Photovoltaic module point hot spot detection method, system, device and photovoltaic power station

Abstract

The present invention discloses a method, system, device and photovoltaic power station for detecting point-like hot spots of photovoltaic components, which relates to the field of photovoltaic power generation, including obtaining an infrared image of a photovoltaic component and determining the grayscale value of the infrared image; constructing a decision surface function; determining the decision surface function value of each pixel point in the infrared image according to the grayscale value of the infrared image; and determining whether the infrared image includes point-like hot spots according to the decision surface function value. Considering that a photovoltaic component cannot convert solar energy into electrical energy after a failure, which leads to point-like hot spots in the infrared image, since the point-like hot spots are highly concentrated and have high grayscale values, a decision surface function that is positively correlated with the concentration of pixels in the infrared image and the grayscale value of pixels in the infrared image is constructed. Determining whether there are point-like hot spots in the infrared image according to the decision surface function can improve the accuracy of detecting point-like hot spots of photovoltaic components in the automatic inspection mode of unmanned aerial vehicles, and the subsequent positioning of the position of the faulty photovoltaic component is more accurate.

Description

Technical Field

[0001] The present invention relates to the field of photovoltaic power generation, and in particular to a method, system, device and photovoltaic power station for detecting point-shaped hot spots of photovoltaic components. Background Art

[0002] The rapid development of the photovoltaic industry has put forward high requirements for the operation and maintenance of photovoltaic power stations. How to achieve efficient operation and maintenance of large-scale power stations and avoid safety problems and economic losses caused by photovoltaic equipment failures has become one of the industry pain points of the current photovoltaic industry. At present, the current inspection method of photovoltaic power stations is still mainly manual inspection. However, due to the large area, large number of equipment and complex installation environment of large-scale ground photovoltaic power stations, manual inspection efficiency is low and difficult to meet the requirements in most scenarios. Distributed rooftop photovoltaic power stations have problems such as scattered distribution, difficulty in accessing the roof and high safety risks. Therefore, in order to solve a series of problems of traditional manual inspection, the intelligent inspection method of drones equipped with infrared sensors is combined with computer vision technology to achieve rapid detection of photovoltaic module failures. In actual operation and maintenance scenarios, the failure of photovoltaic modules will form hot spots in infrared images, which are specifically manifested as a collection of point-like or small-area spot-like pixels with abnormally high temperatures on the photovoltaic module module. Among all hot spot types, point-like hot spots are the most common fault type. Therefore, the accuracy of infrared image point-like hot spot detection is an important factor affecting the intelligent diagnosis results of photovoltaic module failures. Summary of the invention

[0003] The purpose of the present invention is to provide a method, system, device and photovoltaic power station for detecting point-like hot spots of photovoltaic components, which can improve the accuracy of point-like hot spot detection of photovoltaic components in the automatic inspection mode of unmanned aerial vehicles, so that the subsequent location of the faulty photovoltaic components can be more accurate.

[0004] In order to solve the above technical problems, the present invention provides a method for detecting point-shaped hot spots of photovoltaic modules, comprising:

[0005] Acquire an infrared image of a photovoltaic module and determine a grayscale value of the infrared image;

[0006] Constructing a decision surface function, wherein the decision surface function value is positively correlated with the concentration of pixels in the infrared image and the grayscale value of pixels in the infrared image;

[0007] Determining a decision surface function value of each pixel point in the infrared image according to the grayscale value of the infrared image;

[0008] It is determined whether the infrared image includes a point-shaped hot spot according to the decision surface function value.

[0009] On the other hand, after acquiring the infrared image of the photovoltaic module and determining the grayscale value of the infrared image, the method further includes:

[0010] According to the image whitening relationship Obtaining a grayscale value of the whitened infrared image;

[0011] Using a median filter to filter the grayscale value of the whitened infrared image;

[0012] Among them, I white is the gray value of the infrared image after whitening, I is the gray value of the infrared image, mean(I) is the gray mean of the infrared image, and std(I) is the image standard deviation of the infrared image;

[0013] Determining the decision surface function value of each pixel point in the infrared image according to the gray value of the infrared image includes:

[0014] The decision surface function value of each pixel point in the infrared image is determined according to the grayscale value of the infrared image after whitening and filtering.

[0015] On the other hand, after filtering the grayscale value of the whitened infrared image using a median filter, the method further includes:

[0016] According to the grayscale normalization relationship Performing grayscale normalization on the filtered infrared image;

[0017] Among them, I(x,y) is the gray value of the pixel with coordinates (x, y), I min is the minimum grayscale value of the infrared image after filtering, I max is the maximum grayscale value of the infrared image after filtering;

[0018] Determining the decision surface function value of each pixel point in the infrared image according to the grayscale value of the infrared image includes:

[0019] The decision surface function value of each pixel point in the infrared image is determined according to the normalized grayscale value of each pixel point in the infrared image.

