Crystalline silicon solar cell panel invisible defect detection method and device based on ultraviolet Fourier transform spectral imaging

Abstract

The invention provides a crystalline silicon solar cell panel invisible defect detection method and device based on ultraviolet Fourier transform spectral imaging, relates to the technical field of solar cell panels, and solves the technical problem of low detection accuracy of invisible defects of solar cell panels. The method comprises the following steps: extracting ultraviolet fluorescence characteristics and visible light and near-infrared characteristics from an original ultraviolet fluorescence spectrum image through an ultraviolet differential enhancement module to respectively obtain a first key characteristic of ultraviolet fluorescence and a second key characteristic of visible light and near-infrared; performing differential processing on the first key feature and the second key feature to obtain an ultraviolet differential image after defect enhancement, and splicing the ultraviolet differential image with the original ultraviolet fluorescence spectrum image to generate a feature enhancement spectrum image; and performing feature fusion on the spatial dimension and the spectral dimension of the feature enhanced spectral image by using depth separable convolution and space separable convolution through an ultraviolet spectrum defect detection network to obtain a fused feature.

Description

Technical Field

[0001] The present application relates to the technical field of solar cell panels, and in particular to a method and device for detecting invisible defects of crystalline silicon solar cell panels based on ultraviolet Fourier transform spectral imaging. Background Art

[0002] There are often various invisible defects on the surface and inside of crystalline silicon solar panels that are difficult for the human eye to identify, such as hidden cracks, fragments, broken grids, stress, etc. At present, the detection of invisible defects of crystalline silicon solar panels mainly relies on electroluminescence (EL), photoluminescence (PL), infrared thermal imaging and other technologies.

[0003] At present, the existing infrared thermal imaging technology realizes imaging by detecting infrared radiation on the surface of an object. When the solar cell works under bias conditions, the current in the leakage area is too large, causing the local temperature to rise, thereby realizing the detection of leakage defects. Although infrared thermal imaging is sensitive to the leakage location, the resolution is low, and the detection sensitivity of invisible defects such as microcracks that do not cause significant temperature differences is low, resulting in low detection accuracy of invisible defects on crystalline silicon solar panels. Summary of the invention

[0004] The object of the present invention is to provide a method and device for detecting invisible defects of crystalline silicon solar panels based on ultraviolet Fourier transform spectral imaging, so as to solve the technical problem of low detection accuracy of invisible defects on crystalline silicon solar panels.

[0005] In a first aspect, the present application provides a method for detecting invisible defects of crystalline silicon solar panels based on ultraviolet Fourier transform spectral imaging, the method comprising:

[0006] Acquire ultraviolet reflected light and fluorescence signals of a solar panel assembly under the irradiation of the ultraviolet spectrum of sunlight, and generate an original phase plane interference pattern sequence through an ultraviolet imaging spectrometer based on the ultraviolet reflected light and the fluorescence signals;

[0007] Correcting the original phase plane interference pattern sequence to generate a pixel interference cube, and performing Fourier transformation on the pixel interference cube to obtain an original ultraviolet fluorescence spectrum image;

[0008] Extracting the features of ultraviolet fluorescence, and the features of visible light and near infrared from the original ultraviolet fluorescence spectrum image through the ultraviolet differential enhancement module, respectively obtaining the first key feature of ultraviolet fluorescence and the second key feature of visible light and near infrared, performing differential processing on the first key feature and the second key feature to obtain an ultraviolet differential image after defect enhancement, splicing the ultraviolet differential image with the original ultraviolet fluorescence spectrum image, and generating a feature enhanced spectrum image after defect area feature enhancement;

[0009] Using an ultraviolet spectrum defect detection network, the spatial dimension and the spectral dimension of the feature enhanced spectrum image are subjected to feature fusion by using a deep separable convolution and a spatial separable convolution to obtain fused features, and the fused features are subjected to feature encoding by downsampling to obtain an encoded feature map;

[0010] The defect information of the encoded feature map is further enhanced by the ultraviolet difference enhancement module to obtain enhanced features after the defect information is enhanced; the enhanced features combine the features of the spatial dimension and the spectral dimension;

[0011] The enhanced features are decoded by upsampling to obtain a decoded feature map, and defect detection results are obtained based on the decoded features through a two-dimensional convolution layer and an activation function.

[0012] In a possible implementation, the ultraviolet fluorescence feature, the visible light feature, and the near-infrared feature are extracted from the original ultraviolet fluorescence spectrum image by the ultraviolet differential enhancement module to obtain the first key feature of ultraviolet fluorescence and the second key feature of visible light and near-infrared, respectively, including:

[0013] Separating the ultraviolet channel, visible light and near-infrared channels of the original ultraviolet fluorescence spectrum image by an ultraviolet differential enhancement module to obtain a separated feature image, and performing feature extraction on the separated feature image to obtain a first key feature of ultraviolet fluorescence and a second key feature of visible light and near-infrared, respectively;

[0014] The step of performing differential processing between the first key feature and the second key feature to obtain a defect-enhanced ultraviolet differential image includes:

[0015] The first wavelength band of the ultraviolet fluorescence is subtracted from the second wavelength band of the corresponding wavelength to obtain an ultraviolet differential image after defect enhancement; wherein the second wavelength band is the visible light and near-infrared wavelength band.

[0016] In a possible implementation, the resolution of the decoded feature map reaches a specified high resolution;

[0017] The feature decoding of the enhanced features by upsampling to obtain a decoded feature map, and obtaining a defect detection result based on the decoded features by a two-dimensional convolution layer and an activation function, includes:

[0018] Perform feature decoding on the enhanced features by upsampling to restore the features to their original resolution, thereby obtaining a decoded feature map; the original resolution is the resolution of the fused features before downsampling;

[0019] Defect prediction is performed on the decoded feature map using a prediction head composed of a two-dimensional convolutional layer and a Sigmoid activation function to obtain an internal invisible defect detection result of the crystalline silicon solar cell panel.

