awesome_PV_faults_diagnosis

KO | EN

gitlite — search

#hacktoberfest #ai #typescript #go #rust #macos #nodejs #java #windows #awesome-list #vue #pytorch #mcp #open-source

awesome_PV_faults_diagnosis

(★ 25)

literature review renewal of PV faults diagnosis, including papers, datasets, and codes

awesome_PV_faults_diagnosis 최신버젼 다운로드

최종 버전 다운로드 (.zip)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

[![en](https://img.shields.io/badge/lang-en-blue.svg)](https://github.com/574168985/awesome_PV_faults_diagnosis/edit/main/README.md)
[![pt-br](https://img.shields.io/badge/lang-fr-white.svg)](https://github.com/574168985/awesome_PV_faults_diagnosis/edit/main/README.fr.md)
[![es](https://img.shields.io/badge/lang-cn-red.svg)](https://github.com/574168985/awesome_PV_faults_diagnosis/edit/main/README.cn.md)

# awesome_PV_faults_diagnosis
literature review renewal of PV faults diagnosis, including papers, datasets, and codes

[![image](https://github.com/574168985/awesome_PV_faults_diagnosis/assets/52654110/b069499a-c777-4e8e-b935-90b0503bc2c2)](https://www.mdpi.com/1996-1073/16/10/4012)

![数据集大小x准确率](https://github.com/574168985/awesome_PV_faults_diagnosis/assets/52654110/c486d9ff-1f8e-4ed9-8546-bf6a6325d811)

[![image](https://github.com/574168985/awesome_PV_faults_diagnosis/assets/52654110/8dad156e-27a0-443f-a726-099593e9f460)](https://www.sciencedirect.com/science/article/pii/S030626192301499X#fig1)

# PLAN

To start literature review, we focus on:

- [ ] [Image-based diagnosis for PV cells, modules and plants](#IMAGE-BASED-METHODS)
- [ ] [Fusion-based (images and other signals (electrical, thermal, …)](#FUSION-BASED-METHODS)
- [ ] [Find if there exist public datasets with PV images for diagnosis](#PUBLIC-IMAGE-BASED-DATASET-SUMMARY )

# IMAGE BASED METHODS

What kind of images, how they are collected, processed, which kind of faults (type, severity, …), the application level (cell, module, string, plant), the results (simulation or experimental) accuracy and so on

[![image](https://github.com/574168985/awesome_PV_faults_diagnosis/assets/52654110/a5536fd5-a3a1-4d59-bf75-fc20a9d17e93)](https://www.sciencedirect.com/science/article/pii/S1364032123008973)

# FUSION BASED METHODS

What kind of images, how they are collected, processed, how the fusion is done, which kind of faults, the application, the results (simulation or experimental) accuracy and so on

It would be relevant to check if there is any paper review. If not we can be the first to do so.

[![image](https://github.com/574168985/awesome_PV_faults_diagnosis/assets/52654110/d4c4f744-b2de-42a7-9657-82eb7e07fef1)](https://www.sciencedirect.com/science/article/pii/S1364032123008973)

