Photovoltaic array deformation monitoring method, system, equipment and medium

Abstract

The invention discloses a photovoltaic array deformation monitoring method, system and device and a medium. The method comprises the following steps: acquiring photovoltaic array information; dividing and marking the photovoltaic array based on the photovoltaic array information; based on the mark, performing image recognition on the photovoltaic array by using an image recognition algorithm to obtain coordinate data of the photovoltaic module; and according to the coordinate data, calculating to obtain a central point coordinate of the photovoltaic module, and comparing the central point coordinate with a standard coordinate to obtain the deformation size of the photovoltaic array. A monitoring device does not need to be installed on each photovoltaic module, the large-scale monitoring cost is greatly reduced, meanwhile, the monitoring efficiency is improved through multi-view image recognition and automatic data processing, hidden dangers such as support deformation and foundation settlement can be found in time, the risks of power generation efficiency reduction and square matrix collapse are effectively avoided, and the method is suitable for popularization and application. And an intelligent and high-precision solution is provided for safe and efficient operation and maintenance of a photovoltaic power station.

Description

Technical Field

[0001] This invention relates to the field of photovoltaic monitoring technology, and in particular to a method, system, equipment and medium for monitoring the deformation of a photovoltaic array. Background Technology

[0002] In recent years, the global energy structure has accelerated its transition to cleaner energy, with the photovoltaic (PV) power generation industry experiencing explosive growth. As the "dual carbon" goals are further implemented, countries are increasing policy support and investment in PV projects, leading to a continuous increase in PV installed capacity. The safe and stable operation of PV power plants is directly related to the reliability and economy of energy supply, and operation and maintenance (O&M), as a key link in ensuring efficient operation, faces unprecedented challenges. PV power plants typically have large land areas, a large number of modules, and complex geographical distribution. Traditional manual inspection methods are inefficient, costly, and unable to achieve real-time dynamic monitoring of massive amounts of equipment. Therefore, there is an urgent need for intelligent and high-precision O&M technologies to improve management efficiency.

[0003] However, the operation and maintenance of photovoltaic power plants currently faces numerous technical bottlenecks. On the one hand, some companies, in their rush to seize market share or obtain preferential electricity prices, neglect engineering quality during project construction, resulting in structural hazards such as non-standard pile foundation construction, unstable support connections, and component installation deviations. These hazards may cause deformation of the photovoltaic support structure during long-term operation, causing the components to deviate from their optimal angle and exhibit abnormal tilt changes, ultimately leading to a decrease in power generation efficiency. On the other hand, many photovoltaic power plants are located in mountainous or hilly areas with complex terrain due to land resource constraints, facing natural risks such as landslides and foundation settlement. Changes in geological conditions can directly lead to structural instability of the photovoltaic array, or even large-scale collapse accidents, seriously threatening the safety of the power plant. Traditional deformation monitoring methods mainly rely on installing sensors on individual components for real-time monitoring. However, for large photovoltaic arrays with tens or even hundreds of thousands of components, this model suffers from high deployment costs, high maintenance difficulties, and poor system scalability, making it difficult to meet the needs of large-scale applications. In addition, traditional monitoring technologies generally suffer from insufficient accuracy and data lag, making it impossible to capture the development trend of minute deformations in a timely manner. This makes it difficult for operation and maintenance personnel to take timely intervention measures, further exacerbating the safety risks of photovoltaic power plants. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a photovoltaic array deformation monitoring method to solve the problem that traditional deformation monitoring devices cannot be applied to photovoltaic arrays with a large number of components, resulting in high cost and low efficiency in photovoltaic array deformation monitoring.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, the present invention provides a method for monitoring the deformation of a photovoltaic array, comprising:

[0008] Obtain photovoltaic array information;

[0009] Based on the photovoltaic array information, the photovoltaic array is divided and marked;

[0010] Based on the aforementioned markings, an image recognition algorithm is used to perform image recognition on the photovoltaic array to obtain the coordinate data of the photovoltaic modules.

[0011] Based on the coordinate data, the coordinates of the center point of the photovoltaic module are calculated and obtained. The coordinates of the center point are compared with the standard coordinates to obtain the magnitude of the deformation of the photovoltaic array.

