Photovoltaic array operation state monitoring method

Abstract

The invention provides a photovoltaic array operation state monitoring method, which comprises the steps of acquiring an infrared original image of a photovoltaic array, preprocessing the infrared original image to capture hot spot image features, and identifying the image features through a migration network to extract and fuse multi-scale feature information in the infrared original image. Obtaining a target hot spot image according to the fused multi-scale feature information; segmenting the target hot spot image through a semantic segmentation algorithm to obtain a segmented image, and detecting the segmented image according to an edge detection algorithm to obtain a hot spot edge definition image; identifying a hot spot edge defining image, calculating an image pixel difference value to obtain a pixel number proportion of a final hot spot, evaluating the health state of the photovoltaic array according to the pixel number proportion, and monitoring the running state of the photovoltaic array according to the health state of the photovoltaic array; according to the invention, the detection efficiency is improved, and the detection of a large-scale photovoltaic power station in the industry is facilitated.

Description

Technical Field

[0001] The present invention relates to the technical field of photovoltaic array operating status monitoring, and in particular to a photovoltaic array operating status monitoring method. Background Art

[0002] Hot spots are areas of abnormally high temperature within a photovoltaic array that can damage cells or other components. Hot spots can arise from a variety of factors. When a solar cell is shaded or damaged, causing it to operate in reverse, it consumes energy rather than generates it, leading to overheating in that area. Cell mismatch or interconnect failures can lead to uneven current distribution, placing excessive load on certain cells. Diode failures can block reverse current protection, increasing the thermal burden on local cells. Furthermore, structural issues such as cell cracks, weld defects, encapsulation damage, or improper module design can increase local resistance and cause overheating. Partial shading, such as that caused by trees, buildings, or snow, can also cause power reduction in certain areas of a cell, increasing the risk of hot spots. A photovoltaic array consists of multiple modules, and these factors can lead to uneven temperature distribution within the module, ultimately causing hot spots that damage cell performance. Hot spots are one of the most dangerous faults in photovoltaic systems, damaging the cells or encapsulation materials and potentially completely destroying the solar cell in a short period of time, seriously impacting the performance and lifespan of the module. If the photovoltaic system is not monitored and controlled promptly and the temperature rises to a certain level, it can cause a fire, resulting in serious consequences.

[0003] In the existing technology, photovoltaic array surface defect detection methods mainly include manual visual inspection and traditional image optimization methods. Manual visual inspection is inefficient and cannot meet the quality monitoring needs of large-scale photovoltaic power stations; traditional image optimization methods are insensitive to data, have poor feature interpretation capabilities, and have low detection speeds, which cannot meet actual requirements in the industry. Summary of the Invention

[0004] Based on this, the purpose of the present invention is to provide a photovoltaic array operating status monitoring method to solve the technical problem that the photovoltaic array surface detection method in the existing technology has low detection efficiency and cannot meet the detection needs of large-scale photovoltaic power stations in industry.

[0005] In one aspect, the present invention provides a method for monitoring the operating status of a photovoltaic array, comprising: Acquiring an infrared raw image of a photovoltaic array and preprocessing it to capture hot spot image features, including hot spot color and hot spot shape; identifying the image features through a migration network to extract and fuse multi-scale feature information from the infrared raw image; enhancing the image features based on the fused multi-scale feature information to obtain a target hot spot image based on the enhanced image features; the multi-scale feature information including scale, shape, and edge; Performing hot spot contour and shape segmentation on the target hot spot image using a semantic segmentation algorithm to obtain a segmented image, and detecting the segmented image using an edge detection algorithm to extract hot spot edge pixels to obtain a hot spot edge definition image; Identifying the hot spot edge delimiting image and calculating the image pixel difference to obtain a final hot spot pixel number ratio, obtaining an average hot spot area ratio of the photovoltaic array based on the pixel number ratio, evaluating the photovoltaic array health status based on the photovoltaic array hot spot area ratio average, and monitoring the photovoltaic array operation status based on the photovoltaic array health status; The step of enhancing image features according to the fused multi-scale feature information to obtain a target hot spot image according to the enhanced image features includes: Acquire the enhanced image features and identify them to obtain contour pixels, wherein the contour pixels include main pixels and divergent pixels, and the plurality of main pixels form a main contour, identify the divergent pixels and obtain the shortest distance between the divergent pixels and the main contour; Determining whether the shortest distance is less than a distance threshold; If so, the divergent pixels are updated to the main pixels and the main contour is updated according to the updated main pixels, and the target hot spot image is obtained according to the updated main contour; If not, the divergent pixel is an invalid pixel.

