A method for detecting silicon-based photovoltaic modules under sunlight based on a short-wave infrared camera

Abstract

The application discloses a kind of short-wave infrared camera-based photovoltaic module detection method under sunlight.The method comprises: infrared detection is carried out to photovoltaic module to obtain infrared image;Abnormality detection is carried out to obtain the infrared image of abnormal heating area;Positioning and taking abnormal heating area as the region to be detected;Photovoltaic module detection equipment is arranged at the region to be detected, and two-stage shooting mode is used to obtain abnormal heating area image;After using image detection algorithm, defect detection result is obtained.The application can realize the detection and positioning of internal defects of photovoltaic module, help to judge failure cause, realize targeted operation and maintenance, and ensure that photovoltaic module is in the best power generation state;The application is non-contact defect detection, without any electrical contact, can be carried out under the working state of photovoltaic module, has the advantages of simple operation, no additional damage, and does not need to be disassembled and stopped, and can be detected under sunlight, with clear imaging, and can detect electrical faults such as small area circuit breakage.

Description

Technical Field

[0001] This invention relates to a method for detecting silicon-based photovoltaic modules under sunlight, which relates to the field of photovoltaic technology, and specifically to a method for detecting silicon-based photovoltaic modules under sunlight based on a short-wave infrared camera. Background Technology

[0002] In response to environmental pollution and energy shortages caused by the overuse of traditional energy sources, the development of clean energy is being vigorously promoted. Due to the safety, cleanliness, and renewability of solar energy, the solar photovoltaic industry is booming. Photovoltaic modules are the core components of the solar photovoltaic industry; even minor defects can significantly affect their power generation efficiency and lifespan, and may even cause fires. However, photovoltaic module inspection faces challenges such as long working hours and the risk of secondary damage during disassembly. Therefore, regular inspection and maintenance of photovoltaic modules during operation is of paramount importance.

[0003] Photovoltaic module inspection solutions mainly include visible light inspection, infrared thermal imaging inspection, and luminescent imaging inspection. Visible light inspection often uses drones for imaging, offering advantages such as high resolution and a wide inspection area, but it has a very limited range of detectable defect types. Infrared thermal imaging inspection utilizes the abnormal heating phenomenon in defective areas of photovoltaic modules to detect defect locations, but it cannot obtain specific information such as defect type and morphology. Luminescent imaging inspection can be divided into electroluminescence and photoluminescence inspection, requiring an external voltage or specific wavelength of light to excite the photovoltaic module and acquire luminescent images, enabling the detection of invisible defects such as microcracks and black spots.

[0004] Currently, photovoltaic module inspection in operation primarily employs a combination of visible light and infrared thermal imaging, which offers high inspection efficiency but cannot determine the cause and morphology of defects. Traditional luminescent imaging inspection technology requires a dark environment, limiting its application conditions and failing to adequately realize luminescent imaging inspection technology for photovoltaic modules operating under sunlight. Summary of the Invention

[0005] To address the problems existing in the background art, this invention provides a method for detecting silicon-based photovoltaic modules under sunlight based on a short-wave infrared camera. The method utilizes short-wave infrared emission imaging detection technology to detect invisible internal defects in silicon-based photovoltaic modules operating under sunlight.

[0006] The technical solution adopted in this invention is:

[0007] The present invention provides a method for detecting silicon-based photovoltaic modules under sunlight based on a short-wave infrared camera, comprising:

[0008] The initial step involves preparing the necessary testing equipment, including handheld, desktop, or drone-mounted infrared detection devices, as well as the complete set of non-contact testing equipment from this invention. After confirming there are no abnormalities, subsequent testing steps can begin. If the photovoltaic modules are distributed over a small area with low density, handheld or desktop infrared detection devices are used for testing. If the photovoltaic modules are distributed over a wide area with high density, drone-mounted infrared detection devices are used for testing.

[0009] S1. Under sunlight, perform ground or drone infrared detection on several photovoltaic modules in the area to be maintained, which are in operation and connected in series through photovoltaic circuits, to obtain infrared images of each photovoltaic module.

[0010] S2. Perform anomaly detection on the infrared images of each photovoltaic module to obtain the infrared images of photovoltaic modules with abnormal heating areas; locate each photovoltaic module with abnormal heating areas and use the abnormal heating areas in the infrared images of the photovoltaic modules as the areas to be detected.

[0011] If no abnormalities are found, repeat steps S1-S2 to check the remaining photovoltaic modules in turn; if no abnormalities are found in all photovoltaic modules, it is considered that there are no abnormal defects in the photovoltaic modules in the area that could cause overheating, and the operation and maintenance of the area is completed.

[0012] S3. For each photovoltaic module with an abnormal heating area, the photovoltaic module testing equipment is deployed at the testing area of ​​the photovoltaic module. The photovoltaic module testing equipment uses a two-stage shooting method to obtain short-wave infrared emission images of the testing area.

