Method and device for detecting hot spot defect of photovoltaic module, storage medium and electronic device

Through the combination of image acquisition and hot spot defect detection model, multi-scale feature generation and adaptive fusion technology are used to solve the problem of low accuracy in hot spot defect detection of photovoltaic modules, and high-precision detection in complex backgrounds is achieved.

CN118918088APending Publication Date: 2024-11-08HUANENG JILIN POWER GENERATION JIUTAI ELECTRIC FACTORY +2
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411044177.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-31
Publication Date
2024-11-08

AI Technical Summary

Technical Problem

The existing photovoltaic module hot spot defect detection methods are not very accurate, especially in complex backgrounds, which is difficult to effectively extract and identify small defect features, resulting in low detection accuracy and insufficient robustness.

Method used

The image acquisition device is used to collect the photovoltaic module images, and the hot spot defect detection model obtained through training is combined with neighborhood-related feature modules to generate multi-scale feature generation and adaptive fusion to generate initial feature maps, feature extraction and vector fusion, and improve detection accuracy .

Benefits of technology

Effectively detect small defect characteristics, maintain high detection accuracy in complex backgrounds, provide efficient and reliable intelligent inspection solutions for photovoltaic electric fields, and improve the accuracy of hot spot defect detection.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118918088A_ABST
    Figure CN118918088A_ABST
Patent Text Reader

Abstract

The invention discloses a detection method and device for hot spot defects of a photovoltaic module, a storage medium and an electronic device, and the method comprises the steps: carrying out the image collection of a photovoltaic module of a photovoltaic electric field through an image collection device, and obtaining an initial image of the photovoltaic module; an initial feature graph corresponding to the initial image is input into a hot spot defect detection model, the hot spot defect type of the photovoltaic module is determined according to an output result of the hot spot defect detection model, and the hot spot defect detection model takes a historical image of the photovoltaic module as an input sample and is used for detecting the hot spot defect of the photovoltaic module. The hot spot defect type of the historical hot spot defect existing in the photovoltaic module is used as an output sample for training; and determining the hot spot defect of the photovoltaic module based on the hot spot defect type of the photovoltaic module. By adopting the technical scheme, the problem that the detection accuracy of the hot spot defect of the photovoltaic module is not high is solved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of image detection, and in particular, to a method and device for detecting hot spot defects of a photovoltaic module, a storage medium, and an electronic device. Background Art

[0002] In the operation process of the photovoltaic system of modern photovoltaic power plants, if hotspot defects occur, the service life of the photovoltaic modules will be shortened, and may even cause safety accidents such as fires. A large number of hotspot defects will affect the normal operation of the photovoltaic system, and then affect the power generation efficiency and reliability of the photovoltaic system. Therefore, timely and accurate detection and identification of hotspot defects is crucial to maintaining the normal operation of photovoltaic power plants and ensuring system safety.

[0003] At present, traditional defect detection methods mainly rely on manual detection and empirical judgment, which is not only time-consuming and labor-intensive, but also difficult to ensure the accuracy of the detection results. With the development of science and technology, related fields have introduced technologies such as Convolutional Neural Network (CNN) to realize intelligent recognition of hotspot defects. However, the existing CNN still has some shortcomings in photovoltaic hotspot defect detection tasks. For example, due to the increase in network depth, small defect features gradually disappear in the network, resulting in unsatisfactory detection results. In addition, the interference of complex backgrounds also makes the ability to distinguish defect features insufficient, and it is often difficult to effectively extract and identify small defect features. There are low detection accuracy and insufficient robustness, which leads to low defect detection accuracy.

[0004] In the related art, there is a problem that the detection accuracy of hot spot defects of photovoltaic modules is low, and no effective solution has been proposed yet.

[0005] Therefore, it is necessary to improve the related technology to overcome the above-mentioned defects in the related technology. Summary of the invention

[0006] The embodiments of the present application provide a method and device for detecting hot spot defects of a photovoltaic module, a storage medium, and an electronic device, so as to at least solve the problem of low accuracy in detecting hot spot defects of a photovoltaic module.

[0007] According to one aspect of an embodiment of the present application, a method for detecting hotspot defects of a photovoltaic module is provided, comprising: using an image acquisition device to acquire images of a photovoltaic module of a photovoltaic field to obtain an initial image of the photovoltaic module; inputting an initial feature map corresponding to the initial image into a hotspot defect detection model, and determining the hotspot defect type of the photovoltaic module according to an output result of the hotspot defect detection model, wherein the hotspot defect detection model is trained using historical images of the photovoltaic module as input samples and hotspot defect types of historical hotspot defects existing in the photovoltaic module as output samples; and determining the hotspot defect existing in the photovoltaic module based on the hotspot defect type of the photovoltaic module.

