Photovoltaic module image fault classification and identification method based on unmanned aerial vehicle

Abstract

The invention discloses a photovoltaic module image fault classification and identification method based on an unmanned aerial vehicle, and relates to the technical field of photovoltaic module fault detection. According to the method, the pictures are shot and processed in real time, and the thermal imaging pictures are processed, so that the accuracy of fault classification can be ensured, precise division of mechanical faults and electrical faults can be realized, the accuracy of repair and maintenance can be improved, the working efficiency can be improved based on the unmanned aerial vehicle, and time and labor are saved; the infrared thermal imaging device can perform multi-angle scanning on the module, obtain a temperature distribution image, perform comparative analysis in combination with a visible light image, judge the position and reason of a fault battery piece through a temperature abnormal point, and observe the structural integrity in the battery piece to achieve the purpose of hidden crack detection. Based on the thermal imaging technology, the working efficiency can be improved, invalid work is avoided, the maintenance cost can be reduced, and good economic benefits are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of photovoltaic module fault detection, and particularly relates to a method for fault classification and identification of photovoltaic module images based on an unmanned aerial vehicle (UAV). Background Art

[0002] Photovoltaic modules, i.e., solar cell modules, have a low output voltage for a single solar cell. In addition, for an unpackaged battery, the electrodes are prone to falling off due to environmental influences. Therefore, a certain number of single cells must be sealed into a photovoltaic module in series and parallel to prevent the battery electrodes and interconnections from being corroded. Photovoltaic modules are classified into crystalline silicon solar cell modules and thin-film solar cell modules according to the materials of the solar cells.

[0003] Previously, power grid inspections were generally carried out manually, that is, personnel were dispatched to actually inspect the transmission lines and the operation of power equipment along the transmission lines. Among them, solar photovoltaic power generation systems are mainly divided into centralized and distributed types. Centralized power stations generally cover a large area; distributed power stations are generally built on roofs, greenhouses, and large-area water pools. However, after these photovoltaic power stations are connected to the grid, they bring a large amount of operation and maintenance pressure, such as conventional equipment detection and photovoltaic panel inspection. The traditional operation and maintenance method uses manual inspection, which has low efficiency, and most of the equipment faults are judged based on the experience of operation and maintenance personnel, which is extremely prone to deviation. Therefore, we propose a method for fault classification and identification of photovoltaic module images based on an unmanned aerial vehicle (UAV). Summary of the Invention

[0004] The purpose of the present invention is to provide a method for fault classification and identification of photovoltaic module images based on an unmanned aerial vehicle (UAV) to solve the problems mentioned in the above background art.

[0005] The present invention specifically adopts the following technical solutions to achieve the above purpose:

[0006] A method for fault classification of photovoltaic module images based on an unmanned aerial vehicle (UAV) includes the following steps:

[0007] Step 1, obtaining images: obtaining real-time images of photovoltaic modules based on an unmanned aerial vehicle (UAV) and uploading the image data in real time through an internal data transmission module;

[0008] Step 2, image processing: performing step-by-step processing on pictures and thermal imaging pictures;

[0009] Step 3, fault classification: classifying faults according to the processed images into mechanical faults and electrical faults;

[0010] Step 4, data storage: storing the processed pictures and the classified fault information.

[0011] Further, the drone is built - in with a camera and an infrared thermal imaging device. The camera is used to take real - time pictures, and the infrared thermal imaging device is used to detect the infrared radiation energy of the photovoltaic module.

[0012] Further, the image processing includes the following steps:

[0013] Step 211, denoising and smoothing: Eliminate noise interference in the image, such as Gaussian noise, salt - and - pepper noise, etc., restore the clarity of the image, and make the overall image softer by blurring image details and reducing high - frequency information (such as edges, textures).

[0014] Step 212, defect repair: Quickly remove small - area defects such as stains, scratches, and spots in the image. For large - area objects (such as utility poles) that cannot be directly repaired, the content - aware filling function can be used to intelligently match the surrounding environment for coverage, improving the image quality.

[0015] Step 213, color temperature adjustment: Adjust the color temperature of key areas such as eyes separately through curve or level tools to enhance the sense of reality.

[0016] Further, the image processing includes the following steps:

[0017] Step 221, background correction: Reduce the background brightness through a curve adjustment layer to avoid conflict with thermal imaging.

