Photovoltaic panel defect detection method and system driven by mechanism model and image fusion

Abstract

The invention provides a mechanism model and image fusion driven photovoltaic panel defect detection method and system. The method comprises the following steps: collecting photovoltaic panel image data and monitoring data at the same timestamp; the monitoring data comprises illumination intensity, environment temperature, wind speed, current and voltage information; calculating current-voltage experimental data according to the monitoring data; and inputting image data and current-voltage experimental data into the trained explicit and implicit defect layered recognition model to obtain explicit defect and implicit defect recognition results of the photovoltaic panel. The explicit and implicit defect layered identification model is designed in series, so that the model can be more targeted during defect identification, and only when the photovoltaic panel is preliminarily judged to have defects, the explicit defect identification network and the implicit defect identification network are further called for processing, so that the defect identification accuracy is improved. Therefore, undifferentiated processing of all photovoltaic panel related data is avoided, and computing resources and time are saved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent defect detection, and in particular to a photovoltaic panel defect detection method and system driven by mechanism model and image fusion. Background Art

[0002] With the rapid development of the photovoltaic industry, the use of photovoltaic panels is increasing. However, photovoltaic panels are easily affected by environmental factors, manufacturing defects, aging and other factors during long-term operation, resulting in performance degradation and failure. Therefore, timely and accurate detection and diagnosis of photovoltaic panel defects to ensure efficient and stable operation have become important tasks in the maintenance of photovoltaic power generation systems. Traditional photovoltaic panel detection methods mainly rely on manual inspection and simple instrument detection, which is not only inefficient, but also difficult to accurately identify the subtle defect types of photovoltaic panels. Therefore, the study of intelligent and efficient photovoltaic panel defect detection methods has become a key requirement for the maintenance of photovoltaic power generation systems.

[0003] Most existing intelligent detection methods use defect detection methods driven by a single data type. If only image data is used, it is impossible to detect defects inside photovoltaic panels. If only electrical data is used for defect diagnosis, then the changes in electrical data caused by obstructions on or near the photovoltaic panel (such as dust, leaves, shadows, etc.) may interfere with or mislead the accurate diagnosis of defects inside the photovoltaic panel. Therefore, a method that can accurately detect defects in photovoltaic panels is urgently needed. Summary of the invention

[0004] In order to solve the above problems, the present invention proposes a photovoltaic panel defect detection method and system driven by mechanism model and image fusion. The method constructs a defect recognition model composed of an implicit defect recognition network and an explicit defect recognition network based on explicit and implicit hierarchical decision-making. By utilizing the current-voltage data output by the photovoltaic panel mechanism model, the actual current-voltage data and the photovoltaic panel image data as input data for training and detection, and introducing the constraints of the mechanism model, different networks are guided to perform defect recognition training and detection, thereby accurately identifying various defects of the photovoltaic panel.

[0005] In order to achieve the above object, the present invention adopts the following technical solution:

[0006] In a first aspect, the present invention provides a photovoltaic panel defect detection method driven by mechanism model and image fusion, comprising:

[0007] Collecting photovoltaic panel image data and monitoring data with the same time stamp; the monitoring data includes light intensity, ambient temperature, wind speed, current and voltage information;

[0008] Calculate current-voltage experimental data based on monitoring data;

[0009] Input the image data and current-voltage experimental data into a trained layered recognition model for explicit and implicit defects to obtain recognition results of explicit and implicit defects of photovoltaic panels;

[0010] Among them, the hierarchical identification model of explicit and implicit defects includes a defect classification network, an explicit defect recognition network and a implicit defect recognition network connected in series in sequence; the defect classification network is used to identify whether there are defects in the photovoltaic panel based on current-voltage experimental data; if there are defects, the image data corresponding to the current-voltage experimental data is input into the explicit defect recognition network to identify the explicit defects; the current-voltage experimental data corresponding to the image data without explicit defects is input into the implicit defect recognition network to identify the implicit defects.

[0011] Preferably, the calculating of current-voltage experimental data according to the monitoring data specifically includes:

[0012] According to the photovoltaic panel mechanism, the theoretical current-voltage data is calculated based on the light intensity, ambient temperature and wind speed information;

[0013] Based on the current and voltage information, calculate the actual current-voltage data;

[0014] Subtract the theoretical current-voltage data from the actual current-voltage data to obtain the deviation current-voltage data;

[0015] The theoretical current-voltage data and the deviation current-voltage data are added together to obtain the current-voltage experimental data.

