A photovoltaic module automatic positioning method, system and storage medium based on active heating and thermal imaging

Abstract

This invention provides an automatic positioning method, system, and storage medium for photovoltaic modules based on active heating and thermal imaging. The method includes the following steps: S1: Applying a controlled current to a target photovoltaic module or a string containing multiple photovoltaic modules to generate identifiable thermal features; S2: Acquiring a thermal image containing the target photovoltaic module or string using an image acquisition device, the thermal image containing georeferenced information; S3: Processing the thermal image, identifying the thermal features, and determining the identity information of a specific component within the target photovoltaic module or string based on the thermal features; S4: Calculating the precise geographical coordinates of the target photovoltaic module or specific component based on the thermal features and georeferenced information in the thermal image; S5: Binding and storing the identity information with the precise geographical coordinates. This invention can effectively improve the positioning efficiency, accuracy, and security of photovoltaic modules.

Description

Technical Field

[0001] This invention relates to the field of photovoltaic module positioning technology, and in particular to an automatic positioning method, system and storage medium for photovoltaic modules based on active heating and thermal imaging. Background Technology

[0002] During the construction of photovoltaic power plants, the installation and positioning of photovoltaic modules is the basic prerequisite for subsequent operation and maintenance management, fault diagnosis, and power generation efficiency statistics. The core requirement is to establish a precise correspondence between the physical number of the photovoltaic module and its actual installation location (geographic coordinates).

[0003] The most common technical solution used in the industry is the manual labeling and positioning method. The specific implementation process is as follows: when the photovoltaic modules leave the factory or at the installation site, workers manually paste or fix physical labels (such as paper labels, plastic nameplates, etc.) with unique numbers on them to the frame of the photovoltaic modules or designated locations. Then, the label numbers and the corresponding installation area location information are recorded manually, or the geographical coordinates of the area are roughly recorded with the help of handheld GPS devices, and finally a correspondence table between the numbers and the locations is formed.

[0004] However, existing manual labeling and positioning methods have many insurmountable drawbacks, as follows:

[0005] (1) Low efficiency: Large-scale photovoltaic power plants typically contain thousands or even tens of thousands of photovoltaic modules. Manually labeling and recording the location of each module is extremely labor-intensive, requiring a large investment of manpower and a long positioning cycle, which seriously restricts the construction progress of the power plant. Practical data shows that skilled workers can only complete the labeling and location recording of about 500 photovoltaic modules per day.

[0006] (2) Poor accuracy: Errors such as incorrect or missing numbering and location confusion are prone to occur during manual recording. At the same time, the positioning accuracy of handheld GPS devices is usually only at the meter level, which cannot meet the requirements of subsequent refined operation and maintenance for location accuracy. In addition, physical tags are exposed to the outdoor environment for a long time and are easily eroded by wind, rain, ultraviolet rays, sand and dust, which can cause them to fall off or become blurred, resulting in the loss of the correspondence between the number and the location.

[0007] (3) Insufficient safety: Photovoltaic module installation scenarios are mostly complex environments such as high-altitude rooftops, mountains, and water surfaces. Workers are at risk of falling when working at heights, and collisions, electric shocks, and other safety accidents are likely to occur when working in areas with dense photovoltaic modules.

[0008] (4) Poor data compatibility: The location information recorded manually is mostly in paper documents or simple spreadsheets, which is difficult to directly connect to the digital operation and maintenance management platform of smart photovoltaic power stations. It requires secondary input and sorting, which further increases the workload and the probability of errors.

[0009] To address these issues, the industry has attempted several solutions, such as using RFID tags to replace traditional physical tags. However, this still requires manual tag installation, and RFID signals are easily attenuated by the metal frames of photovoltaic modules, limiting the identification distance and making long-distance batch positioning impossible. Another solution involves using drones for visible light photography for positioning, but the highly homogeneous appearance of photovoltaic modules makes it difficult to distinguish the serial numbers and locations of different modules using visible light images, compromising positioning accuracy and reliability.

[0010] Therefore, the lack of a technical solution in the existing technology to automatically, efficiently, accurately and safely locate the corresponding number and position of photovoltaic modules has restricted the development process of intelligent construction of photovoltaic power plants. Summary of the Invention

[0011] The purpose of this invention is to provide an automatic positioning method, system, and storage medium for photovoltaic modules based on active heating and thermal imaging, which can achieve rapid and accurate positioning of photovoltaic modules.

