computer-readable storage medium provided by the present invention comprehensively determine whether it is necessary to clean the photovoltaic modules based on the thermal imaging photos taken by the drone and the on-site power generation data, reducing the cleaning workload of the photovoltaic panels; at the same time, rectify the coordinate system data, and combine the real-time weather information to command the cleaning robot to automatically clean the photovoltaic string more accurately at a suitable time, avoiding cleaning the photovoltaic modules under windy, rainy, thunderstorm, or heavy snow weather conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 It is a schematic flowchart of a method for detecting dust on a photovoltaic panel according to the present invention.

[0064] Figure 2 It is a schematic flowchart of coordinate system data deviation correction in Embodiment 1 of the present invention.

[0065] Figure 3 It is a schematic flowchart of image binarization in Embodiment 1 of the present invention. Detailed implementation manners

[0066] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0067] It should be noted that the following detailed description is exemplary and is intended to provide further illustration of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which the present invention belongs.

[0068] It should be noted that the terms used herein are only for describing specific implementation manners and are not intended to limit the exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0069] Without conflict, the embodiments in the present invention and the features in the embodiments can be combined with each other.

[0070] Embodiment 1

[0071] Refer Figure 1 , this embodiment provides a method for detecting dust on a photovoltaic panel, and the method includes:

[0072] S1. The drone takes a thermal imaging picture and then performs opencv binarization;

[0073] For a photovoltaic power station, under the same conditions, several groups of normal photovoltaic panel strings, that is, those with clean surfaces without dust, are divided. The drone takes thermal imaging pictures of the normal strings, and after opencv binarization processing, a black and white image is obtained as a sample;

[0074] Preferably, the opencv binarization processing is based on adaptive threshold parameters, as Figure 3 shown, and specifically includes:

[0075] S11. Calculate the histogram of the grayscale image, and calculate the number of pixels occupied by each pixel value from 0 to 255. The grayscale image is directly generated by the drone shooting;

[0076] S12. Traverse the threshold from 0 to 255. Pixels less than or equal to the threshold are the background, and pixels greater than the threshold are the foreground;

[0077] S13. Calculate the proportion of the number of background pixels to the total number of pixels and the average value of the background pixels;

[0078] S14. Calculate the proportion of the number of foreground pixels to the total number of pixels and the average value of the foreground pixels;

[0079] S15. Calculate the between-class variance or within-class variance. The threshold that maximizes the between-class variance or minimizes the within-class variance is the optimal threshold;

[0080] The between-class variance represents the degree of difference between the two categories (foreground and background) of the histogram of the grayscale image. When the between-class variance is the largest, it indicates that the difference between the foreground and the background is the largest, and the threshold selected at this time is the best:

[0081]

[0082] In formula (1), is the between-class variance, ω0 and ω1 are the pixel proportions of the foreground and background respectively; μ0 and μ1 are the pixel means of the foreground and background respectively;

[0083] The within-class variance represents the degree of dispersion of the pixel values within each category. When the within-class variance is the smallest, it indicates that the pixel values within each category are more concentrated, and the threshold selected at this time is the best:

[0084]

[0085] In formula (2), is the within-class variance, and are the variances of the foreground and background respectively.

[0086] S16. Use the optimal threshold to perform binary processing on the image.

[0087] S2. The thermal imaging photos taken in other photovoltaic panel detection areas of the photovoltaic power station are subjected to opencv binary operation to convert them into black and white images, and compared with the normal photovoltaic strings under the same conditions.

[0088] S3. Set the pixel threshold. Preferably, use the optimal threshold as the set pixel threshold, compare the pixel values in the image with the pixel threshold, set the pixel values greater than the pixel threshold to 1, and set the pixel values less than the pixel threshold to 0;

[0089] If the number of black pixels in the image of the detection area is less than the pixel threshold, the photovoltaic string in the detection area does not need to be cleaned;

[0090] If the number of black pixels in the detected area image is greater than the pixel threshold, perform the next operation:

[0091] S4. After correcting and positioning the detected area image in the string through the conversion of the picture GPS information coordinates, then correct the coordinate system data, as Figure 2 shown.