[0013] S4. The photovoltaic module inspection equipment uses image detection algorithms to process the short-wave infrared emission images of the two-stage inspection area to obtain defect detection results, thus completing the inspection of photovoltaic modules under sunlight.

[0014] In step S1, for each photovoltaic module in the area to be maintained, an infrared image of the complete surface of the photovoltaic module under sunlight is captured using a handheld or desktop infrared detection device deployed on the ground or an infrared detection device mounted on a drone. The infrared detection device is specifically an infrared thermal imaging camera.

[0015] For photovoltaic module maintenance areas with small area and low distribution density, use handheld or desktop infrared detection equipment. Place the equipment at an appropriate distance, ideally perpendicular to the surface of the photovoltaic module, so that it can clearly capture the entire photovoltaic module without causing shading.

[0016] For photovoltaic module maintenance areas with large areas and high distribution density, drone flight routes are planned and formulated based on the photovoltaic module distribution plan to ensure that the inspection route covers all modules. The drone is equipped with an infrared thermal imaging camera for aerial photography, capturing a complete front view of the photovoltaic module in operation when flying directly above it.

[0017] In step S1, the photovoltaic module includes several solar cell units arranged in a rectangular array. Each solar cell unit is divided into groups, and each group of solar cell units is connected in series with each other. Each group of series-connected units is connected in parallel with each other. The area to be detected includes one or more solar cell units that are abnormally overheating.

[0018] In step S2, for each photovoltaic module's infrared image, during anomaly detection, if the temperature value of a certain area in the infrared image exceeds the temperature threshold, then that area is considered an abnormal heating area.

[0019] In step S3, the photovoltaic module testing equipment includes an image acquisition device, a testing shading cover, a main control host computer, and a control shading cover. The image acquisition device is installed at the center of the testing shading cover and faces the opening of the testing shading cover. The image acquisition device is electrically connected to the main control host computer equipped with an image detection algorithm.

[0020] Two-stage image acquisition is performed using photovoltaic module inspection equipment, targeting the inspection area of ​​each photovoltaic module, as detailed below:

[0021] During the first stage of filming, the photovoltaic module is operating normally. The detection shading cover is placed over each cell unit in the area to be tested. At this time, the image acquisition device faces each cell unit. The main control computer controls the image acquisition device to continuously capture a set of n images of abnormal heating areas in the area to be tested and transmits them to the main control computer. During the second stage of filming, the detection shading cover is always placed over the area to be tested. The control shading cover is placed over one of the cell units connected in series with each cell unit in the area to be tested in other areas of the photovoltaic module. The main control computer then controls the image acquisition device to continuously capture a set of n images of abnormal heating areas in the area to be tested and transmits them to the main control computer.

[0022] The detection light shield includes a fixed horizontal support, four-way adjustable connectors, a length-adjustable horizontal support, a length-adjustable vertical support, four-way adjustable three-way connectors, and a light-shielding cloth. The fixed horizontal support includes four short connecting rods and two long connecting rods. The four short connecting rods are connected to form a rectangular structure by the four-way adjustable connectors. The two long connecting rods are cross-connected to the four-way adjustable connectors. A camera mounting hole is provided at the intersection of the two long connecting rods. The image acquisition device is installed at the camera mounting hole to ensure that the detection light shield can be completely and clearly captured. Image of photovoltaic modules inside the photomask; The length-adjustable longitudinal support includes four first-length-adjustable connecting rods, one end of which is connected to four-way adjustable connectors, and the other end of which is connected to four-way adjustable three-way connectors. The length-adjustable transverse support includes four second-length-adjustable connecting rods, which are connected by four-way adjustable three-way connectors to form a rectangular structure. The length-adjustable transverse support is parallel to the fixed transverse support and parallel to the surface of the photovoltaic module.

[0023] The length-adjustable connecting rod includes an externally threaded connecting rod and an internally threaded connecting rod. The externally threaded connecting rod is threadedly fitted into the internally threaded connecting rod. The length of the length-adjustable connecting rod is adjusted by rotating the threads at the threaded connection between the externally and internally threaded connecting rods. By adjusting the lengths of the length-adjustable horizontal support and the length-adjustable vertical support, the detection light shield can be deformed into an inverted frustum, rectangular, and frustum shape with varying heights, thereby adjusting the area of ​​the length-adjustable horizontal support to change the occlusion area of ​​the detection light shield.

[0024] The light-shielding cloth is placed over a fixed horizontal support and an adjustable vertical support, so that the rectangular opening formed by the adjustable horizontal support serves as the opening of the light-shielding cover, and the image acquisition device faces the center of the adjustable horizontal support.

