Electroluminescence image acquisition system and method for photovoltaic module, and defect detection method

Abstract

The invention provides a photovoltaic module electroluminescent image acquisition system and method, and a defect detection method, which can be used in the technical field of photovoltaic power generation. The system comprises a liquid crystal polarization modulator, an infrared camera and an image processing unit, the liquid crystal polarization modulator is arranged in front of a lens of the infrared camera and used for changing the polarization direction of incident light entering the lens of the infrared camera; the infrared camera is used for shooting the photovoltaic module to be detected so as to be used for collecting an infrared image set of the photovoltaic module; wherein the infrared image set comprises a first infrared image collected during the first polarization direction and a second infrared image collected during the second polarization direction under the same field of view; and the image processing unit is used for carrying out differential processing on the first infrared image and the second infrared image to obtain an electroluminescent image of the photovoltaic module. According to the invention, the electroluminescent signal of the photovoltaic module is extracted through a pure optical polarization modulation means, so that the collection efficiency of the electroluminescent image of the photovoltaic module is greatly improved.

Description

Technical Field

[0001] This application relates to the field of photovoltaic power generation technology, and in particular to a photovoltaic module electroluminescence image acquisition system and method, and a defect detection method. Background Technology

[0002] Photovoltaic module defect inspection is a key link in ensuring the efficient and stable operation of photovoltaic power generation systems. Its role is to accurately identify potential defects inside the modules, such as hidden cracks, broken grids, poor soldering, black cores, and PID attenuation, through precision testing technologies (such as electroluminescence (EL) testing and photoluminescence (PL) testing).

[0003] Electroluminescence (EL) imaging technology for photovoltaic modules is an important method for detecting internal defects in solar cells. Traditional EL detection requires applying a forward voltage or current to the module in a dark environment, causing the solar cells to emit weak near-infrared light, which is then captured by an infrared camera to form an image. EL images can visually reveal defects such as microcracks, poor welding, and failure areas in solar cells, and are of great significance for assessing the health status of the module.

[0004] However, existing technologies require photovoltaic modules to be powered on and a dark environment for practical applications. Even when technologies attempt to extract EL signals during the day, they often require electrical connections or special equipment, resulting in low detection efficiency. This limits the application of EL imaging in on-site photovoltaic module fault inspection, as maintenance personnel cannot quickly obtain information on defects such as microcracks in the modules during the day. Summary of the Invention

[0005] This application provides a photovoltaic module electroluminescence image acquisition system and method, and a defect detection method, to solve the technical problem that the existing photovoltaic module electroluminescence image acquisition technology needs to be operated at night, resulting in low image acquisition efficiency.

[0006] According to the first aspect disclosed in this application, this application provides a photovoltaic module electroluminescence image acquisition system, comprising: A liquid crystal polarization modulator is disposed in front of the lens of an infrared camera to change the polarization direction of incident light entering the lens of the infrared camera; wherein the polarization direction includes an alternately switching first polarization direction and a second polarization direction. An infrared camera is used to photograph the photovoltaic module under test to acquire an infrared image set of the photovoltaic module; wherein the infrared image set includes a first infrared image acquired during a first polarization direction under the same field of view, and a second infrared image acquired during a second polarization direction; An image processing unit is used to perform differential processing on the first infrared image and the second infrared image to obtain the electroluminescent image of the photovoltaic module.

[0007] In one feasible implementation, a narrowband filter is disposed between the lens of the infrared camera and the liquid crystal polarization modulator; the filter can effectively filter out solar background light and unnecessary infrared light, thereby improving the signal-to-noise ratio of the EL signal.

[0008] In one feasible implementation, the liquid crystal polarization modulator is a liquid crystal shutter or a liquid crystal variable waveplate, and the liquid crystal polarization modulator switches the polarization direction electronically and synchronizes with the shooting frequency of the infrared camera.

[0009] In one feasible implementation, the infrared camera is a near-infrared camera that uses an indium gallium arsenide (InGaAs) sensor, which is capable of capturing weak infrared light signals from photovoltaic modules, especially in the 1100-1200nm band.

[0010] In one possible implementation, a mobile platform is also included, which is used to carry the liquid crystal polarization modulator, the infrared camera, and the image processing unit.

[0011] In one feasible implementation, the mobile platform is a ground vehicle or a drone.

[0012] According to the second aspect disclosed in this application, this application provides a method for acquiring electroluminescent images of photovoltaic modules, based on the photovoltaic module electroluminescent image acquisition system as described in any one of the first aspects, comprising: The infrared image set of the photovoltaic module is acquired; wherein the infrared image set includes a first infrared image acquired during a first polarization direction under the same field of view, and a second infrared image acquired during a second polarization direction; Differential processing is performed on the first infrared image and the second infrared image to obtain the electroluminescence image of the photovoltaic module.