[0092] Preferably, the correction of the coordinate system data specifically includes:

[0093] S41. Determining the coordinate system includes determining the source coordinate system and determining the target coordinate system; preferably, the source coordinate system is WGS84; determining the target coordinate system such as GCJ-02 or BD-09;

[0094] Preferably, in this embodiment, GCJ-02 is adopted. Taking the conversion from WGS84 to GCJ-02 (international GPS to Mars coordinate system) as an example, it is applicable to the offset correction of GPS data within the territory of China.

[0095] S42. Extract the GPS information in the image, which is usually related to the Exif metadata. The GPS information of the image is stored in the Exif data of the image;

[0096] S43. Perform latitude and longitude offset on the GPS information, and then perform offset normalization;

[0097] In Java, to extract Exif data, libraries such as metadata-extractor can be used, which provides APIs to parse data such as image Exif information. The GPS information of Exif is usually stored in the format of degrees, minutes, and seconds (DMS), and needs to be converted to the format of decimal degrees (DD):

[0098]

[0099] In formula (3), minutes are minutes, seconds are seconds, and Degress are degrees;

[0100] deltaLat and deltaLon are the offsets of latitude and longitude, which define the latitude and longitude offsets required to encrypt the WGS84 coordinates into GCJ-02:

[0101] deltaLat = transformLat(x, y) (4);

[0102] deltaLon = transformLon(x, y) (5);

[0103] In Formulas (4)-(5), transformLat and transformLon represent the offsets of latitude and longitude calculated through complex non-linear functions, which are usually related to the periodic changes (such as sine and cosine) of latitude and longitude;

[0104] Convert the latitude from degree units to radian units because subsequent calculations (such as trigonometric functions) require the use of the radian system:

[0105]

[0106] Then, correct the influence of the Earth ellipsoid. There is an eccentricity (e 2 ), which is the difference between the equatorial radius and the polar radius. magic: the sine value at the latitude, used to calculate the eccentricity correction on the latitude:

[0107] magic = sin(radLat) (7);

[0108] magic = 1 - e 2 *magic 2 (8);

[0109] Finally, perform offset normalization. a is the semi-major axis (equatorial radius) of the Earth ellipsoid, 1 - e 2 is the eccentricity correction of the Earth ellipsoid, and the normalized latitude adjustedLat and longitude adjustedLon are obtained:

[0110]

[0111]

[0112] S44. Reverse Geocoding: Receive a geographical coordinate, i.e., (latitude, longitude), as input. By comparing the input coordinates with the location information in the geographical database, find the matching address; output the address information corresponding to the input geographical coordinates, including street, city, and country information. Use the database including address information to match the address through coordinate range search to implement reverse geocoding;

[0113] Alternatively, use the ready-made reverse geocoding function of external APIs, such as the ready-made reverse geocoding functions of Google Maps API, Bing Maps, OpenStreetMap, etc.

[0114] S45. Check whether the result meets the expected range, that is, check whether the converted coordinates are within a reasonable range.

[0115] Perform statistical analysis on the corrected data, calculate the statistics of the corrected data, such as mean, standard deviation, and extreme values. If the statistics are within the set range, it meets the expectations.

[0116] S5. After obtaining the string position information, combined with the power generation data of the power station, if the power generation of the target detection string is significantly lower than that of other strings or lower than 80% of the normal photovoltaic string (sample), cleaning is required. Otherwise, the impact on power generation is small and cleaning is not required at present.

[0117] S6. The system automatically commands to clean the photovoltaic strings in a timely manner, and the photovoltaic cleaning robot autonomously plans the cleaning path according to the strings to be cleaned;

[0118] The photovoltaic cleaning robot is an automated device specialized for cleaning dust on the surface of photovoltaic panels. It uses advanced sensor technology and intelligent algorithms to autonomously plan the cleaning path, avoid collisions and missed cleaning, and improve the cleaning efficiency and quality; when it is detected that the photovoltaic string needs to be cleaned, combined with the weather information, select an appropriate time to command the photovoltaic cleaning robot to clean the detected string, and it is prohibited to clean the photovoltaic modules under windy, rainy, thunderstorm or heavy snow weather conditions.

