A photovoltaic panel efficiency monitoring method and system based on image recognition technology

Abstract

The application discloses a kind of photovoltaic panel efficiency monitoring method and system based on image recognition technology, method contains step 1: the historical output power data of photovoltaic panel is obtained with unobstructed photo data, according to historical output power data, historical efficiency output power interval and maximum change slope are obtained by calculation, reference edge feature data is obtained by carrying out edge feature extraction to photo data;Step 2: real-time acquisition target detection area in photovoltaic panel output power data and video stream data;Step 3: whether output power data is out of historical efficiency output power interval or maximum change slope is judged;Step 4: according to edge recognition algorithm, the edge feature data of photovoltaic panel is obtained by processing video stream data;Step 5: the image feature difference amount is obtained by comparing and processing edge feature data with reference edge feature data;Step 6: whether image feature difference amount is out of similarity preset value is judged.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of photovoltaic panel monitoring, in particular to a photovoltaic panel efficiency monitoring method and system based on image recognition technology. BACKGROUND

[0002] Photovoltaic power generation mainly receives light energy through a battery pack and converts it into electrical energy through a photovoltaic battery pack. There are many factors affecting the efficiency of photovoltaic panels, including the inclination angle of the photovoltaic panel, the conversion efficiency of the photovoltaic panel assembly (in the life cycle of the photovoltaic station, there is a problem of gradual reduction of component efficiency and electrical element performance, and the power generation capacity decreases year by year. In addition to these natural aging factors, there are also quality problems of components and inverters, line layout, series-parallel loss, cable loss, etc.), dust shielding, shadow and object shielding. These factors directly affect the photoelectric conversion efficiency.

[0003] Currently, for the factors affecting the photoelectric conversion efficiency of the photovoltaic panel, there are some solutions for dynamically adjusting the inclination angle of the photovoltaic panel. However, for the conversion efficiency of the photovoltaic panel assembly, dust shielding and shadow object shielding factors, monitoring and daily manual inspection are still mainly relied on to monitor the efficiency of the photovoltaic panel, especially relying on manual inspection, which is low in labor cost and work efficiency. SUMMARY

[0004] The technical problem to be solved by the present application is that currently, for the conversion efficiency of the photovoltaic panel assembly, dust shielding and shadow object shielding factors, monitoring and daily manual inspection are still mainly relied on to monitor the efficiency of the photovoltaic panel, especially relying on manual inspection, which is low in labor cost and work efficiency. The present application provides a photovoltaic panel efficiency monitoring method based on image recognition technology. The present application also provides a photovoltaic panel efficiency monitoring system based on image recognition technology. The present application uses image recognition to assist in automatically monitoring the dust shielding and shadow object shielding of the photovoltaic panel, and simultaneously assists in real-time monitoring of the output power of the photovoltaic panel. By combining the two, the efficiency of the photovoltaic panel is monitored for the reference of the monitoring personnel and the arrangement of daily maintenance work, so as to solve the defects caused by the prior art.

[0005] To solve the above technical problems, the present application provides the following technical solutions:

[0006] In a first aspect, a photovoltaic panel efficiency monitoring method based on image recognition technology is provided, which comprises the following steps:

[0007] Step 1: Obtain the historical output power data and unshielded photo data of the photovoltaic panel. Calculate the historical efficiency output power interval [min, max] and the maximum change slope s in a certain period of time according to the historical output power data. Extract the baseline edge feature data according to the photo data.

[0008] Step 2: Real-time acquisition of output power data and video stream data of the photovoltaic panel in the target detection area;

[0009] Step 3: Determine whether the output power data exceeds the historical efficiency output power interval or the maximum change slope;

[0010] If not, perform step 1;

[0011] If yes, perform step 4;

[0012] Step 4: Process the video stream data according to the edge recognition algorithm to obtain the edge feature data of the photovoltaic panel;

[0013] Step 5: Compare the edge feature data with the reference edge feature data to obtain the image feature difference;

[0014] Step 6: Determine whether the image feature difference exceeds the similarity preset value, which is set and adjusted according to actual operation;