[0013] In one feasible implementation, differential processing is performed on the first infrared image and the second infrared image to obtain the electroluminescent image of the photovoltaic module, including: The first infrared image and the second infrared image are input into a pre-constructed differential extraction model to obtain an electroluminescent image output by the differential extraction model; wherein the differential extraction model is obtained based on a deep learning model trained on it.

[0014] In one feasible implementation, after acquiring the infrared image set of the photovoltaic module, the method further includes: The first infrared image and the second infrared image are registered and normalized.

[0015] In one feasible implementation, after acquiring the electroluminescence image of the photovoltaic module, the method further includes: The electroluminescent image is then subjected to image enhancement processing.

[0016] According to a third aspect disclosed in this application, this application provides a method for defect detection of photovoltaic modules, comprising: Acquire an electroluminescence image; wherein the electroluminescence image is obtained based on the photovoltaic module electroluminescence image acquisition method as described in the second aspect; Based on the electroluminescent image, the defect detection results of the photovoltaic module are obtained.

[0017] Compared with the prior art, this application has the following advantages: This application provides a photovoltaic module electroluminescence image acquisition system and method, as well as a defect detection method. It extracts the EL signal of the photovoltaic module through purely optical polarization modulation, enabling the acquisition of electroluminescence images without powering the photovoltaic module. This allows for direct EL inspection of in-service photovoltaic modules under strong daylight conditions. The entire system requires no electrical connection to the photovoltaic module and no current injection, ensuring safety and efficiency while avoiding the hassles and risks associated with on-site wiring. Thus, it efficiently acquires EL images of photovoltaic modules without affecting the normal power generation and operation of the photovoltaic power station. Suitable for large-scale on-site applications and rapid deployment in the field, it significantly improves the acquisition efficiency of photovoltaic module electroluminescence images, offering both safety and high efficiency. Attached Figure Description

[0018] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0019] Figure 1 This is a schematic diagram of the structure of a photovoltaic module electroluminescence image acquisition system provided in an embodiment of this application; Figure 2 This application provides a schematic diagram of the structure of a mobile inspection system for electroluminescent image acquisition of photovoltaic modules. Figure 3 A flowchart illustrating a photovoltaic module electroluminescence image acquisition method provided in this application embodiment; Figure 4 A schematic flowchart of another photovoltaic module electroluminescence image acquisition method provided in this application embodiment; Figure 5 This is a flowchart illustrating a method for detecting defects in photovoltaic modules provided in an embodiment of this application.

[0020] Explanation of reference numerals in the attached figures: 100-Infrared Camera; 200-Narrowband Filter; 300-Liquid Crystal Polarization Modulator; 400 - Image processing unit; 500- Photovoltaic modules; 600 - Unmanned Aerial Vehicle (UAV)

[0021] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0022] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0023] Photovoltaic module defect inspection is a crucial step in ensuring the efficient and stable operation of photovoltaic power generation systems. Its role is to accurately identify potential defects within the modules, such as microcracks, broken grids, poor soldering, black cores, and PID degradation, through sophisticated testing technologies (such as electroluminescence (EL) testing and photoluminescence (PL) testing). This process not only prevents defective modules from causing problems like hot spot effects and power degradation during long-term outdoor operation, but also reduces subsequent maintenance costs, extends module lifespan, and increases the overall power generation of the system.

[0024] Electroluminescence (EL) imaging technology for photovoltaic (PV) modules is an important method for detecting internal defects in solar cells. Traditional EL detection requires applying a forward voltage or current to the module in a dark environment, causing the solar cells to emit weak near-infrared light, which is then captured by an infrared camera to form an image. EL images can visually reveal defects such as microcracks, poor welding, and failure areas in solar cells, and are of great significance for assessing the health status of PV modules. However, existing technologies have the following shortcomings in practical applications: First, photovoltaic modules must be powered on in a dark environment. Traditional electroluminescence testing usually requires nighttime operation or placing the modules in a dark room to avoid interference from strong sunlight. This necessitates shutting down the photovoltaic array and connecting each module to a power source to inject current, a complex and time-consuming process. In existing methods, a bias voltage needs to be applied to the wiring of each module to obtain an electroluminescence image, which is not conducive to the daily operation and maintenance of large-scale photovoltaic power plants.