[0020] On the other hand, the decision surface function is constructed, including:

[0021] The infrared image The pixel point in is taken as the center, and the block image area is constructed according to the scale parameter R and the augmentation constant λ. And the augmented block image area

[0022] Construct a decision surface function, the expression of which is

[0023] Where, B={I(x,y)||ux|≤R,|vy|≤R}, B + ={I(x,y)||ux|≤λR,|vy|≤λR}, f(B) is the decision surface function value, B(i,j) is the gray value of the pixel with coordinates (i, j) in the block image area, and B + (i, j) is the grayscale value of the pixel with coordinates (i, j) in the augmented block image area, a is the relaxation constant, ε is the empirical constant, (u, v) is the coordinate of the central pixel of the block image area, M is the horizontal number of pixels in the infrared image, and N is the horizontal number of pixels in the infrared image.

[0024] On the other hand, determining the decision surface function value of each pixel point in the infrared image according to the gray value of the infrared image includes:

[0025] Taking the pixel point with coordinates (λR+1, λR+1) as the starting point and the pixel point with coordinates (M-λR, N-λR) as the end point, a block image area and an augmented block image area constructed based on the pixel point with coordinates (λR+1, λR+1) as the center slide in the infrared image at a preset step size;

[0026] The decision surface function value of each pixel point in the infrared image is determined according to the grayscale value of the central pixel point of each block image area and the augmented block image area.

[0027] On the other hand, determining whether the infrared image includes a point-shaped hot spot according to the decision surface function value includes:

[0028] When the decision surface function value is not less than 0, determining that the block image area includes a point-shaped hot spot;

[0029] When the decision surface function value is less than 0, determining that the block image area does not include a point-shaped hot spot;

[0030] If the block image area including the point-shaped hot spot exists in the infrared image, determining that the infrared image includes the point-shaped hot spot;

[0031] If the block image area including the point-shaped hot spots does not exist in the infrared image, it is determined that the infrared image does not include the point-shaped hot spots.

[0032] On the other hand, after determining whether the infrared image includes a point-shaped hot spot according to the decision surface function value, the method further includes:

[0033] Marking the coordinates of the central pixel point of the block image area where the point-shaped hot spot exists;

[0034] The position of the faulty photovoltaic module is determined according to the coordinates of the central pixel point.

[0035] In order to solve the above technical problems, the present invention also provides a photovoltaic module point hot spot detection system, comprising:

[0036] A grayscale value acquisition unit, used to acquire an infrared image of the photovoltaic module and determine the grayscale value of the infrared image;

[0037] A construction unit, used to construct a decision surface function, wherein the decision surface function value is positively correlated with the concentration of pixels in the infrared image and the grayscale value of pixels in the infrared image;

[0038] A decision surface function value determining unit, used to determine the decision surface function value of each pixel point in the infrared image according to the gray value of the infrared image;

[0039] The point-shaped hot spot determination unit is used to determine whether the infrared image includes a point-shaped hot spot according to the decision surface function value.

[0040] In order to solve the above technical problems, the present invention also provides a device for detecting point-shaped hot spots of photovoltaic modules, comprising:

[0041] Memory for storing computer programs;

[0042] A processor is used to implement the steps of the above-mentioned method for detecting point-shaped hot spots of photovoltaic components when executing the computer program.

[0043] In order to solve the above technical problems, the present invention further provides a photovoltaic power station, including the above-mentioned photovoltaic component point hot spot detection device, and also includes a photovoltaic component.

[0044] The present invention discloses a method, system, device and photovoltaic power station for detecting point-like hot spots of photovoltaic components, which relates to the field of photovoltaic power generation, including obtaining an infrared image of a photovoltaic component and determining the grayscale value of the infrared image; constructing a decision surface function; determining the decision surface function value of each pixel point in the infrared image according to the grayscale value of the infrared image; and determining whether the infrared image includes point-like hot spots according to the decision surface function value. Considering that a photovoltaic component cannot convert solar energy into electrical energy after a failure, which leads to the generation of point-like hot spots in the infrared image, since the point-like hot spots are highly concentrated and have high grayscale values, a decision surface function that is positively correlated with the concentration of pixels in the infrared image and the grayscale value of pixels in the infrared image is constructed. Determining whether there are point-like hot spots in the infrared image according to the decision surface function can improve the accuracy of detecting point-like hot spots of photovoltaic components in the automatic inspection mode of unmanned aerial vehicles, so that the subsequent location of the faulty photovoltaic component is more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the prior art and the drawings required for use in the embodiments are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0046] Figure 1 A flow chart of a method for detecting point-shaped hot spots of photovoltaic modules provided by the present invention;

[0047] Figure 2 A schematic diagram of an infrared image provided by the present invention;

[0048] Figure 3 A schematic diagram of a block image and an augmented block image sliding window provided by the present invention;

[0049] Figure 4 A flow chart of a photovoltaic module point hot spot detection system provided by the present invention;

[0050] Figure 5 The present invention provides a flow chart of a device for detecting point-shaped hot spots of photovoltaic modules. DETAILED DESCRIPTION

[0051] The core of the present invention is to provide a method, system, device and photovoltaic power station for detecting point-like hot spots of photovoltaic components, which can improve the accuracy of point-like hot spot detection of photovoltaic components in the automatic inspection mode of unmanned aerial vehicles, so that the subsequent positioning of the position of faulty photovoltaic components is more accurate.