[0008] In an exemplary embodiment, before inputting the initial feature map corresponding to the initial image into the hotspot defect detection model, the method also includes: performing image denoising on the initial image to obtain an intermediate image, wherein the number of noise points in the intermediate image is lower than the number of noise points in the initial image; performing image enhancement on the intermediate image to obtain an enhanced image; performing feature extraction on the enhanced image to obtain an image feature vector corresponding to the enhanced image, and generating the initial feature map corresponding to the initial image based on the image feature vector.

[0009] In an exemplary embodiment, the hotspot defect detection model includes at least a neighborhood-related feature module, an initial feature map corresponding to the initial image is input into the hotspot defect detection model, and the hotspot defect type of the photovoltaic module is determined according to the output result of the hotspot defect detection model, including: using the feature generation module of the neighborhood-related feature module to sample the initial feature map to obtain a first scale feature map and a second scale feature map, wherein the number of sampling points corresponding to the first scale feature map is greater than the number of sampling points corresponding to the second scale feature map; a first scale feature vector corresponding to the first scale feature map, a second scale feature vector corresponding to the second scale feature map, and the initial feature map The corresponding third-scale feature vector is input into the feature fusion module of the neighborhood-related feature module, so that the feature fusion module performs vector fusion on the first-scale feature vector, the second-scale feature vector and the third-scale feature vector to obtain a first fusion vector; the first fusion vector is input into the fully connected layer of the hotspot defect detection model, and the first fusion vector is vector updated based on a preset operation in the fully connected layer to obtain an update vector; the first fusion vector and the update vector are fused to obtain a second fusion vector; the second fusion vector is input into the output layer of the hotspot defect detection model, and the hotspot defect type corresponding to the second fusion vector is output through the output layer.

[0010] In an exemplary embodiment, the feature generation module of the neighborhood-related feature module is used to sample the initial feature map to obtain a first-scale feature map and a second-scale feature map, including: upsampling the initial feature map using the feature generation module of the neighborhood-related feature module to obtain the first-scale feature map; and downsampling the initial feature map using the feature generation module of the neighborhood-related feature module to obtain the second-scale feature map.

[0011] In an exemplary embodiment, the feature fusion module performs vector fusion on the first-scale feature vector, the second-scale feature vector and the third-scale feature vector to obtain a first fused vector by using a concatenation operation to merge the first-scale feature vector, the second-scale feature vector and the third-scale feature vector into a first fused vector; the first fused vector X is expressed as follows: X=Cat(X0, X1, X2); wherein X0 represents the first-scale feature vector, X2 represents the second-scale feature vector, and X1 represents the third-scale feature vector.

[0012] In an exemplary embodiment, the first fused vector is updated based on a preset operation in the fully connected layer to obtain an update vector, including: performing a global average pooling operation on the first fused vector in the fully connected layer to obtain a pooling vector; performing a convolution operation on the pooling vector according to a 1×1 convolution kernel to obtain a convolution vector; and performing an activation operation on the convolution vector using an activation function to obtain the update vector, wherein the update vector is expressed as: W g =σ(h(g(X))); where g(·) represents the global average pooling operation, h(·) represents the convolution operation, and σ represents the activation operation.

[0013] According to another aspect of an embodiment of the present application, a device for detecting hotspot defects of a photovoltaic module is also provided, including: an obtaining module, used to use an image acquisition device to acquire images of a photovoltaic module of a photovoltaic field to obtain an initial image of the photovoltaic module; a first determining module, used to input an initial feature map corresponding to the initial image into a hotspot defect detection model, and determine the hotspot defect type of the photovoltaic module according to an output result of the hotspot defect detection model, wherein the hotspot defect detection model is trained using historical images of the photovoltaic module as input samples and hotspot defect types of historical hotspot defects existing in the photovoltaic module as output samples; a second determining module, used to determine the hotspot defects existing in the photovoltaic module based on the hotspot defect type of the photovoltaic module.

[0014] According to another aspect of the embodiments of the present application, a computer-readable storage medium is provided, in which a computer program is stored, wherein the computer program is configured to execute the above-mentioned method for detecting hot spot defects of photovoltaic modules when running.

[0015] According to another aspect of the embodiments of the present application, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the method for detecting hot spot defects of the photovoltaic module through the computer program.