[0020] In a possible implementation, the ultraviolet imaging spectrometer is a total reflection Fourier ultraviolet transform ultraviolet imaging spectrometer, and the total reflection Fourier ultraviolet transform ultraviolet imaging spectrometer includes: a front total reflection imaging mirror, an interference imager, and a rear total reflection imaging mirror;

[0021] The front total reflection imaging mirror is used for focusing and imaging the solar cell panel in the ultraviolet band, and the smallest reflection mirror surface of the front total reflection imaging mirror is plated with screened ultraviolet high-pass, and visible and near-infrared band low-pass filter films; the interference imager is used for decomposing the incident ultraviolet light into images of multiple interference bands; the filter band of the rear total reflection imaging mirror is plated with screened ultraviolet high-pass, and visible and near-infrared band low-pass filter films to enhance the ultraviolet and suppress the visible and near-infrared band images.

[0022] In a possible implementation, the subtracting the first wavelength band of the ultraviolet fluorescence from the second wavelength band of the corresponding wavelength wavelength by wavelength to obtain the ultraviolet differential image after defect enhancement includes subtracting the first wavelength band of the ultraviolet fluorescence from the second wavelength band of the corresponding wavelength wavelength by wavelength to obtain the ultraviolet differential image after defect enhancement by the following formula:

[0023] D(x,y,λ)=IUV(x,y,λ)-IVNIR(x,y,λ);

[0024] Among them, D(x,y,λ) represents the ultraviolet difference image at wavelength λ, IUV(x,y,λ) represents the spectral intensity of the ultraviolet band, and IVNIR(x,y,λ) represents the spectral intensity of the visible light and near-infrared bands.

[0025] In a possible implementation, the feature-enhanced spectral image retains the differential features of the ultraviolet differential image and the original spectral information of the original ultraviolet fluorescence spectral image, and the differential features and the original spectral information are used to characterize defect information of the crystalline silicon solar cell panel.

[0026] In a possible implementation, the spectrum image of the original ultraviolet fluorescence spectrum image covers a spectrum range of 280-1000 nm, and the wavelength interval of the original ultraviolet fluorescence spectrum image is 10 nm.

[0027] In a second aspect, the present application provides a crystalline silicon solar panel invisible defect detection device based on ultraviolet Fourier transform spectral imaging, comprising:

[0028] An acquisition module is used to acquire ultraviolet reflected light and fluorescence signals of a solar panel assembly under the irradiation of the ultraviolet spectrum of sunlight, and generate an original phase plane interference pattern sequence through an ultraviolet imaging spectrometer based on the ultraviolet reflected light and fluorescence signals;

[0029] A transformation module, used for correcting the original phase plane interference pattern sequence to generate a pixel interference cube, and performing Fourier transformation on the pixel interference cube to obtain an original ultraviolet fluorescence spectrum image;

[0030] An ultraviolet differential enhancement module is used to extract the characteristics of ultraviolet fluorescence, and the characteristics of visible light and near infrared from the original ultraviolet fluorescence spectrum image, obtain the first key feature of ultraviolet fluorescence and the second key feature of visible light and near infrared respectively, perform differential processing on the first key feature and the second key feature to obtain an ultraviolet differential image after defect enhancement, and splice the ultraviolet differential image with the original ultraviolet fluorescence spectrum image to generate a feature enhanced spectrum image after defect area feature enhancement;

[0031] An ultraviolet spectral defect detection network is used to perform feature fusion on the spatial dimension and the spectral dimension of the feature enhanced spectral image by using a deep separable convolution and a spatial separable convolution to obtain fused features, and perform feature encoding on the fused features by downsampling to obtain an encoded feature map;

[0032] The ultraviolet difference enhancement module is further used to further enhance the defect information of the encoded feature map to obtain enhanced features after the defect information is enhanced; the enhanced features combine the features of the spatial dimension and the spectral dimension;

[0033] The prediction module is used to decode the enhanced features through upsampling to obtain a decoded feature map, and obtain a defect detection result based on the decoded features through a two-dimensional convolution layer and an activation function.

[0034] In a third aspect, the present application provides an electronic device, comprising a memory and a processor, wherein the memory stores a computer program executable on the processor, and when the processor executes the computer program, the method described in the first aspect is implemented.

[0035] In a fourth aspect, the present application further provides a computer-readable storage medium, which stores computer-executable instructions. When the computer-executable instructions are called and executed by a processor, the computer-executable instructions prompt the processor to execute the method described in the first aspect.

[0036] This application brings the following beneficial effects:

[0037] The present application provides a method and device for detecting invisible defects of crystalline silicon solar panels based on ultraviolet Fourier transform spectral imaging, which can obtain ultraviolet reflected light and fluorescence signals of solar panel components under the irradiation of sunlight ultraviolet spectrum, and generate an original phase plane interference pattern sequence through an ultraviolet imaging spectrometer based on the ultraviolet reflected light and fluorescence signals, correct the original phase plane interference pattern sequence to generate a pixel interference cube, and perform Fourier transform on the pixel interference cube to obtain an original ultraviolet fluorescence spectrum image, and extract the characteristics of ultraviolet fluorescence, the characteristics of visible light and near-infrared from the original ultraviolet fluorescence spectrum image through an ultraviolet differential enhancement module. The first key feature of ultraviolet fluorescence and the second key feature of visible light and near-infrared are obtained respectively, and the first key feature and the second key feature are differentially processed to obtain the ultraviolet difference image after defect enhancement, and the ultraviolet difference image is spliced ​​with the original ultraviolet fluorescence spectrum image to generate a feature enhanced spectrum image after defect area feature enhancement, and the spatial dimension and spectral dimension of the feature enhanced spectrum image are feature fused by the ultraviolet spectrum defect detection network using deep separable convolution and spatial separable convolution to obtain the fused feature, and the fused feature is feature encoded by downsampling to obtain the encoded feature map. The defect information of the encoded feature map is further enhanced by the ultraviolet differential enhancement module to obtain the enhanced features after the defect information is enhanced. The enhanced features combine the features of the spatial dimension and the spectral dimension. The enhanced features are feature decoded by upsampling to obtain the decoded feature map, and the defect detection results are obtained based on the decoded features through a two-dimensional convolution layer and an activation function. In this scheme, the spectral changes can be accurately measured and analyzed through the process of ultraviolet Fourier transform spectral imaging, and then rich spectral information can be extracted from the acquired image, which is convenient for distinguishing defect types and providing their characteristic details. In addition, by comparing the ultraviolet band with the visible Differential calculation is performed on the light and near-infrared bands to construct an ultraviolet differential image. Combined with fluorescence enhancement, the defect feature performance can be significantly improved. The differential image is fused with the ultraviolet fluorescence data cube and then input into the convolutional neural network for supervised training, thereby achieving more accurate detection of invisible defects. Through the defect detection network based on ultraviolet spectral imaging, the sensitivity of the ultraviolet band to the defect characteristics is utilized, and the invisible defects are significantly enhanced through the ultraviolet differential enhancement module. At the same time, the fusion of spatial and spectral features is used to improve the accuracy and robustness of detection, thereby improving the detection accuracy of invisible defects on crystalline silicon solar panels.

[0038] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, preferred embodiments are specifically cited below and described in detail with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the specific implementation methods of the present application or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0040] Figure 1 A schematic diagram of a process for detecting invisible defects of crystalline silicon solar panels based on ultraviolet Fourier transform spectral imaging provided in an embodiment of the present application;

[0041] Figure 2 This is an example of electroluminescence imaging and ultraviolet fluorescence imaging of the same crystalline silicon solar panel;

[0042] Figure 3 Another schematic diagram of the process of the invisible defect detection method of crystalline silicon solar cell panels based on ultraviolet Fourier transform spectral imaging provided in an embodiment of the present application;

[0043] Figure 4 A structural framework diagram of a crystalline silicon solar panel invisible defect detection system based on ultraviolet Fourier transform spectral imaging provided in an embodiment of the present application;

[0044] Figure 5 A schematic diagram of the structure of a crystalline silicon solar panel invisible defect detection device based on ultraviolet Fourier transform spectral imaging provided in an embodiment of the present application;

[0045] Figure 6 A schematic structural diagram of an electronic device provided in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0046] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution of the present application will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application.

[0047] The terms "including" and "having" and any variations thereof mentioned in the embodiments of the present application are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device including a series of steps or units is not limited to the listed steps or units, but may optionally include other steps or units that are not listed, or may optionally include other steps or units that are inherent to these processes, methods, products or devices.

[0048] At present, electroluminescence (EL) technology generates electroluminescence by applying reverse bias to solar cells to stimulate the recombination of electron-hole pairs. The defective areas on solar panels have different luminescence intensities due to different carrier recombination rates, which appear as dark or bright areas in the image. EL technology is highly sensitive to microcracks, metallization defects, etc., and can clearly display the location and shape of defects. However, it requires applying reverse bias to the battery, which may cause certain damage to the battery and is easily affected by external environmental factors such as temperature and humidity. For example, temperature changes can affect the carrier lifetime and recombination rate of semiconductors, thereby affecting the luminescence intensity; high humidity environments may cause water films to form on the surface of the battery, which in turn affects the propagation and absorption of light and affects the detection results.

[0049] Existing photoluminescence (PL) technology can detect defects inside materials non-destructively by stimulating materials to emit light with light of a specific wavelength. It is also highly sensitive to defects that EL technology cannot identify. For example, a photoluminescence imaging system was designed using an 850nm LED linear array light source and a linear array InGaAs camera. However, PL technology has high requirements for light sources and detection equipment, and is easily interfered by background light.

[0050] Existing infrared thermal imaging technology realizes imaging by detecting infrared radiation on the surface of an object. For example, when a solar cell is working under bias conditions, the excessive current in the leakage area causes the local temperature to rise, thereby realizing the detection of leakage defects. Although infrared thermal imaging is sensitive to the leakage location, it has low resolution and limited sensitivity to defects that do not cause significant temperature differences (such as microcracks). In addition, thermal imaging technology is also significantly affected by changes in ambient temperature and requires complex temperature compensation.

[0051] Compared with the high-intensity radiation of natural sunlight, the radiation intensity of electroluminescence and photoluminescence of solar panels is lower, and is mainly concentrated in the short-wave infrared range of 1050-1250nm. To avoid interference from ambient sunlight, existing technologies usually operate indoors or in a dark room. However, cracks in solar panels mainly occur during production, transportation, installation and use, and are caused by factors such as wind load, snow load and thermal mechanical load, especially during the installation stage. The true distribution of cracks often needs to be detected in an installed photovoltaic power station, which affects the adaptability of traditional methods under high outdoor irradiance conditions.

[0052] In addition, the existing imaging technology relies on high-resolution imaging equipment for crack detection, but its ability to identify tiny cracks is limited, resulting in some defects not being discovered and repaired in a timely manner. Existing methods mostly rely on general neural network models or target detection networks for other tasks. However, these methods do not fully consider the impact of imaging methods on defect characteristics, resulting in some potential defects not being effectively identified. At the same time, there is currently no defect extraction algorithm for UV imaging, so the current detection accuracy of invisible defects on crystalline silicon solar panels is low.

[0053] Based on this, the embodiment of the present application provides a method and device for detecting invisible defects of crystalline silicon solar panels based on ultraviolet Fourier transform spectral imaging. This method can solve the technical problem of low detection accuracy of invisible defects on crystalline silicon solar panels.