[0016] Preferably, the method of calculating the current-voltage experimental data based on the monitoring data and inputting the image data and the current-voltage experimental data into the trained layered recognition model for explicit and implicit defects also includes: performing preprocessing operations of image enhancement, position correction, and unifying the image size on the image data; and performing preprocessing of data cleaning and data denoising on the monitoring data.

[0017] Preferably, the dominant defects include cracks, breakages and shielding; the recessive defects include hidden cracks, current leakage and PID effect.

[0018] Preferably, the defect classification network is a binary classification network constructed based on a residual network; the training process of the defect classification network includes:

[0019] Inputting the current-voltage experimental data samples into the defect classification network, and training based on the non-defect label and the defect label; the defect label includes an explicit defect label and an implicit defect label; the explicit defect label includes a crack label, a damage label and an occlusion label, and the implicit defect label includes a hidden crack label, a current leakage label and a PID effect label;

[0020] The first loss function is calculated. When the first loss function is minimized, the defect classification network training is completed.

[0021] Preferably, the training process of the dominant defect recognition network includes:

[0022] Obtaining image data samples corresponding to current-voltage experimental data samples identified as having defects by the defect classification network;

[0023] Inputting the image data sample into an explicit defect recognition network, taking the implicit defect label as a whole as an explicit non-defect label, and performing training based on the explicit defect label and the explicit non-defect label;

[0024] The second loss function is calculated. When the second loss function is minimized, the training of the dominant defect recognition network is completed.

[0025] Preferably, the training process of the hidden defect recognition network includes:

[0026] Acquire a current-voltage experimental data sample corresponding to an image data sample identified as being dominantly defect-free by a dominant defect recognition network;

[0027] Inputting the current-voltage experimental data samples into a hidden defect recognition network and performing training based on hidden defect labels;

[0028] The third loss function is calculated. When the third loss function is minimum, the hidden defect recognition network training is completed.

[0029] In a second aspect, the present invention provides a photovoltaic panel defect detection system driven by mechanism model and image fusion, comprising:

[0030] A data acquisition module is used to collect photovoltaic panel image data and monitoring data with the same time stamp; the monitoring data includes light intensity, ambient temperature, wind speed, current and voltage information;

[0031] Photovoltaic panel mechanism module, used to calculate current-voltage experimental data based on monitoring data;

[0032] A defect recognition module, used for inputting the image data and current-voltage experimental data into a trained layered recognition model for explicit and implicit defects to obtain recognition results of explicit and implicit defects of photovoltaic panels;

[0033] Among them, the hierarchical identification model of explicit and implicit defects includes a defect classification network, an explicit defect recognition network and a implicit defect recognition network connected in series in sequence; the defect classification network is used to identify whether there are defects in the photovoltaic panel based on current-voltage experimental data; if there are defects, the image data corresponding to the current-voltage experimental data is input into the explicit defect recognition network to identify the explicit defects; the current-voltage experimental data corresponding to the image data without explicit defects is input into the implicit defect recognition network to identify the implicit defects.

[0034] In a third aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in a photovoltaic panel defect detection method driven by a mechanism model and image fusion as described in the first aspect.

[0035] In a fourth aspect, the present invention provides a computer 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 steps in the photovoltaic panel defect detection method driven by a mechanism model and image fusion as described in the first aspect are implemented.

[0036] Compared with the prior art, the present invention has the following beneficial effects:

[0037] 1. The present invention constructs a photovoltaic panel mechanism model through environmental data, integrates electrical data and image data, and introduces constraints of the photovoltaic panel mechanism model in the deep neural network training process, which enhances the model's understanding of the physical process and realizes the fusion of model and image to drive the detection of photovoltaic panel defects;

[0038] 2. The present invention uses the explicit and implicit hierarchical decision-making method to first detect the visible explicit defect type of the photovoltaic panel, and then detect the implicit defect type inside the photovoltaic panel. This solves the problem that the electrical data of the explicit defect is misjudged as the implicit defect type, and the image data of the implicit defect is misjudged as defect-free, which greatly improves the model detection accuracy;

[0039] 3. The present invention greatly improves the detection efficiency by realizing an automated photovoltaic panel defect detection process, significantly reduces the reliance on manual detection, and improves the operational management efficiency of large-scale photovoltaic power stations.

[0040] Advantages of additional aspects of the present invention will be given in part in the following description, and in part will become obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The accompanying drawings constituting a part of the present invention are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their description are used to explain the present invention but do not constitute a limitation of the present invention.