[0016] According to another aspect of the present application, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the steps in any one of the above method embodiments are implemented.

[0017] Through this application, an image acquisition device is used to acquire images of a photovoltaic module of a photovoltaic electric field, and an initial image of the photovoltaic module is obtained; an initial feature map corresponding to the initial image is input into a hotspot defect detection model, and the hotspot defect type of the photovoltaic module is determined according to the output result of the hotspot defect detection model, wherein the hotspot defect detection model is trained with the historical image of the photovoltaic module as an input sample and the hotspot defect type of the historical hotspot defect existing in the photovoltaic module as an output sample; based on the hotspot defect type of the photovoltaic module, the hotspot defect existing in the photovoltaic module is determined, which can not only effectively detect small defect features, but also maintain high detection accuracy under complex backgrounds, providing an efficient and reliable solution for the intelligent inspection of photovoltaic electric fields. In addition, the problem of low detection accuracy of hotspot defects of photovoltaic modules is solved, and the detection accuracy of hotspot defects of photovoltaic modules is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0019] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0020] Figure 1 It is a hardware structure block diagram of a mobile terminal of a method for detecting hot spot defects of a photovoltaic module according to an embodiment of the present application;

[0021] Figure 2is a flow chart of a method for detecting hot spot defects of a photovoltaic module according to an embodiment of the present application;

[0022] Figure 3 is a schematic diagram of a connection method of a neighborhood-related feature module according to an embodiment of the present application;

[0023] Figure 4 is a schematic diagram of a structure of adaptive fusion according to an embodiment of the present application;

[0024] Figure 5 It is a structural block diagram of a device for detecting hot spot defects of a photovoltaic module according to an embodiment of the present application. DETAILED DESCRIPTION

[0025] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0026] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0027] The method embodiments provided in the embodiments of the present application can be executed in a mobile terminal, a computer terminal or a similar computing device. Taking running on a mobile terminal as an example, Figure 1 1 is a hardware structure block diagram of a mobile terminal for detecting hot spot defects of a photovoltaic module according to an embodiment of the present application. Figure 1 As shown, the mobile terminal may include one or more ( Figure 1Only one is shown in the figure) a processor 102 (the processor 102 may include but is not limited to a processing device such as a microprocessor (Central Processing Unit, MCU) or a programmable logic device (Field Programmable Gate Array, FPGA)) and a memory 104 for storing data, wherein the mobile terminal may also include a transmission device 106 and an input / output device 108 for communication functions. It can be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the mobile terminal. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations are shown.

[0028] The memory 104 can be used to store computer programs, for example, software programs and modules of application software, such as the computer program corresponding to the method for detecting hot spot defects of photovoltaic modules in the embodiment of the present application. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, that is, to implement the above method. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories can be connected to the mobile terminal via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0029] The transmission device 106 is used to receive or send data via a network. The specific example of the above network may include a wireless network provided by a communication provider of the mobile terminal. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, referred to as NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 can be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0030] In this embodiment, a method for detecting hot spot defects of a photovoltaic module is provided. Figure 2 is a flow chart of a method for detecting hot spot defects of a photovoltaic module according to an embodiment of the present application, such as Figure 2 As shown, the process includes the following steps:

[0031] Step S202: using an image acquisition device to acquire an image of a photovoltaic module of a photovoltaic field to obtain an initial image of the photovoltaic module;

[0032] Step S204: inputting the initial feature map corresponding to the initial image into a hotspot defect detection model, and determining the hotspot defect type of the photovoltaic module according to the output result of the hotspot defect detection model, wherein the hotspot defect detection model is trained with the historical image of the photovoltaic module as an input sample and the hotspot defect type of the historical hotspot defects existing in the photovoltaic module as an output sample;

[0033] Step S206: determining the hotspot defect existing in the photovoltaic module based on the hotspot defect type of the photovoltaic module.

[0034] Through the above steps, the image acquisition device is used to acquire the image of the photovoltaic module of the photovoltaic field to obtain the initial image of the photovoltaic module; the initial feature map corresponding to the initial image is input into the hot spot defect detection model, and the hot spot defect type of the photovoltaic module is determined according to the output result of the hot spot defect detection model, wherein the hot spot defect detection model is trained with the historical image of the photovoltaic module as the input sample and the hot spot defect type of the historical hot spot defect existing in the photovoltaic module as the output sample; the hot spot defect existing in the photovoltaic module is determined based on the hot spot defect type of the photovoltaic module, which can not only effectively detect the characteristics of small defects, but also maintain high detection accuracy under complex backgrounds, providing an efficient and reliable solution for the intelligent inspection of the photovoltaic field. In addition, the problem of low detection accuracy of hot spot defects of photovoltaic modules is solved, and the detection accuracy of hot spot defects of photovoltaic modules is improved.