[0006] The above-mentioned photovoltaic array operation status monitoring method first obtains the target hot spot image, and then uses the semantic segmentation algorithm to segment the target hot spot image according to the hot spot contour and shape to obtain a segmented image, and then detects the segmented image according to the edge detection algorithm to extract the hot spot edge pixels and then obtain the hot spot edge definition image, thereby realizing automatic recognition of the infrared original image and hot spot image extraction and definition, thereby improving detection efficiency; furthermore, the photovoltaic array health status is evaluated according to the hot spot edge definition image, so as to monitor the photovoltaic array operation status according to the photovoltaic array health status, realize automatic monitoring of the photovoltaic array, and thus facilitate the detection of large-scale photovoltaic power stations in industry; and solves the technical problem that the photovoltaic array surface detection method in the existing technology has low detection efficiency and cannot meet the detection needs of large-scale photovoltaic power stations in industry.

[0007] In addition, the photovoltaic array operating status monitoring method according to the present invention may also have the following additional technical features: Furthermore, the step of identifying diverging pixels and obtaining the shortest distance between the diverging pixels and the subject contour includes: Get the number of pixels of the diverging pixels; Determine whether the number of pixels is less than a number threshold; If the number of pixels is less than the number threshold, identifying the diverging pixels and obtaining the shortest distance between the diverging pixels and the outline of the subject; If the number of pixels is not less than the number threshold, then the divergent pixels are identified to obtain the distribution characteristics of all the divergent pixels; Determine whether the divergent pixels are regularly distributed based on the distribution characteristics; If the divergent pixel points are regularly distributed, the divergent pixel points are connected in sequence to construct a secondary contour, and the two end pixel points of the secondary contour are obtained, where the two end pixel points include a starting divergent pixel point and an ending divergent pixel point. According to the starting divergent pixel point and the ending divergent pixel point, the steps of identifying the divergent pixel points and obtaining the shortest distance between the divergent pixel points and the main contour are performed.

[0008] Furthermore, in the step of determining whether the divergent pixels are regularly distributed according to the distribution characteristics, the method includes: Obtaining a position feature and a distance between adjacent pixels of each diverging pixel point, constructing a position database based on the position feature, and clustering the diverging pixels in the position database based on the distance between adjacent pixels to obtain a plurality of unit pixel blocks, each unit pixel block including a plurality of diverging pixels; Each unit pixel block is set as a mass point to obtain a plurality of mass points, and the position database is updated according to the plurality of mass points to obtain a mass point database; Identify multiple particles in a particle database to obtain particle positions, obtain particle distribution characteristics based on the particle positions to obtain edge particles in the particle database, and construct a particle database outline based on the edge particles; It is determined whether the particles are regularly distributed according to the particle database outline to determine whether the divergent pixels are regularly distributed.

[0009] Furthermore, in the step of determining whether the particles are regularly distributed according to the particle database profile to determine whether the divergent pixels are regularly distributed, the regular distribution includes a convergent distribution, and the method includes: Identify a contour of a particle database to obtain end particles and middle particles, and construct a contour centerline based on the end particles and along the middle particles, wherein the end particles determine the starting direction and the ending direction of the contour centerline, and the middle particles determine the shape of the contour centerline; Determine whether the contour midline is a regular line segment, where the regular line segment includes an arc and a straight line; If the contour midline is a regular line segment, the particles are convergently distributed relative to the regular line segment, and the divergent pixels are convergently distributed relative to the regular line segment.

[0010] Furthermore, the step of constructing a contour midline based on the end mass points and along the middle mass points includes: Constructing a preliminary centerline of the particle database contour based on the end particles; Determine whether the number of particles distributed on both sides of the preliminary center line is balanced; If so, the preliminary centerline is the contour centerline; If not, the preliminary center line is adjusted according to the difference in the number of particles on both sides until the number of particles on both sides of the preliminary center line is balanced.

[0011] Furthermore, after the step of determining whether the divergent pixels are regularly distributed according to the distribution characteristics, the method further includes: If the divergent pixels are not regularly distributed, the process returns to executing the step of identifying the image features through the migration network to extract and fuse the multi-scale feature information in the original infrared image.