[0042] Figure 1 A main flow chart of a photovoltaic panel defect detection method driven by a mechanism model and image fusion provided by an embodiment of the present invention;

[0043] Figure 2 A structural diagram of a layered identification model for explicit and implicit defects provided by an embodiment of the present invention;

[0044] Figure 3 A schematic diagram of a photovoltaic panel defect detection system driven by a mechanism model and image fusion provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0045] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0046] Embodiment 1

[0047] like Figure 1 As shown, this embodiment discloses a photovoltaic panel defect detection method driven by mechanism model and image fusion, comprising the following steps:

[0048] S1: Collecting photovoltaic panel image data and monitoring data with the same time stamp; the monitoring data includes light intensity, ambient temperature, wind speed, current and voltage information;

[0049] S2: Collecting photovoltaic panel image data and monitoring data with the same time stamp; the monitoring data includes light intensity, ambient temperature, wind speed, current and voltage information;

[0050] S3: inputting the image data and current-voltage experimental data into the trained layered recognition model for explicit and implicit defects to obtain recognition results of explicit and implicit defects of photovoltaic panels;

[0051] Among them, the hierarchical identification model of explicit and implicit defects includes a defect classification network, an explicit defect recognition network and a implicit defect recognition network connected in series in sequence; the defect classification network is used to identify whether there are defects in the photovoltaic panel based on current-voltage experimental data; if there are defects, the image data corresponding to the current-voltage experimental data is input into the explicit defect recognition network to identify the explicit defects; the current-voltage experimental data corresponding to the image data without explicit defects is input into the implicit defect recognition network to identify the implicit defects.

[0052] Next, combine Figure 1 , a photovoltaic panel defect detection method driven by a mechanism model and image fusion disclosed in this embodiment is described in detail.

[0053] 1. Data Collection

[0054] The data acquisition module is used to collect real-time operating data of photovoltaic panels, including image data and monitoring data, where the monitoring data includes light intensity, ambient temperature, wind speed, etc. For the collected image data and monitoring data, it is necessary to ensure that the data is aligned to a unified timestamp.

[0055] 1. Image acquisition: Use a drone equipped with a high-resolution camera to collect on-site photovoltaic panel image data. Perform pre-processing operations such as image enhancement, position correction, and uniform image size on the collected images, and then convert the image into image Aggregated into the photovoltaic panel defect recognition database.

[0056] 2. Monitoring data collection: Mainly responsible for collecting real-time data of the operating environment of photovoltaic panels, including parameters such as light intensity, ambient temperature, wind speed, current and voltage, and performing pre-processing such as data cleaning and data denoising on the collected monitoring data.

[0057] Among them, the light intensity data G uses a light meter sensor to measure the solar radiation intensity, which is installed in a position parallel to the photovoltaic panel to ensure that the measured light intensity is consistent with the light intensity received by the photovoltaic panel.

[0058] The ambient temperature data T uses a temperature sensor to measure the ambient temperature around the photovoltaic panel. The installation location should be selected to avoid direct sunlight and human interference.

[0059] Wind speed data W uses an anemometer to measure the wind speed around the photovoltaic panels. The anemometer should be installed in a developed area to avoid being affected by photovoltaic panels, buildings and other obstacles.

[0060] The current I and voltage V data use current sensors and voltage sensors to measure the output current and voltage of the photovoltaic panel. The sensor is directly installed at the output end of the photovoltaic panel to record the current and voltage data in real time.

[0061] (II) Labeling

[0062] Before carrying out the labeling task, the labeling strategy is set according to the defect type and characteristics of the photovoltaic panel, and the defects are classified into explicit defects and implicit defects.

[0063] There are three main types of explicit defects: cracks, breakages, and occlusions, which are composed of the defect types visualized in the photovoltaic panel image. The defect types existing in the collected photovoltaic panel visible light image samples are annotated to form a training data set. The annotators use the annotation tools to interactively annotate the explicit defect type labels existing in the photovoltaic panel.

[0064] There are three main types of hidden defects: hidden cracks, current leakage, and PID effects. Their visible light images are normal and defect-free, and internal defects cannot be identified. Therefore, the current and voltage data in the monitoring data can be used to draw the actual current-voltage curve and mark the internal defect types of the photovoltaic panel as hidden defect type labels.