[0035] It should be noted that the above hot spot defect types include but are not limited to obstruction-type hot spot defects, damage-type hot spot defects, fault-type hot spot defects, etc.

[0036] Among them, the reason for the formation of shading type hot spot defects is that the photovoltaic module is exposed to the outdoors for a long time, and it is inevitable that it will be blocked by leaves and dust in the environment. These shielding will cause the local area of ​​the photovoltaic module to not fully receive sunlight, resulting in local overheating. The shading type hot spot defects are, for example, hot spot defects caused by leaf shading and hot spot defects caused by dust shading.

[0037] Damage-type hot spot defects, such as hot spot defects caused by damage to the surface tempered glass, refer to the tempered glass covering the surface of the photovoltaic module to protect the internal battery pack, but under the influence of external impact or long-term wind and rain, the tempered glass may break or wear, thereby affecting the normal operation of the photovoltaic module and causing hot spot defects.

[0038] Fault-type hot spot defects include hot spot defects caused by diode failure, hot spot defects caused by poor junction box welding, hot spot defects caused by internal battery pack failure, and hot spot defects caused by fuse burning or open circuit. The causes of diode failure and poor junction box welding are the same, both of which are diode failure or poor junction box welding in the photovoltaic module, which will lead to uneven current and local overheating. Internal battery pack failure means that if the battery pack inside the photovoltaic module fails, such as short circuit or performance degradation, it will cause abnormal temperature rise in the local area and form a hot spot defect. Hot spot defects caused by fuse burning or open circuit are mainly caused by fuse burning or open circuit in the photovoltaic system, causing abnormal current and excessive temperature in the local area of ​​the photovoltaic module.

[0039] In an exemplary embodiment, before the initial feature map corresponding to the initial image is input into the hotspot defect detection model, the following implementation scheme is further proposed: performing image denoising processing on the initial image to obtain an intermediate image, wherein the number of noise points in the intermediate image is lower than the number of noise points in the initial image; performing image enhancement on the intermediate image to obtain an enhanced image; performing feature extraction operation on the enhanced image to obtain an image feature vector corresponding to the enhanced image, and generating the initial feature map corresponding to the initial image based on the image feature vector.

[0040] In an exemplary embodiment, the hotspot defect detection model includes at least a neighborhood-related feature module, and the specific process of inputting the initial feature map corresponding to the initial image into the hotspot defect detection model and determining the hotspot defect type of the photovoltaic module according to the output result of the hotspot defect detection model includes: using the feature generation module of the neighborhood-related feature module to sample the initial feature map to obtain a first scale feature map and a second scale feature map, wherein the number of sampling points corresponding to the first scale feature map is greater than the number of sampling points corresponding to the second scale feature map; and sampling the first scale feature vector corresponding to the first scale feature map, the second scale feature vector corresponding to the second scale feature map, and the initial feature map. The third-scale feature vector corresponding to the feature graph is input into the feature fusion module of the neighborhood-related feature module, so that the feature fusion module performs vector fusion on the first-scale feature vector, the second-scale feature vector and the third-scale feature vector to obtain a first fusion vector; the first fusion vector is input into the fully connected layer of the hotspot defect detection model, and the first fusion vector is vector updated based on a preset operation in the fully connected layer to obtain an update vector; the first fusion vector and the update vector are fused to obtain a second fusion vector; the second fusion vector is input into the output layer of the hotspot defect detection model, and the hotspot defect type corresponding to the second fusion vector is output through the output layer.

[0041] Based on this embodiment, the photovoltaic electric field hot spot defects can be detected based on the Neighborhood Correlation Feature Module (NCFM). By integrating multi-scale information, the problem that the existing CNN technology causes small defect features to gradually disappear in the deep network can be solved, thereby improving the detection accuracy and robustness. Specifically, the NCFM module generates multi-scale features through up and down sampling, and uses the correlation between features for adaptive fusion, thereby enhancing the expression ability of small defect features. Compared with traditional feature fusion methods, the NCFM module considers the correlation of features in the fusion process, avoids information redundancy, and improves the distinguishing ability of feature representation.