[0012] Furthermore, in the step of identifying the hot spot edge delimiting image and calculating the image pixel difference to obtain the final hot spot pixel quantity ratio, the calculation formula of the pixel quantity ratio is: ; Where, S Indicates the proportion of pixels with hot spots in the photovoltaic array; n Indicates the number of vertices of the photovoltaic array hot spot; ( x i , y i ) represents the coordinates of the i-th vertex, ( x n , y n ) = ( x 0 , y 0).

[0013] Furthermore, in the step of evaluating the health status of the photovoltaic array according to the average value of the hot spot area ratio of the photovoltaic array, the photovoltaic array health status evaluation formula is: ; Where, G It represents the average fraction of the hot spot area of ​​the photovoltaic array collected multiple times; N Indicates the number of times data is collected from the photovoltaic array; S i 、 S Ti Indicates the i The proportion of hot spot pixels in the first data collection and the i The proportion of the number of photovoltaic array pixels.

[0014] Another aspect of the present invention provides a photovoltaic array operating status monitoring system, the system comprising: an acquisition module, configured to acquire an infrared raw image of the photovoltaic array and perform preprocessing to capture hot spot image features, wherein the image features include hot spot color and hot spot shape; identify the image features through a migration network to extract and fuse multi-scale feature information from the infrared raw image; enhance the image features based on the fused multi-scale feature information to obtain a target hot spot image based on the enhanced image features; the multi-scale feature information includes scale, shape, and edge; A segmentation module is used to segment the target hot spot image by hot spot contour and shape using a semantic segmentation algorithm to obtain a segmented image, and detect the segmented image using an edge detection algorithm to extract hot spot edge pixels to obtain a hot spot edge definition image; a monitoring module, configured to identify the hot spot edge delimiting image, calculate the image pixel difference to obtain a final hot spot pixel number ratio, obtain an average hot spot area ratio of the photovoltaic array based on the pixel number ratio, evaluate the photovoltaic array health status based on the photovoltaic array hot spot area ratio average, and monitor the photovoltaic array operation status based on the photovoltaic array health status; The step of enhancing image features according to the fused multi-scale feature information to obtain a target hot spot image according to the enhanced image features includes: Acquire the enhanced image features and identify them to obtain contour pixels, wherein the contour pixels include main pixels and divergent pixels, and the plurality of main pixels form a main contour, identify the divergent pixels and obtain the shortest distance between the divergent pixels and the main contour; Determining whether the shortest distance is less than a distance threshold; If so, the divergent pixels are updated to the main pixels and the main contour is updated according to the updated main pixels, and the target hot spot image is obtained according to the updated main contour; If not, the divergent pixel is an invalid pixel.

[0015] Another aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, which implements the photovoltaic array operating status monitoring method as described above when the program is executed by a processor.

[0016] Another aspect of the present invention provides a data processing device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the photovoltaic array operating status monitoring method as described above when executing the program. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 Flowchart of a photovoltaic array operating status monitoring method according to a first embodiment of the present invention; Figure 2 This is a flowchart of the detailed steps of step S103 in the first embodiment of the present invention; The following specific embodiments will further illustrate the present invention in conjunction with the above-mentioned drawings. DETAILED DESCRIPTION

[0018] To facilitate understanding of the present invention, the present invention will be described more fully below with reference to the accompanying drawings. The drawings illustrate several embodiments of the present invention. However, the present invention may be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to provide a more thorough and comprehensive understanding of the present invention.

[0019] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one skilled in the art to which this invention pertains. The terms used in this specification of the present invention are for the purpose of describing specific embodiments only and are not intended to limit the present invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0020] In order to solve the technical problem that the photovoltaic array surface detection method in the existing technology has low detection efficiency and cannot meet the detection needs of large-scale photovoltaic power stations in industry, the present application provides a photovoltaic array operation status monitoring method, which first obtains the target hot spot image, and then uses a semantic segmentation algorithm to segment the hot spot contour and shape of the target hot spot image to obtain a segmented image, and then detects the segmented image according to the edge detection algorithm to extract the hot spot edge pixels and then obtain a hot spot edge definition image, thereby realizing automatic recognition of the infrared original image and extraction and definition of the hot spot image, thereby improving the detection efficiency; furthermore, the health status of the photovoltaic array is evaluated according to the hot spot edge definition image, so as to monitor the operation status of the photovoltaic array according to the health status of the photovoltaic array, realize automatic monitoring of the photovoltaic array, and thus facilitate the detection of large-scale photovoltaic power stations in industry.