[0025] The inspection shield consists of a cubic frame and a light-shielding cloth. The top of the cubic frame is open to accommodate a short-wave infrared camera; the bottom is open, completely covering the area of ​​the photovoltaic module to be inspected, with all four sides in contact with the surface of the photovoltaic module; the top and four sides are covered with the light-shielding cloth. The cubic frame is composed of adjustable telescopic rods, and the dimensions of each side of the bottom can be dynamically adjusted and fixed within a range of 15cm to 25cm. The side lengths should be adjusted to ensure that the camera can clearly capture images of the photovoltaic module.

[0026] In operation, the photovoltaic module is in a conductive state, with current flowing through the cell units. In the first stage, the photovoltaic module operates normally, and the detection shading cover is positioned over the area to be detected, leaving the remaining surfaces unobstructed. At this stage, the operating current flowing through the cells is relatively high, and the luminous intensity of the target area is relatively weak. In the second stage, the shading cover is used to completely block a single cell unit in the non-detection area of ​​the photovoltaic module, significantly reducing the operating current flowing through the cell and enhancing the luminous intensity of the target area. This invention employs a two-stage imaging method to maximize the intensity of photocarriers within the blocked area and remove interference from sunlight and background light escaping from the edge of the detection shading cover; the shading cover's sunlight filtering effect must reach at least 99.9%.

[0027] The image acquisition device uses a Bobcat 640 short-wave infrared camera paired with a Kowa short-wave infrared lens. At a camera working distance of 20cm, a single pixel represents an actual distance of 0.25mm. Based on the luminescence characteristics of silicon-based photovoltaic modules, whose emission spectrum has a center wavelength of 1150nm, a 1100-1200nm narrow-bandpass filter is fitted in front of the short-wave infrared camera lens to effectively acquire images of the photovoltaic module's emission.

[0028] During the first stage of shooting, a detection light shield is placed over each battery cell unit in the area to be tested. The size of the shading area of ​​the detection light shield is adjusted according to the size of the battery cell units in the area to be tested, so that the area covered by each shading battery cell unit does not exceed 2 / 3 of the battery cell unit's own area, that is, at least 1 / 3 or more of the unit is exposed to sunlight. When there are a large number of battery cell units in the area to be tested, a multiple acquisition method is adopted. In N moving shots, several battery cell units in the area to be tested are blocked by the detection light shield each time, until all battery cell units in the abnormal heat-generating areas have been blocked and images have been captured.

[0029] During the second stage of shooting, the series group of the battery cell unit that is blocked in the area to be tested is determined. Any battery cell unit in the non-test area is selected from the series group. The selected non-test area battery cell unit is completely blocked by the control light shield before the second stage of shooting is carried out.

[0030] In step S4, when only one set of n abnormal heating area images are captured in both the first and second stages, the n abnormal heating area images captured in the first stage and the n abnormal heating area images captured in the second stage are input into the host computer for processing using an image detection algorithm. First, preprocessing is performed: the n abnormal heating area images captured in the first stage are averaged at the pixel level to obtain a single first-stage image; then, the n abnormal heating area images captured in the second stage are averaged at the pixel level to obtain a single second-stage image, which can suppress random noise. The second-stage image and the first-stage image are subtracted at the pixel level to obtain a relatively pure target area luminous image, which can effectively remove interference from the background and sunlight, etc. Then, defect detection is performed on the target area luminous image to finally obtain the defect detection result.

[0031] When N sets of abnormal heating area images are captured in the first stage, N sets of abnormal heating area images are captured in the second stage. When both the first and second stages capture N sets of n abnormal heating area images, the n abnormal heating area images in each set are processed by the same image detection algorithm as described above, and then stitched together according to the position of the captured battery cell unit to finally obtain a complete abnormal heating area image detection result.

[0032] Output the detection image and results, and output and save the detection report. The report includes the mean image obtained from the two stages, a clean luminescent image of the test area, the defect detection algorithm recognition result image, and defect-related information such as defect type and size.

[0033] Conclusion: Complete the output and storage of the inspection report, determine the defect type and location of the abnormal heating area of ​​the photovoltaic module, speculate on its cause, and then determine the treatment method and improvement measures to ultimately achieve the goal of cost reduction and efficiency improvement.

[0034] The defect detection specifically employs morphological processing or the U-Net deep learning algorithm. Relevant parameters, such as thresholds, have been preset to predefined values, and an interface for modification is provided.

[0035] The beneficial effects of this invention are:

[0036] 1) This invention addresses the problem of difficulty in quickly and accurately detecting invisible defects in photovoltaic modules under outdoor sunlight, and develops a highly efficient and targeted defect detection solution. By using various types of infrared thermal imaging devices to locate abnormally heated areas, and then deploying the detection device of this invention accordingly, the detection device can detect and locate internal defects in the photovoltaic module, helping to determine the cause of failure and enabling targeted operation and maintenance. This ensures that the photovoltaic module is in optimal power generation condition, effectively guaranteeing economic benefits.