[0042] In an exemplary embodiment, the implementation process of sampling the initial feature map using the feature generation module of the neighborhood-related feature module to obtain the first scale feature map and the second scale feature map is further described by the following steps: upsampling the initial feature map using the feature generation module of the neighborhood-related feature module to obtain the first scale feature map; and downsampling the initial feature map using the feature generation module of the neighborhood-related feature module to obtain the second scale feature map.

[0043] In an exemplary embodiment, the feature fusion module may perform vector fusion on the first-scale feature vector, the second-scale feature vector and the third-scale feature vector to obtain a first fused vector by using a concatenation operation to merge the first-scale feature vector, the second-scale feature vector and the third-scale feature vector into a first fused vector; the first fused vector X is expressed as follows: X=Cat(X0, X1, X2); wherein X0 represents the first-scale feature vector, X2 represents the second-scale feature vector, and X1 represents the third-scale feature vector.

[0044] In an exemplary embodiment, further, the following technical solution is proposed for performing a vector update on the first fused vector based on a preset operation in the fully connected layer to obtain an updated vector, specifically including: performing a global average pooling operation on the first fused vector in the fully connected layer to obtain a pooling vector; performing a convolution operation on the pooling vector according to a 1×1 convolution kernel to obtain a convolution vector; and performing an activation operation on the convolution vector using an activation function to obtain the update vector, wherein the update vector is expressed as: W g =σ(h(g(X))); where g(·) represents the global average pooling operation, h(·) represents the convolution operation, and σ represents the activation operation.

[0045] Obviously, the embodiments described above are only some embodiments of the present application, not all embodiments. In order to better understand the above method, the above process is described below in conjunction with the embodiments, but it is not intended to limit the technical solutions of the embodiments of the present application, specifically:

[0046] In an optional embodiment, the present application combines Figure 3-4 The process of detecting hot spot defects in photovoltaic electric fields based on the neighborhood correlation feature module is explained. It aims at the problem that small defect features existing in the prior art gradually disappear in the deep network and the feature differentiation ability is insufficient under complex background interference. The present application has made innovations in multi-scale feature generation and feature fusion. Specifically, the NCFM module is introduced, and the NCFM module is used to generate multi-scale features through up and down sampling, and adaptive fusion is performed based on the correlation between features, thereby enhancing the expression ability of small defect features. Compared with the traditional feature fusion method, which usually adopts a simple weighted average or direct addition method, it cannot fully consider the correlation between features of different scales, which easily leads to information redundancy and feature confusion. The NCFM module of the present application fully utilizes the correlation between features of different scales through an adaptive fusion structure, avoids information redundancy, and improves the differentiation ability of feature representation.

[0047] This application aims to use the NCFM module to integrate multi-scale information to solve the problem that small defect features gradually disappear in deep networks. Figure 3-4 The following steps illustrate the specific implementation steps of the NCFM module:

[0048] Step 1, multi-scale feature generation:

[0049] like Figure 3 As shown in the figure, the input feature map is first upsampled and downsampled to generate feature maps of different scales. Through multi-scale feature generation, defect features at different scales can be captured, enhancing the richness of feature representation. For example, let the input feature map (that is, the initial feature map mentioned above) be X, and after upsampling and downsampling, multi-scale features X0 (corresponding to the first scale feature map), X1 and X2 (corresponding to the second scale feature map) are generated, where X1 represents the initial feature map.

[0050] Step 2, feature fusion:

[0051] The correlation between multi-scale features is calculated, and the most relevant feature map is selected for fusion. In this way, information from different scales can be adaptively fused, avoiding information redundancy and enhancing the expressiveness of small defect features.

[0052] like Figure 4As shown in the figure, the multi-scale features X0, X1 and X2 are merged into X by concatenation operation to obtain the fused features. The formula is as follows:

[0053] X=Cat(X0,X1,X2).

[0054] Among them, cat represents the fusion operation.

[0055] Finally, an adaptive fusion structure is used instead of simple feature addition to avoid information confusion and improve the distinguishing ability of feature representation.

[0056] The formula is:

[0057] W g =σ(h(g(X))).

[0058] Among them, g(·) is the global average pooling (GAP), h(·) is a 1×1 convolution operation, and σ is the sigmoid function.

[0059] The updated three scale features can be expressed as:

[0060]

[0061] Among them, Split represents the split operation, Represents element-wise multiplication.

[0062] The final output feature fusion result is:

[0063] F=X0+X1+X2.