[0025] Secondly, field applications are limited. Even if technologies attempt to extract EL signals during the day, electrical connections or specialized equipment are often required. For example, the radiation signal of the photovoltaic module can be modulated by rapidly switching its operating state (e.g., alternating between open and short circuit states), and then image differential analysis can be used to obtain the photoluminescence (PL) and electroluminescence (EL) components. However, this method still requires electrical connections and control of the photovoltaic module and complex phase-locked amplification technology, making it difficult to implement on mobile platforms such as drones. Other methods utilize strong light sources to excite the photovoltaic module to produce photoluminescence (e.g., laser-excited PL imaging), but the signals are still difficult to distinguish against a strong solar background, and the equipment is expensive and complex.

[0026] Third, the detection efficiency is low. Most current EL inspections require taking pictures of each module individually under power outage conditions, making it difficult to inspect operating photovoltaic modules in real time and over a large area. This limits the application of EL imaging in on-site inspection of photovoltaic module faults, and maintenance personnel cannot quickly obtain information on defects such as microcracks in the modules during the day.

[0027] In summary, current technology lacks a solution for directly imaging in-service photovoltaic (PV) modules under strong daylight conditions without applying current to them. How to efficiently acquire EL images of PV modules without affecting the normal power generation of the PV power plant has become an urgent technical problem to be solved.

[0028] To address the aforementioned technical issues, this application proposes a photovoltaic module electroluminescence image acquisition system and method, as well as a defect detection method. It extracts the EL signal of the photovoltaic module through purely optical polarization modulation, acquiring electroluminescence images without powering the photovoltaic module. This allows for efficient acquisition of EL images of the photovoltaic module without affecting the normal power generation and operation of the photovoltaic power station. Suitable for large-scale field applications and rapid deployment in the field, it significantly improves the acquisition efficiency of photovoltaic module electroluminescence images, ensuring safety and high efficiency.

[0029] The technical solutions of the photovoltaic module electroluminescence image acquisition system and method and defect detection method provided in this application are described in detail below through specific embodiments. It should be noted that the following embodiments may exist alone or in combination with each other, and the same or similar content may not be described again in different embodiments.

[0030] Figure 1 A schematic diagram of a photovoltaic module 500 electroluminescent image acquisition system provided in this application embodiment is shown below. Figure 1In some embodiments, the photovoltaic module 500 electroluminescence image acquisition system includes a liquid crystal polarization modulator 300, an infrared camera 100, and an image processing unit 400. The liquid crystal polarization modulator 300 is disposed in front of the lens of the infrared camera 100 to change the polarization direction of the incident light entering the lens of the infrared camera 100. The polarization direction includes an alternately switching first polarization direction and a second polarization direction. The infrared camera 100 is used to capture images of the photovoltaic module 500 under test to acquire an infrared image set of the photovoltaic module 500. The infrared image set includes a first infrared image acquired during the first polarization direction and a second infrared image acquired during the second polarization direction under the same field of view. The image processing unit 400 is used to perform differential processing on the first infrared image and the second infrared image to obtain an electroluminescence image of the photovoltaic module 500.

[0031] Specifically, the infrared camera 100 employs an indium gallium arsenide (InGaAs) sensor, with a lens transmission range covering at least 900–1200 nm, particularly in the 1100–1200 nm band, enabling it to capture weak infrared light signals from photovoltaic modules. The infrared camera 100 is connected to an image processing unit 400 for image storage and processing, allowing for real-time acquisition and storage of image sequences. The high-sensitivity InGaAs infrared camera 100, along with its dedicated lens, is a semiconductor-based optoelectronic imaging device. Its core is the InGaAs absorption layer, a compound of indium (In), gallium (Ga), and arsenic (As) with an adjustable bandgap. This allows for efficient absorption of short-wave infrared (SWIR, wavelength range 900–1700 nm) photons, enabling the infrared camera 100's spectral response to cover the electroluminescence band of photovoltaic silicon cells while remaining insensitive to the visible light band.

[0032] Specifically, the liquid crystal polarization modulator 300 is installed in front of the lens of the infrared camera 100. Under the action of a control signal, it can change the polarization state of the incident light, thereby controlling the infrared camera 100 to alternate between at least two different optical states, thus altering the polarization characteristics of the light entering the camera lens. During shooting, the external environment (such as sunlight intensity and camera exposure settings) is kept as constant as possible, and the optical polarization conditions are changed only by the liquid crystal polarization modulator 300 to achieve rapid modulation of the polarization direction or phase state of the light passing through the lens. The switching frequency of the liquid crystal polarization modulator 300 is synchronized with the camera's frame acquisition frequency.

[0033] Optionally, the liquid crystal polarization modulator 300 can employ linear polarization switching, circular polarization switching, or even multi-level grayscale modulation.

[0034] Specifically, the image processing unit 400 can be an embedded microcomputer or a connected host computer with built-in control and processing software. This image processing unit 400 is connected to the infrared camera 100 and the liquid crystal polarization modulator 300 via lines, and is used to send control commands and receive image data acquired by the camera. On the one hand, it can control the switching of the liquid crystal polarization modulator 300 to synchronize with camera shooting; on the other hand, it executes image processing algorithms to extract electroluminescent images from the acquired polarization sequence images.