[0037] 2) This invention is a non-contact defect detection solution that requires no electrical contact and can be performed while the photovoltaic module is in operation. Compared to existing electroluminescence detection, this detection solution does not require an external power supply and has advantages such as simple operation, no additional damage, and no need for disassembly or shutdown; compared to existing photoluminescence detection, this detection solution can detect under sunlight, provides clear imaging, and can detect electrical faults such as small-area open circuits. Attached Figure Description

[0038] Figure 1 This is a flowchart of the method of the present invention;

[0039] Figure 2 This is a schematic diagram illustrating the detection principle of the method of the present invention, wherein, Figure 2 (a) is a schematic diagram of the principle of photoexcitation of photocarriers by light. Figure 2 (b) is a schematic diagram of the optical carrier diffusion principle. Figure 2 (c) is a schematic diagram of the principle of light emission by photocarriers;

[0040] Figure 3 This is the emission spectrum of the silicon-based photovoltaic module of the present invention;

[0041] Figure 4 This is a side view of the detection device of the present invention;

[0042] Figure 5 This is a schematic diagram of the two working stages of the method of the present invention, wherein, Figure 5 (a) is a schematic diagram of the first stage. Figure 5 (b) is a schematic diagram of the second stage;

[0043] Figure 6 This is a structural diagram of the light shield frame of the present invention;

[0044] Figure 7 This is an installation diagram of the adjustable horizontal support for the light shield of the present invention, wherein, Figure 7 (a) is a schematic diagram of the adjustable horizontal bracket for the sunshade before installation. Figure 7 (b) is a schematic diagram of the adjustable horizontal support bracket after installation.

[0045] Figure 8 This is a diagram showing the detection results of the method of the present invention, wherein... Figure 8 Image (a) shows the results of the first phase of testing. Figure 8 (b) shows the results of the second phase of testing. Figure 8 (c) is the difference image;

[0046] In the diagram: 1. Sun, 2. Photovoltaic module, 3. Area to be tested, 4. Image acquisition equipment, 5. Detection shade, 501. Camera mounting hole, 502. Fixed horizontal support, 503. Four-way adjustable connection, 504. Horizontal support with adjustable length, 5041. External threaded connecting rod, 5042. Internal threaded connecting rod, 5043. Threaded connection, 505. Vertical support with adjustable length, 506. Three-way adjustable connection, 6. Main control computer, 7. Photovoltaic circuit in operation, 8. Control shade. Detailed Implementation

[0047] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0048] like Figure 1 As shown, the present invention provides a method for detecting silicon-based photovoltaic modules under sunlight based on a short-wave infrared camera, comprising:

[0049] The initial step involves preparing the necessary testing equipment, including handheld, desktop, or drone-mounted infrared detection devices, as well as the complete set of non-contact testing equipment from this invention. After confirming there are no abnormalities, subsequent testing steps can begin. If the photovoltaic modules are distributed over a small area with low density, handheld or desktop infrared detection devices are used for testing. If the photovoltaic modules are distributed over a wide area with high density, drone-mounted infrared detection devices are used for testing.

[0050] S1. Under the illumination of the sun 1, perform ground or drone infrared detection on several photovoltaic modules 2 in the area to be maintained, which are in working condition through the photovoltaic circuit 7 in series, and obtain infrared images of each photovoltaic module 2.

[0051] In step S1, for each photovoltaic module 2 in the area to be maintained, an infrared image of the complete surface of the photovoltaic module 2 under sunlight is captured using a handheld or desktop infrared detection device deployed on the ground or an infrared detection device mounted on a drone. The infrared detection device is specifically an infrared thermal imaging camera.

[0052] For photovoltaic module maintenance areas with small area and low distribution density, use handheld or desktop infrared detection equipment. Place the equipment at an appropriate distance, ideally perpendicular to the surface of the photovoltaic module, so that it can clearly capture the entire photovoltaic module without causing shading.

[0053] For photovoltaic module maintenance areas with large areas and high distribution density, drone flight routes are planned and formulated based on the photovoltaic module distribution plan to ensure that the inspection route covers all modules. The drone is equipped with an infrared thermal imaging camera for aerial photography, capturing a complete front view of the photovoltaic module in operation when flying directly above it.

[0054] In step S1, the photovoltaic module 2 includes several solar cell units arranged in a rectangular array. Each solar cell unit is divided into groups, and each group of solar cell units is connected in series with each other. Each group of series-connected units is connected in parallel with each other. The area to be tested 3 includes one or more solar cell units that are abnormally overheating.

[0055] S2. Perform anomaly detection on the infrared images of each photovoltaic module 2 to obtain the infrared images of photovoltaic modules 2 with abnormal heating areas; locate each photovoltaic module 2 with abnormal heating areas and take the abnormal heating areas in the infrared images of photovoltaic modules 2 as the areas to be detected 3.