[0015] If yes, output that the photovoltaic panel accumulates dust or is blocked by a shadow object;

[0016] If not, output that the photovoltaic panel cell component conversion efficiency is abnormal;

[0017] According to the corresponding output evaluation, the operator can reasonably arrange maintenance tasks, specifically whether to arrange photovoltaic panel cleaning or component maintenance tasks;

[0018] When monitoring the photovoltaic panel, the element fault diagnosis technology of the photovoltaic panel cell group can also be introduced. According to the mechanism model of the element itself or certain expert rules, a cell group element fault diagnosis model is established and combined into the monitoring method. When the photovoltaic panel efficiency is abnormal and the photovoltaic panel cleanliness and other object blocking factors on site are not affected, the photovoltaic panel cell group element is diagnosed for fault, which can bring more refined guidance to the monitoring method.

[0019] The above-mentioned photovoltaic panel efficiency monitoring method based on image recognition technology, wherein the method for calculating the historical efficiency output power interval [min, max] and the maximum change slope within a certain period according to the historical output power data is as follows:

[0020] After cleaning the historical output power data, the power data of the photovoltaic panel under normal working conditions is selected and recorded as p;

[0021] The minimum value and the maximum value are calculated by bringing p into formula 1, and the interval [min, max] of the minimum value and the maximum value is the historical efficiency output power;

[0022] Formula 1:

[0023] Wherein, the expected value of X is equal to a linear combination of one or more lag periods, plus a constant term, plus a random error, and c is the constant term, is an autocorrelation coefficient, and ε t is a random error value which is assumed to have an average of 0 and a standard deviation of σ, and σ is assumed to be constant for any t;

[0024] The maximum change slope of the historical output power is derived by using the least square method.

[0025] The specific method of deriving the maximum change slope of the historical output power by using the least square method in the photovoltaic panel efficiency monitoring method based on image recognition technology is as follows:

[0026] The change slope value k and the constant b between each point are obtained by bringing any two points into the formula according to the slope-intercept form y=kx+b, and the coordinate values (x1, y1), (x2, y2), (x3, y3), …, (xn, yn) of n points in the plane. n n

[0027] The slope k=(y2-y1) / (x2-x1) is derived according to the slope formula, and the maximum change slope is obtained by taking the maximum value of the obtained slope. The specific method of obtaining the reference edge feature data by performing edge feature extraction on the photograph data in step 1 of the photovoltaic panel efficiency monitoring method based on image recognition technology is as follows:

[0028] After median filtering and denoising processing of the photograph data, the CNN is used for feature extraction, and the BP neural network is used for feature fusion to obtain the reference edge feature data.

[0029] The specific method of obtaining the edge feature data of the photovoltaic panel by processing the video stream data according to the edge recognition algorithm in step 4 of the photovoltaic panel efficiency monitoring method based on image recognition technology is as follows:

[0030] The video stream data is image processed to obtain a plurality of image data;

[0031] The pixel points of the image data are detected and connected to form contour pixel points.

[0032] ​​The image data with the contour pixel points is subjected to boundary point detection and the boundary points are integrated with the contour pixel points while false pixel points are removed to obtain the edge feature data;

[0033] The photovoltaic panel efficiency monitoring method based on image recognition technology, wherein the image data obtained by image processing on the video stream data is further subjected to noise reduction processing.

[0034] The photovoltaic panel efficiency monitoring method based on image recognition technology, wherein the edge recognition algorithm is Roberts operator, Prewitt operator, Sobel operator, Canny operator or Laplacian operator, and the edge detection of the image is based on the gradient of the image, and the gradient of the image is obtained by using various operators to perform convolution operation on the image.

[0035] In order to achieve the purpose of evaluating the efficiency of the photovoltaic panel, the edge recognition second-order Laplacian operator is preferably used to extract the edge features of the photovoltaic panel. The Laplacian operator is a second-order differential operator with rotational invariance and isotropy. The response obtained by processing individual pixel points is stronger than that obtained by processing edge points. Therefore, it is not suitable for processing images with high noise intensity. If the detected image has high noise, low-pass filtering is required to remove the noise.