[0021] To facilitate understanding of the present invention, several embodiments of the present invention are provided below. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to provide a more thorough and comprehensive disclosure of the present invention.

[0022] Example 1 See also Figure 1 , which shows a photovoltaic array operating status monitoring method in a first embodiment of the present invention, the method includes steps S101 to S105: S101 , obtaining an original infrared image of a photovoltaic array and performing preprocessing to capture hot spot image features, where the image features include hot spot color and hot spot shape.

[0023] S102 , identifying image features through a migration network to extract and fuse multi-scale feature information in the original infrared image, where the multi-scale feature information includes scale, shape, and edge.

[0024] In this example, the EfficientNetB7 migration network is used to identify image features. EfficientNetB7 is a deep neural network consisting of seven convolutional stages. The network first performs preliminary processing on the input infrared image of photovoltaic modules through the initial convolutional layer to capture image features such as the color and shape of hot spots. The data then flows through the MBConv modules in each stage. The MBConv modules have three different network structures: Module 1, Module 2, and Module 3. Each module is composed of three of the five submodules. These modules have variable repetition times, number of filters, and step sizes, enabling flexible extraction and fusion of multi-scale feature information (such as scale, shape, and edges) from infrared images.

[0025] S103 , enhancing image features according to the fused multi-scale feature information to obtain a target hot spot image according to the enhanced image features.

[0026] In this example, see Figure 2 , step S103 specifically includes steps S1031 to S1035: S1031 , acquiring enhanced image features and performing recognition to obtain contour pixels, where the contour pixels include main pixel points and divergent pixel points, and a plurality of main pixel points form a main contour.

[0027] In this embodiment, to more accurately locate the image contour, pixel recognition is performed on the identified image features and the recognition results are divided to obtain main pixels and divergent pixels based on the division results. In this embodiment, main pixels are pixels that can construct the main contour of the image, while divergent pixels are pixels other than the main pixels. To avoid the situation where the main contour is missing due to a large difference between the main contour and the actual contour due to incomplete recognition of divergent pixels, this embodiment requires a special analysis of the divergent pixels to determine whether the main contour needs to be updated based on the analysis results.

[0028] S1032: Identify diverging pixels and obtain the shortest distance between the diverging pixels and the subject contour.

[0029] As a specific example, with the diverging pixel point as the center and an initial radius preset, a circle is drawn according to the initial radius to obtain a diverging point circle. If the diverging point circle does not intersect with the main body outline, the radius is increased to update the initial radius, and the process returns to the step of drawing a circle according to the initial radius to obtain a diverging point circle. This process continues until the diverging point circle intersects with the main body outline and this is the initial intersection point. The initial intersection point is then obtained, and the distance between the initial intersection point and the diverging pixel point is taken as the shortest distance between the diverging pixel point and the main body outline. Specifically, to improve calculation accuracy, the preset value of the initial radius can be a unit pixel value.

[0030] S1033: Determine whether the shortest distance is less than a distance threshold.

[0031] If yes, go to step S1034; if no, go to step S1035; S1034: Update the divergent pixels to the main pixel points and update the main contour according to the updated main pixel points, and obtain the target hot spot image according to the updated main contour.

[0032] S1035: The divergent pixel is an invalid pixel.

[0033] In this embodiment, step S1032 specifically includes: Obtain the number of diverging pixels; determine whether the number of pixels is less than a number threshold; if the number of pixels is less than the number threshold, identify the diverging pixels and obtain the shortest distance between the diverging pixels and the main body contour; if the number of pixels is not less than the number threshold, identify the diverging pixels to obtain the distribution characteristics of all the diverging pixels; determine whether the diverging pixels are regularly distributed based on the distribution characteristics; if the diverging pixels are regularly distributed, sequentially connect the diverging pixels to construct a secondary contour, and obtain the two end pixels of the secondary contour, the two end pixels including the starting diverging pixel and the ending diverging pixel, and perform the steps of identifying the diverging pixels and obtaining the shortest distance between the diverging pixels and the main body contour based on the starting diverging pixel and the ending diverging pixel; if the diverging pixels are not regularly distributed, return to step S102. It should be further explained that the distance threshold and the number threshold in this embodiment can be selected according to actual conditions and are not specifically limited here.