[0052] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0053] Photovoltaic power generation, as a green and clean new energy source, is one of the effective ways of energy utilization by converting solar energy into electrical energy through semiconductor devices. Since the solar energy on the earth is inexhaustible, vigorously developing the photovoltaic industry has become an important means to solve the current global energy crisis and the environmental pollution problems caused by it. As of the whole year of 2021, my country's photovoltaic industry has entered a stage of rapid development. The national photovoltaic grid-connected installed capacity has reached 306GW, and the photovoltaic power generation has reached 325.9 billion kWh. Photovoltaic power generation accounts for about 4% of my country's annual power generation, both of which are historical highs. According to my country's new energy strategic planning for wind power and photovoltaic industries, it is expected that by 2030, the total installed capacity will reach 1.2 billion kilowatts, gradually replacing the dominant position of traditional energy.

[0054] The rapid development of the photovoltaic industry has put forward high requirements for the operation and maintenance of photovoltaic power stations. How to achieve efficient operation and maintenance of large-scale power stations and avoid safety problems and economic losses caused by photovoltaic equipment failures has become one of the industry pain points of the current photovoltaic industry. At present, the current inspection method of photovoltaic power stations is still mainly manual inspection. However, due to the large area, large number of equipment and complex installation environment of large-scale ground photovoltaic power stations, manual inspection efficiency is low and difficult to meet the requirements in most scenarios. Distributed rooftop photovoltaic power stations have problems such as scattered distribution, difficulty in accessing the roof and high safety risks. Therefore, in order to solve a series of problems of traditional manual inspection, the intelligent inspection method of drones equipped with infrared sensors is combined with computer vision technology to achieve rapid detection of photovoltaic module failures. In actual operation and maintenance scenarios, the failure of photovoltaic modules will form hot spots in infrared images, which are specifically manifested as a collection of point-like or small-area spot-like pixels with abnormally high temperatures on the photovoltaic module module. Among all hot spot types, point-like hot spots are the most common fault type. Therefore, the accuracy of infrared image point-like hot spot detection is an important factor affecting the intelligent diagnosis results of photovoltaic module failures.

[0055] Figure 1 A flow chart of a method for detecting a point-shaped hot spot of a photovoltaic module provided by the present invention, the method for detecting a point-shaped hot spot of a photovoltaic module comprising:

[0056] S11: Acquire an infrared image of the photovoltaic module and determine a grayscale value of the infrared image;

[0057] Photovoltaic modules convert solar energy into electrical energy during normal operation. If a photovoltaic module fails, solar energy cannot be converted into electrical energy, which leads to a rise in the temperature of the local area. Infrared images can be used to determine the distribution of heat energy. Faulty photovoltaic modules will form point-like hot spots on the infrared image, and the corresponding grayscale value will also be relatively high. Therefore, when determining whether there are point-like hot spots in the infrared image, it is necessary to first determine the grayscale value of the infrared image.

[0058] S12: construct a decision surface function, the decision surface function value is positively correlated with the concentration of pixels in the infrared image and the gray value of pixels in the infrared image;

[0059] Specifically, it is considered that the local structure of the point-like hot spot exhibits high concentration and high intensity characteristics compared with the background pixels and high-brightness interference pixels. High concentration means that the grayscale distribution of the local structure of the point-like hot spot is mainly concentrated to the center of the point-like hot spot, that is, after the point-like hot spot is imaged in the focal plane by the point-like spread function, the grayscale distribution spreads outward from the center point in the local range, which is referred to as structural singularity in this application. High intensity means that the local structure of the point-like hot spot is higher in grayscale intensity, especially compared to its neighboring pixels. In contrast, the grayscale distribution of the local structure of the background object is more uniform and the grayscale intensity is lower. The local structure of high-grayscale interference pixels such as noise does not have continuity and does not show high concentration characteristics.

[0060] Considering that the point shape will diffuse during the actual shooting process, what is captured is not an independent point, but an area. However, the point-shaped hot spot formed by the failure of the photovoltaic module should have a high grayscale value, and the pixel points in the area are distributed more concentratedly. Therefore, the decision surface function constructed is based on the concentration and grayscale value of the pixel points. If both high concentration and high grayscale value are met, the pixel point is determined to be the corresponding failed photovoltaic module.

[0061] S13: determining a decision surface function value of each pixel point in the infrared image according to the gray value of the infrared image;

[0062] The decision surface function value can characterize the concentration and grayscale value of the pixel point. After determining the decision surface function value of each pixel point according to the grayscale value of each pixel point in the infrared image, it is possible to determine whether there is a point-like hot spot in the infrared image according to the decision surface function value.

[0063] S14: Determine whether the infrared image includes point-shaped hot spots according to the decision surface function value.