[0036] The Laplacian edge extraction process is as follows: 1) Gaussian blur - remove noise; 2) convert to grayscale; 3) Laplace - calculate the second derivative of Laplacian; 4) take absolute value - the edge image can be obtained here; 5) binarization threshold processing - enhance edge features - obtain more obvious edge image.

[0037] The Laplacian operator of the image function is a second-order derivative (f is a second-order differentiable real function), and its definition is as follows:

[0038]

[0039] The discrete form of the Laplacian operator suitable for digital image processing is as follows:

[0040]

[0041] The Laplacian operator generated by the above formula 3 is divided into four-neighborhood operator and eight-neighborhood operator. The four-neighborhood operator calculates the gradient in four directions of the neighborhood of the center pixel, and the eight-neighborhood operator calculates the gradient in eight directions. The operator definition is as follows:

[0042]

[0043] It can be found by Laplacian operator that when the pixel gray value of itself and its neighborhood is the same, the operation result after Laplacian operator processing is zero.

[0044] When the gray value of the pixel point itself is higher than the average gray value of the pixel points in its neighborhood, the operation result is positive; otherwise, it is negative.

[0045] Therefore, the zero-crossing point between the positive peak and the negative peak can be used to determine the edge point of the image.

[0046] Laplacian operator is particularly sensitive to noise, so in order to obtain better edge detection effect, the image needs to be blurred and smoothed to remove high-frequency noise in the image.

[0047] In a second aspect, a photovoltaic panel efficiency monitoring system based on image recognition technology, comprising a data processing module, a data acquisition module, a judgment module, a feature extraction module, and a comparison module.

[0048] The data processing module is configured to obtain historical output power data and unobstructed photo data of the photovoltaic panel, calculate the historical efficiency output power interval [min, max] and the maximum change slope s within a certain period of time according to the historical output power data, and extract the baseline edge feature data according to the photo data.

[0049] The data acquisition module is configured to acquire real-time output power data and video stream data of the photovoltaic panel in the target detection area.

[0050] The judgment module is configured to acquire the output power data, the historical efficiency output power interval, and the maximum change slope, judge whether the output power data exceeds the historical efficiency output power interval or the maximum change slope, and generate feedback data transmitted to the data processing module and the feature extraction module.

[0051] The feature extraction module is configured to process the video stream data according to an edge recognition algorithm to obtain the edge feature data of the photovoltaic panel.

[0052] The comparison module is configured to compare the edge feature data with the baseline edge feature data to obtain the image feature difference.

[0053] The judgment module is further configured to judge whether the image feature difference exceeds a preset similarity value, and if it does, output that the photovoltaic panel accumulates dust or is obstructed by a shadow object, and if it does not, output that the photovoltaic panel battery assembly conversion efficiency is abnormal.

[0054] In a third aspect, a computer readable storage medium, wherein the computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the steps of the method in any one of the first aspect.

[0055] The photovoltaic panel efficiency monitoring method and system based on image recognition technology provided by the application has the following technical effects:

[0056] The application adopts image recognition to assist in automatically monitoring dust and shadow objects that shield the photovoltaic panel, and to assist in real-time monitoring of the output power of the photovoltaic panel. By combining the two, the efficiency of the photovoltaic panel is monitored, which provides a reference for monitoring personnel and arranges routine maintenance work. BRIEF DESCRIPTION OF DRAWINGS

[0057] Figure 1 A flowchart of a photovoltaic panel efficiency monitoring method based on image recognition technology;

[0058] Figure 2 A structural schematic diagram of a photovoltaic panel efficiency monitoring system based on image recognition technology.

[0059] Wherein, the reference signs are as follows:

[0060] The data processing module 100, the data acquisition module 200, the judgment module 300, the feature extraction module 400, and the comparison module 500. DETAILED DESCRIPTION

[0061] In order to make the technical means, creative features, purposes and effects achieved by the application easy to understand, the technical solutions in the embodiments of the application are described clearly and completely in combination with specific drawings. Obviously, the described embodiments are part of the embodiments of the application, not all.

[0062] Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the application.