[0034] Furthermore, in the step of judging whether the divergent pixel points are regularly distributed according to the distribution characteristics, the photovoltaic array operation status monitoring method also includes: obtaining the position characteristics of each divergent pixel point and the distance between adjacent pixels, constructing a position database according to the position characteristics, clustering the divergent pixel points in the position database according to the distance between adjacent pixels to obtain multiple unit pixel blocks, each unit pixel block includes multiple divergent pixel points; setting each unit pixel block to a particle to obtain multiple particles, updating the position database according to the multiple particles to obtain a particle database; identifying multiple particles in the particle database to obtain particle positions, obtaining particle distribution characteristics according to the particle positions to obtain edge particles of the particle database, and constructing a particle database outline according to the edge particles; judging whether the particles are regularly distributed according to the particle database outline to judge whether the divergent pixel points are regularly distributed.

[0035] In this embodiment, in order to improve the processing efficiency of the system, the divergent pixel points are clustered to obtain multiple unit pixel blocks, each unit pixel block includes multiple divergent pixel points; each unit pixel block is set to a corresponding particle to obtain multiple particles. By analyzing the particles, the analysis amount is reduced, thereby improving the analysis efficiency of the system.

[0036] Furthermore, in the step of judging whether the particles are regularly distributed according to the particle database contour to judge whether the divergent pixel points are regularly distributed, the regular distribution includes a convergent distribution, and the method also includes: identifying the particle database contour to obtain end particles and middle particles, constructing a contour centerline according to the end particles and along the middle particles, wherein the end particles determine the starting direction and the end direction of the contour centerline, and the middle particles determine the shape of the contour centerline; judging whether the contour centerline is a regular line segment, and regular line segments include arcs and straight lines; if the contour centerline is a regular line segment, the particles are convergently distributed relative to the regular line segment, and then the divergent pixel points are convergently distributed relative to the regular line segment.

[0037] Furthermore, the step of constructing a contour centerline based on the end particles and along the middle particles includes: constructing a preliminary centerline of the particle database contour based on the end particles; judging whether the number of particles distributed on both sides of the preliminary centerline by the middle particles is balanced; if the number of particles distributed on both sides of the preliminary centerline by the middle particles is balanced, the preliminary centerline is the contour centerline; if the number of particles distributed on both sides of the preliminary centerline by the middle particles is unbalanced, adjusting the preliminary centerline according to the difference in the number of particles on both sides until the number of particles on both sides of the preliminary centerline is balanced.

[0038] S104, performing hot spot contour and shape segmentation on the target hot spot image using a semantic segmentation algorithm to obtain a segmented image, detecting the segmented image using an edge detection algorithm to extract hot spot edge pixels and thereby obtain a hot spot edge definition image.

[0039] Based on the above-mentioned EfficientNetB7 migration network, in order to achieve pixel-level segmentation of the hot spot contour and shape of the photovoltaic array, this application uses the DeepLabV3+ semantic segmentation method to perform hot spot contour and shape segmentation on the target hot spot image to obtain a segmented image.

[0040] S105. Identify the hot spot edge delimiting image and calculate the image pixel difference to obtain the final hot spot pixel number ratio, obtain the average hot spot area ratio of the photovoltaic array based on the pixel number ratio, evaluate the photovoltaic array health status based on the average hot spot area ratio of the photovoltaic array, and monitor the photovoltaic array operation status based on the photovoltaic array health status.

[0041] Based on the above-mentioned photovoltaic array hot spot pixel-level segmentation results, in this embodiment, a hot spot edge extraction method based on the Canny algorithm is used to achieve hot spot edge definition. Specifically, the calculation formula for the pixel number ratio is: ; Where, S Indicates the proportion of pixels with hot spots in the photovoltaic array; n Indicates the number of vertices of the photovoltaic array hot spot; ( x i , y i ) represents the coordinates of the i-th vertex, ( x n , y n ) = ( x 0 , y 0).