[0056] If no abnormalities are found, repeat steps S1-S2 to check the remaining photovoltaic modules in turn; if no abnormalities are found in all photovoltaic modules, it is considered that there are no abnormal defects in the photovoltaic modules in the area that could cause overheating, and the operation and maintenance of the area is completed.

[0057] In step S2, for each photovoltaic module 2, during anomaly detection, if the temperature value of a certain area in the infrared image exceeds the temperature threshold, then that area is considered an abnormal heating area.

[0058] S3. For each photovoltaic module 2 with an abnormal heating area, the photovoltaic module testing equipment is deployed at the testing area 3 of the photovoltaic module 2, such as... Figure 4 As shown, a short-wave infrared image of the area to be tested at location 3 is obtained using a two-stage imaging method with photovoltaic module testing equipment.

[0059] In step S3, the photovoltaic module testing equipment includes an image acquisition device 4, a testing shading cover 5, a main control host computer 6, and a control shading cover 8. The image acquisition device 4 is installed at the center of the testing shading cover 5 and faces the opening direction of the testing shading cover 5. The image acquisition device 4 is electrically connected to the main control host computer 6, which is equipped with an image detection algorithm.

[0060] Two-stage image acquisition is performed using photovoltaic module inspection equipment, specifically for the inspection area 3 of each photovoltaic module 2, as follows:

[0061] During the first stage of filming, the photovoltaic module 2 operates normally, and the detection shading cover 5 is used to cover each cell unit in the area to be tested 3. At this time, the image acquisition device 4 is oriented towards each cell unit, and the main control host computer 6 controls the image acquisition device 4 to continuously capture a set of n abnormal heating area images of the area to be tested 3 and transmit them to the main control host computer 6. During the second stage of filming, the detection shading cover 5 is always used to cover the area to be tested 3, and the control shading cover 8 is used to cover one of the cell units connected in series with each cell unit in the area to be tested 3 in other areas of the photovoltaic module 2. Then, the main control host computer 6 controls the image acquisition device 4 to continuously capture a set of n abnormal heating area images of the area to be tested 3 and transmit them to the main control host computer 6. The value of n is between 30 and 100.

[0062] The light emission detection principle of the silicon-based photovoltaic module involved in this invention is as follows: Figure 2 a, Figure 2 b and Figure 2 As shown in Figure c, this is a schematic diagram of the light-emitting principle. When the solar cell units within photovoltaic module 2 are illuminated, each unit is excited to generate photocarriers. During their transition back to the ground state, they emit light of a fixed wavelength, with a central wavelength of 1150 nm. Figure 3 As shown. If the illumination in a single solar cell is uneven or some areas are not illuminated, the photocarriers generated in the illuminated areas will quickly diffuse to the unilluminated areas and eventually be evenly distributed throughout the entire solar cell.

[0063] This invention utilizes the principle that photocarriers can diffuse within a solar cell unit. A designed detection device blocks the portion of the solar cell unit to be tested, ensuring the light intensity in the blocked area is less than 1 / 1000th of the intensity in the unblocked area. The unblocked area, excited by sunlight, generates a high concentration of photocarriers, which immediately diffuse and become uniformly distributed throughout the blocked area of ​​the solar cell unit under test. Within the blocked area, a considerable intensity of photocarriers undergoes a ground-state transition, emitting light with a center wavelength of 1150 nm, with almost no interference from other light sources. A short-wave infrared camera captures the light in the 1100-1200 nm wavelength range, obtaining an image of the emitted light from the blocked area.

[0064] like Figure 6As shown, the detection light shield 5 includes a fixed horizontal support 502, a four-way adjustable four-way connector 503, a length adjustable horizontal support 504, a length adjustable vertical support 505, a four-way adjustable three-way connector 506, and a light-shielding cloth. The fixed horizontal support 502 includes four short connecting rods and two long connecting rods. The four short connecting rods are connected by the four-way adjustable four-way connector 503 to form a rectangular structure. The two long connecting rods are cross-connected to the four-way adjustable four-way connector 503. A camera mounting hole 501 is provided at the intersection of the two long connecting rods. The image acquisition device 4 is installed at the camera mounting hole 501 to ensure complete and clear imaging. Image of photovoltaic module 2 inside the light shield 5 is obtained; the length adjustable longitudinal support 505 includes four first length adjustable connecting rods, one end of each of the four first length adjustable connecting rods is connected to four directional adjustable four-way connectors 503, and the other end of each of the four first length adjustable connecting rods is connected to four directional adjustable three-way connectors 506; the length adjustable transverse support 504 includes four second length adjustable connecting rods, the four second length adjustable connecting rods are connected by four directional adjustable three-way connectors 506 to form a rectangular structure; the length adjustable transverse support 504 is parallel to the fixed transverse support 502 and parallel to the surface of photovoltaic module 2.