[0035] Optionally, the image processing unit 400 can further perform defect identification and analysis of the photovoltaic module 500. Simultaneously, the image processing unit 400 can also display the EL defect images from the identification results to the user in real time, or store them for subsequent offline analysis.

[0036] In this embodiment, a system for electroluminescence imaging acquisition without current injection is provided, specifically including a near-infrared camera 100, a liquid crystal polarization modulator 300, and an image processing device. The liquid crystal polarization modulator 300 is placed in the optical path of the infrared camera 100 to alternately modulate the polarization direction or phase of the incident light entering the lens of the infrared camera 100. The image processing unit 400 controls the infrared camera 100 and the polarization modulator to work synchronously, performs differential processing on the acquired infrared image, and outputs the electroluminescence image of the photovoltaic module 500.

[0037] The photovoltaic module 500 under test is typically a silicon crystal solar module, operating in an open-circuit state under normal sunlight. Because the photovoltaic module 500 exhibits weak spontaneous emission under natural light, equivalent to a very low-excitation electroluminescence (EL) signal, the transmittance of this signal varies under different polarization states through the optical system. Meanwhile, the intensity of strong solar background scattered light remains essentially constant across both polarization states. Based on this principle, comparing two infrared images can eliminate most of the background components that do not change with polarization, highlighting the weak EL signal that is related to the polarization state.

[0038] Unlike traditional electroluminescence (EL) image acquisition technologies that rely on electrical excitation, existing EL detection methods are limited by the need for wiring and power, and can only be operated at night. This embodiment extracts the EL signal of the photovoltaic (PV) module 500 under natural sunlight without power, using purely optical polarization modulation. This achieves the acquisition of EL images without powering the PV module 500. This technology enables direct EL detection of in-service PV modules 500 in strong daylight conditions. The entire system requires no electrical connection to the PV module 500 and no current injection, making it safe and efficient, avoiding the hassles and risks associated with on-site wiring. Thus, it efficiently acquires EL images of the PV module 500 without affecting the normal power generation and operation of the PV power plant. It is suitable for large-scale on-site applications and rapid deployment in the field, significantly improving the acquisition efficiency of PV module 500 EL images, and is both safe and efficient.

[0039] Optionally, a narrowband filter 200 is provided between the lens of the infrared camera 100 and the liquid crystal polarization modulator 300. This filter can effectively filter out solar background light and unnecessary infrared light, thereby improving the signal-to-noise ratio of the EL signal.

[0040] The narrowband filter 200 can filter out visible light and most of the solar infrared background, allowing only narrowband light close to the silicon bandgap radiation wavelength to pass through, thereby improving the signal-to-noise ratio of the EL signal relative to the background and further enhancing the detection accuracy of electroluminescence images.

[0041] By combining the liquid crystal polarization modulator 300 and the narrowband filter 200, the ambient light interference is suppressed by narrowband filtering and polarization modulation. The EL signal can be extracted under strong solar radiation at noon, which is suitable for strong light environment during the day. There is no need to wait until night, which greatly expands the usage time window of EL imaging and improves the inspection efficiency of photovoltaic module 500.

[0042] Specifically, the center wavelength of the narrowband filter 200 is preferably about 1100–1150 nm, with a bandwidth of about 20 nm. It only allows narrowband spectrum to pass through, which is equivalent to providing a selective window for the infrared camera 100 to see through the near-infrared spontaneous emission of the photovoltaic module 500, allowing the bandgap radiation of the silicon cell to pass through while blocking other stray light, so that only the EL signal band emitted by the module can be transmitted. It has an out-of-band cutoff capability of OD4 or higher to maximize the suppression of solar background and improve the detection contrast of weak EL signals.

[0043] In addition, the center wavelength of the narrowband filter 200 can also be adjusted according to the solar cell material.

[0044] Optionally, the liquid crystal polarization modulator 300 is a liquid crystal shutter or a liquid crystal variable waveplate. The liquid crystal polarization modulator switches the polarization direction electronically and synchronizes with the shooting frequency of the infrared camera. Installed at the front end of the camera's optical path, it is used to electronically control the switching of polarization states without changing the external lighting. Devices such as liquid crystal shutters and liquid crystal variable waveplates can be used to achieve rapid modulation of the polarization direction or phase state of the transmitted light. Its switching frequency and timing are set by the control unit and synchronized with the camera's frame acquisition.