[0064] Considering that the decision surface function value is positively correlated with the concentration of pixels in the infrared image and the grayscale value of pixels in the infrared image, the larger the decision surface function value is, the greater the possibility that the pixel is a point hot spot. Therefore, the threshold can be set according to the actual situation to determine the point hot spot.

[0065] Figure 2 A schematic diagram of an infrared image provided by the present invention;

[0066] Figure 2 The three bright dots in the middle are point-shaped hot spots. The purpose of this application is to determine the faulty photovoltaic strings among multiple photovoltaic strings, and then determine the location of the photovoltaic strings to achieve more accurate and efficient operation and maintenance.

[0067] In order to solve the problem that the prior art directly uses the temperature threshold determination method to detect point-like hot spots, resulting in low detection accuracy, the present invention can effectively characterize the signal pattern of point-like hot spots in the image by considering the local structural characteristics of the point-like hot spots.

[0068] The present invention discloses a detection method, system, device and photovoltaic power station of photovoltaic component point hot spots, which relates to the field of photovoltaic power generation, including obtaining an infrared image of a photovoltaic component and determining the grayscale value of the infrared image; constructing a decision surface function; determining the decision surface function value of each pixel point in the infrared image according to the grayscale value of the infrared image; and determining whether the infrared image includes point hot spots according to the decision surface function value. Considering that the photovoltaic component cannot convert solar energy into electrical energy after a failure, which leads to the generation of point hot spots in the infrared image, and since the point hot spots are highly concentrated and have high grayscale values, the present application constructs a decision surface function that is positively correlated with the concentration of the pixel points in the infrared image and the grayscale value of the pixel points in the infrared image. Determining whether there are point hot spots in the infrared image according to the decision surface function can improve the accuracy of detecting point hot spots of photovoltaic components in the automatic inspection mode of unmanned aerial vehicles, so that the subsequent location of the faulty photovoltaic component is more accurate.

[0069] Based on the above embodiments:

[0070] In some embodiments, after acquiring the infrared image of the photovoltaic assembly and determining the grayscale value of the infrared image, the method further includes:

[0071] According to the image whitening relationship Get the gray value of the whitened infrared image;

[0072] A median filter is used to filter the grayscale value of the whitened infrared image;

[0073] Among them, I white is the gray value of the infrared image after whitening, I is the gray value of the infrared image, mean(I) is the gray mean of the infrared image, and std(I) is the image standard deviation of the infrared image;

[0074] The decision surface function value of each pixel in the infrared image is determined according to the gray value of the infrared image, including:

[0075] The decision surface function value of each pixel in the infrared image is determined according to the gray value of the infrared image after whitening and filtering.

[0076] In some embodiments, after filtering the grayscale value of the whitened infrared image using a median filter, the method further includes:

[0077] According to the grayscale normalization relationship Normalize the grayscale of the filtered infrared image;

[0078] Among them, I(x,y) is the gray value of the pixel with coordinates (x, y), I min is the minimum grayscale value of the filtered infrared image, I max is the maximum grayscale value of the filtered infrared image;

[0079] The decision surface function value of each pixel in the infrared image is determined according to the gray value of the infrared image, including:

[0080] The decision surface function value of each pixel in the infrared image is determined according to the normalized gray value of each pixel in the infrared image.

[0081] The infrared thermal imaging process is affected by many factors such as weather, lighting, ground environment, and imaging sensor parameters. The acquired infrared images are affected by factors such as light intensity, weather conditions, imaging angle, and sensor thermal noise, resulting in different grayscale distribution characteristics of infrared images taken at different times and in different scenes. The interference of the above factors on infrared images is mainly uneven grayscale distribution interference and salt and pepper noise interference. Image whitening, median filtering, and image normalization can effectively remove the influence of uneven grayscale distribution and suppress salt and pepper noise. After filtering the whitened image with a 3X3 filter template, pixel grayscale normalization is further performed.

[0082] Specifically, after acquiring the captured infrared image, the infrared image is first whitened, and the whitened image is filtered using a median filter template to obtain a denoised infrared image, and then the pixel grayscale of the denoised image is normalized.

[0083] This application preprocesses the obtained infrared images in three ways, and involves the preprocessed infrared images in the subsequent decision surface function value calculation, which can more accurately determine the point hot spots and improve the accuracy of operation and maintenance.

[0084] The point-shaped hot spot determination method based on decision surface function proposed in the present invention avoids the problem of mismatch between the specific distribution modeling of background and noise distribution and the actual distribution, thereby improving the robustness of detection.

[0085] In some embodiments, constructing a decision surface function includes:

[0086] Infrared image The pixel point in is taken as the center, and the block image area is constructed according to the scale parameter R and the augmentation constant λ. And the augmented block image area

[0087] Construct a decision surface function. The expression of the decision surface function is

[0088] Where, B={I(x,y)||ux|≤R,|vy|≤R}, B + ={I(x,y)||ux|≤λR,|vy|≤λR}, f(B) is the decision surface function value, B(i,j) is the gray value of the pixel with coordinates (i, j) in the block image area, and B + (i, j) is the gray value of the pixel with coordinates (i, j) in the augmented block image area, a is the relaxation constant, ε is the empirical constant, (u, v) is the coordinate of the central pixel of the block image area, M is the horizontal number of pixels in the infrared image, and N is the horizontal number of pixels in the infrared image.