[0065] The same image is annotated multiple times, and the implicit labels are no longer annotated for the images that are annotated with explicit labels. The annotated images are saved in the same format and stored together with the labels in the photovoltaic panel defect recognition and annotation database. Since the monitoring data and image data are aligned in time, the defect labels annotated on the images are used as the labels of the monitoring data.

[0066] The data with category labels stored in the annotation database is used as training data and input into the deep neural network defect recognition module to train the defect recognition model.

[0067] (III) Photovoltaic panel mechanism model

[0068] The photovoltaic panel mechanism model module builds a mechanism model based on the physical properties and working principles of the photovoltaic panel. It mainly builds the photovoltaic panel mechanism model through information such as light intensity (G), ambient temperature (T) and wind speed (W), and calculates the current-voltage characteristic curve of the photovoltaic panel under normal working conditions.

[0069] The photovoltaic mechanism model is constructed using a single diode model, which is mainly composed of the photocurrent I ph , diode saturation current I0, series resistance R s , parallel resistance R sh and diode ideal factor n to describe the output characteristics of the photovoltaic panel. The output current I m and output voltage V m Satisfies the following equation:

[0070]

[0071] in, is the thermal voltage, k is the Boltzmann constant, and q is the electron charge.

[0072] The calculation process of the five parameters is as follows:

[0073] 1. First calculate the operating temperature T of the photovoltaic panel c , which is affected by light intensity, ambient temperature and wind speed, and is calculated as follows:

[0074]

[0075] Among them, G ref It is the light intensity under standard test conditions (STC), usually 1000W / m 2; T NOCT The nominal operating temperature is 22°C at ambient temperature and 800W / m 2 , measured under windless conditions; τ wind speed influence coefficient, which indicates the influence of wind speed on the temperature of photovoltaic panels.

[0076] 2. Calculate the five parameters I under STC conditions ph ,I0,R s , R sh and n, where the photocurrent I ph and series resistance R s With light intensity G and operating temperature T c Related; diode saturation current I0 and diode ideal factor n and operating temperature T c Related; Parallel resistance R sh Related to the light intensity G.

[0077] 3. Substitute the five parameters into the current-voltage equation of the photovoltaic panel mechanism model to calculate the theoretical current-voltage curve of the photovoltaic panel under the input of light intensity, ambient temperature and wind speed data at a certain moment.

[0078] (IV) Hierarchical identification model for explicit and implicit defects

[0079] This embodiment builds a defect recognition model composed of an implicit defect recognition network and an explicit defect recognition network based on the explicit and implicit hierarchical decision method. By using the current-voltage data output by the photovoltaic panel mechanism model, the actual current-voltage data, and the photovoltaic panel image data as the input data for training and detection, the explicit and implicit hierarchical decision method is adopted, and the constraints of the mechanism model are introduced to guide different networks to perform defect recognition training and detection, thereby accurately identifying various defects of photovoltaic panels. The overall process is as follows Figure 2 As shown:

[0080] 1. Defect Classification Network and Training

[0081] The defect classification network adopts a mechanism-constrained residual network (MCResNet) architecture, which embeds the mechanism laws of the mechanism model as constraints into the ResNet training process.

[0082] The defect classification network includes a defect recognition binary classification unit, which is used to distinguish between defect-free photovoltaic panels θ normal and photovoltaic panel defect θ abnormal Among them, the photovoltaic panel has no defects θ normal The corresponding label is represented by y=0, and the photovoltaic panel defect θ abnormal The corresponding label is represented by y=1. Photovoltaic panel defect θ abnormal The category includes three types of dominant defects {θ d1 ,θ d2 ,θd3}、Three types of hidden defects {θ r1 ,θ r2 ,θ r3} and a class of dominantly defect-free and recessively defective (i.e., the case classified as non-defective in the explicit defect recognition network, which includes three types of implicit defects); where θ d1 ,θ d2 ,θ d3 They are crack, damage, and occlusion, respectively, and the corresponding labels are represented by y=2,3,4; For explicit no defect and implicit defect, the corresponding label is y=5; θ r1 ,θ r2 ,θ r3 They are hidden crack, current leakage, and PID effect, and the corresponding labels are y=6, 7, and 8.

[0083] The specific training steps are as follows:

[0084] S321: Deviation data processing, according to the output current I of the mechanism model m and output voltage V m Draw the current-voltage curve m , the theoretical current-voltage curve l m The actual measured current-voltage curve l0 is then fused to calculate the deviation current-voltage data l res = l m -l0, and then the deviation current-voltage data l res Output current-voltage data with mechanism model m Add the current-voltage experimental data l = l res +l m , and divide l into training set l train and the test set l test .