[0045] Specifically, a liquid crystal shutter is an optical control device based on the electro-optic effect of liquid crystal materials. It dynamically adjusts the alignment of liquid crystal molecules by applying voltage to achieve rapid switching of light transmittance.

[0046] Specifically, a liquid crystal variable waveplate is an optical element that uses the birefringence properties of liquid crystal to achieve dynamic control of the polarization state of light. Its core structure is a transparent cell filled with nematic liquid crystal, and the surface is coated with a transparent conductive film to apply voltage.

[0047] For example, the liquid crystal shutter can switch between polarization state A (first polarization direction) and polarization state B (second polarization direction). Corresponding to different polarization orientations, the infrared camera 100 synchronously acquires image frames of the corresponding field of view in state A and state B respectively.

[0048] Optional, the infrared camera 100 is a near-infrared camera.

[0049] Among them, a near-infrared camera is a special imaging device that can capture light in the near-infrared spectrum (wavelengths are usually between 700 and 2500 nanometers), thereby enabling it to detect weak infrared light signals of photovoltaic modules at around 1100nm.

[0050] Optionally, a near-infrared camera using an InGaAs sensor has high sensitivity to weak light with a wavelength of around 1100nm. When used with a low F-number wide-angle lens (e.g., 5mm focal length / F1.4, spectral response 400-1700nm), it can cover the entire photovoltaic module for imaging.

[0051] In some embodiments, the photovoltaic module 500 electroluminescent image acquisition system further includes a mobile platform for mounting a liquid crystal polarization modulator 300, an infrared camera 100, and an image processing unit 400.

[0052] In this embodiment, the liquid crystal polarization modulator 300, infrared camera 100, and image processing unit 400 can be integrated onto a mobile platform. The entire device is small in size, lightweight, compact, portable, and easily deployed, enabling automated inspection of a large-area photovoltaic module array 500. This deployment method fully leverages the system's advantages of non-contact rapid detection, achieving long-distance, rapid scanning and inspection. Dozens to hundreds of photovoltaic modules 500 can be inspected within one hour, covering an area far exceeding that of manual nighttime image capture, significantly saving labor and time costs.

[0053] Optionally, the mobile platform can be a ground vehicle or a drone 600.

[0054] The system can be flexibly installed on platforms such as UAVs 600 and ground vehicles. Without external power supply or electrical contacts, it can scan photovoltaic arrays by flying or moving to obtain EL images of 500 photovoltaic modules in batches. It is very suitable for online inspection of large-scale photovoltaic power plants and other scenarios where electrical excitation cannot be easily implemented, providing a highly efficient non-contact inspection tool for photovoltaic operation and maintenance.

[0055] For example, the system can be installed on the gimbal of a drone 600 and fly over a photovoltaic power station along a predetermined route. The drone 600 or the detection device can then point the infrared camera 100 at the front of the photovoltaic module 500, capture infrared images of the photovoltaic module 500 in different polarization states one by one, and process them in real time. This allows for the acquisition of EL defect distribution maps of each photovoltaic module 500 without power interruption or wiring.

[0056] Figure 3 A flowchart illustrating a photovoltaic module electroluminescence image acquisition method provided in this application embodiment is shown below. Figure 3 In some embodiments, the photovoltaic module electroluminescence image acquisition method, based on the aforementioned photovoltaic module electroluminescence image acquisition system, includes the following steps: S301, Acquire an infrared image set of the photovoltaic module; wherein the infrared image set includes a first infrared image acquired during a first polarization direction under the same field of view, and a second infrared image acquired during a second polarization direction.

[0057] Due to abundant sunlight, each cell inside the photovoltaic module generates extremely low-intensity radiative recombination emission at its junction temperature. This emission is bandgap radiation from silicon, with wavelengths mainly concentrated around 1100-1200 nm, but its intensity is far lower than the background radiation caused by sunlight reflection and scattering. To extract this weak signal, the image processing unit controls the liquid crystal polarization modulator to alternately switch between two states at a certain frequency. Simultaneously, an infrared camera synchronously acquires infrared images, simultaneously exposing and obtaining image frames under corresponding polarization states A (first polarization direction) and B (second polarization direction). For example, switching every 20 ms, in polarization state A, the liquid crystal polarization modulator makes the incident light polarized horizontally, and in polarization state B, the polarization direction is vertical.

[0058] Specifically, infrared images are acquired through infrared cameras.

[0059] Specifically, the first polarization direction and the second polarization direction are switched alternately by a liquid crystal polarization modulator.

[0060] S302, perform differential processing on the first infrared image and the second infrared image to obtain the electroluminescent image of the photovoltaic module.

[0061] In this process, after extracting multiple frames of images under different polarization states, differential operations are performed on the infrared images under different polarization states. This can eliminate most of the background components that do not change with polarization, highlight the luminous signals caused by the difference in polarization transmittance, and thus form a preliminary EL image.