[0065] like Figure 7 a and Figure 7 As shown in b, the length-adjustable connecting rod includes an externally threaded connecting rod 5041 and an internally threaded connecting rod 5042. The externally threaded connecting rod 5041 is threadedly fitted into the internally threaded connecting rod 5042. The length of the length-adjustable connecting rod is adjusted by rotating the thread at the threaded connection 5043 of the externally threaded connecting rod 5041 and the internally threaded connecting rod 5042. By adjusting the length of the length-adjustable transverse support 504 and the length-adjustable longitudinal support 505, the detection light shield 5 is deformed into an inverted frustum, a rectangle, and a frustum shape with varying heights, thereby adjusting the area of ​​the length-adjustable transverse support 504 to change the blocking area of ​​the detection light shield 5.

[0066] The light-shielding cloth is completely covered on the fixed horizontal support 502 and the length-adjustable vertical support 505, so that the rectangular opening formed by the length-adjustable horizontal support 504 serves as the opening of the detection light-shielding cover 5, and the image acquisition device 4 faces the center position of the length-adjustable horizontal support 504.

[0067] The detection shield 5 consists of a cubic frame and a light-shielding cloth. The top surface of the cubic frame is open to cooperate with a short-wave infrared camera; the bottom surface is open, completely covering the area 3 to be tested of the photovoltaic module 2, and the four sides are in contact with the surface of the photovoltaic module 2; the top and four sides are covered with the light-shielding cloth. The cubic frame consists of adjustable telescopic rods, and the dimensions of each side of the bottom surface can be dynamically adjusted and fixed within the range of 15cm to 25cm. The side lengths should be designed to ensure that the camera can clearly capture images of the photovoltaic module 2.

[0068] In operation, the photovoltaic module 2 is in a conductive state, with current flowing through the cell units. In the first stage, the photovoltaic module 2 operates normally, and the detection shading cover 5 covers the area to be detected 3, leaving the remaining surfaces unobstructed. At this time, the operating current flowing through the cells is relatively large, and the luminous intensity of the target area is relatively weak. In the second stage, the control shading cover 8 completely blocks a single cell unit in the non-detection area of ​​the photovoltaic module 2, significantly reducing the operating current flowing through the cell and enhancing the luminous intensity of the target area. This invention employs a two-stage imaging method to maximize the intensity of photocarriers within the blocked area and remove interference from sunlight and background light escaping from the edge of the detection shading cover 5; the shading cover's sunlight filtering effect must reach over 99.9%.

[0069] Image acquisition device 4 uses a Bobcat 640 short-wave infrared camera paired with a Kowa short-wave infrared lens. At a camera working distance of 20cm, a single pixel represents an actual distance of 0.25mm. Based on the luminescence characteristics of silicon-based photovoltaic modules, whose emission spectrum has a center wavelength of 1150nm, a 1100-1200nm narrow-bandpass filter is fitted in front of the short-wave infrared camera lens to effectively acquire images of the photovoltaic module's emission.

[0070] During the first stage of shooting, the detection light shield 5 is placed on each battery cell unit in the area to be tested 3. The size of the shading area of ​​the detection light shield 5 is adjusted according to the size of the battery cell units in the area to be tested 3, so that the area covered by each shading battery cell unit does not exceed 2 / 3 of the battery cell unit's own area, that is, at least 1 / 3 or more of the unit is exposed to sunlight. When there are many battery cell units in the area to be tested 3, a multiple acquisition method is adopted. In N moving shots, several battery cell units in the area to be tested 3 are blocked by the detection light shield 5 each time, until all battery cell units in the abnormal heat area have been blocked and images have been captured.

[0071] During the second stage of shooting, the series group of the battery cell unit that is blocked in the test area 3 is determined, and any one of the battery cell units in the non-test area is selected from the series group. The selected battery cell unit in the non-test area is completely blocked by the control light shield 8, and then the second stage of shooting is carried out.

[0072] like Figure 5 a and Figure 5 As shown in b, in specific implementation, each row of solar cell units on the selected photovoltaic module 2 is connected in series. When there are abnormal heat generation areas in four of the solar cell units, a detection shading cover 5 is used to block part of the four solar cell units. During the second stage of shooting, a control shading cover 8 is used to block the entire area of ​​two of the solar cell units in the non-detection area connected in series with the four solar cell units.

[0073] S4. The photovoltaic module inspection equipment uses an image detection algorithm to process the short-wave infrared images of the two-stage inspection area 3 to obtain the defect detection results, thus completing the inspection of the photovoltaic module 2 under sunlight.