[0018] Step 222, image fusion: Intercept the corresponding areas of the thermal image and the visible - light image according to the temperature threshold, perform seamless stitching, and align and fuse the thermal image and the visible - light image through image registration technology (such as ENVI software).

[0019] Step 223, optimization processing: Use a sharpening tool to enhance the edge contrast, or enhance the three - dimensional sense through color adjustment, and perform dynamic range adjustment.

[0020] Further, the mechanical failures are divided into: physical damage and installation - related failures. Among them, the physical damage is divided into cracks, fractures, and encapsulation material aging, and the installation - related failures are divided into bracket problems and junction box failures.

[0021] Further, the electrical failures are divided into: power attenuation and hot - spot effect. Among them, the power attenuation is caused by material aging, encapsulation failure, or ultraviolet radiation, resulting in a decrease in output power, and the hot - spot effect is that local shading or poor soldering causes the solar cell to overheat, accelerating aging and even causing a fire.

[0022] A method for fault identification of photovoltaic module images based on a drone, includes the following steps:

[0023] Step 1, mechanical fault identification: Detect and analyze by combining pictures and infrared thermal imaging pictures. Check whether there are surface cracks and hidden cracks on its surface, whether the glass is broken, whether the encapsulation material is aged, whether the bracket is modified or deformed, whether the diode is damaged, whether the solder joints are loose or the wires are broken;

[0024] Step 2, electrical fault identification: Analyze the temperature distribution of the components using thermal imaging technology. Determine whether there is local heating by observing whether there is a hot spot effect or snail pattern;

[0025] Step 3, fault location: Record the coordinates according to the positioning unit of the drone to accurately locate the fault point information;

[0026] Step 4, fault handling: Upload the completed fault information and location information to the control center, and the staff will formulate maintenance and repair tasks.

[0027] Furthermore, the drone is also internally provided with a positioning unit, a data transceiver module, a central processing unit and a driving system. The positioning unit is used to locate the position of the shooting point. The data transceiver module is used to receive the start / stop commands issued by the remote controller and transmit image data. The central processing unit is used to receive the commands issued via the data transceiver module and issue start / stop commands to the driving system. The driving system is used to drive the flight of the drone.

[0028] Furthermore, the electrical fault identification includes the following steps:

[0029] Step 21, infrared thermal imaging scanning: Scan the components from multiple angles to obtain the temperature distribution image and perform comparative analysis in combination with the visible light image;

[0030] Step 22, hot spot identification: Determine the position and cause of the faulty solar cell through temperature abnormal points;

[0031] Step 23, hidden crack detection: Observe the integrity of the internal structure of the solar cell;

[0032] Step 24, thermal trend analysis: Long-term monitor the temperature change trend and evaluate the risk of fault development.

[0033] The beneficial effects of the present invention are as follows:

[0034] 1. By taking and processing pictures in real time and cooperating with the processing of thermal imaging pictures, the present invention can ensure the accuracy of fault classification, can accurately divide mechanical faults and electrical faults, can improve the accuracy of repair and maintenance, and can improve work efficiency based on the drone, saving time and effort.

[0035] 2. The present invention can perform multi-angle scanning on components through an infrared thermal imaging device, obtain a temperature distribution image, conduct comparative analysis in combination with a visible light image, and then determine the location and cause of a faulty cell through temperature abnormal points. Moreover, observing the structural integrity inside the cell can achieve the purpose of detecting hidden cracks. Based on thermal imaging technology, the work efficiency can be improved, ineffective work can be avoided, the maintenance cost can be reduced, and better economic benefits can be obtained. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 is a flowchart of a method for fault classification of photovoltaic module images based on an unmanned aerial vehicle;

[0037] Figure 2 is a flowchart of the processing of pictures taken by a camera in the present invention;

[0038] Figure 3 is a flowchart of the processing of thermal imaging pictures in the present invention;

[0039] Figure 4 is a flowchart of a method for fault identification of photovoltaic module images based on an unmanned aerial vehicle;

[0040] Figure 5 is a flowchart of the electrical fault identification in the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0041] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0042] Please refer to Figure 1 - Figure 3 , the present invention provides a method for fault classification of photovoltaic module images based on an unmanned aerial vehicle, including the following steps:

[0043] Step 1, obtaining images: obtaining real-time images of photovoltaic modules based on an unmanned aerial vehicle and uploading the image data in real time through the internal data transmission module thereof;

[0044] Step 2, image processing: performing step-by-step processing on pictures and thermal imaging pictures;

[0045] Step 3, fault classification: classifying faults according to the processed images into mechanical faults and electrical faults;

[0046] Step 4, data storage: storing the processed pictures and the classified fault information.