[0063] It should be noted that the structures, proportions, sizes, etc. shown in the drawings attached to the specification are only used to understand and read the content disclosed by the specification for those skilled in the art, and do not define the limiting conditions for the implementation of the application, so they do not have technical significance. Any modification of structure, change of proportion relationship or adjustment of size, which does not affect the effects and purposes that can be achieved by the application, should still fall within the scope of the technical content disclosed by the application.

[0064] Meanwhile, the terms such as "upper", "lower", "left", "right", "middle", and "one" cited in the present specification are merely for clear understanding of the description, and are not intended to limit the scope of the present application, and the change or adjustment of the relative relationship is also considered as the scope of the present application without substantial change of the technical content.

[0065] The present application provides a photovoltaic panel efficiency monitoring method and system based on image recognition technology, which aims to automatically monitor the dust and shadow object blocking of the photovoltaic panel through image recognition assistance, and to assist in real-time monitoring of the output power of the photovoltaic panel. Through the combination of the two, the efficiency of the photovoltaic panel is monitored for reference by the monitoring personnel and arrangement of daily maintenance work.

[0066] As shown in Figure 1 A photovoltaic panel efficiency monitoring method based on image recognition technology, comprising the following steps:

[0067] Step 1: Obtain the historical output power data and unblocked photo data of the photovoltaic panel, calculate the historical efficiency output power interval [min, max] and the maximum change slope s within a certain period of time according to the historical output power data, and extract the reference edge feature data according to the photo data;

[0068] Step 2: Real-time acquisition of the output power data and video stream data of the photovoltaic panel in the target detection area;

[0069] Step 3: Determine whether the output power data exceeds the historical efficiency output power interval or the maximum change slope;

[0070] If not, execute step 1;

[0071] If yes, execute step 4;

[0072] Step 4: Process the video stream data according to the edge recognition algorithm to obtain the edge feature data of the photovoltaic panel;

[0073] Step 5: Compare the edge feature data with the reference edge feature data to obtain the image feature difference;

[0074] Step 6: Determine whether the image feature difference exceeds the similarity preset value, which is set and adjusted according to the actual operation;

[0075] If yes, output the photovoltaic panel accumulated dust or shadow object blocking;

[0076] If not, output the abnormal conversion efficiency of the photovoltaic panel battery assembly;

[0077] According to the corresponding output evaluation, the operator can reasonably arrange the maintenance task, specifically, whether to arrange the photovoltaic panel cleaning or component maintenance task work;

[0078] When monitoring the photovoltaic panel, the element fault diagnosis technology of the photovoltaic panel battery pack can also be introduced, and the battery pack element fault diagnosis model is established according to the mechanism model of the element itself or certain expert rules, combined with the monitoring method, when the photovoltaic panel efficiency is abnormal and the on-site photovoltaic panel cleanliness and no other object shielding influence factors, the photovoltaic panel battery pack element is diagnosed for fault, which can bring more refined guidance for the monitoring method.

[0079] The above-mentioned photovoltaic panel efficiency monitoring method based on image recognition technology, wherein the method for calculating the historical efficiency output power interval [min, max] and the maximum change slope in a certain period according to the historical output power data is as follows:

[0080] After cleaning the historical output power data, the power data of the photovoltaic panel under normal working condition is screened and recorded as p;

[0081] The p is brought into formula 1 to calculate the minimum value and the maximum value, and the interval [min, max] of the minimum value and the maximum value is the historical efficiency output power;

[0082] Formula 1:

[0083] Wherein: the expected value of X is equal to the linear combination of one or more lag periods, plus a constant term, plus a random error, c is the constant term, is the autocorrelation coefficient, and ε t is a random error value assumed to have an average of 0 and a standard deviation of σ; σ is assumed to be constant for any t;

[0084] The maximum change slope of the historical output power is derived by using the least square method.