[0074] In step S4, when only one set of n abnormal heating area images are captured in both the first and second stages, the n abnormal heating area images captured in the first and second stages are input into the main control computer 6 for image detection algorithm processing. First, preprocessing is performed: the n abnormal heating area images captured in the first stage are averaged at the pixel level to obtain a single first-stage image; then, the n abnormal heating area images captured in the second stage are averaged at the pixel level to obtain a single second-stage image, which suppresses random noise. The second-stage image and the first-stage image are then subtracted at the pixel level to obtain a relatively clean target area luminous image, effectively removing interference from the background and sunlight. Then, defect detection is performed on the target area luminous image, finally obtaining the defect detection result. Specifically, morphological processing or the U-Net deep learning algorithm is used for defect detection. Relevant parameters such as thresholds have been preset, and a modification interface is retained.

[0075] When N sets of abnormal heating area images are captured in the first stage, N sets of abnormal heating area images are captured in the second stage. When both the first and second stages capture N sets of n abnormal heating area images, the n abnormal heating area images in each set are processed by the same image detection algorithm as described above, and then stitched together according to the position of the captured battery cell unit to finally obtain a complete abnormal heating area image detection result.

[0076] Output the detection image and results, and output and save the detection report. The report includes the mean image obtained from the two stages, a clean luminescent image of the test area, the defect detection algorithm recognition result image, and defect-related information such as defect type and size.

[0077] Conclusion: Complete the output and storage of the inspection report, determine the defect type and location of the abnormal heating area of ​​the photovoltaic module, speculate on its cause, and then determine the treatment method and improvement measures to ultimately achieve the goal of cost reduction and efficiency improvement.

Claims

1. A method for detecting silicon-based photovoltaic modules under sunlight based on a short-wave infrared camera, characterized in that, include: S1. Under the illumination of the sun (1), perform ground or drone infrared detection on several photovoltaic modules (2) connected in series by the working photovoltaic circuit (7) in the area to be maintained, and obtain infrared images of each photovoltaic module (2). S2. Perform anomaly detection on the infrared images of each photovoltaic module (2) to obtain the infrared images of the photovoltaic module (2) with abnormal heating areas; locate each photovoltaic module (2) with abnormal heating areas and take the abnormal heating areas in the infrared images of the photovoltaic module (2) as the areas to be detected (3). S3. For each photovoltaic module (2) with an abnormal heating area, the photovoltaic module detection equipment is set up at the location of the area to be detected (3) of the photovoltaic module (2). The abnormal heating area at the location of the area to be detected (3) is obtained by using the photovoltaic module detection equipment in a two-stage shooting method. S4. The photovoltaic module testing equipment will use the image detection algorithm to process the abnormal heating area image of the two-stage test area (3) to obtain the defect detection result and complete the detection of the photovoltaic module (2) under sunlight. The photovoltaic module (2) includes several cell units arranged in a rectangular array; In step S3, the photovoltaic module testing equipment includes an image acquisition device (4), a testing shading cover (5), a main control host computer (6), and a control shading cover (8). The image acquisition device (4) is installed at the center of the testing shading cover (5) and faces the opening direction of the testing shading cover (5). The image acquisition device (4) is electrically connected to the main control host computer (6) which is equipped with an image detection algorithm. Two-stage image acquisition is performed using photovoltaic module inspection equipment. For the inspection area (3) of each photovoltaic module (2), the details are as follows: During the first stage of shooting, the photovoltaic module (2) operates normally, and the detection shading cover (5) is placed on each cell unit in the area to be tested (3). At this time, the image acquisition device (4) faces each cell unit, and the host computer (6) controls the image acquisition device (4) to continuously capture a set of n abnormal heating area images of the area to be tested (3) and transmit them to the host computer (6). During the second stage of shooting, the detection shading cover (5) is always placed on the area to be tested (3), and the control shading cover (8) is placed on one of the cell units connected in series with each cell unit in the area to be tested (3) in other areas of the photovoltaic module (2). Then, the host computer (6) controls the image acquisition device (4) to continuously capture a set of n abnormal heating area images of the area to be tested (3) and transmit them to the host computer (6). In step S4, when only one set of n abnormal heating area images are captured in both the first and second stages, the n abnormal heating area images captured in the first stage and the n abnormal heating area images captured in the second stage are input into the host computer (6) for processing using an image detection algorithm. First, preprocessing is performed, and the n abnormal heating area images captured in the first stage are averaged at the pixel level to obtain a single first stage image. Then, the n abnormal heating area images captured in the second stage are averaged at the pixel level to obtain a single second stage image. The second stage image and the first stage image are subtracted at the pixel level to obtain the target area luminous image. Then, defect detection is performed on the target area luminous image to finally obtain the defect detection result. When N sets of n abnormal heating area images are captured in both the first and second stages, the images of each set of n abnormal heating area images are processed by the same image detection algorithm as described above, and then stitched together according to the position of the captured battery cell unit to finally obtain a complete abnormal heating area image detection result.