[0085] It should be understood that the calculations related to the curves can be performed by those skilled in the art.

[0086] S322: Build a residual network model, initialize network weights, and train current-voltage data l train Input into the defect recognition binary classification unit for training and output the classification result.

[0087] S323: Calculate the loss function to update the ResNet model parameters, and embed the photovoltaic panel mechanism model equation as a constraint into the network model training so that the network model satisfies the physical laws. IV Including network model loss terms and mechanism model loss terms, the first loss function is calculated as follows:

[0088]

[0089] Among them, λ1 and λ2 are the weights of the network model loss term and the mechanism model loss term; N is the number of samples; C is the number of categories, which are two categories, namely, photovoltaic panels without defects and photovoltaic panels with defects; is the classification probability; y ij is the true fault type label of the i-th sample. If sample i belongs to the j-th class, then y ij =1, otherwise y ij =0;1(y i =j) is the indicator function, y i is the true label of the i-th sample, when y i =c, the value is 1, otherwise it is 2; l res_i is the deviation current-voltage curve of the ith sample, and the deviation data is used as the difference, that is, only the relative change is considered to prevent the value of the loss term of the mechanism model from being too large; l target_c is the target deviation current-voltage curve corresponding to category c; defined according to previous knowledge or physical laws, l target_c Desirable target_0 and l target_1 :When l res_i For a sample without defects, select l target_0 Calculate, l target_0 is the target deviation current-voltage curve for the non-defective category of photovoltaic panels; when l res_i For defective samples, select l target_1 Calculate, l target_1 Target deviation current-voltage curve for PV panel defect categories.

[0090] In this embodiment, the selection of current-voltage data is considered in detail. For the samples in network training, current-voltage experimental data is needed, because when the photovoltaic panel is defect-free, the actual measured current-voltage curve and the theoretical current-voltage curve generated by the mechanism model may be the same, and the deviation may be 2 after subtracting the deviation. The data of 2 does not help the training network, so the experimental data is the ideal data generated by the mechanism model plus the deviation data. As for the calculation of the loss function, the purpose is to make a difference with a target deviation data to form a loss term of the mechanism constraint. Therefore, there is no need to use experimental data, and the calculation of the mechanism constraint loss can be completed using the deviation data; at the same time, the deviation data value is small, which can effectively prevent the loss of the mechanism constraint term from being too large, thereby affecting the loss function to adjust and update the model parameters.

[0091] This embodiment constructs the mechanism model loss term by using the deviation data curve between the mechanism model output and the actual measured value, introduces mechanism constraints, and better adapts to classification tasks and training network models.

[0092] S324: Use test set l test The trained model is evaluated to assess whether the defect classification network can correctly classify defects based on the six defect categories and obtain defect detection classification results.

[0093] In this embodiment, the current-voltage curve output by the mechanism model is an ideal value generated under the current environment. By performing deviation processing on the actual measured current-voltage data, it is equivalent to realizing the fusion of the two data to obtain current-voltage curve data with rich feature information.

[0094] Furthermore, by using the mechanism model in the current environment to generate an ideal value, this value can be used as a benchmark value in the current environment and time and space. By performing deviation processing on the measured value, a deviation value is obtained, and then the deviation value is added to the ideal value and input into the classification detection model, so that the input feature information is richer and helps defect detection.

[0095] In this embodiment, the photovoltaic panel mechanism model equation is constructed based on the principles of physics and the working principle of photovoltaic panels, and can accurately describe the behavior of photovoltaic panels under different conditions. These equations are embedded in the network model as constraints to ensure that the network model follows the laws of physics while learning. Since the network model is constrained by the laws of physics during the training process, it can reduce the prediction error caused by data noise, model overfitting, etc., making the prediction results of the model more reliable and improving the physical accuracy of the model output.

[0096] 2. Overt defect recognition network and training

[0097] When the defect classification network classifies the defect as a defective type, the explicit defect recognition network is activated to detect defects in the photovoltaic panel image data. The Vision Transformer (ViT) model architecture is used for the explicit defect recognition network, and the network model is trained using labeled image data. The specific training steps are as follows:

[0098] S311: Image data preprocessing, including background removal, image size adjustment, normalization and other operations, and dividing the processed images into training sets and test sets. r1 ,θ r2 ,θ r3} as a new tag, i.e. This label corresponds to an explicit defect-free classification result for the training image.