[0012] As a preferred embodiment of the photovoltaic array deformation monitoring method of the present invention, the photovoltaic array is divided and marked based on the photovoltaic array information, including:

[0013] The photovoltaic array is divided into units of box-shaped units;

[0014] The transformer substation is marked with numbers, and then an image recognition algorithm is used to identify the markings.

[0015] As a preferred embodiment of the photovoltaic array deformation monitoring method described in this invention, the method includes: using an image recognition algorithm to perform image recognition on the photovoltaic array and obtain photovoltaic module coordinate data, including:

[0016] Acquire photovoltaic array images from different angles, perform image recognition on the photovoltaic array images, and obtain the original coordinate data of the four corners of each photovoltaic module;

[0017] By integrating the original coordinate data of the four corners of the photovoltaic module from different shooting angles, a measurement dataset is constructed.

[0018] The beneficial effects of this preferred technical solution are as follows: By acquiring photovoltaic array images from different angles and integrating the original coordinate data, multi-source perspective information can be used to eliminate occlusion or measurement blind spots under a single perspective, improving the integrity and accuracy of the coordinate data. The constructed measurement dataset, combined with the high-precision positioning capability of the UAV, provides spatial geometric constraints for subsequent three-sphere intersection calculations, enabling the calculation accuracy of the photovoltaic module corner coordinates to reach the millimeter level, ensuring the reliability of deformation monitoring.

[0019] As a preferred embodiment of the photovoltaic array deformation monitoring method of the present invention, the method includes: calculating and obtaining the coordinates of the center point of the photovoltaic module, comparing the center point coordinates with standard coordinates to obtain the magnitude of the photovoltaic array deformation, including:

[0020] Based on the shooting location coordinates and the measurement dataset, the coordinate data of the four corners of each photovoltaic module are obtained using the three-sphere intersection method;

[0021] Based on the four corner coordinate data, a first equation is established, and the first equation is solved to obtain the coordinates of the center point of the photovoltaic module.

[0022] The magnitude of the photovoltaic array deformation is obtained by comparing the coordinates of the center point with the standard coordinates.

[0023] The beneficial effects of this preferred technical solution are as follows: Based on the mathematical model established by the principle of three-sphere intersection, it can obtain the spatial coordinates of the photovoltaic module corner points using the three-dimensional coordinates and ranging data from UAVs at different locations. By utilizing redundant calculations from geometric intersection to automatically eliminate abnormal data, the impact of single sensor errors on the results is reduced. Furthermore, it does not rely on complex ground control points, making it suitable for rapid scanning and deformation analysis of large-area photovoltaic arrays.

[0024] As a preferred embodiment of the photovoltaic array deformation monitoring method described in this invention, the first equation is expressed as:

[0025]

[0026] Where (X1,Y1,Z1), (X2,Y2,Z2), and (X3,Y3,Z3) are coordinate data of different shooting angles, (r A1 ,r B1 ,r C1 ,r D1 ), (r A2 ,r B2 ,r C2 ,r D2 ) and (r A3 ,r B3 ,r C3 ,r D3 (x1,y1,z1), (x2,y2,z2), (x3,y3,z3), and (x4,y4,z4) are the coordinates of the four corners of the photovoltaic module.

[0027] As a preferred embodiment of the photovoltaic array deformation monitoring method described in this invention, the coordinates of the center point of the photovoltaic module are obtained as follows:

[0028]

[0029] Where (x1,y1,z1), (x2,y2,z2), (x3,y3,z3), and (x4,y4,z4) are the coordinate data of the four corners of the photovoltaic module.

[0030] As a preferred embodiment of the photovoltaic array deformation monitoring method described in this invention, image recognition includes:

[0031] The image recognition algorithm is the OpenCV algorithm;

[0032] Images of the numbered markings on the transformer substation are collected, and the images are preprocessed using the OpenCV algorithm, including grayscale conversion and noise filtering.

[0033] The contour and character features of the transformer substation number mark are extracted using a feature detection algorithm. The contour and character features are then matched with a preset number mark template in the database to determine the transformer substation number mark.