[0062] Because the reflection of solar glare by the glass surface of photovoltaic modules and the metal grid lines of the cells exhibits polarization characteristics, while the spontaneous emission from within the cells shows no significant preference in any polarization direction, by selecting an appropriate polarization orientation, the intensity of background reflected light can be made similar in two frames, while the emission signal of the module itself will show subtle differences between the two frames. For example, polarization state A (first polarization direction) may cause some surface reflections of the module to be suppressed by the polarizer, while polarization state B (second polarization direction) has the opposite effect, causing the background component to increase or decrease in one frame. In contrast, the EL emission from within the module is weaker and unpolarized, and the difference in transmittance between the two frames is mainly caused by changes in the transmittance of the polarizer. By comparing these two frames, the image processing unit can attribute the differences to the emission signal of the module itself.

[0063] Optionally, the image difference method obtains a difference image by subtracting the pixels of two infrared images obtained under polarization state A and polarization state B one by one. Since the background light is approximately the same in both frames, the difference will suppress the background and highlight the luminescence signal caused by the difference in polarization transmittance, thus forming a preliminary electroluminescent image.

[0064] Then, the difference image D = Image A - Image B is calculated. In the difference result, if a pixel belongs to a defective area of ​​the component, because the solar cell in that area emits less light, it will show a negative dark difference value after subtracting the reflected background under the two polarization states. Conversely, the intact area, due to its strong light emission, may show a positive bright difference value in the difference, thus highlighting the defect. For a globally uniform solar background, since the values ​​are approximately equal in both frames, they cancel each other out after the difference, greatly reducing background noise.

[0065] Specifically, the differential processing of the first infrared image and the second infrared image is completed by the image processing unit.

[0066] In this embodiment, a method for acquiring electroluminescence (EL) images of photovoltaic (PV) modules under natural sunlight conditions is provided without energizing the PV module under test. By introducing a liquid crystal polarization modulator into the optical path of an infrared camera, the polarization state of the light entering the infrared camera is alternately changed, and at least two images of the same field of view are acquired. The infrared images under different polarization states are then compared to extract the spontaneous emission signal of the PV module related to the polarization change, reconstructing an EL image reflecting internal defects of the PV module. This method utilizes optical modulation to achieve EL signal separation, and the PV module maintains its natural operating state throughout the process, requiring no external power source.

[0067] This research pioneers a novel approach to extracting EL signals using optical modulation, achieving complete separation from the traditional electrical excitation + image processing path. Utilizing at least two frames of infrared images under different polarization states, the weak EL emission signal component of the photovoltaic (PV) module is extracted. By extracting extremely weak signals through polarization difference, the EL emission signal of the PV module itself can be extracted from a distance under strong daylight, enabling rapid, non-contact detection of internal defects in PV modules. This yields clear EL images and offers advantages such as high detection accuracy and strong anti-interference capabilities. It overcomes many limitations of existing EL detection methods in field applications, enabling rapid, non-invasive diagnosis of hidden defects within PV modules. It is particularly suitable for scenarios where access to power is inconvenient, such as drone inspection of large PV arrays, inspection of building-integrated photovoltaic (BIPV) panels, and on-orbit monitoring of space solar cell arrays, demonstrating broad application prospects.

[0068] exist Figure 3 Based on the embodiments shown, the following is combined with Figure 4 The technical solution of the above-mentioned electroluminescence image acquisition method for photovoltaic modules will be further introduced.

[0069] Figure 4 A flowchart illustrating another photovoltaic module electroluminescence image acquisition method provided in this application embodiment is shown below. Figure 4 In some embodiments, the process of the photovoltaic module electroluminescence image acquisition method includes the following steps: S401, Acquire an infrared image set of the photovoltaic module; wherein the infrared image set includes a first infrared image acquired during a first polarization direction under the same field of view, and a second infrared image acquired during a second polarization direction.

[0070] S402, perform registration and grayscale normalization processing on the first infrared image and the second infrared image.

[0071] Specifically, the infrared images of polarization state A (first polarization method) and state B (second polarization method) are registered and normalized to ensure that the two images can be subtracted pixel by pixel.

[0072] Specifically, registration is a crucial step in image processing and analysis, aiming to spatially align multiple images acquired at different times, from different viewpoints, or from different sensors, ensuring they are geometrically consistent. Its core principle involves finding similarity metrics between images (such as mutual information or cross-correlation) and combining them with specific transformation models (such as rigid transformation, affine transformation, or nonlinear transformation) to perform coordinate transformations on one or more images, thereby eliminating differences in translation, rotation, scaling, or deformation between the images.