[0064] Through the above steps, multi-scale feature generation and adaptive fusion are realized based on the NCFM module, which can effectively solve the problem that small defect features gradually disappear in the deep network. This method has a strong feature differentiation ability under complex backgrounds, improves the accuracy and robustness of detection, and provides an efficient and reliable solution for the intelligent inspection of photovoltaic fields. Combining the NCFM module with the deep learning network, the hot spot defect detection model is trained, which provides an efficient and accurate photovoltaic field hot spot defect detection method, which can not only be applied to the hot spot defect detection of photovoltaic fields, but also can be extended to other industrial fields that require high-precision defect detection, providing strong technical support for the intelligent inspection and maintenance of photovoltaic fields.

[0065] Through the description of the above implementation methods, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods of each embodiment of the present application.

[0066] In this embodiment, a device for detecting hot spot defects of a photovoltaic module is also provided, and the device is used to implement the above-mentioned embodiments and preferred implementation modes, and the descriptions that have been made are not repeated here. As used below, the term "module" can implement a combination of software and / or hardware of a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0067] Figure 5 According to a structural block diagram of a device for detecting hot spot defects of a photovoltaic module according to an embodiment of the present application, the device includes:

[0068] An obtaining module 52 is used to use an image acquisition device to acquire an image of a photovoltaic module of a photovoltaic field to obtain an initial image of the photovoltaic module;

[0069] A first determination module 54 is used to input the initial feature map corresponding to the initial image into a hotspot defect detection model, and determine the hotspot defect type of the photovoltaic module according to the output result of the hotspot defect detection model, wherein the hotspot defect detection model is trained with the historical image of the photovoltaic module as an input sample and the hotspot defect type of the historical hotspot defect existing in the photovoltaic module as an output sample;

[0070] The second determination module 56 is configured to determine the hot spot defect existing in the photovoltaic module based on the hot spot defect type of the photovoltaic module.

[0071] Through the above device, the image acquisition device is used to acquire the image of the photovoltaic module of the photovoltaic field to obtain the initial image of the photovoltaic module; the initial feature map corresponding to the initial image is input into the hot spot defect detection model, and the hot spot defect type of the photovoltaic module is determined according to the output result of the hot spot defect detection model, wherein the hot spot defect detection model is trained with the historical image of the photovoltaic module as the input sample and the hot spot defect type of the historical hot spot defect existing in the photovoltaic module as the output sample; the hot spot defect existing in the photovoltaic module is determined based on the hot spot defect type of the photovoltaic module, which can not only effectively detect the characteristics of small defects, but also maintain high detection accuracy under complex backgrounds, providing an efficient and reliable solution for the intelligent inspection of the photovoltaic field. In addition, the problem of low detection accuracy of hot spot defects of photovoltaic modules is solved, and the detection accuracy of hot spot defects of photovoltaic modules is improved.

[0072] In an exemplary embodiment, the first determination module is also used to: perform image denoising on the initial image to obtain an intermediate image, wherein the number of noise points in the intermediate image is lower than the number of noise points in the initial image; perform image enhancement on the intermediate image to obtain an enhanced image; perform feature extraction on the enhanced image to obtain an image feature vector corresponding to the enhanced image, and generate an initial feature map corresponding to the initial image based on the image feature vector.

[0073] In an exemplary embodiment, the hotspot defect detection model includes at least a neighborhood-related feature module and a first determination module, and is further used to: use the feature generation module of the neighborhood-related feature module to sample the initial feature map to obtain a first-scale feature map and a second-scale feature map, wherein the number of sampling points corresponding to the first-scale feature map is greater than the number of sampling points corresponding to the second-scale feature map; input the first-scale feature vector corresponding to the first-scale feature map, the second-scale feature vector corresponding to the second-scale feature map, and the third-scale feature vector corresponding to the initial feature map into the feature fusion module of the neighborhood-related feature module, so that the feature fusion module performs vector fusion on the first-scale feature vector, the second-scale feature vector and the third-scale feature vector to obtain a first fusion vector; input the first fusion vector into the fully connected layer of the hotspot defect detection model, and perform vector update on the first fusion vector based on a preset operation in the fully connected layer to obtain an update vector; fuse the first fusion vector and the update vector to obtain a second fusion vector; input the second fusion vector into the output layer of the hotspot defect detection model, and output the hotspot defect type corresponding to the second fusion vector through the output layer.

[0074] In an exemplary embodiment, the first determination module is further used to: upsample the initial feature map using the feature generation module of the neighborhood-related feature module to obtain the first-scale feature map; and downsample the initial feature map using the feature generation module of the neighborhood-related feature module to obtain the second-scale feature map.