[0089] Furthermore, in order to characterize the structural singularity of the two-dimensional image, the one-dimensional structure is analyzed first. Ideally, the grayscale distribution of the point target after the point spread function presents a binary Gaussian distribution, and its distribution along each direction presents a univariate Gaussian distribution, that is, X~N(μ,σ 2 ), where μ represents the statistical mean of the variable and σ represents the statistical standard deviation. The one-dimensional Gaussian distribution satisfies the "three standard deviations" principle, which can effectively describe the high concentration characteristics. The "three standard deviations" rule of the one-dimensional Gaussian distribution is Prob(μ-σ≤X≤μ+σ)≈68%, Prob(μ-2σ≤X≤μ+2σ)≈95%, Prob(μ-3σ≤X≤μ+3σ)≈99.7%, where Prob represents the probability density function.

[0090] More than 68% of the majority energy is concentrated in the central part corresponding to [μ-σ, μ+σ], and almost all (more than 99.7%) is distributed in the local range of [μ-3σ, μ+3σ]. Therefore, the standard deviation σ controls the concentration of the normal distribution. The smaller σ is, the more concentrated the distribution is. Based on the above considerations, this rule can be used to construct a mathematical expression of one-dimensional structural singularity, which is specifically described as: given a one-dimensional random variable Y~Φ and a cumulative density threshold τ>0, Φ is called a structural singularity distribution if and only if there exists and σ U >0 satisfies the formula

[0091] Analogous to the one-dimensional distribution case, the ideal point-shaped hot spot can be regarded as a two-dimensional Gaussian distribution after passing through the point spread function of the optical focal plane. Considering that within the spatial range defined by the Mahalanobis distance, the cumulative probability density of the two-dimensional Gaussian distribution also satisfies the "three standard deviations" rule, on this basis, the mathematical expression of the two-dimensional structural singularity is derived by analogy with the one-dimensional structural singularity.

[0092] The two-dimensional Gaussian distribution is Z = (X, Y) ~ N (μ, Σ), where Z represents a two-dimensional random variable, μ = (μX ,μ Y ) T represents the mean of a multivariate variable, represents the covariance matrix, where |α|<1 represents the correlation coefficient between variables X and Y, and σ X With σ Y represents the statistical standard deviation of X and Y. The Mahalanobis distance space of the two-dimensional variable is

[0093] In the space defined by the Mahalanobis distance, the cumulative probability density of the two-dimensional Gaussian distribution is Where Ω R =||z-z0|| Σ ≤R represents the spatial region where the Mahalanobis distance is within R>0. The "three standard deviations" rule of the two-dimensional Gaussian distribution is Prob(Ω 1.52 )≈68%, Prob(Ω 2.49 )≈95% and Prob(Ω 3.44 )≈99.7%.

[0094] The mathematical expression of the two-dimensional structural singularity is described as follows: given a two-dimensional random variable Z~Ψ and a cumulative density threshold τ>0, Ψ is called a structural singular distribution if and only if there exists and Ω R =||z-z0|| Σ ≤R satisfies Where R>0 indicates the upper bound of the concentration of the target cumulative density in the Mahalanobis distance space.

[0095] According to the defined two-dimensional structural singularity, the ideal target spatial distribution is approximately a binary Gaussian distribution, satisfying the regional constraints of high concentration and high grayscale. The pixel point in is taken as the center, and the block image area is constructed according to the scale parameter R and the augmentation constant λ. And the augmented block image area a is a relaxation constant set to avoid the denominator being equal to 0.

[0096] Furthermore, considering that the area size of the point-shaped hot spot is usually within the range of 15×15, and the grayscale of the hot spot structure is concentrated in the central range of 7×7, in the specific embodiment, R is 3 and a is 10. -5 , ε is taken as 0.05. The "three standard deviation" rule of the two-dimensional Gaussian distribution, one standard deviation range corresponds to the area determined by R, and the three standard deviation ranges are determined by λR, and satisfy λR / R=3.44 / 1.52. Therefore, in the specific embodiment, λ≈2.26.

[0097] The mathematical expression of structural singularity proposed in the present invention fully characterizes the structural differences between point-shaped hot spots and background and noise in local space, thereby effectively suppressing the interference of background and noise on the detection results and improving the detection accuracy.

[0098] Figure 3 A schematic diagram of a block image and an augmented block image sliding window provided by the present invention;

[0099] In some embodiments, determining the decision surface function value of each pixel in the infrared image according to the gray value of the infrared image includes:

[0100] Taking the pixel point with coordinates (λR+1, λR+1) as the starting point and the pixel point with coordinates (M-λR, N-λR) as the end point, the block image area and the augmented block image area constructed based on the pixel point with coordinates (λR+1, λR+1) as the center slide in the infrared image with a preset step size;

[0101] The decision surface function value of each pixel point in the infrared image is determined according to the gray value of the central pixel point of each block image area and the augmented block image area.