[0012] To achieve the above objectives, the solution of the present invention is as follows: In a first aspect, the present invention provides an automatic positioning method for photovoltaic modules based on active heating and thermal imaging, comprising the following steps: S1: Apply a controlled current to the target photovoltaic module or a string containing multiple photovoltaic modules to generate identifiable thermal characteristics in the target photovoltaic module or string. S2: Acquire a thermal image containing the target photovoltaic module or string using an image acquisition device. The thermal image contains georeferenced information. S3: Process thermal images, identify thermal features, and determine the identity information of specific components in the target photovoltaic module or string based on the thermal features; S4: Calculate the precise geographic coordinates of the target photovoltaic module or a specific module based on the thermal features and georeferenced information in the thermal image; S5: Bind and store identity information with precise geographic coordinates.

[0013] In a preferred embodiment, in step S1, when current is applied to the target photovoltaic module, the controlled current is a pulse current modulated according to a preset coding rule, so that the thermal characteristics form a dynamic thermal signal. In step S3, identifying thermal features includes performing time-domain analysis on the thermal image sequence to decode the identity information corresponding to the preset encoding rules.

[0014] In the preferred embodiment, in step S1, the same current is applied to the strings; In step S3, thermal features are identified and the identity information of specific components in the string is determined, including: S31: Acquire a visible light image of the target area through the visible light channel of the image acquisition device, and generate a spatial mask grid of multiple photovoltaic modules in the string based on the visible light image; S32: Spatial registration of the thermal image with the spatial mask grid; S33: Determine the overlap between the heated areas in the thermal image and each grid in the spatial mask grid; S34: Based on the overlap exceeding a preset threshold, lock the grid set corresponding to the string that is currently generating heat.

[0015] In the preferred embodiment, step S3 further includes: S35: Count the grids in the locked grid set along a preset direction; S36: When the count reaches the target sequence number, the target mesh corresponding to that sequence number is determined as the specific component; In step S4, the precise geographic coordinates of a specific component are calculated based on the location of the target grid.

[0016] In a preferred embodiment, the calculation is based on the inverse projection model of the collinearity equation: ,in, For geographic coordinates, Let be the ray scale factor from the target to the camera's optical center. Let be the rotation transformation matrix of the image acquisition device relative to the world coordinate system. These are the coordinates of the thermal features in the camera coordinate system. This is the position vector of the image acquisition device in the world coordinate system.

[0017] In a preferred embodiment, the image acquisition device is any one of a drone equipped with an RTK positioning module and an infrared thermal imager, a fixedly deployed dual-lens camera, or a ground mobile robot.

[0018] The preferred solution includes the following steps: S71: Apply a controlled current to a target photovoltaic module or a string containing multiple photovoltaic modules to generate identifiable thermal characteristics in the target photovoltaic module or string. S72: Acquire thermal images containing the target photovoltaic modules or strings using an image acquisition device; S73: Process thermal images, identify thermal features, and display an augmented reality interface containing thermal images and / or visible light images on a mobile terminal. The augmented reality interface includes suggestions for identity information automatically identified by the system. Receive user confirmation or correction instructions for the identity information suggestions and determine the final identity information based on the instructions.

[0019] Secondly, the present invention provides an automatic positioning system for photovoltaic modules based on active heating and thermal imaging. This system is used to implement the aforementioned automatic positioning method for photovoltaic modules based on active heating and thermal imaging. The system includes: A heating control module is used to apply a controlled current to a target photovoltaic module or a string containing multiple photovoltaic modules to generate identifiable thermal characteristics. The imaging acquisition module is used to acquire thermal images containing target photovoltaic modules or strings and georeferenced information. The intelligent analysis module is used to process thermal images to identify thermal features and identity information, calculate the precise geographic coordinates of specific components in the target photovoltaic module or string, and bind the identity information to the geographic coordinates.

[0020] In a preferred embodiment, the heating control module is configured to send heating commands to the microinverter, string inverter, or maximum power optimizer connected to the target photovoltaic module or string via power line carrier communication or a wireless mesh network. The imaging acquisition module includes a visible light camera and an infrared thermal imager; The intelligent parsing module includes: The image processing unit is used to generate a spatial mask grid for photovoltaic modules based on visible light images and to preprocess thermal images. Feature recognition unit, used to identify thermal features and fuse them with spatial mask mesh for analysis; Coordinate calculation unit, used to calculate geographic coordinates based on the inverse projection model of the collinearity equation; The system also includes a database with pre-set three-dimensional geographic information of photovoltaic arrays. The intelligent parsing module is also used to match the calculated geographic coordinates with the three-dimensional geographic information and update the component ledger. The intelligent parsing module is also configured as follows: Generate augmented reality interface data containing thermal and / or visible light images, which is labeled with suggested identity information automatically identified by the system; Receive user confirmation or correction instructions from mobile terminals and update identity information based on the instructions.