[0042] Based on the pixel numerical calculation within the image edge, the hot spot status of the photovoltaic modules in the photovoltaic array is further quantitatively evaluated. The average of the hot spot area ratio of the photovoltaic modules in N data sets is taken to evaluate the health status of the photovoltaic array. In this embodiment, when the following G value is not less than 95 points, the photovoltaic array is in a healthy state. Furthermore, the photovoltaic array health status evaluation formula is as follows: ; Where, G It represents the average fraction of the hot spot area of ​​the photovoltaic array collected multiple times; N Indicates the number of times data is collected from the photovoltaic array; S i 、 S Ti Indicates the i The proportion of hot spot pixels in the first data collection and the i The proportion of the number of photovoltaic array pixels.

[0043] In summary, the photovoltaic array operation status monitoring method in the above-mentioned embodiment of the present invention first obtains the target hot spot image, and then uses the semantic segmentation algorithm to segment the target hot spot image according to the hot spot contour and shape to obtain a segmented image, and then detects the segmented image according to the edge detection algorithm to extract the hot spot edge pixels and then obtain the hot spot edge definition image, thereby realizing automatic recognition of the infrared original image and extraction and definition of the hot spot image, thereby improving the detection efficiency; furthermore, the photovoltaic array health status is evaluated according to the hot spot edge definition image, so as to monitor the photovoltaic array operation status according to the photovoltaic array health status, realize automatic monitoring of the photovoltaic array, and thus facilitate the detection of large-scale photovoltaic power stations in industry; and solves the technical problem that the photovoltaic array surface detection method in the prior art has low detection efficiency and cannot meet the detection needs of large-scale photovoltaic power stations in industry.

[0044] Example 2 A second embodiment of the present invention provides a photovoltaic array operating status monitoring system, the system comprising: an acquisition module, configured to acquire an infrared raw image of the photovoltaic array and perform preprocessing to capture hot spot image features, wherein the image features include hot spot color and hot spot shape; identify the image features through a migration network to extract and fuse multi-scale feature information from the infrared raw image; enhance the image features based on the fused multi-scale feature information to obtain a target hot spot image based on the enhanced image features; the multi-scale feature information includes scale, shape, and edge; A segmentation module is used to segment the target hot spot image by hot spot contour and shape using a semantic segmentation algorithm to obtain a segmented image, and detect the segmented image using an edge detection algorithm to extract hot spot edge pixels to obtain a hot spot edge definition image; a monitoring module, configured to identify the hot spot edge delimiting image, calculate the image pixel difference to obtain a final hot spot pixel number ratio, obtain an average hot spot area ratio of the photovoltaic array based on the pixel number ratio, evaluate the photovoltaic array health status based on the photovoltaic array hot spot area ratio average, and monitor the photovoltaic array operation status based on the photovoltaic array health status; The step of enhancing image features according to the fused multi-scale feature information to obtain a target hot spot image according to the enhanced image features includes: Acquire the enhanced image features and identify them to obtain contour pixels, wherein the contour pixels include main pixels and divergent pixels, and the plurality of main pixels form a main contour, identify the divergent pixels and obtain the shortest distance between the divergent pixels and the main contour; Determining whether the shortest distance is less than a distance threshold; If so, the divergent pixels are updated to the main pixels and the main contour is updated according to the updated main pixels, and the target hot spot image is obtained according to the updated main contour; If not, the divergent pixel is an invalid pixel.

[0045] In summary, the photovoltaic array operation status monitoring system in the above-mentioned embodiment of the present invention first obtains the target hot spot image, and then uses the semantic segmentation algorithm to segment the target hot spot image according to the hot spot contour and shape to obtain a segmented image, and then detects the segmented image according to the edge detection algorithm to extract the hot spot edge pixels and then obtain the hot spot edge definition image, thereby realizing automatic recognition of the infrared original image and extraction and definition of the hot spot image, thereby improving the detection efficiency; furthermore, the photovoltaic array health status is evaluated according to the hot spot edge definition image, so as to monitor the photovoltaic array operation status according to the photovoltaic array health status, realize automatic monitoring of the photovoltaic array, and thus facilitate the detection of large-scale photovoltaic power stations in industry; and solves the technical problem that the photovoltaic array surface detection method in the prior art has low detection efficiency and cannot meet the detection needs of large-scale photovoltaic power stations in industry.

[0046] In addition, an embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, which implements the steps of the method in the above embodiment when the program is executed by a processor.

[0047] In addition, an embodiment of the present invention further provides a data processing device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method in the above embodiment when executing the program.