# PUBLIC IMAGE BASED DATASET SUMMARY

todo

## 【2023】IMAGE BASED METHODS

- [![Type de document](https://img.shields.io/badge/Journal-8A2BE2)|International Journal of Photoenergy，Q4] [![Type d'image](https://img.shields.io/badge/ImageType-2b83e2)|RGB] [![收集方法](https://img.shields.io/badge/CollectionMethod-e28a2b)|UAV] [![处理方法](https://img.shields.io/badge/ImageProcessing-83e22b)|Blur, flipping, rotation, noise, translation, zoom, and distortion] [![故障类型](https://img.shields.io/badge/FaultType-ff0000)|Normal plates, snail marks, layer cracks, broken glass, discoloration, and burns] [![应用层次](https://img.shields.io/badge/ApplicationLevel-009999)|pannel] [![数据规模](https://img.shields.io/badge/DataSize-990066)|Initially 600, extended to 3,150] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|CNN+Classifier (without using the AI classifier)] [![准确率](https://img.shields.io/badge/Metrics-660066)|Pretend to use DenseNet-201 with k-nearest neighbors on a self-constructed dataset to achieve 100] [Enhancing Photovoltaic Module Fault Diagnosis with Unmanned Aerial Vehicles and Deep Learning-Based Image Analysis](https://www.hindawi.com/journals/ijp/2023/8665729/) **[Report|In multi-classification tasks, k-nearest neighbors is the best among non-fully connected classifiers.]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|IEEE Transactions on Industry Applications，Q2] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|IR+EL] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|11/2] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|array] [![DataSize](https://img.shields.io/badge/DataSize-990066)|20000/2,624] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|MPViT（Multi-Path Vision Transformer for Dense Prediction）+ Self-Attention Layer] [![Metrics](https://img.shields.io/badge/Metrics-660066)|The claimed accuracies for binary classification and multi-class classification experiments on dataset1 and dataset2 are 94.1% and 88.5%, and 90.7% and 86.4%, respectively.] [Photovoltaic Panel Fault Detection and Diagnosis Based on a Targeted Transformer-Style Model](https://ieeexplore.ieee.org/abstract/document/10274093) **[Report|The local detailed features can enhance the discriminative capability of the model for different categories of solar panel images.]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|Aswan University Journal of Sciences and Technology] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|IR] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|lab] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|array] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|Claimed to achieve 99.98% accuracy on a self-built dataset.] [Solar Cell Anomaly Detection Based on Wavelet Scattering Transform and Artificial Intelligence](https://aujst.journals.ekb.eg/article_312683.html) **[Report|Using Wavelet Scattering Transform (WST) to better represent image features for multi-class anomaly detection in infrared images.]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|Information Systems Research in Vietnam] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|EL] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|array] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Case Study: Utilising of Deep Learning Models for Fault Detection and Diagnosis of Photovoltaic Modules to Improve Solar Energy Constructions’ O&M Activities Quality](https://link.springer.com/chapter/10.1007/978-981-99-4792-8_5) **[Report|The use of unsupervised learning for detecting and diagnosing visual faults and defects within the module.]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|Computational Intelligence and Neuroscience no longer accepts submissions] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|EL] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|4] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|Cell] [![DataSize](https://img.shields.io/badge/DataSize-990066)|2,624] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Fault-Level Grading of Photovoltaic Cells Employing Lightweight Deep Learning Models](https://www.hindawi.com/journals/cin/2023/2663150/) **[Report|CPU-based model]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|IET Renewable Power Generation，Q4] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|IR] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|array] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|MobileNet-V3] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Automatic defect identification of PV panels with IR images through unmanned aircraft](https://ietresearch.onlinelibrary.wiley.com/doi/full/10.1049/rpg2.12831) **[Report|MobileNet V3 > Traditional Convolutional Neural Networks (CNNs)]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|Sustainability，Q3] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|IR] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|The seven defects include sand accumulation on PV modules, covered PV modules, cracked PV modules, degradation, dirty PV modules, short-circuited PV modules, and overheated bypass diodes.] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|array] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|Hybrid CNN–ML Model] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Embedded Hybrid Model (CNN–ML) for Fault Diagnosis of Photovoltaic Modules Using Thermographic Images](https://www.mdpi.com/2071-1050/15/10/7811) **[Report|Deploying Raspberry Pi 4, this embedded solution is feasible in the diagnosis of PV modules.]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|MULTIMEDIA SYSTEMS，Q4] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|IR] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|hotspot] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|array] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|Using an improved semantic segmentation model (Attention DeepLab) for object detection tasks.] [![Metrics](https://img.shields.io/badge/Metrics-660066)|89.6%] [Hotspot defect detection for photovoltaic modules under complex backgrounds](https://link.springer.com/article/10.1007/s00530-023-01176-w) **[Report|"Add a prediction head for detecting minor hotspots in the predictive network. / Modify the path aggregation network in the feature fusion network by integrating multi-scale feature maps to enhance semantic information. / Apply an efficient channel attention module in the feature fusion network to eliminate confusion.]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|Soft Computing，Q3] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|RGB] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|array] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|YOLOv7-GX] [![Metrics](https://img.shields.io/badge/Metrics-660066)|93.4%] [YOLOv7-GX-based defect detection for PV panels](https://www.researchsquare.com/article/rs-3118568/v1) **[Report|In the training process, allocate different weights to various images and categories to address the issue of imbalanced sample distribution.]**