[0034] The beneficial effects of this preferred technical solution are as follows: the OpenCV algorithm is used to identify the transformer substation number markings, the image quality is improved through grayscale conversion and noise filtering, and the transformer substation number can be identified quickly and accurately by combining feature detection and template matching technology.

[0035] Secondly, the present invention provides a photovoltaic array deformation monitoring system, comprising: a data acquisition module for acquiring photovoltaic array information;

[0036] The division and marking module is used to divide and mark the photovoltaic array based on the photovoltaic array information;

[0037] The image recognition module is used to perform image recognition on the photovoltaic array based on the markers using an image recognition algorithm, and to obtain the coordinate data of the photovoltaic modules.

[0038] The comparison module is used to calculate the coordinates of the center point of the photovoltaic module based on the coordinate data, compare the center point coordinates with the standard coordinates, and obtain the magnitude of the photovoltaic array deformation.

[0039] In a third aspect, the present invention provides an electronic device, comprising:

[0040] Memory and processor;

[0041] The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions, which, when executed by the processor, implement the steps of the photovoltaic array deformation monitoring method.

[0042] Fourthly, the present invention provides a computer-readable storage medium storing computer-executable instructions that, when executed by a processor, implement the steps of the photovoltaic array deformation monitoring method.

[0043] Compared with existing technologies, the advantages of this invention are as follows: This invention divides the photovoltaic array into units of transformer boxes and marks them with numbers. It uses OpenCV image recognition algorithms to locate the transformer boxes, obtains the corner distance data of the photovoltaic modules, and then calculates the coordinates of the module's center point using the three-sphere intersection principle and compares them with standard values, thus realizing deformation monitoring of the photovoltaic array. This invention eliminates the need to install monitoring devices on each photovoltaic module, significantly reducing the cost of large-scale monitoring. At the same time, it improves monitoring efficiency by utilizing multi-view image recognition and automated data processing, enabling timely detection of potential hazards such as support deformation and foundation settlement. This effectively avoids the risk of decreased power generation efficiency and array collapse, providing an intelligent and high-precision solution for the safe and efficient operation and maintenance of photovoltaic power plants. Attached Figure Description

[0044] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0045] Figure 1 This is a schematic diagram of the overall process of a photovoltaic array deformation monitoring method according to an embodiment of the present invention.

[0046] Figure 2 This is a schematic diagram of the photovoltaic array division in an embodiment of the photovoltaic array deformation monitoring method of the present invention.

[0047] Figure 3 This is a schematic diagram of UAV three-sphere positioning for a photovoltaic array deformation monitoring method according to an embodiment of the present invention.

[0048] Figure 4 This is a schematic diagram illustrating the calculation of the photovoltaic module position in a photovoltaic array deformation monitoring method according to an embodiment of the present invention. Detailed Implementation

[0049] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0050] Example 1, referring to Figure 1 As an embodiment of the present invention, a method for monitoring the deformation of a photovoltaic array is provided, comprising:

[0051] S100: Obtain photovoltaic array information;

[0052] S102: Based on the photovoltaic array information, the photovoltaic array is divided and marked;

[0053] S104: Based on the marker, use image recognition algorithms to identify the photovoltaic array and obtain the coordinate data of the photovoltaic modules;

[0054] S106: Based on the coordinate data, calculate and obtain the coordinates of the center point of the photovoltaic module, compare the center point coordinates with the standard coordinates, and obtain the magnitude of the photovoltaic array deformation.

[0055] It should be noted that photovoltaic arrays are large-scale, and acquiring images and coordinate data of the photovoltaic array from different angles for image recognition saves costs and shortens monitoring time. This invention breaks down large-scale photovoltaic arrays into manageable units by dividing and marking them into transformer substations. Combined with image recognition algorithms, it automatically matches the substation numbers, improving the targeting of monitoring. By using multi-source image data to calculate the center point coordinates of photovoltaic modules and comparing them with standard values, it achieves quantitative analysis of photovoltaic array deformation. There is no need to install monitoring devices on each module, reducing hardware costs and deployment complexity. At the same time, through automated data processing, the monitoring cycle is shortened, enabling timely detection of potential hazards such as photovoltaic support deformation and foundation settlement. This provides accurate decision-making basis for power plant operation and maintenance, ensuring the safe and stable operation and power generation efficiency of the photovoltaic array.