[0099] S312: Construct a ViT network model, initialize network parameters, input the image training set into the network for training, and obtain classification results.

[0100] S313: Calculate the model loss function and update the model parameters. The cross entropy loss function is used as the optimization target. The second loss function is calculated as follows:

[0101]

[0102] Among them, G is the number of categories, there are four categories M is the sample size; is the classification probability; y pq is the true label.

[0103] S314: Use the image test set to evaluate the trained model and obtain the defect detection classification result. When the explicit defect recognition network classifies the image as defect-free, the implicit defect recognition multi-classification network is activated, and the current-voltage experimental data corresponding to the image is input into the implicit defect recognition multi-classification network to detect the implicit defects of the photovoltaic panel.

[0104] 3. Hidden defect recognition network and training

[0105] The structure of the hidden defect recognition network is similar to that of the defect classification network, except that the defect recognition binary classification unit is replaced by a hidden defect recognition multi-classification unit.

[0106] The multi-classification unit for hidden defect recognition performs multi-classification tasks, corresponding to three hidden defect categories {θ r1 ,θ r2 ,θ r3}. The corresponding third loss function is as follows:

[0107]

[0108] Among them, K is the number of samples; C is the number of categories, which are three categories, namely hidden cracks, current leakage, and PID effect; l target_c is the target deviation current-voltage curve corresponding to category c, defined according to prior knowledge or physical laws, l target_6 is the target deviation current-voltage curve of the crack category, l target_7 is the target deviation current-voltage curve for the current leakage category, l target_8 The target deviation current-voltage curve for the PID effect category.

[0109] The specific implementation process will not be repeated here. The test set l is also used. test The trained model is evaluated to see whether the hidden defect recognition network can accurately classify the three hidden defect categories and obtain the defect detection classification results.

[0110] The defect recognition binary classification unit of the defect classification network and the latent defect recognition multi-classification network of the latent defect recognition network in this embodiment are used to solve the situation where the current-voltage curves of explicit defects and implicit defects are similar, and avoid defect recognition errors caused by similar curves. Since explicit defects and implicit defects may show similar characteristics on the current-voltage curve, the traditional single classification method is prone to misjudgment. By introducing a multi-classification latent defect recognition network, this embodiment can more finely distinguish different types of latent defects on the basis of determining that there are no explicit defects, reduce recognition errors caused by similar curves, and improve the accuracy of defect recognition.

[0111] As a specific implementation method, the explicit and implicit decision fusion method mainly uses the explicit and implicit defect recognition network to make hierarchical decisions to obtain classification results. The photovoltaic panel defect detection task has explicit defects and implicit defects. First, an overall defect detection is performed to detect whether the photovoltaic panel has defects. If there are defects, the defect data is explicitly detected, and the detection is stopped when the detection result is an explicit defect. If no explicit defects are detected, the implicit defects are further detected. The main steps of this method are:

[0112] (1) Overall defect detection: First, a defect classification network is used to detect X IV The data is used for defect detection, and the classification categories are defect-free, θ normal and defective θ abnormal ={θ d1 ,θ d2 ,θ d3 ,θ r1 ,θ r2 ,θ r3}, if the test result is no defect, then the decision result y is output binary , if there is a defect, the explicit defect recognition network is activated;

[0113] (2) Overt defect detection: The defective X IV_abnormal The image data corresponding to the data X image Input to the explicit defect recognition network for defect detection. If the detection result is defective θ' normal ={θ d1 ,θ d2 ,θ d3}, then there are three types of dominant defects {θ d1 ,θ d2 ,θ d3}, output the decision result y explicit , if the inspection result shows no defect θ a ' bnormal ={θ r1 ,θ r2 ,θ r3} then activate the hidden defect recognition network;

[0114] (3) Hidden defect detection: Transform defect-free image data into image_normal Corresponding The data is input into the hidden defect recognition network for defect detection. The detection results are three types of hidden defects {θ r1 ,θ r2 ,θ r3}, output the decision result y multi ;

[0115] (4) Decision completion: The decision results of the three networks are summarized and the final decision result y is output.