- [![文献类型](https://img.shields.io/badge/arxiv-8A2BE2)|arxiv] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|RGB] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|UAV] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|Hotspot, Snail Trail, Fissure] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|array] [![DataSize](https://img.shields.io/badge/DataSize-990066)|Simulated and Real Image Datasets] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|Unsupervised Sensor Algorithm and 3D Augmented Reality Visualization] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Unveiling the Invisible: Enhanced Detection and Analysis of Deteriorated Areas in Solar PV Modules Using Unsupervised Sensing Algorithms and 3D Augmented Reality](https://arxiv.org/abs/2307.05136) **[Report|Unsupervised Sensor Algorithm and a Novel Integration with 3D Augmented Reality Visualization]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|JOURNAL OF CIRCUITS SYSTEMS AND COMPUTERS，Q4] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|RGB] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|UAV] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|Defect Detection and Color Classification] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|U-Net] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [A Deep Learning-Based Surface Defects Detection and Color Classification Method for Solar Cells](https://www.worldscientific.com/doi/abs/10.1142/S0218126623501566) **[Report|Segmenting small objects allows for the detection of surface defects.]**

- [![文献类型](https://img.shields.io/badge/Conference-8A2BE2)|Chinese Control Conference (CCC)] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|UAV] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|In YOLOv5, CPP-Neck has been substituted for the PANet neck network. CPP-Neck consists of CBPN and PA structures, aimed at enhancing the fusion of feature information in the backbone network while improving the precision of defect bounding box localization in complex backgrounds.] [![Metrics](https://img.shields.io/badge/Metrics-660066)|The claim states a 6.54% improvement in mAP50.] [YOLOv5-CPP: Improved YOLOv5-Based Defect Detection for Photovoltaic Panels](https://ieeexplore.ieee.org/abstract/document/10241184) **[Report|The target detection model is capable of addressing challenges in photovoltaic infrared images, such as weak target texture features, complex image backgrounds, and low contrast.]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|IEEE Access，Q3] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|IR] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|The thermal imaging dataset provided by AI Hub.] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|To address the issue of edge information loss when using thermal images, a new contrast enhancement module has been introduced, employing contrast enhancement techniques.] [![Metrics](https://img.shields.io/badge/Metrics-660066)|Claiming an accuracy rate of 96.17%.] [CECvT: Initial Diagnosis of Anomalies in Thermal Images](https://ieeexplore.ieee.org/abstract/document/10264866) **[Report|To address the common issue of degradation in edge information within thermal images, a contrast enhancement module has been introduced.]**

- [![文献类型](https://img.shields.io/badge/Conference-8A2BE2)|10th International Maritime Science Conference (IMSC 2023)] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|RGB] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|The short-circuit current (Isc) and maximum power point power (Pmpp) of photovoltaic (PV) panels.] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|Gaussian Process Regression (GPR), Support Vector Machine (SVM), Random Forest (RF), and Neural Network (NN)] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Using Machine Learning Techniques for Predicting Electrical Data of PV Panels from RGB Images](https://www.researchgate.net/profile/Ilija-Knezevic-2/publication/371672199_Using_Machine_Learning_Techniques_for_Predicting_Electrical_Data_of_PV_Panels_from_RGB_Images/links/648eb2fac41fb852dd0dada8/Using-Machine-Learning-Techniques-for-Predicting-Electrical-Data-of-PV-Panels-from-RGB-Images.pdf) **[Report|Using machine learning (ML) techniques to predict the electrical data of PV (photovoltaic) cells from RGB images.]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|Journal of Optics-India] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|RGB] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|Mechanical defects such as cracks and pinholes.] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|Defect detection using image processing and fuzzy logic techniques. Image processing techniques involve thresholding, mathematical morphology, and edge detection operators.] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Defect detection on solar cells using mathematical morphology and fuzzy logic techniques](https://link.springer.com/article/10.1007/s12596-023-01162-5) **[Report|Enhancing the Recognition Accuracy of Individual and Collective Defects Using the Mamdani Fuzzy Model.]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|Journal of Optics-India] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|Utilize multiple morphological operations and Canny edge detection with adaptive parameters to extract and highlight objects on solar panels. Subsequently, employ their features and a classification algorithm to categorize the detected objects, identifying various defect types and components.] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Automated visual inspection of solar cell images using adapted morphological and edge detection algorithms](https://link.springer.com/article/10.1007/s12596-023-01284-w) **[Report|The visual detection method is a non-contact approach, superior to contact-based methods.]**