[0021] Thirdly, the present invention provides a storage medium storing computer program instructions, which, when executed by a processor, implement the above-mentioned automatic positioning method for photovoltaic modules based on active heating and thermal imaging.

[0022] After adopting the above solution, the beneficial effects of the present invention are as follows: 1. This invention, through an automated process of "active heating to excite thermal features - thermal imaging acquisition and identification - automatic calculation and binding", can effectively replace the traditional inefficient and cumbersome manual labeling and recording methods, which is conducive to improving efficiency, shortening the construction and operation and maintenance cycle of photovoltaic power plants, realizing unmanned operation in complex and dangerous environments, and significantly reducing labor costs and safety risks.

[0023] 2. By introducing identifiable thermal features as a medium, this invention establishes a unique and reliable correspondence between photovoltaic modules and spatial locations, eliminating errors such as misrecording, omissions, and confusion that are unavoidable in manual recording. Compared with traditional methods, the accuracy and stability of the positioning results have achieved a qualitative leap.

[0024] 3. The final output of this invention is digital data containing accurate geographic coordinates and component identity information. Its format is standard and can be seamlessly connected to the digital platform of photovoltaic power plants without any secondary input or format conversion, thus solving the pain point of poor data compatibility in existing technologies.

[0025] 4. The entire positioning process of this invention does not require maintenance personnel to enter the photovoltaic array area or perform high-altitude operations. It is completed entirely through remote control, which fundamentally avoids safety hazards such as electric shock, falls, and collisions. It is particularly suitable for complex application scenarios such as large-scale ground power stations, mountain power stations, water surface power stations, and distributed rooftops. Attached Figure Description

[0026] Figure 1 This is a framework diagram of the positioning system in an embodiment of the present invention; Figure 2 This is a schematic diagram of a micro inverter with MPPT function connected to a photovoltaic module in an embodiment of the present invention. Figure 3 This is a schematic diagram of a string inverter with MPPT function connected to multiple photovoltaic modules in an embodiment of the present invention; Figure 4 This is a schematic diagram of the physical network mask of the photovoltaic module array under the visible light image during dual-light fusion mask positioning in an embodiment of the present invention; Figure 5 This is a schematic diagram of the entire heat dissipation zone under infrared thermal imaging during dual-light fusion mask positioning in an embodiment of the present invention; Figure 6 This is a schematic diagram of the overlap between dual-light fusion and IoU during dual-light fusion mask positioning in an embodiment of the present invention; Figure 7 This is a schematic diagram of vector sorting and target locking during dual-light fusion mask positioning in an embodiment of the present invention; Figure 8 This is a schematic diagram illustrating how the image acquisition device calculates the geographic coordinates of a photovoltaic module based on a collinearity equation-based inverse projection model in an embodiment of the present invention. Detailed Implementation

[0027] The present invention will now be further described in conjunction with the accompanying drawings and specific embodiments.

[0028] This embodiment provides an automatic positioning method for photovoltaic modules based on active heating and thermal imaging, including the following steps: S1: Apply a controlled current to the target photovoltaic module or a string containing multiple photovoltaic modules to generate identifiable thermal characteristics in the target photovoltaic module or string. S2: Acquire a thermal image containing the target photovoltaic module or string using an image acquisition device. The thermal image contains georeferenced information. S3: Process thermal images, identify thermal features, and determine the identity information of specific components in the target photovoltaic module or string based on the thermal features; S4: Calculate the precise geographic coordinates of the target photovoltaic module or a specific module based on the thermal features and georeferenced information in the thermal image; S5: Bind and store identity information with precise geographic coordinates.

[0029] Furthermore, in step S1, when current is applied to the target photovoltaic module, the controlled current is a pulse current modulated according to a preset coding rule, so that the thermal characteristics form a dynamic thermal signal. In step S3, identifying thermal features includes performing time-domain analysis on the thermal image sequence to decode the identity information corresponding to the preset encoding rules.