[0073] Specifically, grayscale normalization is a technique for adjusting the distribution range of grayscale values ​​in an image. Its purpose is to eliminate inconsistencies in grayscale levels caused by differences in imaging conditions (such as light intensity and sensor sensitivity) or image sources, thus giving the image a uniform grayscale scale. Its core methods include linear normalization and nonlinear normalization.

[0074] S403, input the first infrared image and the second infrared image into the pre-constructed differential extraction model to obtain the electroluminescent image output by the differential extraction model; wherein, the differential extraction model is obtained based on the training of a deep learning model.

[0075] The deep learning approach involves taking the original infrared images in polarization state A (first polarization method) and state B (second polarization method) as input and feeding them into a pre-trained deep learning model. The model then outputs a predicted EL defect image based on the differences in image features before and after the polarization switch. This method can more robustly extract extremely weak EL signals and reduce the effects of noise and uneven illumination.

[0076] An EL signal is extracted by introducing a trained deep learning model. The model's training samples are EL images taken at night with the camera powered on from several photovoltaic modules with typical defects, used as labels. These images are compared with infrared images acquired during the day using the polarization method described in this embodiment, and the model is trained in a supervised manner to improve detection reliability. After training, in field applications, the model can automatically learn the subtle differences between the first and second infrared images, filtering out background noise and retaining only the EL feature signal. In this process, no manual adjustment of the differential threshold is required, and the deep learning model has stronger adaptability to complex lighting changes. The introduction of the deep learning model helps improve robustness to interference from noise and non-uniform lighting, and the obtained defect information is comparable to the results obtained under laboratory conditions with the camera powered on.

[0077] In addition to pixel differencing and deep learning extraction, image processing can also employ phase-locked loop (PLL) accumulation methods with more frames to improve the signal-to-noise ratio.

[0078] S404 performs image enhancement processing on the electroluminescent image.

[0079] After obtaining the EL image, necessary algorithmic processing is applied, such as filtering, denoising, and histogram equalization, to improve the signal-to-noise ratio and clarity. For example, adaptive filtering can be used to remove isolated noise points, and logarithmic enhancement can be used to highlight dark details. The final output image is the EL image reflecting defects in the photovoltaic module, making defects such as microcracks, fragments, and failed solder joints in the cells clearly visible.

[0080] The extracted EL images undergo necessary image enhancement and analysis. Enhancement algorithms such as denoising and histogram equalization can be applied to the differentially derived EL images to improve the signal-to-noise ratio and clarity. Then, based on features such as dark spots and crack patterns displayed in the EL images, the type and location of defects within the photovoltaic module can be determined. The reconstructed EL images are consistent with those obtained through traditional power-on methods and can be used to assess the module's solder joint connection status, microcrack distribution, and cell failure areas.

[0081] S405, based on electroluminescent images, determines the defect detection results of photovoltaic modules.

[0082] Among these methods, the defect detection results of photovoltaic modules are identified using an electroluminescence model, which can be performed manually or through artificial intelligence. Finally, based on features such as dark spots and crack patterns displayed in the EL image, the type and location of defects within the photovoltaic module are determined.

[0083] Specifically, in defect recognition, a defect detection and classification model can be trained. A convolutional neural network is constructed to analyze the input electroluminescent image and output the identified defect type and its location in the image.

[0084] For the defect detection and classification model, it can be trained using a dataset of electroluminescent images labeled with defects, so that the model can learn to distinguish defect patterns from the distribution of luminescence intensity.

[0085] In this embodiment, the differential extraction model built using a deep learning model demonstrates, through actual testing, that it extracts defects with higher contrast, detects even fine cracks, and exhibits better robustness to changes in external conditions. This deep learning-enhanced approach provides reliable technical support for large-scale automated inspection. Traditional image differential methods perform pixel-level differential operations on images acquired under two polarization states to cancel out most of the constant background light, obtaining the weak EL emission distribution of the components from the differential results. Furthermore, by inputting polarized infrared images into a pre-trained deep learning model, the model automatically separates the EL signal components and outputs an electroluminescent image, effectively improving the signal-to-noise ratio and stability of EL signal extraction under strong light backgrounds.

[0086] Figure 5This is a flowchart illustrating a defect detection method for photovoltaic modules provided in an embodiment of this application. (See attached diagram.) Figure 5 In some embodiments, the defect detection method for the photovoltaic module includes the following steps: S501, acquire an electroluminescence image; wherein, the electroluminescence image is obtained based on the electroluminescence image acquisition method of the voltage module.

[0087] S502, based on electroluminescent images, obtains defect detection results for photovoltaic modules.