[0075] In an exemplary embodiment, the first determination module is further used to: implement the feature fusion module to perform vector fusion on the first scale feature vector, the second scale feature vector and the third scale feature vector in the following manner to obtain a first fused vector: use a series operation to merge the first scale feature vector, the second scale feature vector and the third scale feature vector into a first fused vector; the first fused vector X is expressed as follows: X=Cat(X0,X1,X2); wherein X0 represents the first scale feature vector, X2 represents the second scale feature vector, and X1 represents the third scale feature vector.

[0076] In an exemplary embodiment, the first determination module is further used to: perform a global average pooling operation on the first fused vector in the fully connected layer to obtain a pooling vector; perform a convolution operation on the pooling vector according to a 1×1 convolution kernel to obtain a convolution vector; and perform an activation operation on the convolution vector using an activation function to obtain the update vector, wherein the update vector is expressed as: W g =σ(h(g(X))); where g(·) represents the global average pooling operation, h(·) represents the convolution operation, and σ represents the activation operation.

[0077] An embodiment of the present application further provides a computer-readable storage medium, in which a computer program is stored, wherein the computer program is configured to execute the steps of any of the above method embodiments when running.

[0078] Optionally, in this embodiment, the storage medium may be configured to store a computer program for performing the following steps:

[0079] S1, using an image acquisition device to acquire an image of a photovoltaic module of a photovoltaic field to obtain an initial image of the photovoltaic module;

[0080] S2, inputting the initial feature map corresponding to the initial image into a hotspot defect detection model, and determining the hotspot defect type of the photovoltaic module according to the output result of the hotspot defect detection model, wherein the hotspot defect detection model is trained with the historical image of the photovoltaic module as an input sample and the hotspot defect type of the historical hotspot defects existing in the photovoltaic module as an output sample;

[0081] S3: Determine the hotspot defect existing in the photovoltaic module based on the hotspot defect type of the photovoltaic module.

[0082] In an exemplary embodiment, the computer-readable storage medium may include, but is not limited to, various media that can store computer programs, such as a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk.

[0083] For specific examples in this embodiment, reference may be made to the examples described in the above embodiments and exemplary implementation modes, and this embodiment will not be described in detail herein.

[0084] An embodiment of the present application further provides an electronic device, including a memory and a processor, wherein a computer program is stored in the memory, and the processor is configured to run the computer program to execute the steps in any one of the above method embodiments.

[0085] Optionally, in this embodiment, the processor may be configured to perform the following steps through a computer program:

[0086] S1, using an image acquisition device to acquire an image of a photovoltaic module of a photovoltaic field to obtain an initial image of the photovoltaic module;

[0087] S2, inputting the initial feature map corresponding to the initial image into a hotspot defect detection model, and determining the hotspot defect type of the photovoltaic module according to the output result of the hotspot defect detection model, wherein the hotspot defect detection model is trained with the historical image of the photovoltaic module as an input sample and the hotspot defect type of the historical hotspot defects existing in the photovoltaic module as an output sample;

[0088] S3: Determine the hotspot defect existing in the photovoltaic module based on the hotspot defect type of the photovoltaic module.

[0089] In an exemplary embodiment, the electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor, and the input / output device is connected to the processor.

[0090] An embodiment of the present application further provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the steps in any one of the above method embodiments are implemented.

[0091] An embodiment of the present application further provides another computer program product, including a non-volatile computer-readable storage medium, wherein the non-volatile computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in any of the above method embodiments are implemented.

[0092] An embodiment of the present application also provides a computer program, which includes computer instructions, which are stored in a computer-readable storage medium; a processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device performs the steps in any one of the above method embodiments.

[0093] For specific examples in this embodiment, reference may be made to the examples described in the above embodiments and exemplary implementation modes, and this embodiment will not be described in detail herein.

[0094] Obviously, those skilled in the art should understand that the above modules or steps of the present application can be implemented by a general computing device, they can be concentrated on a single computing device, or distributed on a network composed of multiple computing devices, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, and in some cases, the steps shown or described can be executed in a different order from that herein, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. Thus, the present application is not limited to any specific combination of hardware and software.

[0095] The above is only a preferred implementation of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A method for detecting hot spot defects of a photovoltaic module, characterized in that: include: Using an image acquisition device to acquire an image of a photovoltaic module of a photovoltaic field to obtain an initial image of the photovoltaic module; Inputting the initial feature map corresponding to the initial image into a hotspot defect detection model, and determining the hotspot defect type of the photovoltaic module according to the output result of the hotspot defect detection model, wherein the hotspot defect detection model is trained with the historical images of the photovoltaic module as input samples and the hotspot defect types of the historical hotspot defects existing in the photovoltaic module as output samples; A hot spot defect existing in the photovoltaic module is determined based on the hot spot defect type of the photovoltaic module.