## 【2022】IMAGE BASED METHODS

- [![文献类型](https://img.shields.io/badge/Conference-8A2BE2)|Chinese Control and Decision Conference (CCDC)] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|IR] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|UAV] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|Positioning and extraction, horizontal rotation, perspective distortion removal, cluster segmentation. Significance.] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|hotspot] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|array] [![DataSize](https://img.shields.io/badge/DataSize-990066)|"The image labeled 135 was captured, and 1020 photovoltaic modules were extracted from it. Among them, there are 96 positive samples, indicating the presence of hotspots in these PV module samples. Additionally, there are 924 negative samples, representing normal PV modules.] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|Grid Segmentation -> Cluster Segmentation] [![Metrics](https://img.shields.io/badge/Metrics-660066)|The claim is to have achieved 99.71% accuracy on a self-built dataset.] [Hotspot detection of photovoltaic modules in infrared thermal image based on saliency analysis](https://ieeexplore.ieee.org/abstract/document/10033497) **[Report|In binary classification tasks, clustering segmentation is superior to grid segmentation.]**

- [![文献类型](https://img.shields.io/badge/Conference-8A2BE2)|ICESEP-2022] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|IR] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|UAV] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|Change the position and size of the hotspot, crop the portion with the hotspot, and align sizes. Report rotation and flipping are not useful.] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|hotspot] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|array] [![DataSize](https://img.shields.io/badge/DataSize-990066)|The initial set comprises 39 images with hotspots and 235 images without hotspots, augmented to a total of 292 images with hotspots and 292 images without hotspots.] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|CNN+Classifier (Fully Connected Classifier)] [![Metrics](https://img.shields.io/badge/Metrics-660066)|The claim is to achieve 96.58% on a self-built dataset.] [Hot Spot Detection of Photovoltaic Module Infrared Near-field Image based on Convolutional Neural Network](https://iopscience.iop.org/article/10.1088/1742-6596/2310/1/012076/meta) **[Report|In binary classification tasks, a fully connected classifier is superior to a naive Bayes classifier.]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|IEEE Journal of Photovoltaics，Q3] [ImageType|EL] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|laser line scanning] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|Image stack -> Classification into background, electroluminescent, or laser pixels -> Reconstruction of non-contact traditional electroluminescence, highlighting defective electroluminescence -> Calibration] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|array] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Reconstruction and Calibration of Contactless Electroluminescence Images From Laser Line Scanning of Photovoltaic Modules](https://ieeexplore.ieee.org/abstract/document/9736329) **[Report|A contactless electroluminescence (EL) image processing method based on laser line scanning is proposed, generating two types of images: CCEL corresponding to voltage-biased EL (EBEL) and CHEL highlighting defects.]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|Sustainable Energy Grids & Networks，Q3] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|EL] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|darkroom] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|"Defective (a power loss exceeding 3% of the initial power output is considered as a defect), cracking, corrosion.] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|array] [![DataSize](https://img.shields.io/badge/DataSize-990066)|Prepared two datasets, 2624/1028.] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|CNN+Classifier(SVM)] [![Metrics](https://img.shields.io/badge/Metrics-660066)|The claim is to achieve 99.49% on a self-built dataset.] [A combined convolutional neural network model and support vector machine technique for fault detection and classification based on electroluminescence images of photovoltaic modules](https://www.sciencedirect.com/science/article/pii/S2352467722001916) **[Report|"1. Propose two databases, D1 (defects and normal) and D2 (normal, cracks, and corrosion), for training and evaluation. These databases comprise electroluminescence images of photovoltaic cells. 2. Defects located at the corners of the cells are considered as part of the PV cell rather than as defects.]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|IEEE Transactions on Industrial Informatics，Q1] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|EL] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|manufacturing] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|10] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|Cell] [![DataSize](https://img.shields.io/badge/DataSize-990066)|36,543] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|Claimed to achieve 99.49% on a self-built dataset.] [PVEL-AD: A Large-Scale Open-World Dataset for Photovoltaic Cell Anomaly Detection](https://ieeexplore.ieee.org/abstract/document/9744494/algorithmstabFilter=dataset#algorithms) **[Report|Provide public datasets. The first one offers a framework truth for photovoltaic solar cell anomaly detection (target detection task), a public dataset.]** HIGHLIGHT

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|mdpi data] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|RGB] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|Dynamic Reconfiguration and Maximum Power Point Tracking] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|Cell] [![DataSize](https://img.shields.io/badge/DataSize-990066)|5211] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|Estimating the Electrical Characteristics of Photovoltaic Panels Using Image-Based Approaches.] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Dataset for Detecting the Electrical Behavior of Photovoltaic Panels from RGB Images](https://www.mdpi.com/2306-5729/7/6/82) **[Report|The dataset introduces a correlation between RGB images and photovoltaic panel electrical data under various irradiance and shadow conditions. Additionally, the dataset includes supplementary weather data and other image features to support the training of estimation models. Specifically, this dataset is designed to facilitate the development of image-based models for estimating electrical data, serving as an alternative to extensive sensor arrays.]**