[0102] Construct a local sliding window to traverse the infrared image. Considering that the structural singularity of the point-like hot spots is mainly reflected in the local space, all the pixels of the infrared image to be detected are traversed, and a local sliding window is constructed with the selected pixel as the center to determine the position of the point-like hot spots in the infrared image. Specifically in the embodiment, a local sliding window is constructed starting from the pixel point with image coordinates (λR+1, λR+1) as the center point. The pixel coordinates of the upper left corner of the starting local sliding window are (1,1), and the pixel coordinates of the lower right corner are (2λR+1, 2λR+1). It is traversed from left to right (horizontally) and from top to bottom (vertically) with a step size of 1 until it traverses to the pixel point with image coordinates (M-λR, N-λR), as shown in the example. Figure 3 shown.

[0103] In some embodiments, determining whether the infrared image includes a point-shaped hot spot according to the decision surface function value includes:

[0104] When the decision surface function value is not less than 0, it is determined that the block image area includes a point-shaped hot spot;

[0105] When the decision surface function value is less than 0, it is determined that the block image area does not include a point-shaped hot spot;

[0106] If there is a block image area including a point-shaped hot spot in the infrared image, it is determined that the infrared image includes the point-shaped hot spot;

[0107] If there is no block image area including the point-shaped hot spots in the infrared image, it is determined that the infrared image does not include the point-shaped hot spots.

[0108] According to the defined decision surface function, it is determined whether the infrared image to be detected contains point-shaped hot spots. By setting appropriate block image B and augmented block image B + scale, from left to right on the infrared image with a certain step length (such as Figure 3 horizontal direction as shown), from top to bottom (as shown Figure 3 If a sliding window is performed in the vertical direction shown in the figure, the judgment of whether the block image contains point-like hot spots is that when the decision surface function value is not less than 0, it is determined that the block image area includes point-like hot spots; when the decision surface function value is less than 0, it is determined that the block image area does not include point-like hot spots.

[0109] In some embodiments, after determining whether the infrared image includes a point-shaped hot spot according to the decision surface function value, the method further includes:

[0110] Mark the coordinates of the central pixel point of the block image area where the point-shaped hot spot exists;

[0111] The location of the faulty PV module is determined according to the coordinates of the central pixel point.

[0112] Different marking methods may be used for the center point where a point-like hot spot exists and the center point where a point-like hot spot does not exist, so as to subsequently determine the position of the point-like hot spot.

[0113] Specifically, after the traversal is completed, the local sliding window center point determined to contain a point-shaped hot spot according to the decision surface function is marked with a white pixel, and the local sliding window center point determined not to contain a point-shaped hot spot is marked with a black pixel.

[0114] In summary, compared with the prior art, the method of the embodiments of the present invention can effectively characterize the signal pattern of point hot spots in the image by considering the local structural characteristics of the point hot spots; the proposed point hot spot determination method based on the decision surface function avoids the problem of mismatch between the specific distribution modeling of the background and noise distribution and the actual distribution, thereby improving the robustness of detection; the proposed mathematical expression of structural singularity fully characterizes the structural differences between the point hot spots and the background and noise in the local space, thereby effectively suppressing the interference of the background and noise on the detection results and improving the detection accuracy.

[0115] Figure 4 The present invention provides a flow chart of a photovoltaic module point hot spot detection system, the photovoltaic module point hot spot detection system comprising:

[0116] A gray value acquisition unit 41 is used to acquire an infrared image of a photovoltaic module and determine a gray value of the infrared image;

[0117] A construction unit 42 is used to construct a decision surface function, wherein the decision surface function value is positively correlated with the concentration of pixels in the infrared image and the grayscale value of pixels in the infrared image;

[0118] A decision surface function value determining unit 43, used to determine the decision surface function value of each pixel in the infrared image according to the gray value of the infrared image;

[0119] The point-shaped hot spot determining unit 44 is used to determine whether the infrared image includes a point-shaped hot spot according to the decision surface function value.

[0120] Based on the above embodiments:

[0121] Also includes:

[0122] Whitening unit, used to whiten the image according to the relationship Get the gray value of the whitened infrared image;

[0123] A filtering unit, used for filtering the grayscale value of the whitened infrared image using a median filter;

[0124] Among them, I white is the gray value of the infrared image after whitening, I is the gray value of the infrared image, mean(I) is the gray mean of the infrared image, and std(I) is the image standard deviation of the infrared image;

[0125] The decision surface function value determining unit 43 is specifically configured to determine the decision surface function value of each pixel in the infrared image according to the grayscale value of the infrared image after whitening and filtering.