[0085] The above-mentioned photovoltaic panel efficiency monitoring method based on image recognition technology, wherein the specific method for deriving the maximum change slope of the historical output power by using the least square method is as follows:

[0086] According to the slope-intercept form y=kx+b, the coordinate values (x1, y1), (x2, y2), (x3, y3), …, (x n , y n ) of n points in the plane are calculated, and the slope value k and the constant b between each point are obtained by bringing any two points into the formula;

[0087] According to the slope formula, k=(y2-y1) / (x2-x1) is derived, and the maximum value of the obtained slope is the maximum change slope. The specific method of obtaining the reference edge feature data by performing edge feature extraction on the photograph data in step 1 is as follows:

[0088] After the photograph data is subjected to median filtering denoising processing, the CNN is used for feature extraction, and the BP neural network is used for feature fusion to obtain the reference edge feature data.

[0089] The specific method of obtaining the edge feature data of the photovoltaic panel by processing the video stream data according to the edge recognition algorithm in step 4 is as follows:

[0090] The video stream data is subjected to image processing to obtain a plurality of image data;

[0091] The pixel points of the image data are detected and connected to form contour pixel points;

[0092] The image data with contour pixel points is subjected to boundary point detection and integrated with the contour pixel points, and false pixel points are removed to obtain edge feature data;

[0093] The specific method of obtaining the edge feature data of the photovoltaic panel by processing the video stream data according to the edge recognition algorithm in step 4 is as follows:

[0094] The specific method of obtaining the edge feature data of the photovoltaic panel by processing the video stream data according to the edge recognition algorithm in step 4 is as follows:

[0095] In order to achieve the purpose of evaluating the efficiency of the photovoltaic panel, it is preferred to use the edge recognition second-order Laplacian operator to extract the edge feature of the photovoltaic panel. The Laplacian operator is a rotationally invariant isotropic second-order differential operator. The response obtained by processing individual pixel points is stronger than that obtained by processing edge points, so it is not suitable for processing images with high noise intensity. If the detected image has high noise, low-pass filtering is required to remove the noise.

[0096] Laplacian edge extraction process: 1) Gaussian blur - remove noise; 2) grayscale conversion; 3) Laplacian - second derivative calculation Laplacian; 4) take absolute value - here the edge image can be obtained; 5) thresholding again - enhance edge features - get a more obvious edge image;

[0097] The Laplacian operator of the image function is a second derivative (f is a second derivative of a real function), which is defined as:

[0098]

[0099] The expression of the discrete form suitable for digital image processing is:

[0100]

[0101] The Laplacian operator generated by the above formula 3 is divided into four-neighborhood operator and eight-neighborhood operator, the four-neighborhood operator is to calculate the gradient in four directions of the neighborhood of the center pixel, and the eight-neighborhood operator is to calculate the gradient in eight directions, and the operator is defined as follows:

[0102]

[0103] It can be found through the Laplacian operator that when the gray value of the pixel itself and the gray value in the neighborhood are the same, the operation result after the Laplacian operator processing is zero.

[0104] When the gray value of the pixel itself is higher than the average gray value of the pixel points in the neighborhood, the operation result is positive; otherwise, it is negative.

[0105] Therefore, the zero-crossing point between the positive peak and the negative peak can be used to determine the edge point of the image.

[0106] The Laplacian operator is particularly sensitive to noise, so in order to obtain better edge detection effect, it is necessary to perform blur smoothing processing on the image to remove high-frequency noise in the image.

[0107] As shown in Figure 2 The second aspect is a photovoltaic panel efficiency monitoring system based on image recognition technology, which comprises a data processing module 100, a data acquisition module 200, a judgment module 300, a feature extraction module 400, and a comparison module 500.

[0108] The data processing module 100 is used for acquiring historical output power data and unobstructed photo data of the photovoltaic panel, and is also used for calculating the historical efficiency output power interval [min, max] and the maximum change slope s in a certain period of time according to the historical output power data, and is also used for extracting the reference edge feature data according to the photo data.

[0109] The data acquisition module 200 is configured to acquire output power data and video stream data of the photovoltaic panel in the target detection area in real time.

[0110] The judging module 300 is configured to acquire the output power data, the historical efficiency output power interval, the maximum change slope, and judge whether the output power data exceeds the historical efficiency output power interval or the maximum change slope, and generate feedback data transmitted to the data processing module 100 and the feature extraction module 400.