- [![文献类型](https://img.shields.io/badge/Conference-8A2BE2)|Chinese Automation Congress (CAC)] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|EL] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|Cracks and Fragments] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|Cell] [![DataSize](https://img.shields.io/badge/DataSize-990066)|402] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|Improve YOLOv5s by integrating spatial and channel information to focus the network on key regions in the image; reduce overlapping effects through enhanced channel attention.] [![Metrics](https://img.shields.io/badge/Metrics-660066)|The claimed average precision (mAP) achieved on the constructed defective image dataset is 0.923, showing a 3.3% improvement in accuracy compared to the original YOLOv5s algorithm.] [Defect detection for PV Modules based on the improved YOLOv5s](https://ieeexplore.ieee.org/abstract/document/10055183) **[Report|The cross-stage partial network operations units and spatial pyramid pooling operations units in the YOLOv5s network were improved by introducing a visual attention mechanism.]**

- [![文献类型](https://img.shields.io/badge/Conference-8A2BE2)|IEEE International Conference on Power and Power and Renewable Energy (ICPRE)] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|IR] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|Perform grayscale processing and data augmentation on the existing public dataset of solar cell faults.] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|10] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|36,543] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|Integrate the SENet attention mechanism into the lightweight network ShuffleNetV2.] [![Metrics](https://img.shields.io/badge/Metrics-660066)|The claim to achieve 84.06% on public datasets.] [Fault Diagnosis Method for Photovoltaic Panels Based on Improved ShuffleNet V2 and Infrared Images](https://ieeexplore.ieee.org/abstract/document/9960396) **[Report|The diverse types of faults in photovoltaic modules, coupled with their subtle differences in characteristics, result in a less effective diagnosis of certain fault types.]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|Elsevier Energy，Q1] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|IR] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|Transformation and correction, silver screen removal, non-linear interpolation, equivalent segmentation, and clustering.] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|Uneven dust accumulation on the panel surface.] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|Cell] [![DataSize](https://img.shields.io/badge/DataSize-990066)|36,543] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|A new error assessment method, called 'error loop,' has been proposed for analyzing the consistency between measurement results and experimental outcomes.] [![Metrics](https://img.shields.io/badge/Metrics-660066)|The claimed R2 and Mean Absolute Error (MAE) are 78.7% and 3.67, respectively.] [A deep residual neural network identification method for uneven dust accumulation on photovoltaic (PV) panels](https://www.sciencedirect.com/science/article/pii/S0360544221025500) **[Report|"A new method has been proposed for the identification of uneven dust accumulation on the surface of photovoltaic panels, aimed at quantitatively analyzing the dust state (concentration and distribution). However, it is still necessary to consider the effects of dust concentration and types, as well as different lighting conditions.]**