[0056] Example 2, refer to Figures 1-4 As an embodiment of the present invention, a method for monitoring the deformation of a photovoltaic array is provided based on the above embodiment.

[0057] In one optional implementation, in step S100, photovoltaic array information can be obtained through photovoltaic power plant design drawings or geographic information system databases, including structured data such as the overall layout of the array, the distribution of transformer locations, and the row and column coordinates of photovoltaic modules, to provide basic geographic information for subsequent division and marking.

[0058] In another optional implementation, in step S100, photovoltaic array information can also be obtained through historical monitoring data ledgers, such as the recorded initial position coordinates of the components and the corresponding relationship of the transformer box numbers, so as to facilitate the quick retrieval of historical benchmark data and comparison and analysis with the current monitoring results.

[0059] In this embodiment of the invention, step S102, which involves dividing and marking the photovoltaic array based on the photovoltaic array information, further includes sub-steps A1-A2:

[0060] A1: Divide the photovoltaic array into units of box-shaped units;

[0061] A2: Mark the transformer with numbers, and then use an image recognition algorithm to identify the markings.

[0062] The photovoltaic (PV) modules generate direct current (DC), which is converted into approximately 400V alternating current (AC) by an inverter. This AC power is then collected at a transformer substation, stepped up to 35kV, and finally transmitted to a substation via collector lines. Generally, the capacity of a PV power station is the sum of the capacities of the transformer substations. Each transformer substation corresponds to a PV array composed of all the PV modules below it, and the transformer substation number matches the PV array number. The PV allocation results are as follows: Figure 2 As shown.

[0063] In one optional implementation, the photovoltaic array can be divided as follows: according to the electrical wiring topology of the photovoltaic power station, the photovoltaic modules connected to the same transformer are divided into an independent array, with the low-voltage side incoming line range of each transformer as the boundary, to ensure the uniqueness of the electrical connection of each array and facilitate the correspondence between subsequent monitoring data and power transmission path.

[0064] In another alternative implementation, the photovoltaic array can be divided as follows: based on geographical distribution characteristics, for photovoltaic fields with complex terrain, arrays are divided according to geographical units such as hillside orientation and flat areas, with each array corresponding to one or more transformer substations. Differentiated monitoring strategies are formulated in combination with terrain characteristics, and the monitoring frequency is increased for arrays in high-risk hillside areas.

[0065] In one optional implementation, the photovoltaic array can be marked by methods such as spraying numbers, affixing reflective signs, or embedding RFID tags. For example, spraying numbers uses weather-resistant paint to spray conspicuous Arabic numerals onto the top of the transformer substation, suitable for typical open-air environments; affixing reflective signs uses an aluminum alloy base plate with reflective film, and the numbering information can be clearly identified by an infrared camera during nighttime drone inspections; embedding RFID tags involves embedding a radio frequency identification chip inside the transformer substation, allowing the RFID reader on the drone to read the numbering information from a distance, suitable for automated identification in harsh environments such as high dust and high humidity.

[0066] It should be noted that by dividing the photovoltaic array into units based on transformer substations, and marking these substations with numbers and using image recognition algorithms to identify the numbers, large-scale photovoltaic arrays are broken down into manageable units corresponding to transformer substations. This simplifies the monitoring targets and reduces system complexity by leveraging the electrical connection characteristics between the transformer substations and photovoltaic modules. Simultaneously, by combining numbering and image recognition, a precise mapping relationship of "transformer substation number - array - module" is established, enabling drones and other equipment to quickly locate target arrays. This facilitates accurate location of abnormal areas by maintenance personnel and allows for joint analysis of electrical operation data. Furthermore, this division and marking method is flexibly adaptable to power plants of different sizes; adding or removing arrays only requires updating the transformer substation number information, enhancing system scalability and reducing subsequent maintenance costs.

[0067] In this embodiment of the invention, step S104, which involves using an image recognition algorithm to perform image recognition on the photovoltaic array based on the markers to obtain the coordinate data of the photovoltaic modules, also includes sub-steps B1-B2:

[0068] B1: Obtain images of the photovoltaic array from different angles, perform image recognition on the photovoltaic array images, and obtain the original coordinate data of the four corners of each photovoltaic module;

[0069] B2: Integrate the original coordinate data of the four corners of the photovoltaic module from different shooting angles to construct a measurement dataset.