[0126] Normalization unit, used to normalize the relationship according to the grayscale Normalize the grayscale of the filtered infrared image;

[0127] Among them, I(x,y) is the gray value of the pixel with coordinates (x, y), I min is the minimum grayscale value of the filtered infrared image, I max is the maximum grayscale value of the filtered infrared image;

[0128] The decision surface function value determining unit 43 is specifically configured to determine the decision surface function value of each pixel point in the infrared image according to the normalized grayscale value of each pixel point in the infrared image.

[0129] An image area determination unit is used to determine the infrared image area. The pixel point in is taken as the center, and the block image area is constructed according to the scale parameter R and the augmentation constant λ. And the augmented block image area

[0130] The construction unit 42 is specifically used to construct a decision surface function. The expression of the decision surface function is:

[0131] Where, B={I(x,y)||ux|≤R,|vy|≤R}, B + ={I(x,y)||ux|≤λR,|vy|≤λR}, f(B) is the decision surface function value, B(i,j) is the gray value of the pixel with coordinates (i, j) in the block image area, and B + (i, j) is the gray value of the pixel with coordinates (i, j) in the augmented block image area, a is the relaxation constant, ε is the empirical constant, (u, v) is the coordinate of the central pixel of the block image area, M is the horizontal number of pixels in the infrared image, and N is the horizontal number of pixels in the infrared image.

[0132] The decision surface function value determining unit 43 is specifically used to use the pixel point with coordinates (λR+1, λR+1) as the starting point and the pixel point with coordinates (M-λR, N-λR) as the end point, and to slide the block image area and the augmented block image area constructed based on the pixel point with coordinates (λR+1, λR+1) as the center in the infrared image with a preset step size;

[0133] The decision surface function value determining unit 43 is specifically used to determine the decision surface function value of each pixel point in the infrared image according to the grayscale value of the central pixel point of each block image area and the augmented block image area.

[0134] A point-shaped hot spot determining unit 44, specifically configured to determine that the block image region includes a point-shaped hot spot when the decision surface function value is not less than 0;

[0135] When the decision surface function value is less than 0, it is determined that the block image area does not include a point-shaped hot spot;

[0136] If there is a block image area including a point-shaped hot spot in the infrared image, it is determined that the infrared image includes the point-shaped hot spot;

[0137] If there is no block image area including the point-shaped hot spots in the infrared image, it is determined that the infrared image does not include the point-shaped hot spots.

[0138] A marking unit, used to mark the coordinates of the central pixel point of the block image area where the point-shaped hot spot exists;

[0139] The position determination unit is used to determine the position of the faulty photovoltaic module according to the coordinates of the central pixel point.

[0140] For an introduction to the photovoltaic module point hot spot detection system provided in the present application, please refer to the above-mentioned embodiments, which will not be described in detail here.

[0141] Figure 5 A flow chart of a device for detecting a point-shaped hot spot of a photovoltaic module provided by the present invention, the device for detecting a point-shaped hot spot of a photovoltaic module comprising:

[0142] Memory for storing computer programs;

[0143] The processor is used to implement the steps of the above-mentioned method for detecting point-shaped hot spots of photovoltaic components when executing a computer program.

[0144] For the introduction of the detection device for point-shaped hot spots of photovoltaic modules provided in the present application, please refer to the above-mentioned embodiments, which will not be described in detail here.

[0145] The present application provides a photovoltaic power station, including the above-mentioned photovoltaic component point hot spot detection device, and also includes a photovoltaic component.

[0146] For an introduction to the photovoltaic power station provided in this application, please refer to the above embodiments, which will not be repeated here.

[0147] It should also be noted that, in this specification, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "comprise", "include" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the statement "comprises a ..." does not exclude the presence of other identical elements in the process, method, article or device including the element.

[0148] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0149] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for detecting point-shaped hot spots of photovoltaic modules, characterized in that: include: Acquire an infrared image of a photovoltaic module and determine a grayscale value of the infrared image; Constructing a decision surface function, wherein the decision surface function value is positively correlated with the concentration of pixels in the infrared image and the grayscale value of pixels in the infrared image; Determining a decision surface function value of each pixel point in the infrared image according to the grayscale value of the infrared image; Determining whether the infrared image includes a point-shaped hot spot according to the decision surface function value; Construct a decision surface function, including: The infrared image The pixel point in is taken as the center, and the block image area is constructed according to the scale parameter R and the augmentation constant λ. And the augmented block image area Construct a decision surface function, the expression of which is Where, B={I(x,y)||ux|≤R,|vy|≤R}, B + ={I(x,y)||ux|≤λR,|vy|≤λR}, f(B) is the decision surface function value, B(i,j) is the gray value of the pixel with coordinates (i, j) in the block image area, and B + (i, j) is the gray value of the pixel with coordinates (i, j) in the augmented block image area, a is the relaxation constant, ε is the empirical constant, (u, v) is the coordinate of the central pixel of the block image area, M is the number of horizontal pixels of the infrared image, and N is the number of horizontal pixels of the infrared image; Determining the decision surface function value of each pixel point in the infrared image according to the gray value of the infrared image includes: Taking the pixel point with coordinates (λR+1, λR+1) as the starting point and the pixel point with coordinates (M-λR, N-λR) as the end point, a block image area and an augmented block image area constructed based on the pixel point with coordinates (λR+1, λR+1) as the center slide in the infrared image at a preset step size; The decision surface function value of each pixel point in the infrared image is determined according to the grayscale value of the central pixel point of each block image area and the augmented block image area.