[0030] Furthermore, in step S1, the same current is applied to the strings; In step S3, thermal features are identified and the identity information of specific components in the string is determined, including: S31: Acquire a visible light image of the target area through the visible light channel of the image acquisition device, and generate a spatial mask grid of multiple photovoltaic modules in the string based on the visible light image; S32: Spatial registration of the thermal image with the spatial mask grid; S33: Determine the overlap between the heated areas in the thermal image and each grid in the spatial mask grid; S34: Based on the overlap exceeding a preset threshold, lock the grid set corresponding to the string that is currently generating heat.

[0031] Furthermore, step S3 also includes: S35: Count the grids in the locked grid set along a preset direction; S36: When the count reaches the target sequence number, the target mesh corresponding to that sequence number is determined as the specific component; In step S4, the precise geographic coordinates of a specific component are calculated based on the location of the target grid.

[0032] The active heating principle of this embodiment is to generate Joule heat by modulating the DC-side current of the inverter and utilizing the internal resistance of the photovoltaic module cells. This allows photovoltaic modules to exhibit identifiable features in the infrared band.

[0033] Specifically, depending on the architecture of the photovoltaic power station, there are two control scenarios: Scenario 1: Independent control of micro inverter (with MPPT function) like Figure 2 As shown, in this architecture, each photovoltaic module is connected to an independent micro-inverter. Heating commands are sent to the micro-inverter of the target photovoltaic module via a communication link (such as PLC power line carrier or wireless mesh network).

[0034] Because of its point-to-point control capabilities, individual photovoltaic modules can be heated independently. However, simply heating up a single photovoltaic module will appear as a static "hot spot" in an infrared thermal image. However, factors such as sunlight reflection off glass, localized shading, and even uneven heat dissipation within the photovoltaic module itself can all create similar localized high-temperature spots. Relying solely on temperature difference identification in a complex background makes it difficult to definitively determine if the hot spot is controlled, easily leading to misjudgments.

[0035] Therefore, the heating command in this embodiment includes a preset encoding rule, such as using binary encoding "101" to correspond to a "on-off-on" pulse sequence. The micro-inverter applies a modulated pulse current to the target photovoltaic module according to the command, causing the photovoltaic module to generate a dynamic thermal signal. Since each photovoltaic module can be controlled independently, different encoding sequences can be used to assign unique thermal characteristic identifiers to different photovoltaic modules. Specifically, temporal analysis can be used to filter out static or random thermal noise in the environment, locking only targets that conform to a specific "flickering pattern." By controlling the time interval and duration of heating, it is ensured that the identity information of different photovoltaic modules can be distinguished in the thermal imaging field of view.

[0036] Scenario 2: String inverter control mode (with MPPT function) In a traditional string inverter architecture, multiple photovoltaic modules are connected to the DC side of the inverter. The inverter needs to have four-quadrant operation capability on the DC side or reverse current injection function under anti-islanding protection to ensure that it can draw current from the AC side or energy storage side during non-power generation periods (night or cloudy days).

[0037] like Figure 3As shown, when a photovoltaic power station adopts a string inverter architecture, multiple photovoltaic modules are connected in series to form a string and share a single string inverter. To address the challenge of similar heating characteristics caused by multiple modules connected in series on the DC side of a string inverter and having the same current, a positioning technology based on dual-spectral loads—namely, dynamic mapping between a visible topological mask and infrared dynamic features—can be used; this is known as dual-spectral fusion mask positioning.

[0038] Specifically, the visible light image of the target area is acquired through the visible light channel of the image acquisition device. Then, a deep learning instance segmentation algorithm (such as Mask R-CNN) is used to automatically identify and extract the physical boundaries of all photovoltaic modules, generating a static spatial mask grid in the algorithm's memory. Each grid corresponds to one photovoltaic module, and the grid number corresponds one-to-one with the physical location of the photovoltaic module.

[0039] Spatially register the thermal image with the visible light image to ensure that the pixel coordinates in the thermal image are aligned with the grid coordinates in the mask grid.

[0040] The system calculates the intersection over Union (IoU) between the heated regions in the thermal image (obtained through temperature threshold segmentation) and each grid in the mask grid. When the IoU value of a certain grid set exceeds a preset threshold, the system determines that the string corresponding to that grid set is heating up. This is well known and achievable by those skilled in the art.