[0111] The feature extraction module 400 is configured to process the video stream data according to an edge recognition algorithm to obtain edge feature data of the photovoltaic panel.

[0112] The comparison module 500 is configured to compare the edge feature data with the reference edge feature data to obtain an image feature difference value.

[0113] The judging module 300 is further configured to judge whether the image feature difference value exceeds a preset similarity value, and if yes, output that the photovoltaic panel accumulates dust or is blocked by a shadow object, and if not, output that the conversion efficiency of the photovoltaic panel is abnormal.

[0114] In a third aspect, a computer readable storage medium is provided, and a computer program is stored in the computer readable storage medium. The computer program is executed by a processor to implement the steps of any method in the first aspect.

[0115] In the embodiments of the present application, the disclosed system, device and method can be implemented in other manners;

[0116] For example, the division of the units or modules is only a logical function division, and there can be another division manner in actual implementation;

[0117] For example, a plurality of units or modules or components can be combined or integrated into another system;

[0118] In addition, each functional unit or module in the embodiments of the present application can be integrated into one processing unit or module, or can be a separate physical entity, etc.

[0119] It should be understood that, in various embodiments of the present application, the magnitude of the serial number of each process does not mean the execution order, and the execution order of each process should be determined according to its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0120] If the functions are realized in the form of software function units and sold or used as independent products, they can be stored in a machine readable storage medium;

[0121] Therefore, the technical scheme of the present application can be embodied in the form of a software product, which can be stored in a machine readable storage medium and can include a plurality of instructions to cause an electronic device to execute all or part of the processes of the technical scheme described in the embodiments of the present application.

[0122] The storage medium can include ROM, RAM, removable disks, hard disks, magnetic disks, or optical disks, and various media that can store program codes.

[0123] In summary, the photovoltaic panel efficiency monitoring method and system based on image recognition technology can automatically monitor the dust and shadow objects that block the photovoltaic panel through image recognition assistance, and can assist in real-time monitoring of the output power of the photovoltaic panel. By combining the two, the efficiency of the photovoltaic panel is monitored for reference by panel personnel and daily maintenance work is arranged.

[0124] The specific embodiments of the application are described above. It should be understood that the application is not limited to the specific embodiments described above, and that the devices and structures not fully described should be understood as being implemented in a common manner in the art; those skilled in the art can make various modifications or changes within the scope of the claims, and make several simple deductions, modifications or substitutions, which do not affect the essential content of the application.

Claims

1. A photovoltaic panel efficiency monitoring method based on image recognition technology, characterized in that: The following steps are involved: Step 1: Obtain historical output power data of the photovoltaic panel and unobstructed photo data. Calculate the historical efficiency output power range [min, max] and the maximum change slope of the historical output power within a certain period of time based on the historical output power data, denoted as s. Perform edge feature extraction based on the photo data to obtain baseline edge feature data. Step 2: Real-time collection of output power data and video stream data of photovoltaic panels within the target detection area; Step 3: Determine whether the output power data exceeds the historical efficiency output power range or the maximum change slope; If not exceeded, go to step 1; If exceeded, go to step 4; Step 4: Processing the video stream data according to an edge recognition algorithm to obtain edge feature data of the photovoltaic panel; Step 5: Compare the edge feature data with the reference edge feature data to obtain an image feature difference; Step 6: Determine whether the difference in the image features exceeds a preset similarity value; If exceeded, it is output that the photovoltaic panel is covered by dust or shadow objects; If it does not exceed the limit, the conversion efficiency of the photovoltaic panel battery assembly is output as abnormal. The method for calculating the historical efficiency output power range [min, max] and the maximum change slope within a certain period of time based on the historical output power data is as follows: After cleaning the historical output power data, the power data of the photovoltaic panel under normal operation is obtained by screening and recorded as p; Substitute p into Formula 1 to calculate the minimum and maximum values. The interval [min, max] between the minimum and maximum values ​​is the historical efficiency output power. Formula 1: ; Where: The expected value of X is equal to the linear combination of one or more lag periods, plus a constant term, plus random error, c is a constant term, is the autocorrelation coefficient, is the random error value with mean equal to 0 and standard deviation equal to σ; σ is assumed to be constant for any t; The maximum change slope of the historical output power is derived using a least squares method.