[0054] The embodiments of the present invention are further described below in conjunction with the accompanying drawings.

[0055] Figure 1 The present invention provides a flow chart of a method for detecting invisible defects of crystalline silicon solar panels based on ultraviolet Fourier transform spectral imaging. Figure 1 As shown, the method includes:

[0056] Step S110, obtaining ultraviolet reflected light and fluorescence signals of the solar panel assembly under the irradiation of the ultraviolet spectrum of sunlight, and generating an original phase plane interferogram sequence through an ultraviolet imaging spectrometer based on the ultraviolet reflected light and the fluorescence signals.

[0057] Under the irradiation of the ultraviolet spectrum of sunlight, the interior and surface of the solar panel components will reflect part of the ultraviolet light and produce fluorescence. The ultraviolet reflected light and the fluorescence signal pass through the ultraviolet imaging spectrometer to generate a sequence of original phase plane interference patterns. Specifically, under the irradiation of the ultraviolet spectrum, the organic molecules in the encapsulation material (such as EVA) of the solar panel will absorb ultraviolet light and produce fluorescence. The strength and distribution of the fluorescence signal are closely related to the material composition, structure and defects. By comparing the ultraviolet reflected light and fluorescence, the ultraviolet imaging equipment can detect invisible defects in the solar panel.

[0058] The ultraviolet fluorescence imaging technology of the ultraviolet imaging spectrometer can effectively detect the crack area. The packaging material around the crack will change due to the photochemical reaction, resulting in the enhancement or weakening of the fluorescence signal, which will appear as a more obvious shadow or bright spot in the image, making tiny cracks easier to identify.

[0059] For example, Figure 2As shown in the figure, electroluminescence imaging and ultraviolet fluorescence imaging were performed on the same crystalline silicon solar panel, respectively. It can be clearly seen that ultraviolet fluorescence imaging has significant advantages in the detection of tiny cracks. The packaging material around the crack causes a significant change in the fluorescence signal due to the photochemical reaction, forming a strong contrast in the ultraviolet fluorescence image. In addition, ultraviolet fluorescence imaging is less affected by external light sources, has low background noise, and weak fluorescence signals are easier to detect, thereby improving the sensitivity and accuracy of detection. The ultraviolet imaging technology of the ultraviolet imaging spectrometer can detect invisible defects such as tiny cracks, bubbles or material degradation, which are usually not found through conventional visual inspection, greatly improving the detection accuracy and reliability.

[0060] As a possible implementation, the ultraviolet imaging spectrometer is a total reflection Fourier ultraviolet transform ultraviolet imaging spectrometer, which includes: a front total reflection imaging mirror, an interference imager, and a rear total reflection imaging mirror; the front total reflection imaging mirror is used to focus and image the solar cell panel in the ultraviolet band, and the minimum reflection mirror surface of the front total reflection imaging mirror is coated with screened ultraviolet high-pass and visible and near-infrared band low-pass filters; the interference imager is used to decompose the incident ultraviolet light into images of multiple interference bands; the filter band of the rear total reflection imaging mirror is coated with screened ultraviolet high-pass and visible and near-infrared band low-pass filters to enhance the ultraviolet and suppress the visible and near-infrared band images.

[0061] For example, Figure 3 As shown, the structure of the total reflection Fourier transform ultraviolet imaging spectrometer is composed of a front total reflection imaging mirror, an interference imager, a rear total reflection imaging mirror, a high-sensitivity ultraviolet detector, a signal acquisition module and other components. Among them, the front total reflection imaging mirror is used to focus and image the solar panel in the ultraviolet band, and the smallest reflective mirror surface of the front total reflection imaging mirror is coated with a screened ultraviolet high-pass and visible-near infrared low-pass filter film; the interference imager (spectrographic interference system) is used to decompose the incident ultraviolet light into images of multiple interference bands; the filter band setting of the rear total reflection imaging mirror is the same as that of the front total emission imaging mirror, which further enhances the ultraviolet and suppresses the visible-near infrared band image; the high-sensitivity ultraviolet detector is responsible for detecting the full-band signal; the signal acquisition module is used to collect the signal output by the detector and convert it into a digital signal for further processing;

[0062] By designing and applying bandpass filters coated with ultraviolet high-pass and visible-near-infrared low-pass in the fully reflective front and rear imaging mirrors, high-sensitivity ultraviolet-enhanced full-band hyperspectral images are obtained.

[0063] Step S120, correcting the original phase plane interference pattern sequence to generate a pixel interference cube, and performing Fourier transform on the pixel interference cube to obtain an original ultraviolet fluorescence spectrum image.

[0064] In this step, the original phase plane interferogram sequence is corrected to generate a pixel interference cube, and the pixel interference cube is Fourier transformed (preprocessed) to obtain an ultraviolet fluorescence spectrum data cube. In a possible implementation, the data processing system generates an ultraviolet spectrum data cube by inverting the original phase plane interferogram sequence.

[0065] In actual applications, when cracks occur in crystalline silicon solar panels, the packaging materials around the cracks will undergo photochemical changes. The crack area usually shows enhanced fluorescence due to lack of oxygen, especially in the rear packaging material containing ultraviolet absorbers. The absorber may diffuse through the cracks to the front packaging area, forming a significant fluorescence reaction. In addition, different types of defects may also cause fluorescence quenching or changes in spectral characteristics, such as the appearance of new absorption peaks or emission peaks. Ultraviolet Fourier transform spectral imaging technology can accurately measure and analyze these spectral changes, extract rich spectral information from the acquired two-dimensional images, and use it to distinguish defect types and provide their characteristic details.