[0047] In this embodiment, preferably, the drone has a built-in camera and an infrared thermal imaging device, the camera is used to take real-time pictures, and the infrared thermal imaging device is used to detect the infrared radiation energy of the photovoltaic module.

[0048] In this embodiment, preferably, the image processing includes the following steps:

[0049] Step 211, denoising and smoothing: eliminating noise interference in the image, such as Gaussian noise, salt and pepper noise, etc., restoring the clarity of the image, blurring the image details, reducing high-frequency information (such as edges, textures), and making the image as a whole softer;

[0050] Step 212, defect repair: quickly remove small area defects such as stains, scratches, spots, etc. in the image. For large area objects that cannot be directly repaired (such as telephone poles), the content recognition filling function can be used to intelligently match the surrounding environment to cover and improve image quality;

[0051] Step 213, color temperature adjustment: Use the curve or levels tool to adjust the color temperature of key areas such as the eyes to enhance realism.

[0052] In this embodiment, preferably, the image processing includes the following steps:

[0053] Step 221, background correction: reduce the background brightness through the curve adjustment layer to avoid conflict with thermal imaging;

[0054] Step 222, image fusion: intercept corresponding areas of the thermal image and the visible light image according to the temperature threshold, perform seamless stitching, and align the thermal image and the visible light image through image registration technology (such as ENVI software) and then fuse them;

[0055] Step 223, optimization processing: Use the sharpening tool to increase edge contrast, or enhance the three-dimensional effect through color adjustment, and perform dynamic range adjustment.

[0056] In this embodiment, preferably, mechanical failures are divided into: physical damage and installation-related failures, wherein physical damage is divided into cracks, breakages and aging of packaging materials, and installation-related failures are divided into bracket problems and junction box failures.

[0057] In this embodiment, preferably, electrical faults are divided into: power attenuation and hot spot effect, wherein power attenuation is the decrease in output power due to material aging, packaging failure or ultraviolet radiation, and hot spot effect is the overheating of the battery cell due to local shading or poor welding, which accelerates aging and even causes fire.

[0058] The working principle and use process of this embodiment:

[0059] Step 1, image acquisition: The real-time image of the photovoltaic module is acquired based on the drone, and the image data is uploaded in real time through its internal data transmission module; the drone has a built-in camera and infrared thermal imaging equipment, the camera is used to take real-time pictures, and the infrared thermal imaging equipment is used to detect the infrared radiation energy of the photovoltaic module.

[0060] Step 2, image processing: perform step-by-step processing of pictures and thermal imaging pictures;

[0061] The following processing is performed on the pictures taken by the camera:

[0062] Step 211, denoising and smoothing: eliminating noise interference in the image, such as Gaussian noise, salt and pepper noise, etc., restoring the clarity of the image, blurring the image details, reducing high-frequency information (such as edges, textures), and making the image as a whole softer;

[0063] Step 212, defect repair: quickly remove small area defects such as stains, scratches, spots, etc. in the image. For large area objects that cannot be directly repaired (such as telephone poles), the content recognition filling function can be used to intelligently match the surrounding environment to cover and improve image quality;

[0064] Step 213, color temperature adjustment: Use the curve or levels tool to adjust the color temperature of key areas such as the eyes to enhance realism.

[0065] The thermal imaging images are processed as follows:

[0066] Step 221, background correction: reduce the background brightness through the curve adjustment layer to avoid conflict with thermal imaging;

[0067] Step 222, image fusion: intercept corresponding areas of the thermal image and the visible light image according to the temperature threshold, perform seamless stitching, and align the thermal image and the visible light image through image registration technology (such as ENVI software) and then fuse them;

[0068] Step 223, optimization processing: Use the sharpening tool to increase edge contrast, or enhance the three-dimensional effect through color adjustment, and perform dynamic range adjustment.