[0041] Furthermore, in step S4, the calculation is based on the inverse projection model of the collinearity equation: ,in, For geographic coordinates, Let be the ray scale factor from the target to the camera's optical center. Let be the rotation transformation matrix of the image acquisition device relative to the world coordinate system. These are the coordinates of the thermal features in the camera coordinate system. This is the position vector of the image acquisition device in the world coordinate system.

[0042] like Figure 8 As shown, firstly, the pixel coordinates are projected into normalized camera coordinates using the camera intrinsic parameter matrix. Secondly, by combining the rotation transformation matrix R of the image acquisition device relative to the world coordinate system with the position vector T, an inverse projection model is constructed, which is the collinearity equation pointing from the optical center of the camera to the target photovoltaic module. Finally, a ground height constraint is introduced. ( The scale factor is obtained by finding the intersection of the inverse projection model of the collinearity equation and the plane at that altitude (where the photovoltaic module is located). Thus, the geographic coordinates of the thermal features are determined. .

[0043] This mathematical model allows each numbered hotspot pixel (u,v) in an image to be reconstructed into its real-world latitude and longitude coordinates (Lat,Lon).

[0044] Furthermore, the image acquisition device can be any one of the following: a drone equipped with an RTK positioning module and an infrared thermal imager, a fixedly deployed dual-lens camera, or a ground mobile robot.

[0045] The image acquisition device in this embodiment can use one of the following three carriers: (1) Industrial-grade UAV: ​​Equipped with RTK positioning module, infrared thermal imager and visible light camera, it inspects photovoltaic module arrays through preset routes, collects thermal and visible light images in real time, and records accurate POS data (position, attitude and timestamp).

[0046] (2) Fixed deployment of dual-lens PTZ cameras: Dual-lens PTZ cameras (visible light + infrared) are fixedly deployed at key locations in photovoltaic power plants to monitor the target area by preset angles. This is suitable for small-scale, high-frequency positioning needs.

[0047] (3) Ground mobile robot: equipped with dual-light camera and positioning module, it moves along a preset path and is suitable for ground photovoltaic power stations or complex terrain.

[0048] Regardless of the medium used, the acquired thermal images contain georeferenced information, including: RTK positioning coordinates (longitude, latitude, elevation), camera attitude angles (pitch angle, roll angle, yaw angle), timestamps, etc.

[0049] The image acquisition devices in this embodiment must all be equipped with high-sensitivity long-wave infrared (LWIR) sensors, with a thermal sensitivity (NETD) of less than 50 mK. The core of the recognition capability depends on the ground sampling distance (GSD). To ensure that the imaging algorithm can stably extract the geometric features of the photovoltaic module and suppress environmental thermal noise, the effective coverage pixels of the imaging system over the target photovoltaic module should not be less than 6×6 pixels, preferably 10×10 pixels or more. This requirement is related to the ground sampling distance (GSD), which is determined by the flight altitude (or deployment distance) of the carrier, the lens focal length, and the size of the photosensitive element. For a standard 2m×1m photovoltaic module, the GSD of the infrared channel must be ≤20cm / pixel. At this point, approximately 10 pixels can be allocated to a single photovoltaic module along its long side and approximately 5 pixels along its short side, thus ensuring that the physical gaps between modules can be clearly distinguished during spatial topological sorting, achieving accurate counting of the photovoltaic modules.

[0050] Furthermore, in other embodiments, if it is a simple photovoltaic power station, it can be identified through a semi-automatic manual assistance method, including the following steps: S71: Apply a controlled current to a target photovoltaic module or a string containing multiple photovoltaic modules to generate identifiable thermal characteristics in the target photovoltaic module or string. S72: Acquire thermal images containing the target photovoltaic modules or strings using an image acquisition device; S73: Process thermal images, identify thermal features, and display an augmented reality interface containing thermal images and / or visible light images on a mobile terminal. The augmented reality interface includes suggestions for identity information automatically identified by the system. Receive user confirmation or correction instructions for the identity information suggestions and determine the final identity information based on the instructions.

[0051] Specifically, the acquired thermal and / or visible light images are processed to generate an augmented reality (AR) interface with suggested numbers on the operations and maintenance app. Human intervention is then required only for final online logical confirmation to quickly bind the numbers to locations, eliminating the need for on-site visits.

[0052] Furthermore, for simple photovoltaic power plants, the photovoltaic modules are already equipped with a location framework diagram during the early deployment and construction. Therefore, after the final logical confirmation is performed online to quickly bind the number and location, the precise geographical coordinates of the target photovoltaic module or specific module in step S4 and the binding and storage of identity information and precise geographical coordinates in step S5 mentioned above are not required, making the operation simpler.