2. The photovoltaic panel efficiency monitoring method based on image recognition technology according to claim 1, characterized in that: The specific method of deriving the maximum change slope of the historical output power using the least squares method is as follows: According to the slope-intercept formula y=kx+b, the calculation is carried out through the coordinate values ​​of n points in the plane (x1,y1), (x2,y2), (x3,y3), ... (x n ,y n ), substitute any two points into the formula to obtain the slope value k and constant b between each point; According to the slope formula, we can deduce k=(y2-y1) / (x2-x1), and take the maximum value of the slope as the maximum change slope.

3. The photovoltaic panel efficiency monitoring method based on image recognition technology according to claim 2, characterized in that: The specific method for extracting edge features from the photo data in step 1 to obtain reference edge feature data is as follows: The photo data is subjected to median filtering and denoising, and then feature extraction is performed using CNN, followed by feature fusion using a BP neural network to obtain the reference edge feature data.

4. The photovoltaic panel efficiency monitoring method based on image recognition technology according to claim 3, characterized in that: The specific method of processing the video stream data according to the edge recognition algorithm in step 4 to obtain the edge feature data of the photovoltaic panel is as follows: Performing image processing on the video stream data to obtain a plurality of image data; Performing pixel detection on the image data and connecting the pixel points to form contour pixel points; Boundary point detection is performed on the image data having the contour pixel points, and the boundary points are integrated with the contour pixel points, while false pixel points are removed to obtain the edge feature data.

5. The photovoltaic panel efficiency monitoring method based on image recognition technology according to claim 4, characterized in that: After the video stream data is processed into images to obtain a plurality of image data, the image data needs to be subjected to noise reduction processing.

6. A photovoltaic panel efficiency monitoring method based on image recognition technology according to any one of claims 1 to 5, characterized in that: The edge recognition algorithm is Roberts operator, Prewitt operator, Sobel operator, Canny operator or Laplacian operator.

7. A photovoltaic panel efficiency monitoring system based on image recognition technology, characterized in that: Contains data processing module, data acquisition module, judgment module, feature extraction module, and comparison module; The data processing module is used to obtain historical output power data of the photovoltaic panel and unobstructed photo data, and is also used to calculate the historical efficiency output power range [min, max] and the maximum change slope of the historical output power within a certain period of time according to the historical output power data, recorded as s, and is also used to extract edge features according to the photo data to obtain reference edge feature data; The data acquisition module is used to collect the output power data and video stream data of the photovoltaic panels in the target detection area in real time; The judgment module is used to obtain the output power data, the historical efficiency output power interval, and the maximum change slope, and judge whether the output power data exceeds the historical efficiency output power interval or the maximum change slope, and generate feedback data transmitted to the data processing module and the feature extraction module; The feature extraction module is used to process the video stream data according to the edge recognition algorithm to obtain the edge feature data of the photovoltaic panel; The comparison module is used to compare the edge feature data with the reference edge feature data to obtain an image feature difference; The judgment module is further configured to determine whether the image feature difference exceeds a preset similarity value. If so, outputting that the photovoltaic panel is dust-accumulated or obscured by a shadow object is output; if not, outputting that the photovoltaic panel battery assembly conversion efficiency is abnormal. A method for calculating the historical efficiency output power range [min, max] and the maximum change slope within a certain period of time based on the historical output power data is as follows: After cleaning the historical output power data, the power data of the photovoltaic panel under normal operation is obtained by screening and recorded as p; Substitute p into Formula 1 to calculate the minimum and maximum values. The interval [min, max] between the minimum and maximum values ​​is the historical efficiency output power. Formula 1: ; Where: The expected value of X is equal to the linear combination of one or more lag periods, plus a constant term, plus random error, c is a constant term, is the autocorrelation coefficient, is the random error value with mean equal to 0 and standard deviation equal to σ; σ is assumed to be constant for any t; The maximum change slope of the historical output power is derived using a least squares method.

8. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the method according to any one of claims 1 to 6.

Images