[0069] Step 3, fault classification: Fault classification is performed based on the processed images, which are divided into mechanical faults and electrical faults; among them, mechanical faults are divided into: physical damage and installation-related faults, among which physical damage is divided into cracks, breakages and aging of packaging materials, and installation-related faults are divided into bracket problems and junction box failures; electrical faults are divided into: power attenuation and hot spot effect, among which power attenuation is the decrease in output power due to material aging, packaging failure or ultraviolet radiation, and hot spot effect is the overheating of the battery cell due to local shading or poor welding, which accelerates aging and even causes fire.

[0070] Step 4, Data storage: Store the processed pictures and the classified fault information.

[0071] Please refer to Figure 4 - Figure 5 , the present invention also provides a method for fault identification of photovoltaic module images based on drones, including the following steps:

[0072] Step 1, Mechanical fault identification: Combine pictures and infrared thermal imaging pictures for detection and analysis, and check whether there are surface cracks and hidden cracks on its surface, whether there is glass breakage, whether there is encapsulation material aging, whether there is bracket modification or deformation, whether there is diode damage, solder joint loosening or wire breakage;

[0073] Step 2, Electrical fault identification: Use thermal imaging technology to analyze the temperature distribution of the module, and determine whether there is local heating by observing whether there is hot spot effect or snail pattern;

[0074] Step 3, Fault location: Record the coordinates according to the positioning unit of the drone, and accurately locate the fault point information;

[0075] Step 4, Fault handling: Upload the identified fault information and location information to the control center, and the staff will formulate maintenance and repair tasks.

[0076] In this embodiment, preferably, the drone is also built-in with a positioning unit, a data transceiver module, a central processor, and a drive system. The positioning unit is used to realize the position positioning of the shooting point, the data transceiver module is used to receive the start / stop commands issued by the remote controller and the transmission of image data, the central processor is used to receive the commands issued via the data transceiver module and issue start / stop commands to the drive system, and the drive system is used to realize the flight drive of the drone.

[0077] In this embodiment, preferably, the electrical fault identification includes the following steps:

[0078] Step 21, Infrared thermal imaging scanning: Scan the module from multiple angles to obtain a temperature distribution image, and perform comparative analysis in combination with the visible light image;

[0079] Step 22, Hot spot identification: Determine the position and cause of the faulty cell by the temperature anomaly point;

[0080] Step 23, Hidden crack detection: Observe the integrity of the internal structure of the cell;

[0081] Step 24, Thermal trend analysis: Long-term monitor the temperature change trend and evaluate the risk of fault development.

[0082] The working principle and usage process of this embodiment:

[0083] Step 1. Mechanical fault identification: Detect and analyze by combining pictures and infrared thermal imaging pictures. Check whether there are surface cracks and hidden cracks on its surface, whether the glass is broken, whether the encapsulation material is aged, whether there are bracket modifications or deformations, whether the diode is damaged, solder joints are loose or wires are broken; The drone is also built-in with a positioning unit, a data transceiver module, a central processor and a drive system. The positioning unit is used to locate the position of the shooting point, the data transceiver module is used to receive the start / stop commands sent by the remote controller and transmit image data, the central processor is used to receive the commands issued via the data transceiver module and issue start / stop commands to the drive system, and the drive system is used to drive the flight of the drone.

[0084] Step 2. Electrical fault identification: Analyze the temperature distribution of components using thermal imaging technology, and determine whether there is local heating by observing whether there is a hot spot effect or snail pattern; The electrical fault identification includes the following steps:

[0085] Step 21. Infrared thermal imaging scanning: Scan the components from multiple angles to obtain a temperature distribution image, and perform comparative analysis in combination with visible light images;

[0086] Step 22. Hot spot identification: Determine the position and cause of the faulty cell through temperature abnormal points;

[0087] Step 23. Hidden crack detection: Observe the integrity of the internal structure of the cell;

[0088] Step 24. Thermal trend analysis: Long-term monitor the temperature change trend and evaluate the risk of fault development.

[0089] Step 3. Fault location: Record the coordinates according to the positioning unit of the drone and accurately locate the fault point information;

[0090] Step 4. Fault handling: Upload the completed identified fault information and location information to the control center, and the staff will formulate maintenance and repair tasks.

[0091] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but will be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A fault classification method for photovoltaic module images based on drones, characterized in that, It includes the following steps: Step 1, Image acquisition: Obtain real-time images of photovoltaic modules based on a drone, and upload the image data in real time through the internal data transmission module thereof; Step 2, Image processing: Perform step-by-step processing on the pictures and thermal imaging pictures; Step 3, Fault classification: Classify the faults according to the processed images, which are divided into mechanical faults and electrical faults; Step 4, Data storage: Store the processed pictures and the classified fault information.