[0088] Optionally, the electroluminescent image is input into a pre-built defect classification and detection model to obtain the defect detection results output by the defect classification and detection model; wherein, the defect classification and detection model is trained based on a deep learning model, and the defect detection results include defect location, defect type, and damage degree.

[0089] In terms of defect identification, a defect detection and classification model can be trained. A convolutional neural network is constructed to analyze the input electroluminescent image, outputting the identified defect type, its location in the image, and the extent of damage. This reduces reliance on human experience and improves the digitalization level of operations and maintenance. Compared to traditional manual screening and analysis methods, it significantly reduces the workload of defect detection.

[0090] Specifically, for defect detection and classification models, deep learning models can be trained using a dataset of electroluminescent images labeled with defects, enabling the deep learning models to learn to identify defect patterns from the distribution of luminescence intensity.

[0091] Specifically, the deep learning model is preferably an image processing network such as a multiplicative neural network (CNN), YOLO, or U-Net; this defect classification and detection model can automatically extract spatial morphological features from the image and classify and accurately locate the defects based on their topological structure (hidden cracks, broken grids, black cores, process contamination, etc.).

[0092] In this embodiment, daytime detection of hidden defects in photovoltaic modules can be achieved, greatly improving the convenience and timeliness of defect detection in photovoltaic modules under daytime conditions.

[0093] In the description of this application, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "clockwise", "counterclockwise", "axial", "radial", "circumferential", etc., indicating the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this application.

[0094] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0095] In this application, unless otherwise expressly specified and limited, the terms "installation," "connection," "joining," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components, unless otherwise expressly limited. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.

[0096] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0097] In the above embodiments, the descriptions of each embodiment have their own emphasis. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments. The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as the combination of these technical features does not contradict each other, it should be considered within the scope of this specification.

[0098] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this application are indicated by the appended claims.

[0099] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this application is limited only by the appended claims.

Claims

1. A photovoltaic module electroluminescence image acquisition system, characterized in that, include: A liquid crystal polarization modulator is disposed in front of the lens of an infrared camera to change the polarization direction of incident light entering the lens of the infrared camera; wherein the polarization direction includes an alternately switching first polarization direction and a second polarization direction. An infrared camera is used to photograph the photovoltaic module under test to acquire an infrared image set of the photovoltaic module; wherein the infrared image set includes a first infrared image acquired during a first polarization direction under the same field of view, and a second infrared image acquired during a second polarization direction; An image processing unit is used to perform differential processing on the first infrared image and the second infrared image to obtain the electroluminescent image of the photovoltaic module.

2. The photovoltaic module electroluminescence image acquisition system according to claim 1, characterized in that, A narrowband filter is disposed between the lens of the infrared camera and the liquid crystal polarization modulator.

3. The photovoltaic module electroluminescence image acquisition system according to claim 1, characterized in that, The liquid crystal polarization modulator is a liquid crystal shutter or a liquid crystal variable waveplate.

4. The photovoltaic module electroluminescence image acquisition system according to claim 1, characterized in that, The infrared camera is a near-infrared camera.

5. The photovoltaic module electroluminescence image acquisition system according to any one of claims 1-4, characterized in that, It also includes a mobile platform for mounting the liquid crystal polarization modulator, the infrared camera, and the image processing unit.

6. The photovoltaic module electroluminescence image acquisition system according to claim 5, characterized in that, The mobile platform is either a ground vehicle or a drone.

7. A method for acquiring electroluminescence images of photovoltaic modules, characterized in that, The photovoltaic module electroluminescence image acquisition system based on any one of claims 1-6 includes: The infrared image set of the photovoltaic module is acquired; wherein the infrared image set includes a first infrared image acquired during a first polarization direction under the same field of view, and a second infrared image acquired during a second polarization direction; Differential processing is performed on the first infrared image and the second infrared image to obtain the electroluminescence image of the photovoltaic module.

8. The photovoltaic module electroluminescence image acquisition method according to claim 7, characterized in that, Differential processing is performed on the first infrared image and the second infrared image to obtain the electroluminescence image of the photovoltaic module, including: The first infrared image and the second infrared image are input into a pre-constructed differential extraction model to obtain an electroluminescent image output by the differential extraction model; wherein the differential extraction model is obtained based on a deep learning model trained on it.

9. The photovoltaic module electroluminescence image acquisition method according to claim 7, characterized in that, After acquiring the infrared image set of the photovoltaic module, the method further includes: The first infrared image and the second infrared image are registered and normalized.

10. A method for defect detection in photovoltaic modules, characterized in that, include: Acquire an electroluminescence image; wherein the electroluminescence image is obtained based on the photovoltaic module electroluminescence image acquisition method as described in any one of claims 7-9; Based on the electroluminescent image, the defect detection results of the photovoltaic module are obtained.

Images