[0119] The SLAM (Simultaneous Localization and Mapping) algorithm allows the robot to simultaneously perform localization and map construction in an unknown environment. The robot continuously obtains environmental information using sensors (such as IMU (Inertial Measurement Unit), LiDAR) and uses the algorithm to update its own position and the environment. The SLAM (Simultaneous Localization and Mapping) algorithm allows the robot to simultaneously perform localization and map construction in an unknown environment.

[0120] Sensor technology can provide instant motion information, can quickly estimate the speed and direction of an object. Provide necessary positioning support when GPS signals are weak or unavailable. Can dynamically construct an environmental map and adjust the path in real time. Can still work effectively even if the environment is unknown or dynamically changing.

[0121] Embodiment 2

[0122] This embodiment provides a system for implementing a method for detecting dust on a photovoltaic panel, and the system includes:

[0123] At least one processor; and

[0124] A memory, the memory stores instructions, when the instructions are executed by the at least one processor, the at least one processor is caused to execute a method for detecting dust on a photovoltaic panel as described above.

[0125] In this embodiment, the electronic device includes, but is not limited to: personal computer, server computer, workstation, desktop computer, laptop computer, notebook computer, mobile computing device, smart phone, tablet computer, cellular phone, personal digital assistant (PDA), handheld system, messaging device, wearable computing device, consumer electronic device, etc.

[0126] Embodiment 3

[0127] This embodiment also provides a computer-readable storage medium storing executable instructions that, when executed, cause the machine to execute a method for detecting dust on a photovoltaic panel as described above.

[0128] Specifically, a system or a system equipped with a readable storage medium can be provided, on which software program codes for implementing the functions of any one of the above embodiments are stored, and the computer or processor of the system or the system is caused to read and execute the instructions stored in the readable storage medium.

[0129] In this case, the program code read from the readable medium itself can implement the functions of any one of the above embodiments, so the computer-readable code and the readable storage medium storing the computer-readable code constitute a part of this specification.

[0130] Examples of the readable storage medium include floppy disk, hard disk, magneto-optical disk, optical disk (such as CD-ROM, CD-R, CD-RW, DVD-ROM, DVD-RAM, DVD-RW, DVD-RW), magnetic tape, non-volatile memory card, and ROM. Optionally, the program code can be downloaded from a server computer or a cloud via a communication network.

[0131] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0132] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processors of general-purpose computers, special-purpose computers, embedded processors, or other programmable data processing devices to generate a machine, such that the instructions executed by the processors of the computer or other programmable data processing devices generate a system for implementing the functions specified in the Figure 1 one or more flows and / or blocks Figure 1 function specified in one or more blocks.

[0133] These computer program instructions can also be stored in a computer-readable memory capable of guiding a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including an instruction system that implements the functions specified in the Figure 1 one or more flows and / or blocks Figure 1 function specified in one or more blocks.

[0134] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in the Figure 1 one or more flows and / or blocks Figure 1 function specified in one or more blocks.

[0135] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the technical solutions of the present invention, rather than limitations on the specific implementation manners of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the claims of the present invention shall be included in the protection scope of the claims of the present invention.

Claims

1. A method for detecting dust on a photovoltaic panel, characterized in that, The method includes: S1. For a photovoltaic power station, under the same conditions, several groups of photovoltaic panel strings with clean and dust-free surfaces are divided. A drone is used to take thermal imaging pictures of the normal strings. After binaryzation processing by opencv, a black-and-white image is obtained as a comparison sample. S2. The thermal imaging photos taken in other photovoltaic panel detection areas of the photovoltaic power station are subjected to opencv binaryzation operation and converted into black-and-white images as a data set. S3. A pixel threshold is set, and the pixel values of the images in the data set are compared with the pixel threshold. Pixel values greater than the pixel threshold are set to 1, and pixel values less than the threshold are set to 0. If the number of black pixels in the image of the detection area is less than the pixel threshold, the photovoltaic string in the detection area does not need to be cleaned. If the number of black pixels in the image of the detection area is greater than the pixel threshold, the next step is performed. S4. After coordinate conversion and correction through the GPS information coordinates of the picture, the string where the detection area image is located is positioned, and then the coordinate system data is corrected for deviation. S5. After obtaining the string position information, combined with the power generation data of the power station, if the power generation of the target detection string is lower than 80% of that of the normal photovoltaic string, it needs to be cleaned. Otherwise, the impact on power generation is small and it does not need to be cleaned at present. S6. The photovoltaic cleaning robot uses sensor technology and intelligent algorithms to autonomously plan the cleaning path according to the string to be cleaned.