2. The fault classification method for photovoltaic module images based on drones according to claim 1, wherein: The drone is built-in with a camera and an infrared thermal imaging device. The camera is used to take real-time pictures, and the infrared thermal imaging device is used to detect the infrared radiation energy of the photovoltaic module.

3. A fault classification method for photovoltaic module images based on drones according to claim 1, characterized in that, The image processing includes the following steps: Step 211, Denoising and smoothing: Eliminate noise interference in the image, such as Gaussian noise, salt and pepper noise, etc., restore the clarity of the image, and make the overall image softer by blurring the image details and reducing high-frequency information (such as edges, textures); Step 212, Defect repair: Quickly remove small-area defects such as stains, scratches, and spots in the image. For large-area objects (such as utility poles) that cannot be directly repaired, the content recognition filling function can be used to intelligently match the surrounding environment for coverage to improve the image quality; Step 213, Color temperature adjustment: Adjust the color temperature of key areas such as eyes separately through curve or level tools to enhance the sense of reality.

4. A fault classification method for photovoltaic module images based on drones according to claim 1, characterized in that, The image processing includes the following steps: Step 221, Background correction: Reduce the background brightness through a curve adjustment layer to avoid conflicts with thermal imaging; Step 222, Image fusion: Intercept the corresponding areas of the thermal image and the visible light image according to the temperature threshold, perform seamless splicing, and align and fuse the thermal image and the visible light image through image registration technology (such as ENVI software); Step 223, Optimization processing: Use a sharpening tool to enhance the edge contrast, or enhance the three-dimensional sense through color adjustment, and perform dynamic range adjustment.

5. A fault classification method for photovoltaic module images based on drones according to claim 1, characterized in that: The mechanical faults are divided into: physical damage and installation-related faults. Among them, the physical damage is divided into cracks, fractures, and encapsulation material aging, and the installation-related faults are divided into bracket problems and junction box faults.

6. The fault classification method for photovoltaic module images based on an unmanned aerial vehicle according to claim 1, wherein: The electrical faults are divided into: power attenuation and hot spot effect. Among them, the power attenuation is caused by material aging, encapsulation failure, or ultraviolet irradiation resulting in a decrease in output power, and the hot spot effect is that local shading or poor soldering causes the solar cell to overheat, accelerating aging and even causing a fire.

7. A fault identification method for photovoltaic module images based on drones, characterized in that, It includes the following steps: Step 1, Mechanical fault identification: Combine pictures and infrared thermal imaging pictures for detection and analysis, and check whether there are surface cracks and hidden cracks on its surface, whether there is glass fracture, whether there is encapsulation material aging, whether there is bracket modification or deformation, whether there is diode damage, solder joint looseness or wire breakage; Step 2, Electrical fault identification: Use thermal imaging technology to analyze the temperature distribution of the component, and determine whether there is local heating by observing whether there is a hot spot effect or snail pattern; Step 3, Fault location: Record the coordinates according to the positioning unit of the drone to accurately locate the fault point information; Step 4, Fault handling: Upload the completed identified fault information and location information to the control center, and let the staff formulate maintenance and repair tasks.

8. A fault identification method for photovoltaic module images based on drones according to claim 7, characterized in that: The drone is also built-in with a positioning unit, a data transceiver module, a central processor, and a driving system. The positioning unit is used to locate the position of the shooting point. The data transceiver module is used to receive the start / stop commands sent by the remote controller and transmit image data. The central processor is used to receive the commands issued via the data transceiver module and issue start / stop commands to the driving system. The driving system is used to drive the flight of the drone.

9. The fault identification method for photovoltaic module images based on an unmanned aerial vehicle according to claim 7, wherein, The electrical fault identification includes the following steps: Step 21, infrared thermal imaging scanning: Scan the components from multiple angles to obtain a temperature distribution image, and perform comparative analysis in combination with the visible light image; Step 22, hot spot identification: Determine the position and cause of the faulty solar cell through the temperature anomaly point; Step 23, crack detection: Observe the integrity of the internal structure of the solar cell; Step 24, thermal trend analysis: Long-term monitor the temperature change trend and evaluate the risk of fault development.

Images