[0053] Combination Figures 4 to 7 Taking a string of photovoltaic modules as an example, this paper further introduces the collaborative process of active heating and thermal imaging positioning.

[0054] Before initiating active heating, the image acquisition device uses the visible light channel to acquire visible light images of the target area. Using deep learning instance segmentation algorithms (such as Mask R-CNN), it automatically identifies and extracts the physical boundaries of all photovoltaic modules, generating a static spatial mask grid in the algorithm's memory. Each grid corresponds to one photovoltaic module, and the grid number corresponds one-to-one with the physical location of the photovoltaic module, such as... Figure 4 As shown.

[0055] A string inverter injects a constant current into the DC side, causing the specific string to produce a identifiable thermal characteristic.

[0056] like Figure 5 and Figure 6As shown, after the visible light channel of the image acquisition device captures the bright "heat wave," it performs a spatial intersection-over-union (IoU) operation on the pixel coordinates of the heat wave and the visible light mask grid. When the overlap between a certain column of the mask grid and the heat wave signal exceeds a threshold (e.g., 90%), the logical range of that group of strings is automatically locked.

[0057] like Figure 7 As shown, a one-dimensional spatial vector coordinate axis is established with the starting point as the origin, along the direction of the visible light mask arrangement. On this spatial vector axis, the mask grids covering thermal features are automatically counted starting from the origin. For example, if the third mask needs to be located, when the counter reaches "3", the geometric center pixel coordinates (u, v) of the third grid are extracted. Combining the RTK location information (latitude, longitude, and altitude) recorded in real time by the image acquisition device and the gimbal angle parameters, the pixel coordinates are mapped to real-world WGS84 geographic coordinates using the inverse projection model of the collinearity equation. Finally, the third element of the physical ID (UID) of this string is strongly bound to the parsed latitude and longitude.

[0058] For a simple photovoltaic power station, a semi-automatic, manually-assisted identification method can be used. The acquired thermal and / or visible light images are processed and then used to generate an augmented reality (AR) interface with suggested numbers on the operation and maintenance app. Human intervention is then required only for final logical confirmation online, enabling rapid binding of numbers to locations without needing to visit the site.

[0059] For a large photovoltaic power station, a combination of active heating, image capture, and automatic coordinate calculation can be used to obtain accurate and rapid positioning results. Deep learning models (such as Mask R-CNN) are used to automatically extract the edges of all photovoltaic modules from infrared images and assign them temporary spatial index numbers.

[0060] Using image acquisition devices, high-precision orthophoto maps (DOM) and digital surface models (DSM) of the entire field are pre-generated. All physical supports and photovoltaic module locations of the photovoltaic power station are pre-modeled, and each location is assigned a unique "physical space UID".

[0061] During the active heating inspection process, the real-time resolved geographical coordinates of the photovoltaic modules are matched with the pre-set 3D model for nearest neighbor matching.

[0062] The information of successfully matched components will be automatically filled into the electronic ledger, including: inverter number, string number, component position within the station, precise latitude and longitude, installation tilt angle, and thermal health index during inspection.

[0063] Furthermore, this embodiment not only completes the positioning but also establishes a feedback loop. If an abnormal temperature rise is detected in the "third" photovoltaic module during heating (such as hot spots or no temperature rise), a fault report will be generated simultaneously. This report is directly linked to the aforementioned precise geographic coordinates, guiding maintenance and inspection robots or repair personnel to achieve precise "point-to-point" troubleshooting.

[0064] The core technical idea of ​​this embodiment is: to actively heat the photovoltaic module to generate a unique thermal feature, to use an image acquisition device to collect the thermal image and geographic coordinate information of the photovoltaic module, to identify the thermal feature and match the photovoltaic module number, and finally to establish a correspondence between the number and the location.

[0065] This embodiment utilizes an automated process of "active heating to excite thermal features - thermal imaging acquisition and recognition - automatic calculation and binding" to effectively replace the traditional, inefficient, and cumbersome manual labeling and recording methods. This improves efficiency, shortens the construction and operation and maintenance cycle of photovoltaic power plants, enables unmanned operation in complex and dangerous environments, and significantly reduces labor costs and safety risks.

[0066] This embodiment establishes a unique and reliable correspondence between photovoltaic modules and spatial locations by introducing identifiable thermal features as a medium. This eliminates errors such as misrecording, omissions, and confusion that are unavoidable in manual recording from the source. Compared with traditional methods, the accuracy and stability of the positioning results have achieved a qualitative leap.