[0048] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0049] More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic devices), a portable computer disk cartridge (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and programmable read-only memory (EPROM or flash memory), fiber optic devices, and portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting, or processing it in another suitable manner as necessary, and then storing it in a computer memory.

[0050] It should be understood that various components of the present invention may be implemented using hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods may be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof may be used: a discrete logic circuit having logic gate circuits for implementing logic functions on data signals, an application-specific integrated circuit having suitable combinational logic gate circuits, a programmable gate array (PGA), a field-programmable gate array (FPGA), etc.

[0051] Throughout this specification, reference to terms such as "one embodiment," "some embodiments," "examples," "specific examples," or "some examples" means that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, schematic representations of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0052] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to the embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the claims and their equivalents.

Claims

1. A method for monitoring the operating status of a photovoltaic array, characterized in that: include: Acquiring an infrared raw image of a photovoltaic array and preprocessing it to capture hot spot image features, including hot spot color and hot spot shape; identifying the image features through a migration network to extract and fuse multi-scale feature information from the infrared raw image; enhancing the image features based on the fused multi-scale feature information to obtain a target hot spot image based on the enhanced image features; the multi-scale feature information including scale, shape, and edge; Performing hot spot contour and shape segmentation on the target hot spot image using a semantic segmentation algorithm to obtain a segmented image, and detecting the segmented image using an edge detection algorithm to extract hot spot edge pixels to obtain a hot spot edge definition image; Identifying the hot spot edge delimiting image and calculating the image pixel difference to obtain a final hot spot pixel number ratio, obtaining an average hot spot area ratio of the photovoltaic array based on the pixel number ratio, evaluating the photovoltaic array health status based on the photovoltaic array hot spot area ratio average, and monitoring the photovoltaic array operation status based on the photovoltaic array health status; The step of enhancing image features according to the fused multi-scale feature information to obtain a target hot spot image according to the enhanced image features includes: Acquire the enhanced image features and identify them to obtain contour pixels, wherein the contour pixels include main pixels and divergent pixels, and the plurality of main pixels form a main contour, identify the divergent pixels and obtain the shortest distance between the divergent pixels and the main contour; Determining whether the shortest distance is less than a distance threshold; If so, the divergent pixels are updated to the main pixels and the main contour is updated according to the updated main pixels, and the target hot spot image is obtained according to the updated main contour; If not, the divergent pixel is an invalid pixel.

2. The photovoltaic array operating status monitoring method according to claim 1, characterized in that: The steps of identifying diverging pixels and obtaining the shortest distance between the diverging pixels and the subject contour include: Get the number of pixels of the diverging pixels; Determine whether the number of pixels is less than a number threshold; If the number of pixels is less than the number threshold, identifying diverging pixels and obtaining the shortest distance between the diverging pixels and the outline of the subject; If the number of pixels is not less than the number threshold, then the divergent pixels are identified to obtain the distribution characteristics of all the divergent pixels; Determine whether the divergent pixels are regularly distributed based on the distribution characteristics; If the divergent pixel points are regularly distributed, the divergent pixel points are connected in sequence to construct a secondary contour, and the two end pixel points of the secondary contour are obtained, where the two end pixel points include a starting divergent pixel point and an ending divergent pixel point. According to the starting divergent pixel point and the ending divergent pixel point, the steps of identifying the divergent pixel points and obtaining the shortest distance between the divergent pixel points and the main contour are performed.

3. The photovoltaic array operating status monitoring method according to claim 2, characterized in that: In the step of determining whether the divergent pixels are regularly distributed according to the distribution characteristics, the method includes: Obtaining a position feature and a distance between adjacent pixels of each diverging pixel point, constructing a position database based on the position feature, and clustering the diverging pixels in the position database based on the distance between adjacent pixels to obtain a plurality of unit pixel blocks, each unit pixel block including a plurality of diverging pixels; Each unit pixel block is set as a mass point to obtain a plurality of mass points, and the position database is updated according to the plurality of mass points to obtain a mass point database; Identify multiple particles in a particle database to obtain particle positions, obtain particle distribution characteristics based on the particle positions to obtain edge particles in the particle database, and construct a particle database outline based on the edge particles; It is determined whether the particles are regularly distributed according to the particle database outline to determine whether the divergent pixels are regularly distributed.