2. The method for detecting dust on a photovoltaic panel according to claim 1, characterized in that, The opencv binaryzation is based on adaptive threshold parameters, specifically including: S11. Calculate the histogram of the grayscale image, and calculate the number of pixels occupied by each pixel value from 0 to 255. The grayscale image is directly generated by the drone shooting. S12. Traverse the threshold from 0 to 255. Pixels less than or equal to the threshold are the background, and pixels greater than the threshold are the foreground. S13. Calculate the proportion of the number of background pixels in the total number of pixels and the average value of the background pixels. S14. Calculate the proportion of the number of foreground pixels in the total number of pixels and the average value of the foreground pixels. S15. Calculate the between-class variance or within-class variance. The threshold that maximizes the between-class variance or minimizes the within-class variance is used as the optimal threshold. When the between-class variance is the largest, the difference between the foreground and the background is the largest, and the threshold selected at this time is the best. In formula (1), is the between-class variance, ω0 and ω1 are the pixel ratios of the foreground and background respectively; μ0 and μ1 are the pixel means of the foreground and background respectively. When the within-class variance is the smallest, the pixel values within each category are more concentrated, and the threshold selected at this time is the best. In formula (2), is the within-class variance, and are the variances of the foreground and background, respectively; S16. Use the optimal threshold to perform binaryzation processing on the image.

3. The method for detecting dust on a photovoltaic panel according to claim 2, characterized in that, The correction of the coordinate system data specifically includes: S41. Determining the coordinate system includes determining the source coordinate system and determining the target coordinate system. S42. Extracting the GPS information in the image is usually related to the Exif metadata, and the GPS coordinates of the image are stored in the Exif data of the image. S43. Perform latitude and longitude offset on the GPS information, and then perform offset normalization: Convert the format of the picture GPS information to the decimal DD format: In formula (3), minutes is minutes, seconds is seconds, and Degress is degrees. deltaLat and deltaLon are the offsets of latitude and longitude, which define the latitude and longitude offset. deltaLat = transformLat(x,y) (4); deltaLon = transformLon(x, y) (5); In formulas (4)-(5), transformLat and transformLon represent the offsets of latitude and longitude calculated through complex non-linear functions; Convert the latitude from angular units to radian units: Then, correct for the effect of the Earth ellipsoid, with eccentricity e 2 That is, the difference between the equatorial radius and the polar radius, and magic is the sine value at the latitude, used to calculate the eccentricity correction at the latitude: magic = sin(radLat) (7); magic = 1 - e 2 *magic 2 (8); Finally, perform offset normalization. Let \(a\) be the semi-major axis of the Earth ellipsoid, and \(1 - e\) 2 be the eccentricity correction of the Earth ellipsoid, to obtain the normalized latitude \(adjustedLat\) and longitude \(adjustedLon\): S44. Reverse geocoding: Receive a geographical coordinate, i.e., (latitude, longitude) as input, and find the matching address by comparing the input coordinates with the location information in the geographical database; Output the address information corresponding to the input geographical coordinates, including street, city, and country information. Use the database including address information to match the address through coordinate range search to implement reverse geocoding; Alternatively, use the ready-made reverse geocoding function of an external API to implement reverse geocoding; S45. Check whether the rectification result meets the expected range. If it meets, the rectification is completed; if not, rectify again.

4. A method for detecting dust on a photovoltaic panel according to claim 3, characterized in that, The specific method for checking whether the rectification result meets the expected range is as follows: Conduct statistical analysis on the rectified data, calculate the statistic of the data in the rectified coordinate system. If the statistic is within the set statistical threshold, it meets the expectation.

5. A method for detecting dust on a photovoltaic panel according to claim 2, characterized in that, The set pixel threshold is the optimal threshold.

6. An apparatus for a method of detecting dust on a photovoltaic panel, characterized in that, The device includes: A processor; A memory, on which a computer program that can run on the processor is stored; Wherein, when the computer program is executed by the processor, it implements the steps of a method for detecting dust on a photovoltaic panel as described in any one of claims 1 to 5.

7. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the method as described in any one of claims 1 to 5.

Images