2. The method for detecting silicon-based photovoltaic modules under sunlight based on a short-wave infrared camera according to claim 1, characterized in that: In step S1, for each photovoltaic module (2) in the area to be maintained, an infrared image of the complete surface of the photovoltaic module (2) under sunlight is captured by a handheld or desktop infrared detection device deployed on the ground or an infrared detection device carried by a drone.

3. The method for detecting silicon-based photovoltaic modules under sunlight based on a short-wave infrared camera according to claim 1, characterized in that: In step S1, each battery cell unit is divided into groups, each group of battery cell units is connected in series with each other, and each group of series units is connected in parallel. The area to be tested (3) includes one or more battery cell units that are abnormally overheating.

4. The method for detecting silicon-based photovoltaic modules under sunlight based on a short-wave infrared camera according to claim 1, characterized in that: In step S2, for each photovoltaic module (2) infrared image, during anomaly detection, if the temperature value of a certain area in the infrared image exceeds the temperature threshold, then that area is regarded as an abnormal heating area.

5. The method for detecting silicon-based photovoltaic modules under sunlight based on a short-wave infrared camera according to claim 1, characterized in that: The detection light shield (5) includes a fixed horizontal support (502), a four-way adjustable four-way connector (503), a length adjustable horizontal support (504), a length adjustable vertical support (505), a four-way adjustable three-way connector (506), and a light shield. The fixed horizontal support (502) includes four short connecting rods and two long connecting rods. The four short connecting rods are connected by the four-way adjustable four-way connector (503) to form a rectangular structure. The two long connecting rods are cross-connected to the four-way adjustable four-way connector (503). A camera mounting hole (501) is provided at the intersection of the two long connecting rods. The image acquisition device (4) is installed in the camera mounting hole. (501) The length adjustable longitudinal support (505) includes four first length adjustable connecting rods, one end of each of the four first length adjustable connecting rods is connected to four four-way adjustable connectors (503), and the other end of each of the four first length adjustable connecting rods is connected to four three-way adjustable connectors (506). The length adjustable transverse support (504) includes four second length adjustable connecting rods, which are connected by four three-way adjustable connectors (506) to form a rectangular structure. The length adjustable transverse support (504) is parallel to the fixed transverse support (502) and parallel to the surface of the photovoltaic module (2). The length-adjustable connecting rod includes an external threaded connecting rod (5041) and an internal threaded connecting rod (5042). The external threaded connecting rod (5041) is threadedly fitted into the internal threaded connecting rod (5042). The length of the length-adjustable connecting rod is adjusted by rotating the thread at the threaded connection (5043) of the external threaded connecting rod (5041) and the internal threaded connecting rod (5042). The detection light shield (5) is deformed into a height-varying inverted frustum, rectangular and frustum shapes by adjusting the length of the length-adjustable transverse support (504) and the length-adjustable longitudinal support (505). This adjusts the area of ​​the length-adjustable transverse support (504) to change the blocking area of ​​the detection light shield (5). The light-shielding cloth is completely covered on the fixed horizontal support (502) and the length-adjustable vertical support (505), so that the rectangular opening formed by the length-adjustable horizontal support (504) serves as the opening of the detection light-shielding cover (5), and the image acquisition device (4) faces the center position of the length-adjustable horizontal support (504).

6. The method for detecting silicon-based photovoltaic modules under sunlight based on a short-wave infrared camera according to claim 5, characterized in that: The image acquisition device (4) uses a short-wave infrared camera and is equipped with an 1100-1200nm narrow-band pass filter in front of the lens of the short-wave infrared camera.

7. The method for detecting silicon-based photovoltaic modules under sunlight based on a short-wave infrared camera according to claim 1, characterized in that: During the first stage of shooting, the detection light shield (5) covers each battery cell unit in the area to be tested (3). The size of the area covered by the detection light shield (5) is adjusted according to the size of the battery cell unit in the area to be tested (3) so that the area covered by each battery cell unit does not exceed 2 / 3 of the area of ​​the battery cell unit itself. When there are many battery cell units in the area to be tested (3), multiple acquisitions are adopted. In N moving shots, several battery cell units in the area to be tested (3) are covered by the detection light shield (5) each time until all battery cell units in the abnormal heating area have been covered and images have been captured. During the second stage of shooting, determine the series group in which the blocked battery cell unit in the area to be tested (3) is located, select any battery cell unit in the series group that is not in the testing area, use the control light shield (8) to completely block the selected battery cell unit in the non-testing area, and then perform the second stage of shooting.

8. The method for detecting silicon-based photovoltaic modules under sunlight based on a short-wave infrared camera according to claim 1, characterized in that: The defect detection specifically employs morphological processing or the U-Net deep learning algorithm.

Images