[0066] In a possible implementation, the spectral image of the original ultraviolet fluorescence spectral image covers a spectral range of 280-1000nm, and the wavelength interval of the original ultraviolet fluorescence spectral image is 10nm. Compared with the traditional electroluminescence or photoluminescence method, the ultraviolet Fourier transform spectral imaging device in the embodiment of the present application avoids the interfering spectral region of the short-wave infrared range (1050nm to 1250nm) of the crystalline silicon material, effectively reducing the influence of outdoor sunlight on ultraviolet imaging, thereby achieving high-resolution invisible defect detection.

[0067] Step S130, extracting the characteristics of ultraviolet fluorescence, visible light and near-infrared from the original ultraviolet fluorescence spectrum image through the ultraviolet differential enhancement module, obtaining the first key feature of ultraviolet fluorescence and the second key feature of visible light and near-infrared respectively, and performing differential processing on the first key feature and the second key feature to obtain an ultraviolet differential image after defect enhancement, splicing the ultraviolet differential image with the original ultraviolet fluorescence spectrum image, and generating a feature enhanced spectrum image after defect area feature enhancement.

[0068] In the embodiment of the present application, the feature-enhanced spectral image retains the differential features of the ultraviolet differential image and the original spectral information of the original ultraviolet fluorescence spectral image, and the differential features and the original spectral information are used to characterize the defect information of the crystalline silicon solar cell panel.

[0069] Since the ultraviolet fluorescence image appears as a more obvious shadow or bright spot in the image, such as Figure 4As shown, the ultraviolet differential enhancement module provided in the embodiment of the present application performs spectral differential operation on the ultraviolet band and the visible light-near infrared band, thereby enhancing the defect characteristics hidden in the ultraviolet band.

[0070] Specifically, the UV differential enhancement module first separates the UV channel and visible-near infrared channel of the input spectral image, extracts features from the separated feature images, and obtains the key features of the UV and visible-near infrared, respectively. Subsequently, the differential operation between the UV features and the visible-near infrared features is used to further highlight the invisible defects in the UV band. Finally, the enhanced differential features are spliced ​​with the original spectral image to generate a feature-enhanced image.

[0071] As an optional implementation, the above-mentioned extraction of ultraviolet fluorescence characteristics and visible light and near-infrared characteristics from the original ultraviolet fluorescence spectrum image through the ultraviolet differential enhancement module to obtain the first key characteristics of ultraviolet fluorescence and the second key characteristics of visible light and near-infrared respectively may include the following steps: separating the ultraviolet channel and the visible light and near-infrared channels of the original ultraviolet fluorescence spectrum image through the ultraviolet differential enhancement module to obtain a separated feature image, and performing feature extraction on the separated feature image to obtain the first key characteristics of ultraviolet fluorescence and the second key characteristics of visible light and near-infrared respectively.

[0072] The above-mentioned differential processing between the first key feature and the second key feature to obtain the defect-enhanced ultraviolet differential image can specifically include the following steps: subtracting the first band of ultraviolet fluorescence from the second band of corresponding wavelength wavelength by wavelength to obtain the defect-enhanced ultraviolet differential image; wherein the second band is the band of visible light and near-infrared.

[0073] The above-mentioned subtracting the first wavelength band of ultraviolet fluorescence from the second wavelength band of corresponding wavelength by wavelength to obtain the ultraviolet differential image after defect enhancement may specifically include subtracting the first wavelength band of ultraviolet fluorescence from the second wavelength band of corresponding wavelength by wavelength by using the following formula to obtain the ultraviolet differential image after defect enhancement:

[0074] D(x,y,λ)=IUV(x,y,λ)-IVNIR(x,y,λ);

[0075] Among them, D(x,y,λ) represents the ultraviolet difference image at wavelength λ, IUV(x,y,λ) represents the spectral intensity of the ultraviolet band, and IVNIR(x,y,λ) represents the spectral intensity of the visible light and near-infrared bands.

[0076] Exemplarily, sub-spectral data of the ultraviolet band (280-360nm) and the visible-near infrared band (400-1000nm) are extracted from the preprocessed spectral cube (the above-mentioned ultraviolet fluorescence spectral data cube). Subtraction operation is performed on the ultraviolet band and the visible-near infrared band of the corresponding wavelength wavelength by wavelength to form an ultraviolet differential image.

[0077] By splicing the differential image D(x, y, λ) with the original spectral image (the above-mentioned ultraviolet fluorescence spectral data cube), an enhanced spectral data cube is formed, so that the spectral data cube retains the differential features and the original spectral information, providing rich defect representation for the network input.

[0078] Step S140, using the ultraviolet spectrum defect detection network to fuse the spatial dimension and the spectral dimension of the feature enhanced spectral image using deep separable convolution and spatial separable convolution to obtain fused features, and feature encoding the fused features through downsampling to obtain an encoded feature map.

[0079] In this step, depth-wise separable convolution and spatially separable convolution are used to fuse the spatial and spectral dimensions of the enhanced spectral data cube, and then the fused features are downsampled to complete the encoding operation.

[0080] Exemplarily, the spectral image first passes through the ultraviolet difference enhancement module to enhance the defect area features, and then the spatial dimension and spectral dimension are fused through deep separable convolution and spatial separable convolution, and down-sampled to complete feature encoding. The encoded features are again enhanced through the ultraviolet difference enhancement module to enhance the defect information, and further combine the spatial and spectral features. Next, the features are decoded by upsampling and restored to the original resolution. Finally, the decoded features are passed through a prediction head composed of a two-dimensional convolution layer and a Sigmoid activation function to generate defect detection results.

[0081] The complete architecture of the ultraviolet spectrum defect detection network in the embodiment of the present application is as follows Figure 4 As shown in the figure, the ultraviolet spectrum defect detection network is used to analyze the spectral image to achieve high-precision detection of invisible defects. The ultraviolet spectrum defect detection network achieves significant enhancement of invisible defects through the ultraviolet differential enhancement module, and at the same time improves the accuracy and robustness of detection by fusion of spatial and spectral features, providing a new method for accurate identification of invisible defects in solar panels.