2. The method for detecting point-shaped hot spots of photovoltaic modules according to claim 1, characterized in that: After acquiring the infrared image of the photovoltaic module and determining the grayscale value of the infrared image, the method further includes: According to the image whitening relationship Obtaining a grayscale value of the whitened infrared image; Using a median filter to filter the grayscale value of the whitened infrared image; Among them, I white is the gray value of the infrared image after whitening, I is the gray value of the infrared image, mean(I) is the gray mean of the infrared image, and std(I) is the image standard deviation of the infrared image; Determining the decision surface function value of each pixel point in the infrared image according to the gray value of the infrared image includes: The decision surface function value of each pixel point in the infrared image is determined according to the grayscale value of the infrared image after whitening and filtering.

3. The method for detecting point-shaped hot spots of photovoltaic modules according to claim 2, characterized in that: After filtering the grayscale value of the whitened infrared image using a median filter, the method further includes: According to the grayscale normalization relationship Performing grayscale normalization on the filtered infrared image; Among them, I(x,y) is the gray value of the pixel with coordinates (x, y), I min is the minimum grayscale value of the infrared image after filtering, I max is the maximum grayscale value of the infrared image after filtering; Determining the decision surface function value of each pixel point in the infrared image according to the gray value of the infrared image includes: The decision surface function value of each pixel point in the infrared image is determined according to the normalized grayscale value of each pixel point in the infrared image.

4. The method for detecting a point-shaped hot spot of a photovoltaic module according to claim 1, characterized in that: Determining whether the infrared image includes a point-shaped hot spot according to the decision surface function value includes: When the decision surface function value is not less than 0, determining that the block image area includes a point-shaped hot spot; When the decision surface function value is less than 0, determining that the block image area does not include a point-shaped hot spot; If the block image area including the point-shaped hot spot exists in the infrared image, determining that the infrared image includes the point-shaped hot spot; If the block image area including the point-shaped hot spots does not exist in the infrared image, it is determined that the infrared image does not include the point-shaped hot spots.

5. The method for detecting point-shaped hot spots of photovoltaic modules according to claim 1, characterized in that: After determining whether the infrared image includes a point-shaped hot spot according to the decision surface function value, the method further includes: Marking the coordinates of the central pixel point of the block image area where the point-shaped hot spot exists; The position of the faulty photovoltaic module is determined according to the coordinates of the central pixel point.

6. A photovoltaic module point hot spot detection system, characterized in that: include: A grayscale value acquisition unit, used to acquire an infrared image of the photovoltaic module and determine the grayscale value of the infrared image; A construction unit, used for constructing a decision surface function, wherein a value of the decision surface function is positively correlated with a concentration of pixels in the infrared image and a gray value of pixels in the infrared image; A decision surface function value determining unit, used to determine the decision surface function value of each pixel point in the infrared image according to the gray value of the infrared image; a point-shaped hot spot determination unit, configured to determine whether the infrared image includes a point-shaped hot spot according to the decision surface function value; Also includes: An image area determination unit is used to determine the infrared image area. The pixel point in is taken as the center, and the block image area is constructed according to the scale parameter R and the augmentation constant λ. And the augmented block image area The construction unit is specifically used to construct a decision surface function, and the expression of the decision surface function is: Where, B={I(x,y)||ux|≤R,|vy|≤R}, B + ={I(x,y)||ux|≤λR,|vy|≤λR}, f(B) is the decision surface function value, B(i,j) is the gray value of the pixel with coordinates (i, j) in the block image area, and B + (i, j) is the gray value of the pixel with coordinates (i, j) in the augmented block image area, a is the relaxation constant, ε is the empirical constant, (u, v) is the coordinate of the central pixel of the block image area, M is the number of horizontal pixels in the infrared image, and N is the number of horizontal pixels in the infrared image; The decision surface function value determination unit is specifically used to use the pixel point with coordinates (λR+1, λR+1) as the starting point and the pixel point with coordinates (M-λR, N-λR) as the end point, and to slide the block image area and the augmented block image area constructed based on the pixel point with coordinates (λR+1, λR+1) as the center in the infrared image with a preset step size; The decision surface function value of each pixel point in the infrared image is determined according to the gray value of the central pixel point of each block image area and the augmented block image area.

7. A device for detecting point-shaped hot spots of photovoltaic modules, characterized in that: include: Memory for storing computer programs; A processor is used to implement the steps of the method for detecting point-shaped hot spots of photovoltaic components as described in any one of claims 1 to 5 when executing the computer program.

8. A photovoltaic power station, characterized in that: The device for detecting point-shaped hot spots of photovoltaic components as described in claim 7 also includes photovoltaic components.

Images