[0070] In this embodiment of the invention, a UAV based on BeiDou high-precision positioning is used to perform lidar point cloud scanning: the UAV flies above the transformer substation, identifies the transformer substation number mark through image recognition function, and performs high-precision point cloud scanning of the photovoltaic array corresponding to the transformer substation using the high-precision lidar point cloud scanner. The distance from the UAV to the four corners of each photovoltaic module, marked as A, B, C, and D, is measured and the measurement data is recorded.

[0071] In this embodiment of the invention, the image recognition algorithm is the OpenCV algorithm; images of the numbering marks on the transformer substation are acquired, and the images are preprocessed using the OpenCV algorithm, including grayscale conversion and noise filtering; the contour and character features of the transformer substation numbering marks are extracted using a feature detection algorithm, and the contour and character features are matched with preset numbering mark templates in the database to determine the transformer substation numbering marks.

[0072] Specifically, the camera on the drone captures images of the numbered markings on the top of the transformer. Due to changes in lighting and mechanical vibrations in the shooting environment, the image preprocessing stage sequentially performs grayscale conversion, Gaussian blurring, and adaptive threshold binarization. Converting the RGB image to grayscale reduces color channel interference; filtering high-frequency noise while preserving the outline features of the number; and automatically determining the threshold to generate a binary image ensures that the numbered characters can be clearly separated under different lighting conditions.

[0073] All contours in the image are extracted, and candidate regions are selected based on contour area and aspect ratio. For each candidate region, polygon approximation is performed to obtain the minimum bounding rectangle for tilt correction. Character segmentation uses the vertical projection method, calculating the projection histogram of each column of pixels, and determining character boundaries through valley detection. To improve recognition robustness, morphological processing is performed on the segmented characters to repair broken strokes and eliminate adhesion.

[0074] In the template matching stage, the preprocessed character images are compared with preset standard character templates in the database. The template library contains standard characters of different fonts and sizes, and each character template has its Hu moments, HOG features, and SIFT descriptors pre-extracted. During matching, normalized cross-correlation coefficients, structural similarity indices, and Euclidean distances are calculated simultaneously. A similarity threshold is set for initial screening, and feature voting is performed on candidate characters. Finally, posterior validation is performed based on the context rules of the numbering, and the numbering result with the highest confidence is output.

[0075] In one alternative implementation, the image recognition algorithm can be a deep learning-based convolutional neural network. For example, a pre-trained YOLOv8 or Faster R-CNN model can be used, with customized fine-tuning training for a photovoltaic power station scenario. A large-scale sample library is constructed, collecting numbered images of different types of boxes under various lighting conditions, and data augmentation techniques are used to simulate complex environments. During annotation, polygonal bounding boxes are used to accurately define the numbered regions, and each character is classified and labeled. During the training phase, the bottom convolutional layers of the pre-trained model are frozen, and only the detection head and classification layers are fine-tuned, using the Adam optimizer and a multi-scale training strategy. Upon deployment, the model can quickly scan panoramic images collected by a drone using a sliding window mechanism, outputting the coordinates and confidence scores of the numbered regions, and using a non-maximum suppression algorithm to eliminate overlapping detection boxes. To further improve recognition accuracy, post-processing rules are introduced: valid numbers are confirmed only after three consecutive detections of the same number, and verification is performed using number length and character combination logic.

[0076] In another alternative implementation, the image recognition algorithm can also be a Transformer-based visual model. The photovoltaic array image is divided into fixed-size patches, each of which is flattened and converted into a token sequence through linear projection. The model architecture employs a multi-layer Transformer encoder, with each layer containing a multi-head self-attention mechanism and a feedforward neural network. A global attention mechanism captures long-distance dependencies between patches. To preserve the spatial structure information of the image, learnable positional encoding is introduced, and an absolute positional bias is added during the encoding process. For the transformer substation number recognition task, a dual-branch output head is designed: one branch is used to locate the number region, and the other branch combines OCR technology to achieve character recognition. During training, a hybrid loss function is used, and knowledge distillation technology is introduced to accelerate convergence. To handle perspective changes and occlusion issues, partial occlusion simulations are manually added to the training data, and contrastive learning enhances the model's understanding of incomplete numbering. During deployment, the model can fuse feature maps from different levels through a feature pyramid network to improve the detection capability of small targets. Simultaneously, attention heatmaps are used to visualize the key areas of focus for the model, assisting in manual verification.