[0082] Step S150, further enhancing the defect information of the encoded feature map through the ultraviolet difference enhancement module to obtain enhanced features after the defect information is enhanced.

[0083] The enhanced features combine the features of the spatial dimension and the spectral dimension. In the embodiment of the present application, the encoded features are further enhanced by the ultraviolet differential enhancement module to further enhance the defect information, and the spatial and spectral dimension features are combined to complete the extraction of deep-level feature representation.

[0084] Step S160, feature decoding is performed on the enhanced features by upsampling to obtain a decoded feature map, and defect detection results are obtained based on the decoded features through a two-dimensional convolution layer and an activation function.

[0085] In this step, the features are restored to the original resolution by upsampling to form a high-resolution feature map. Then, the prediction head composed of a two-dimensional convolutional layer and a Sigmoid activation function predicts the high-resolution feature map to generate the final defect detection result map.

[0086] As an optional implementation, the resolution of the decoded feature map reaches a specified high resolution; the above-mentioned feature decoding of the enhanced features through upsampling to obtain a decoded feature map, and the defect detection results are obtained based on the decoded features through a two-dimensional convolution layer and an activation function. Specifically, the following steps may be included: feature decoding of the enhanced features through upsampling to restore the features to the original resolution to obtain a decoded feature map; the original resolution is the resolution of the fused features before downsampling; and defect prediction of the decoded feature map using a prediction head composed of a two-dimensional convolution layer and a Sigmoid activation function to obtain internal invisible defect detection results of the crystalline silicon solar cell panel.

[0087] Through the process of ultraviolet Fourier transform spectral imaging, the spectral changes can be accurately measured and analyzed, and rich spectral information can be extracted from the acquired images, which is convenient for distinguishing defect types and providing their characteristic details. Moreover, by performing differential calculations in the ultraviolet band (280-360nm) and the visible light and near-infrared bands, an ultraviolet differential image is constructed, and the defect feature performance is significantly improved by combining fluorescence enhancement technology. The differential image is fused with the ultraviolet fluorescence data cube and input into the convolutional neural network (CNN) for supervised training, thereby achieving more accurate detection of invisible defects. Furthermore, through the defect detection network based on ultraviolet spectral imaging, the sensitivity of the ultraviolet band to the defect characteristics is utilized, and the invisible defects are significantly enhanced through the ultraviolet differential enhancement module. At the same time, the accuracy and robustness of detection are improved by using the fusion of spatial and spectral features, that is, the accuracy and robustness of defect detection are improved by feature enhancement, which not only makes up for the shortcomings of the existing technology, so that the detection accuracy of invisible defects on crystalline silicon solar panels is improved, but also provides an efficient and robust solution for the detection of invisible defects of solar panels under complex lighting environments, and realizes more efficient identification of internal invisible defects of solar panels.

[0088] In the embodiment of the present application, there is no need to apply voltage or physical contact as in the prior art, thereby avoiding damage to the solar cell and achieving non-destructive testing. Furthermore, the ultraviolet band is short and is less affected by visible light and near-infrared light, which is suitable for testing under strong light conditions, improving the stability of use in high-irradiance sunlight environments in outdoor environments, and thus being able to adapt to strong light environments.

[0089] Figure 5 A schematic diagram of the structure of a crystalline silicon solar panel invisible defect detection device based on ultraviolet Fourier transform spectral imaging is provided. Figure 5 As shown, the crystalline silicon solar panel invisible defect detection device 500 based on ultraviolet Fourier transform spectrum imaging includes:

[0090] The acquisition module 501 is used to acquire the ultraviolet reflected light and the fluorescence signal of the solar panel assembly under the irradiation of the ultraviolet spectrum of sunlight, and generate an original phase plane interference pattern sequence through an ultraviolet imaging spectrometer based on the ultraviolet reflected light and the fluorescence signal;

[0091] The transformation module 502 is used to correct the original phase plane interference pattern sequence to generate a pixel interference cube, and perform Fourier transformation on the pixel interference cube to obtain an original ultraviolet fluorescence spectrum image;

[0092] The ultraviolet differential enhancement module 503 is used to extract the characteristics of ultraviolet fluorescence, and the characteristics of visible light and near infrared from the original ultraviolet fluorescence spectrum image, respectively obtain the first key feature of ultraviolet fluorescence and the second key feature of visible light and near infrared, and perform differential processing on the first key feature and the second key feature to obtain an ultraviolet differential image after defect enhancement, and splice the ultraviolet differential image with the original ultraviolet fluorescence spectrum image to generate a feature enhanced spectrum image after defect area feature enhancement;

[0093] The ultraviolet spectrum defect detection network 504 is used to perform feature fusion on the spatial dimension and the spectral dimension of the feature enhanced spectrum image by using a depth-separable convolution and a space-separable convolution to obtain fused features, and perform feature encoding on the fused features by downsampling to obtain an encoded feature map;

[0094] The ultraviolet difference enhancement module is further used to further enhance the defect information of the encoded feature map to obtain enhanced features after the defect information is enhanced; the enhanced features combine the features of the spatial dimension and the spectral dimension;

[0095] The prediction module 505 is used to decode the enhanced features through upsampling to obtain a decoded feature map, and obtain a defect detection result based on the decoded features through a two-dimensional convolution layer and an activation function.

[0096] The crystalline silicon solar panel invisible defect detection device based on ultraviolet Fourier transform spectral imaging provided in the embodiment of the present application has the same technical features as the crystalline silicon solar panel invisible defect detection method based on ultraviolet Fourier transform spectral imaging provided in the above embodiment, so it can also solve the same technical problems and achieve the same technical effects.