4. The photovoltaic array operating status monitoring method according to claim 3, characterized in that: In the step of determining whether the particles are regularly distributed according to the particle database outline to determine whether the divergent pixels are regularly distributed, the regular distribution includes a convergent distribution, and the method includes: Identify a contour of a particle database to obtain end particles and middle particles, and construct a contour centerline based on the end particles and along the middle particles, wherein the end particles determine the starting direction and the ending direction of the contour centerline, and the middle particles determine the shape of the contour centerline; Determine whether the contour midline is a regular line segment, where the regular line segment includes an arc and a straight line; If the contour midline is a regular line segment, the particles are convergently distributed relative to the regular line segment, and the divergent pixels are convergently distributed relative to the regular line segment.

5. The photovoltaic array operating status monitoring method according to claim 4, characterized in that: The step of constructing the contour midline according to the end mass points and along the middle mass points comprises: Constructing a preliminary centerline of the particle database contour based on the end particles; Determine whether the number of particles distributed on both sides of the preliminary center line is balanced; If so, the preliminary centerline is the contour centerline; If not, the preliminary center line is adjusted according to the difference in the number of particles on both sides until the number of particles on both sides of the preliminary center line is balanced.

6. The photovoltaic array operating status monitoring method according to claim 2, characterized in that: After the step of determining whether the divergent pixels are regularly distributed according to the distribution characteristics, the following steps are further included: If the divergent pixels are not regularly distributed, the process returns to executing the step of identifying the image features through the migration network to extract and fuse the multi-scale feature information in the original infrared image.

7. The photovoltaic array operating status monitoring method according to claim 1, characterized in that: In the step of identifying the hot spot edge delimiting image and calculating the image pixel difference to obtain the final pixel quantity ratio of the hot spot, the calculation formula of the pixel quantity ratio is: ； Where, S Indicates the proportion of pixels with hot spots in the photovoltaic array; n Indicates the number of vertices of the photovoltaic array hot spot; ( x i , y i ) represents the coordinates of the i-th vertex, ( x n , y n ) = ( x 0 , y 0).

8. The photovoltaic array operating status monitoring method according to claim 7, characterized in that: In the step of evaluating the health status of the photovoltaic array according to the average value of the hot spot area ratio of the photovoltaic array, the photovoltaic array health status evaluation formula is: ； Where, G It represents the average fraction of the hot spot area of ​​the photovoltaic array collected multiple times; N Indicates the number of times data is collected from the photovoltaic array; S i 、 S Ti Indicates the i The proportion of hot spot pixels in the first data collection and the i The proportion of the number of photovoltaic array pixels.

9. A photovoltaic array operating status monitoring system, characterized in that: The system comprises: an acquisition module, configured to acquire an infrared raw image of the photovoltaic array and perform preprocessing to capture hot spot image features, wherein the image features include hot spot color and hot spot shape; identify the image features through a migration network to extract and fuse multi-scale feature information from the infrared raw image; enhance the image features based on the fused multi-scale feature information to obtain a target hot spot image based on the enhanced image features; the multi-scale feature information includes scale, shape, and edge; A segmentation module is used to segment the target hot spot image by hot spot contour and shape using a semantic segmentation algorithm to obtain a segmented image, and detect the segmented image using an edge detection algorithm to extract hot spot edge pixels to obtain a hot spot edge definition image; a monitoring module, configured to identify the hot spot edge delimiting image, calculate the image pixel difference to obtain a final hot spot pixel number ratio, obtain an average hot spot area ratio of the photovoltaic array based on the pixel number ratio, evaluate the photovoltaic array health status based on the photovoltaic array hot spot area ratio average, and monitor the photovoltaic array operation status based on the photovoltaic array health status; The step of enhancing image features according to the fused multi-scale feature information to obtain a target hot spot image according to the enhanced image features includes: Acquire the enhanced image features and identify them to obtain contour pixels, wherein the contour pixels include main pixels and divergent pixels, and the plurality of main pixels form a main contour, identify the divergent pixels and obtain the shortest distance between the divergent pixels and the main contour; Determining whether the shortest distance is less than a distance threshold; If so, the divergent pixels are updated to the main pixels and the main contour is updated according to the updated main pixels, and the target hot spot image is obtained according to the updated main contour; If not, the divergent pixel is an invalid pixel.

10. A data processing device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the photovoltaic array operating status monitoring method according to any one of claims 1 to 8 is implemented.

Images