[0077] It should be noted that by utilizing the image recognition function on the UAV to accurately identify the transformer substation number markings, and combining it with its BeiDou high-precision positioning capability, the target photovoltaic array can be quickly located. By acquiring images of photovoltaic modules from different angles and extracting the original coordinate data of the corner points, and then integrating multi-source data to construct a measurement dataset, and simultaneously using lidar point cloud scanning to obtain high-precision distance data from the UAV to the corner points of the modules, a multi-dimensional measurement of the spatial position of the photovoltaic modules can be achieved. This method, through the dual constraints of multi-view image fusion and lidar ranging, eliminates the measurement blind spots and errors of a single sensor, enabling the accuracy of the module corner point coordinates to reach the millimeter level. The constructed measurement dataset provides sufficient geometric solution conditions for subsequent three-sphere intersection calculations, and can accurately obtain the coordinates of the module center point and compare them with standard values, thereby quantifying the degree of deformation of the photovoltaic array, effectively improving the accuracy and reliability of deformation monitoring. At the same time, the automated scanning mode based on the UAV significantly shortens the data acquisition time.

[0078] In this embodiment of the invention, step S106, which calculates and obtains the coordinates of the center point of the photovoltaic module based on coordinate data, compares the center point coordinates with the standard coordinates to obtain the magnitude of the photovoltaic array deformation, also includes sub-steps C1-C3:

[0079] C1: Based on the shooting location coordinates and measurement dataset, the coordinate data of the four corners of each photovoltaic module are obtained using the three-sphere intersection method;

[0080] C2: Based on the coordinate data of the four corners, establish the first equation, solve the first equation, and obtain the coordinates of the center point of the photovoltaic module;

[0081] C3: Compare the center point coordinates with the standard coordinates to obtain the magnitude of the photovoltaic array deformation.

[0082] In an embodiment of the present invention, the first equation is expressed as:

[0083]

[0084] Where (X1,Y1,Z1), (X2,Y2,Z2), and (X3,Y3,Z3) are coordinate data of different shooting angles, (r A1 ,r B1 ,r C1 ,r D1 ), (r A2 ,r B2 ,r C2 ,r D2 ) and (r A3 ,r B3 ,r C3 ,r D3(x1,y1,z1), (x2,y2,z2), (x3,y3,z3), and (x4,y4,z4) are the coordinates of the four corners of the photovoltaic module.

[0085] In this embodiment of the invention, the drone's position is adjusted, and three sets of distance measurements are taken. Since the drone's position is known, the coordinates of the four corners of each photovoltaic module can be calculated using the principle of the intersection of three spheres. Figure 3 As shown; the coordinates of point A can be obtained by solving the first equation, and the coordinates of points B, C, and D can be calculated similarly, as follows. Figure 4 As shown.

[0086] In this embodiment of the invention, the coordinates of the center point of the photovoltaic module are represented as follows:

[0087]

[0088] Where (x1,y1,z1), (x2,y2,z2), (x3,y3,z3), and (x4,y4,z4) are the coordinate data of the four corners of the photovoltaic module.

[0089] In one optional implementation, the method for obtaining the magnitude of photovoltaic array deformation can be absolute deformation value analysis. Specifically, the system compares the measured coordinates of the center point of each photovoltaic module with the design standard coordinates point by point, calculates the straight-line distance between the two in three-dimensional space, and uses this as the absolute deformation of the module. To ensure the accuracy of the analysis, the system performs three independent measurements and takes the average value to reduce the error of a single measurement. Subsequently, the calculated absolute deformation is compared with a preset threshold, and modules exceeding the threshold are marked as abnormal modules. To further improve the accuracy of the analysis, the system performs a second verification of abnormal modules by adjusting the measurement angle and increasing the number of measurement points to confirm the authenticity of the deformation. This method establishes a module location database to achieve precise quantification of the deformation degree of each module and generates a three-dimensional deformation distribution map, which intuitively displays the deformation state of each module in the array. Maintenance personnel can use this distribution map to quickly locate abnormal modules, view their historical deformation data, and analyze the deformation development trend.