[0097] An electronic device provided in an embodiment of the present application is Figure 6 As shown, the electronic device 600 includes a processor 602 and a memory 601, wherein the memory stores a computer program that can be run on the processor, and the processor implements the steps of the method provided in the above embodiment when executing the computer program.

[0098] See also Figure 6 The electronic device further includes: a bus 603 and a communication interface 604, a processor 602, a communication interface 604 and a memory 601 are connected via the bus 603; the processor 602 is used to execute an executable module stored in the memory 601, such as a computer program.

[0099] The memory 601 may include a high-speed random access memory (RAM), and may also include a non-volatile memory, such as at least one disk storage. The communication connection between the system network element and at least one other network element is realized through at least one communication interface 604 (which may be wired or wireless), and the Internet, wide area network, local area network, metropolitan area network, etc. may be used.

[0100] The bus 603 may be an ISA bus, a PCI bus, or an EISA bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 6 Only one bidirectional arrow is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0101] Among them, the memory 601 is used to store programs, and the processor 602 executes the program after receiving the execution instruction. The method executed by the device defined by the process disclosed in any embodiment of the present application can be applied to the processor 602 or implemented by the processor 602.

[0102] The processor 602 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the hardware integrated logic circuit or software instructions in the processor 602. The above processor 602 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present application can be implemented or executed. The general processor can be a microprocessor or the processor can also be any conventional processor, etc. The steps of the method disclosed in the embodiments of the present application can be directly embodied as a hardware decoding processor to execute, or the hardware and software modules in the decoding processor are combined to execute. The software module can be located in a storage medium mature in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory 601, and the processor 602 reads the information in the memory 601 and completes the steps of the above method in combination with its hardware.

[0103] Corresponding to the above-mentioned method for detecting invisible defects of crystalline silicon solar panels based on ultraviolet Fourier transform spectral imaging, an embodiment of the present application also provides a computer-readable storage medium, which stores computer-executable instructions. When the computer-executable instructions are called and executed by a processor, the computer-executable instructions prompt the processor to execute the steps of the above-mentioned method for detecting invisible defects of crystalline silicon solar panels based on ultraviolet Fourier transform spectral imaging.

[0104] The crystalline silicon solar panel invisible defect detection device based on ultraviolet Fourier transform spectral imaging provided in the embodiment of the present application can be specific hardware on the device or software or firmware installed on the device. The device provided in the embodiment of the present application, its implementation principle and the technical effect produced are the same as those in the aforementioned method embodiment. For the sake of brief description, the parts not mentioned in the device embodiment can refer to the corresponding contents in the aforementioned method embodiment. Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working processes of the systems, devices and units described above can all refer to the corresponding processes in the aforementioned method embodiment, and will not be repeated here.

[0105] In the embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.

[0106] For another example, the flowchart and block diagram in the accompanying drawings show the possible architecture, function and operation of the device, method and computer program product according to multiple embodiments of the present application. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of the code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or the flowchart, and the combination of the boxes in the block diagram and / or the flowchart can be implemented with a dedicated hardware-based system that performs a specified function or action, or can be implemented with a combination of dedicated hardware and computer instructions.

[0107] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0108] In addition, each functional unit in the embodiments provided in the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0109] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the invisible defect detection method of crystalline silicon solar panels based on ultraviolet Fourier transform spectral imaging described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, referred to as ROM), random access memory (Random Access Memory, referred to as RAM), disk or optical disk and other media that can store program codes.

[0110] It should be noted that similar numbers and letters represent similar items in the following figures. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. In addition, the terms "first", "second", "third", etc. are only used to distinguish the description and are not to be understood as indicating or implying relative importance.

[0111] Finally, it should be noted that the above-described embodiments are only specific implementation methods of the present application, which are used to illustrate the technical solution of the present application, rather than to limit it. The protection scope of the present application is not limited thereto. Although the present application is described in detail with reference to the aforementioned embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solution recorded in the aforementioned embodiments within the technical scope disclosed in the present application, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solution deviate from the scope of the technical solution of the embodiment of the present application. They should all be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

Claims

1. A method for detecting invisible defects of crystalline silicon solar panels based on ultraviolet Fourier transform spectral imaging, characterized in that: The method comprises: Acquire ultraviolet reflected light and fluorescence signals of a solar panel assembly under the irradiation of the ultraviolet spectrum of sunlight, and generate an original phase plane interference pattern sequence through an ultraviolet imaging spectrometer based on the ultraviolet reflected light and the fluorescence signals; Correcting the original phase plane interference pattern sequence to generate a pixel interference cube, and performing Fourier transformation on the pixel interference cube to obtain an original ultraviolet fluorescence spectrum image; Extracting the features of ultraviolet fluorescence, and the features of visible light and near infrared from the original ultraviolet fluorescence spectrum image through the ultraviolet differential enhancement module, respectively obtaining the first key feature of ultraviolet fluorescence and the second key feature of visible light and near infrared, performing differential processing on the first key feature and the second key feature to obtain an ultraviolet differential image after defect enhancement, splicing the ultraviolet differential image with the original ultraviolet fluorescence spectrum image, and generating a feature enhanced spectrum image after defect area feature enhancement; Using an ultraviolet spectrum defect detection network, the spatial dimension and the spectral dimension of the feature enhanced spectrum image are subjected to feature fusion by using a deep separable convolution and a spatial separable convolution to obtain fused features, and the fused features are subjected to feature encoding by downsampling to obtain an encoded feature map; The defect information of the encoded feature map is further enhanced by the ultraviolet difference enhancement module to obtain enhanced features after the defect information is enhanced; the enhanced features combine the features of the spatial dimension and the spectral dimension; The enhanced features are decoded by upsampling to obtain a decoded feature map, and defect detection results are obtained based on the decoded features through a two-dimensional convolution layer and an activation function.