[0067] The final output of this embodiment is digital data containing precise geographic coordinates and component identification information. Its format is standard, requiring no secondary input or format conversion, and can be seamlessly connected to the digital platform of photovoltaic power plants, solving the pain point of poor data compatibility in existing technologies.

[0068] The entire positioning process in this embodiment does not require maintenance personnel to enter the photovoltaic array area or perform high-altitude operations. It is completed entirely through remote control, fundamentally avoiding safety hazards such as electric shock, falls, and collisions. It is particularly suitable for complex application scenarios such as large ground power stations, mountain power stations, water surface power stations, and distributed rooftops.

[0069] This embodiment also provides an automatic positioning system for photovoltaic modules based on active heating and thermal imaging, such as... Figure 1 As shown, this system is used to implement the above-mentioned automatic positioning method for photovoltaic modules based on active heating and thermal imaging. The system includes: A heating control module is used to apply a controlled current to a target photovoltaic module or a string containing multiple photovoltaic modules to generate identifiable thermal characteristics. The imaging acquisition module is used to acquire thermal images containing target photovoltaic modules or strings and georeferenced information. The intelligent analysis module is used to process thermal images to identify thermal features and identity information, calculate the precise geographic coordinates of specific components in the target photovoltaic module or string, and bind the identity information to the geographic coordinates.

[0070] Furthermore, the heating control module is configured to send heating commands to the microinverter, string inverter, or maximum power optimizer connected to the target photovoltaic module or string via power line carrier communication or wireless mesh network; The imaging acquisition module includes a visible light camera and an infrared thermal imager; The intelligent parsing module includes: The image processing unit is used to generate a spatial mask grid for photovoltaic modules based on visible light images and to preprocess thermal images. Feature recognition unit, used to identify thermal features and fuse them with spatial mask mesh for analysis; Coordinate calculation unit, used to calculate geographic coordinates based on the inverse projection model of the collinearity equation; The system also includes a database with pre-set three-dimensional geographic information of photovoltaic arrays. The intelligent parsing module is also used to match the calculated geographic coordinates with the three-dimensional geographic information and update the component ledger. The intelligent parsing module is also configured as follows: Generate augmented reality interface data containing thermal and / or visible light images, which is labeled with suggested identity information automatically identified by the system; Receive user confirmation or correction instructions from mobile terminals and update identity information based on the instructions.

[0071] The system workflow in this embodiment is as follows: The heating control module issues heating commands based on the positioning task; The imaging acquisition module acquires thermal images and georeferenced information; The intelligent analysis module processes images, identifies thermal features, and calculates coordinates; Ultimately, the identity information is bound to the coordinates and stored.

[0072] This embodiment also provides a storage medium storing computer program instructions. When these instructions are executed by a processor, they implement the aforementioned automatic positioning method for photovoltaic modules based on active heating and thermal imaging. The storage medium can be any medium capable of storing program code, such as a USB flash drive, portable hard drive, read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0073] The above description is only a preferred embodiment of the present invention and is not intended to limit the design of this case. All equivalent changes made based on the key design features of this case shall fall within the protection scope of this case.

Claims

1. An automatic positioning method for photovoltaic modules based on active heating and thermal imaging, characterized in that: Includes the following steps: S1: Apply a controlled current to the target photovoltaic module or a string containing multiple photovoltaic modules to generate identifiable thermal characteristics in the target photovoltaic module or string. S2: Acquire a thermal image containing the target photovoltaic module or string using an image acquisition device. The thermal image contains georeferenced information. S3: Process thermal images, identify thermal features, and determine the identity information of specific components in the target photovoltaic module or string based on the thermal features; S4: Calculate the precise geographic coordinates of the target photovoltaic module or a specific module based on the thermal features and georeferenced information in the thermal image; S5: Bind and store identity information with precise geographic coordinates.

2. The automatic positioning method for photovoltaic modules based on active heating and thermal imaging as described in claim 1, characterized in that: In step S1, when current is applied to the target photovoltaic module, the controlled current is a pulse current modulated according to a preset coding rule, so that the thermal characteristics form a dynamic thermal signal. In step S3, identifying thermal features includes performing time-domain analysis on the thermal image sequence to decode the identity information corresponding to the preset encoding rules.