[0090] In another optional implementation, the method for obtaining the magnitude of photovoltaic array deformation can be regional deformation gradient analysis. Specifically, the system first analyzes the relative positional changes between the center points of adjacent photovoltaic modules, calculates their coordinate differences in three-dimensional space, and forms a deformation gradient matrix within the region. To ensure the accuracy of gradient calculation, the system automatically identifies and excludes abnormal data points caused by measurement errors. Subsequently, by setting a gradient threshold, the system automatically identifies regions with abnormal deformation gradients. To further analyze regional deformation characteristics, the system performs gridding processing on the abnormal regions, calculates the average deformation gradient and direction within each grid, and forms a regional deformation feature vector. This method analyzes the positional correlation between adjacent modules, detects consistent deformation trends caused by regional factors such as foundation settlement and landslides, and generates a regional risk assessment report. The report details the location, deformation characteristics, and development trends of high-risk areas, providing a scientific basis for preventative maintenance. Maintenance personnel can prioritize reinforcement of high-risk areas based on the report content to prevent the problem from escalating.

[0091] It should be noted that this invention divides the photovoltaic array into units of transformer substations and marks them with numbers. It uses OpenCV image recognition algorithms to locate the transformer substations, obtains the corner distance data of the photovoltaic modules, and then calculates the coordinates of the module's center point using the three-sphere intersection principle and compares them with standard values, thus achieving deformation monitoring of the photovoltaic array. This invention eliminates the need to install monitoring devices on each photovoltaic module, significantly reducing the cost of large-scale monitoring. Simultaneously, it improves monitoring efficiency through multi-view image recognition and automated data processing, enabling timely detection of potential hazards such as support deformation and foundation settlement. This effectively avoids the risk of decreased power generation efficiency and array collapse, providing an intelligent and high-precision solution for the safe and efficient operation and maintenance of photovoltaic power plants.

[0092] It should also be noted that the magnitude of deformation can be determined by comparing the coordinates of the center point of each photovoltaic module with the corresponding standard position in the database. Since the high-precision positioning accuracy of UAVs can reach the millimeter level, and the accuracy of lidar point cloud ranging can also reach the millimeter level, the positioning accuracy of the center point of each photovoltaic module can reach the millimeter level.

[0093] Example 3 illustrates a schematic scheme for a photovoltaic array deformation monitoring method. It should be noted that the technical solution of this photovoltaic array deformation monitoring system belongs to the same concept as the aforementioned photovoltaic array deformation monitoring method. Details not described in detail in this embodiment can be found in the description of the aforementioned photovoltaic array deformation monitoring method.

[0094] This embodiment also provides a photovoltaic array deformation monitoring system, including:

[0095] The data acquisition module is used to acquire information about the photovoltaic array.

[0096] The partitioning and marking module is used to partition and mark the photovoltaic array based on the photovoltaic array information;

[0097] The image recognition module is used to perform image recognition on the photovoltaic array based on the markers and to obtain the coordinate data of the photovoltaic modules.

[0098] The comparison module is used to calculate and obtain the center point coordinates of the photovoltaic module based on the coordinate data, compare the center point coordinates with the standard coordinates, and obtain the magnitude of the photovoltaic array deformation.

[0099] This embodiment also provides an electronic device suitable for photovoltaic array deformation monitoring, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to realize the photovoltaic array deformation monitoring method proposed in the above embodiment.

[0100] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, it implements the photovoltaic array deformation monitoring method proposed in the above embodiments.

[0101] The storage medium proposed in this embodiment and the photovoltaic array deformation monitoring method proposed in the above embodiments belong to the same inventive concept. Technical details not described in detail in this embodiment can be found in the above embodiments, and this embodiment has the same beneficial effects as the above embodiments.

[0102] Based on the above description of the implementation methods, those skilled in the art will clearly understand that the present invention can be implemented using software and necessary general-purpose hardware, and of course, it can also be implemented using hardware. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk, or optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of the various embodiments of the present invention.