[0116] y={y binary ||y explicit ||y multi} (6)

[0117] The main embodiment adopts a trained deep neural network defect recognition module to perform real-time diagnosis of photovoltaic panel defects on site. By deploying a data acquisition module on site, the image information, environmental information and current-voltage information of the photovoltaic panel are collected, and the three types of information are preprocessed. The preprocessed environmental information is input into the photovoltaic panel mechanism model to obtain the corresponding theoretical current-voltage data, and the theoretical current-voltage data and the measured current-voltage data are subjected to deviation data processing to obtain the current-voltage experimental data. The current-voltage experimental data and image data are input into the trained explicit and implicit defect hierarchical recognition model, and the current-voltage experimental data is first used for overall defect diagnosis. If a defect is diagnosed, the image data input is introduced, and the explicit defect diagnosis is activated. In the explicit defect diagnosis, if no defect is detected, the corresponding current-voltage experimental data is introduced again, and the implicit defect diagnosis is activated for diagnostic operation. The method provided in this embodiment significantly improves the accuracy and efficiency of fault detection, and effectively avoids the misjudgment that may be caused by a single information source.

[0118] When an obvious or hidden fault is diagnosed, a fault alarm can be issued, and the operation and maintenance personnel can promptly perform maintenance on the photovoltaic panels. At the same time, the diagnosis results and maintenance records are uploaded to the photovoltaic panel diagnostic database to facilitate subsequent management and analysis of fault types and maintenance methods.

[0119] Embodiment 2

[0120] like Figure 3 As shown, this embodiment provides a photovoltaic panel defect detection system driven by mechanism model and image fusion, including:

[0121] A data acquisition module is used to collect photovoltaic panel image data and monitoring data with the same time stamp; the monitoring data includes light intensity, ambient temperature, wind speed, current and voltage information;

[0122] Photovoltaic panel mechanism module, used to calculate current-voltage experimental data based on monitoring data;

[0123] A defect recognition module, used for inputting the image data and current-voltage experimental data into a trained layered recognition model for explicit and implicit defects to obtain recognition results of explicit and implicit defects of photovoltaic panels;

[0124] Among them, the hierarchical identification model of explicit and implicit defects includes a defect classification network, an explicit defect recognition network and a implicit defect recognition network connected in series in sequence; the defect classification network is used to identify whether there are defects in the photovoltaic panel based on current-voltage experimental data; if there are defects, the image data corresponding to the current-voltage experimental data is input into the explicit defect recognition network to identify the explicit defects; the current-voltage experimental data corresponding to the image data without explicit defects is input into the implicit defect recognition network to identify the implicit defects.

[0125] The explicit and implicit defect hierarchical recognition model provided in this specific embodiment realizes the refined classification of photovoltaic panel defects by sequentially connecting a defect classification network, an explicit defect recognition network and a implicit defect recognition network in series.

[0126] First, the defect classification network preliminarily determines whether the photovoltaic panel has defects based on the current-voltage experimental data, which helps to quickly screen out photovoltaic panels that need further inspection and reduce invalid inspections; then, for photovoltaic panels with defects, the model will input the image data into the explicit defect recognition network or the implicit defect recognition network for further analysis based on its current-voltage experimental data. This targeted detection method improves the accuracy and efficiency of detection. The design of the series network enables the model to be more targeted when processing image data. Only when the photovoltaic panel is preliminarily judged to have defects will the explicit defect recognition network or the implicit defect recognition network be further called for processing, thereby avoiding indiscriminate processing of all photovoltaic panel image data and saving computing resources and time.

[0127] Embodiment 3

[0128] This embodiment provides a computer-readable storage medium having a computer program stored thereon. When the program is executed by a processor, the steps in the photovoltaic panel defect detection method driven by a mechanism model and image fusion as described in the first embodiment above are implemented.

[0129] Embodiment 4

[0130] This embodiment provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the steps in the photovoltaic panel defect detection method driven by a mechanism model and image fusion as described in the first embodiment above are implemented.

[0131] The steps or modules involved in the above embodiments 2 to 4 correspond to those in embodiment 1. For the specific implementation, please refer to the relevant description of embodiment 1. The term "computer-readable storage medium" should be understood as a single medium or multiple media including one or more instruction sets; it should also be understood to include any medium that can store, encode or carry an instruction set for execution by a processor and enable the processor to execute any method in the present invention.

[0132] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A photovoltaic panel defect detection method driven by mechanism model and image fusion, characterized in that: include: Collecting photovoltaic panel image data and monitoring data with the same time stamp; the monitoring data includes light intensity, ambient temperature, wind speed, current and voltage information; Calculate current-voltage experimental data based on monitoring data; Input the image data and current-voltage experimental data into a trained layered recognition model for explicit and implicit defects to obtain recognition results of explicit and implicit defects of photovoltaic panels; Among them, the hierarchical identification model of explicit and implicit defects includes a defect classification network, an explicit defect recognition network and a implicit defect recognition network connected in series in sequence; the defect classification network is used to identify whether there are defects in the photovoltaic panel based on current-voltage experimental data; if there are defects, the image data corresponding to the current-voltage experimental data is input into the explicit defect recognition network to identify the explicit defects; the current-voltage experimental data corresponding to the image data without explicit defects is input into the implicit defect recognition network to identify the implicit defects.