3. The automatic positioning method for photovoltaic modules based on active heating and thermal imaging as described in claim 1, characterized in that: In step S1, the same current is applied to the strings; In step S3, thermal features are identified and the identity information of specific components in the string is determined, including: S31: Obtain a visible light image of the target area through the visible light channel of the image acquisition device, and generate a spatial mask grid of multiple photovoltaic modules in the string based on the visible light image; S32: Spatial registration of the thermal image with the spatial mask grid; S33: Determine the overlap between the heated areas in the thermal image and each grid in the spatial mask grid; S34: Based on the overlap exceeding a preset threshold, lock the grid set corresponding to the string that is currently generating heat.

4. The automatic positioning method for photovoltaic modules based on active heating and thermal imaging as described in claim 3, characterized in that: Step S3 also includes: S35: Count the grids in the locked grid set along a preset direction; S36: When the count reaches the target sequence number, the target mesh corresponding to that sequence number is determined as the specific component; In step S4, the precise geographic coordinates of a specific component are calculated based on the location of the target grid.

5. The automatic positioning method for photovoltaic modules based on active heating and thermal imaging as described in claim 1, characterized in that: In step S4, the calculation is based on the inverse projection model of the collinearity equation: ,in, For geographic coordinates, Let be the ray scale factor from the target to the camera's optical center. Let be the rotation transformation matrix of the image acquisition device relative to the world coordinate system. These are the coordinates of the thermal features in the camera coordinate system. This is the position vector of the image acquisition device in the world coordinate system.

6. The automatic positioning method for photovoltaic modules based on active heating and thermal imaging as described in claim 1, characterized in that: The image acquisition device can be any one of the following: a drone equipped with an RTK positioning module and an infrared thermal imager, a fixed dual-lens camera, or a ground mobile robot.

7. The automatic positioning method for photovoltaic modules based on active heating and thermal imaging as described in any one of claims 1, characterized in that: Includes the following steps: S71: Apply a controlled current to a target photovoltaic module or a string containing multiple photovoltaic modules to generate identifiable thermal characteristics in the target photovoltaic module or string. S72: Acquire thermal images containing the target photovoltaic modules or strings using an image acquisition device; S73: Process thermal images, identify thermal features, and display an augmented reality interface containing thermal images and / or visible light images on a mobile terminal. The augmented reality interface includes suggestions for identity information automatically identified by the system. Receive user confirmation or correction instructions for the identity information suggestions and determine the final identity information based on the instructions.

8. An automatic positioning system for photovoltaic modules based on active heating and thermal imaging, characterized in that: This system is used to implement the automatic positioning method for photovoltaic modules based on active heating and thermal imaging as described in any one of claims 1-6, and the system includes: A heating control module is used to apply a controlled current to a target photovoltaic module or a string containing multiple photovoltaic modules to generate identifiable thermal characteristics. The imaging acquisition module is used to acquire thermal images containing target photovoltaic modules or strings and georeferenced information. The intelligent analysis module is used to process thermal images to identify thermal features and identity information, calculate the precise geographic coordinates of specific components in the target photovoltaic module or string, and bind the identity information to the geographic coordinates.

9. The photovoltaic module automatic positioning system based on active heating and thermal imaging as described in claim 8, characterized in that: The heating control module is configured to send heating commands to the micro-inverter, string inverter, or maximum power optimizer connected to the target photovoltaic module or string via power line carrier communication or wireless mesh network. The imaging acquisition module includes a visible light camera and an infrared thermal imager; The intelligent parsing module includes: The image processing unit is used to generate a spatial mask grid for photovoltaic modules based on visible light images and to preprocess thermal images. Feature recognition unit, used to identify thermal features and fuse them with spatial mask mesh for analysis; Coordinate calculation unit, used to calculate geographic coordinates based on the inverse projection model of the collinearity equation; The system also includes a database with pre-set three-dimensional geographic information of photovoltaic arrays. The intelligent parsing module is also used to match the calculated geographic coordinates with the three-dimensional geographic information and update the component ledger. The intelligent parsing module is also configured to: Generate augmented reality interface data containing thermal and / or visible light images, which is labeled with suggested identity information automatically identified by the system; Receive user confirmation or correction instructions from mobile terminals and update identity information based on the instructions.

10. A storage medium, characterized in that: The storage medium stores computer program instructions, which, when executed by a processor, implement an automatic positioning method for photovoltaic modules based on active heating and thermal imaging as described in any one of claims 1-6.

Images