2. The method according to claim 1, characterized in that Before inputting the initial feature map corresponding to the initial image into the hotspot defect detection model, the method further includes: Performing image noise reduction processing on the initial image to obtain an intermediate image, wherein the number of noise points in the intermediate image is lower than the number of noise points in the initial image; Performing image enhancement on the intermediate image to obtain an enhanced image; A feature extraction operation is performed on the enhanced image to obtain an image feature vector corresponding to the enhanced image, and an initial feature map corresponding to the initial image is generated based on the image feature vector.

3. The method according to claim 1, characterized in that: The hotspot defect detection model at least includes a neighborhood-related feature module, inputs an initial feature map corresponding to the initial image into the hotspot defect detection model, and determines the hotspot defect type of the photovoltaic module according to an output result of the hotspot defect detection model, including: The feature generation module of the neighborhood-related feature module is used to sample the initial feature map to obtain a first scale feature map and a second scale feature map, wherein the number of sampling points corresponding to the first scale feature map is greater than the number of sampling points corresponding to the second scale feature map; Inputting a first scale feature vector corresponding to the first scale feature map, a second scale feature vector corresponding to the second scale feature map, and a third scale feature vector corresponding to the initial feature map into a feature fusion module of the neighborhood-related feature module, so that the feature fusion module performs vector fusion on the first scale feature vector, the second scale feature vector, and the third scale feature vector to obtain a first fused vector; Inputting the first fused vector into a fully connected layer of the hotspot defect detection model, and performing a vector update on the first fused vector based on a preset operation in the fully connected layer to obtain an updated vector; Fusing the first fused vector and the update vector to obtain a second fused vector; The second fusion vector is input into the output layer of the hotspot defect detection model, and the hotspot defect type corresponding to the second fusion vector is output through the output layer.

4. The method according to claim 3, characterized in that The initial feature map is sampled using the feature generation module of the neighborhood-related feature module to obtain a first scale feature map and a second scale feature map, including: Up-sampling the initial feature map using the feature generation module of the neighborhood-related feature module to obtain the first scale feature map; And down-sample the initial feature map using the feature generation module of the neighborhood-related feature module to obtain the second-scale feature map.

5. The method according to claim 4, characterized in that The feature fusion module performs vector fusion on the first-scale feature vector, the second-scale feature vector and the third-scale feature vector to obtain a first fused vector by: Using a concatenation operation, the first scale feature vector, the second scale feature vector and the third scale feature vector are merged into a first fused vector; The first fusion vector X is expressed as follows: X=Cat(X0,X1,X2); Among them, X0 represents the first-scale feature vector, X2 represents the second-scale feature vector, and X1 represents the third-scale feature vector.

6. The method according to claim 5, characterized in that The method of performing a vector update on the first fusion vector based on a preset operation in the fully connected layer to obtain an updated vector includes: Performing a global average pooling operation on the first fusion vector in the fully connected layer to obtain a pooled vector; Performing a convolution operation on the pooled vector according to a 1×1 convolution kernel to obtain a convolution vector; The convolution vector is activated using an activation function to obtain the update vector, wherein the update vector is expressed as: g =σ(h(g(X))); Among them, g(·) represents the global average pooling operation, h(·) represents the convolution operation, and σ represents the activation operation.

7. A device for detecting hot spot defects of photovoltaic modules, characterized in that: include: An obtaining module is used to use an image acquisition device to acquire an image of a photovoltaic module of a photovoltaic electric field to obtain an initial image of the photovoltaic module; A first determination module is used to input the initial feature map corresponding to the initial image into a hotspot defect detection model, and determine the hotspot defect type of the photovoltaic module according to the output result of the hotspot defect detection model, wherein the hotspot defect detection model is trained with the historical image of the photovoltaic module as an input sample and the hotspot defect type of the historical hotspot defect existing in the photovoltaic module as an output sample; The second determination module is used to determine the hot spot defect existing in the photovoltaic module based on the hot spot defect type of the photovoltaic module.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein the program executes the method according to any one of claims 1 to 6 when executed.

9. An electronic device comprising a memory and a processor, characterized in that: A computer program is stored in the memory, and the processor is configured to execute the method according to any one of claims 1 to 6 through the computer program.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.