2. The method according to claim 1, characterized in that The ultraviolet fluorescence feature, the visible light feature and the near infrared feature are extracted from the original ultraviolet fluorescence spectrum image by the ultraviolet differential enhancement module, and the first key feature of ultraviolet fluorescence and the second key feature of visible light and near infrared are obtained respectively, including: Separating the ultraviolet channel, visible light and near-infrared channels of the original ultraviolet fluorescence spectrum image by an ultraviolet differential enhancement module to obtain a separated feature image, and performing feature extraction on the separated feature image to obtain a first key feature of ultraviolet fluorescence and a second key feature of visible light and near-infrared, respectively; The step of performing differential processing between the first key feature and the second key feature to obtain a defect-enhanced ultraviolet differential image includes: The first wavelength band of the ultraviolet fluorescence is subtracted from the second wavelength band of the corresponding wavelength to obtain an ultraviolet differential image after defect enhancement; wherein the second wavelength band is the visible light and near-infrared wavelength band.

3. The method according to claim 1, characterized in that The resolution of the decoded feature map reaches a specified high resolution; The feature decoding of the enhanced features by upsampling to obtain a decoded feature map, and obtaining a defect detection result based on the decoded features by a two-dimensional convolution layer and an activation function, includes: Perform feature decoding on the enhanced features by upsampling to restore the features to their original resolution, thereby obtaining a decoded feature map; the original resolution is the resolution of the fused features before downsampling; Defect prediction is performed on the decoded feature map using a prediction head composed of a two-dimensional convolutional layer and a Sigmoid activation function to obtain an internal invisible defect detection result of the crystalline silicon solar cell panel.

4. The method according to claim 1, characterized in that: The ultraviolet imaging spectrometer is a total reflection Fourier ultraviolet transform ultraviolet imaging spectrometer, and the total reflection Fourier ultraviolet transform ultraviolet imaging spectrometer comprises: a front total reflection imaging mirror, an interference imager, and a rear total reflection imaging mirror; The front total reflection imaging mirror is used for focusing and imaging the solar cell panel in the ultraviolet band, and the smallest reflection mirror surface of the front total reflection imaging mirror is plated with screened ultraviolet high-pass, and visible and near-infrared band low-pass filter films; the interference imager is used for decomposing the incident ultraviolet light into images of multiple interference bands; the filter band of the rear total reflection imaging mirror is plated with screened ultraviolet high-pass, and visible and near-infrared band low-pass filter films to enhance the ultraviolet and suppress the visible and near-infrared band images.

5. The method according to claim 2, characterized in that: Subtracting the first wavelength band of the ultraviolet fluorescence from the second wavelength band of the corresponding wavelength wavelength by wavelength to obtain the ultraviolet differential image after defect enhancement, including subtracting the first wavelength band of the ultraviolet fluorescence from the second wavelength band of the corresponding wavelength wavelength by wavelength to obtain the ultraviolet differential image after defect enhancement by the following formula: D(x,y,λ)=I UV (x,y,λ)-I VNIR (x,y,λ); Where D(x,y,λ) represents the ultraviolet difference image at wavelength λ, I UV (x, y, λ) represents the spectral intensity of the ultraviolet band, I VNIR (x, y, λ) represents the spectral intensity of visible light and near-infrared bands.

6. The method according to claim 1, characterized in that The feature-enhanced spectral image retains the differential features of the ultraviolet differential image and the original spectral information of the original ultraviolet fluorescence spectral image, and the differential features and the original spectral information are used to characterize the defect information of the crystalline silicon solar cell panel.

7. The method according to claim 1, characterized in that The spectral image of the original ultraviolet fluorescence spectral image covers a spectral range of 280-1000nm, and the wavelength interval of the original ultraviolet fluorescence spectral image is 10nm.

8. A crystalline silicon solar cell panel invisible defect detection device based on ultraviolet Fourier transform spectral imaging, characterized in that: include: An acquisition module is used to acquire ultraviolet reflected light and fluorescence signals of a solar panel assembly under the irradiation of the ultraviolet spectrum of sunlight, and generate an original phase plane interference pattern sequence through an ultraviolet imaging spectrometer based on the ultraviolet reflected light and fluorescence signals; A transformation module, used for correcting the original phase plane interference pattern sequence to generate a pixel interference cube, and performing Fourier transformation on the pixel interference cube to obtain an original ultraviolet fluorescence spectrum image; An ultraviolet differential enhancement module is used to extract the characteristics of ultraviolet fluorescence, and the characteristics of visible light and near infrared from the original ultraviolet fluorescence spectrum image, obtain the first key feature of ultraviolet fluorescence and the second key feature of visible light and near infrared respectively, perform differential processing on the first key feature and the second key feature to obtain an ultraviolet differential image after defect enhancement, and splice the ultraviolet differential image with the original ultraviolet fluorescence spectrum image to generate a feature enhanced spectrum image after defect area feature enhancement; An ultraviolet spectral defect detection network is used to perform feature fusion on the spatial dimension and the spectral dimension of the feature enhanced spectral image by using a deep separable convolution and a spatial separable convolution to obtain fused features, and perform feature encoding on the fused features by downsampling to obtain an encoded feature map; The ultraviolet difference enhancement module is further used to further enhance the defect information of the encoded feature map to obtain enhanced features after the defect information is enhanced; the enhanced features combine the features of the spatial dimension and the spectral dimension; The prediction module is used to decode the enhanced features through upsampling to obtain a decoded feature map, and obtain a defect detection result based on the decoded features through a two-dimensional convolution layer and an activation function.

9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program that can be run on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions. When the computer-executable instructions are called and executed by a processor, the computer-executable instructions prompt the processor to execute the method according to any one of claims 1 to 7.

Images