[0103] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for monitoring the deformation of a photovoltaic array, characterized in that, include: Obtain photovoltaic array information; Based on the photovoltaic array information, the photovoltaic array is divided and marked; Based on the aforementioned markings, an image recognition algorithm is used to perform image recognition on the photovoltaic array to obtain the coordinate data of the photovoltaic modules. Based on the coordinate data, the coordinates of the center point of the photovoltaic module are calculated and obtained. The coordinates of the center point are compared with the standard coordinates to obtain the magnitude of the deformation of the photovoltaic array.

2. The photovoltaic array deformation monitoring method as described in claim 1, characterized in that, Based on the photovoltaic array information, the photovoltaic array is divided and marked, including: The photovoltaic array is divided into units of box-shaped units; The transformer substation is marked with numbers, and then an image recognition algorithm is used to identify the markings.

3. The photovoltaic array deformation monitoring method as described in claim 2, characterized in that, Image recognition algorithms are used to perform image recognition on the photovoltaic array to obtain the coordinate data of the photovoltaic modules, including: Acquire photovoltaic array images from different angles, perform image recognition on the photovoltaic array images, and obtain the original coordinate data of the four corners of each photovoltaic module; By integrating the original coordinate data of the four corners of the photovoltaic module from different shooting angles, a measurement dataset is constructed.

4. The photovoltaic array deformation monitoring method as described in claim 3, characterized in that, Calculate and obtain the coordinates of the center point of the photovoltaic module, compare the center point coordinates with standard coordinates, and obtain the magnitude of the photovoltaic array deformation, including: Based on the shooting location coordinates and the measurement dataset, the coordinate data of the four corners of each photovoltaic module are obtained using the three-sphere intersection method; Based on the four corner coordinate data, a first equation is established, and the first equation is solved to obtain the coordinates of the center point of the photovoltaic module. The magnitude of the photovoltaic array deformation is obtained by comparing the coordinates of the center point with the standard coordinates.

5. The photovoltaic array deformation monitoring method as described in claim 4, characterized in that, The first equation is expressed as: Where (X1,Y1,Z1), (X2,Y2,Z2), and (X3,Y3,Z3) are coordinate data of different shooting angles, (r A1 ,r B1 ,r C1 ,r D1 ), (r A2 ,r B2 ,r C2 ,r D2 ) and (r A3 ,r B3 ,r C3 ,r D3 (x1,y1,z1), (x2,y2,z2), (x3,y3,z3), and (x4,y4,z4) are the coordinates of the four corners of the photovoltaic module.

6. The photovoltaic array deformation monitoring method as described in claim 2, characterized in that, The coordinates of the center point of the photovoltaic module are represented as follows: Where (x1,y1,z1), (x2,y2,z2), (x3,y3,z3), and (x4,y4,z4) are the coordinate data of the four corners of the photovoltaic module.

7. The photovoltaic array deformation monitoring method as described in claim 5, characterized in that, Image recognition includes: The image recognition algorithm is the OpenCV algorithm; Images of the numbered markings on the transformer substation are collected, and the images are preprocessed using the OpenCV algorithm, including grayscale conversion and noise filtering. The contour and character features of the transformer substation number mark are extracted using a feature detection algorithm. The contour and character features are then matched with a preset number mark template in the database to determine the transformer substation number mark.

8. A photovoltaic array deformation monitoring system, using the method described in any one of claims 1-7, characterized in that, include: The data acquisition module is used to acquire information about the photovoltaic array. The division and marking module is used to divide and mark the photovoltaic array based on the photovoltaic array information; The image recognition module is used to perform image recognition on the photovoltaic array based on the markers using an image recognition algorithm, and to obtain the coordinate data of the photovoltaic modules. The comparison module is used to calculate the coordinates of the center point of the photovoltaic module based on the coordinate data, compare the center point coordinates with the standard coordinates, and obtain the magnitude of the photovoltaic array deformation.

9. An electronic device, characterized in that, include: Memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, they implement the steps of the photovoltaic array deformation monitoring method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, It stores computer-executable instructions, which, when executed by a processor, implement the steps of the photovoltaic array deformation monitoring method according to any one of claims 1 to 7.

Images