2. The photovoltaic panel defect detection method driven by mechanism model and image fusion as claimed in claim 1, characterized in that: The calculation of current-voltage experimental data according to the monitoring data specifically includes: According to the photovoltaic panel mechanism, the theoretical current-voltage data is calculated based on the light intensity, ambient temperature and wind speed information; Based on the current and voltage information, calculate the actual current-voltage data; Subtract the theoretical current-voltage data from the actual current-voltage data to obtain the deviation current-voltage data; The theoretical current-voltage data and the deviation current-voltage data are added together to obtain the current-voltage experimental data.

3. The photovoltaic panel defect detection method driven by mechanism model and image fusion as claimed in claim 1, characterized in that: The method further includes: performing preprocessing operations of image enhancement, position correction, and uniform image size on the image data; and performing preprocessing of data cleaning and data denoising on the monitoring data.

4. The photovoltaic panel defect detection method driven by mechanism model and image fusion as claimed in claim 1, characterized in that: The explicit defects include cracks, breakages and obstructions; the implicit defects include hidden cracks, current leakage and PID effects.

5. The photovoltaic panel defect detection method driven by mechanism model and image fusion as claimed in claim 1, characterized in that: The defect classification network is a binary classification network constructed based on a residual network; The training process of the defect classification network includes: Inputting the current-voltage experimental data samples into the defect classification network, and training based on the non-defect label and the defect label; the defect label includes an explicit defect label and an implicit defect label; the explicit defect label includes a crack label, a damage label and an occlusion label, and the implicit defect label includes a hidden crack label, a current leakage label and a PID effect label; The first loss function is calculated. When the first loss function is minimized, the defect classification network training is completed.

6. The photovoltaic panel defect detection method driven by mechanism model and image fusion as claimed in claim 1, characterized in that: The training process of the dominant defect recognition network includes: Obtaining image data samples corresponding to current-voltage experimental data samples identified as having defects by the defect classification network; Inputting the image data sample into an explicit defect recognition network, taking the implicit defect label as a whole as an explicit non-defect label, and performing training based on the explicit defect label and the explicit non-defect label; The second loss function is calculated. When the second loss function is minimized, the training of the dominant defect recognition network is completed.

7. The photovoltaic panel defect detection method driven by mechanism model and image fusion as claimed in claim 1, characterized in that: The training process of the hidden defect recognition network includes: Acquire a current-voltage experimental data sample corresponding to an image data sample identified as being dominantly defect-free by a dominant defect recognition network; Inputting the current-voltage experimental data samples into a hidden defect recognition network and performing training based on hidden defect labels; The third loss function is calculated. When the third loss function is minimum, the hidden defect recognition network training is completed.

8. A photovoltaic panel defect detection system driven by mechanism model and image fusion, characterized in that: include: A data acquisition module is used to collect photovoltaic panel image data and monitoring data with the same time stamp; the monitoring data includes light intensity, ambient temperature, wind speed, current and voltage information; Photovoltaic panel mechanism module, used to calculate current-voltage experimental data based on monitoring data; A defect recognition module, used for inputting the image data and current-voltage experimental data into a trained layered recognition model for explicit and implicit defects to obtain recognition results of explicit and implicit defects of photovoltaic panels; Among them, the hierarchical identification model of explicit and implicit defects includes a defect classification network, an explicit defect recognition network and a implicit defect recognition network connected in series in sequence; the defect classification network is used to identify whether there are defects in the photovoltaic panel based on current-voltage experimental data; if there are defects, the image data corresponding to the current-voltage experimental data is input into the explicit defect recognition network to identify the explicit defects; the current-voltage experimental data corresponding to the image data without explicit defects is input into the implicit defect recognition network to identify the implicit defects.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps in a photovoltaic panel defect detection method driven by a mechanism model and image fusion as described in any one of claims 1 to 7 are implemented.

10. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps in the photovoltaic panel defect detection method driven by mechanism model and image fusion as described in any one of claims 1-7 are implemented.

Images