- [![文献类型](https://img.shields.io/badge/Book-8A2BE2)|onlinelibrary.wiley] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|IR] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|Edge Detection and Hough Transform] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|10] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|Cell] [![DataSize](https://img.shields.io/badge/DataSize-990066)|36,543] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|The algorithm is divided into two stages. In the first stage, localization of solar panels is carried out, along with feature extraction and analysis. In the second stage, the impact of surface area degradation of photovoltaic panels on the power output of the photovoltaic system is analyzed.] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Photovoltaic Module Fault. Part 1: Detection with Image Processing Approaches-Chapter 3](https://onlinelibrary.wiley.com/doi/abs/10.1002/9781119873785.ch3) **[Report|Analyzing the Impact of Photovoltaic Panel Surface Area Degradation on Photovoltaic System Power Output.]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|Journal of Quality Technology，Q3] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|IR] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|UAV] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|Utilizing the Transform Invariant Low-rank Textures (TILT) method to address the non-smooth background in photovoltaic images, followed by applying edge detection for cropping and aligning the images; post-processing procedures for image denoising and segmentation.] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|20 normal (without anomalies) training samples and 100 test samples.] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|The Robust Principal Component Analysis (RPCA) method is employed, which is capable of separating sparse corrupted anomaly components from a low-rank background. This enables simultaneous detection and isolation of anomalies.] [![Metrics](https://img.shields.io/badge/Metrics-660066)|The method claims a recall rate of 0.80 on a self-constructed dataset, with a maximum recall rate of 1 for detecting critical anomalies. In terms of F1 score, the approach outperforms two baseline methods by 44.5% and 114.3%, respectively.] [Online automatic anomaly detection for photovoltaic systems using thermography imaging and low rank matrix decomposition](https://www.tandfonline.com/doi/full/10.1080/00224065.2021.1948372) **[Report|The bottom edges of an image typically have normal hotspots that should not be detected as anomalies.]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|International Journal of Photoenergy，Q4] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|IR+RGB] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|10] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Infrared Thermal Images of Solar PV Panels for Fault Identification Using Image Processing Technique](https://www.hindawi.com/journals/ijp/2022/6427076/) **[Report|The thermal imaging can rapidly locate hotspots using Hough transform technology.]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|International Journal of Photoenergy] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|IR] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|10] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|Cell] [![DataSize](https://img.shields.io/badge/DataSize-990066)|36,543] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|Claimed to achieve 99.49% on a self-built dataset.] [Machine learning framework for photovoltaic module defect detection with infrared images](https://link.springer.com/article/10.1007/s13198-021-01544-7) **[Report|]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|International Journal of Photoenergy，Q3] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|IR] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|Utilize edge detection technology based on fuzzy logic to identify abnormalities in the orientation of PV modules. Additionally, employ gray-level co-occurrence matrices for extracting texture features from images. These extracted features are labeled and used to train a support vector machine classifier for the classification of fault types in PV modules.] [![Metrics](https://img.shields.io/badge/Metrics-660066)|Claimed 94.4%] [Deep learning-based model for fault classification in solar modules using infrared images](https://www.sciencedirect.com/science/article/pii/S221313882200162X) **[Report|Defect Detection in Photovoltaic (PV) Modules Using Texture Feature Analysis and Supervised Machine Learning Methods]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|International Journal of Photoenergy] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|IR] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|The thermal imaging camera mounted on a tripod.] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [An efficient optical inspection of photovoltaic modules deploying edge detectors and ancillary techniques](https://ijece.iaescore.com/index.php/IJECE/article/view/28435) **[Report|The experimental system utilizes an infrared thermal imaging camera installed on a tripod, concurrently integrating real-time electrical parameters obtained from tests conducted within the laboratory.]**

- [![文献类型](https://img.shields.io/badge/Conference-8A2BE2)|Privacy and Data Analytics: Select Proceedings of ISPDA] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|36,543] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|DenseNet] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Defect Analysis of Faulty Regions in Photovoltaic Panels Using Deep Learning Method](https://link.springer.com/chapter/10.1007/978-981-16-9089-1_5) **[Report|Adopting the DenseNet architecture to demonstrate its superior efficiency compared to CNN and ResNet-50.]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|Energy Reports，Q2] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|IR] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|hotspot] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|panel] [![DataSize](https://img.shields.io/badge/DataSize-990066)|36,543] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|改进的YOLOv5] [![Metrics](https://img.shields.io/badge/Metrics-660066)|The claimed mean average precision (mAP) has reached 87.8%, with an average recall rate of 89.0%, and an F1 score of 88.9%.] [A novel detection method for hot spots of photovoltaic (PV) panels using improved anchors and prediction heads of YOLOv5 network](https://www.sciencedirect.com/science/article/pii/S235248472201575X) **[Report|"By employing the K-means clustering algorithm on the length-width ratio of data annotation boxes, a set of smaller numerical anchor points is added to optimize the clustering process. This is aimed at improving the detection accuracy of hotspots on photovoltaic (PV) panels at different scales for the YOLOv5 network.]**

- [![文献类型](https://img.shields.io/badge/Conference-8A2BE2)|Advances in Science and Engineering Technology International Conferences (ASET)] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|hundreds] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|Transfer Learning, MobileNetV2 as Feature Extractor] [![Metrics](https://img.shields.io/badge/Metrics-660066)|In the elpv-dataset, an 85% accuracy rate in fault classification was reported, while a 92% accuracy rate in fault classification was reported for both the magnetic tile and Kolektor Surface-Defect datasets.] [Fault Detection and Identification Based on Image Processing and Deep Learning](https://ieeexplore.ieee.org/abstract/document/9734799) **[Report|Promotion in various fault detection issues such as power inverters, solar modules, and magnetic tiles, leading to a universal fault detection and localization model.]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|Sustainable Energy, Grids and Networks，Q3] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|I designed a PV system in the MATLAB/Simulink environment, utilizing real-time irradiance and temperature data obtained from the photovoltaic system at the Agartala branch of the National Institute of Technology. The constructed dataset is used for feature extraction, including a novel metric, to train algorithms in Python 3.7.] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|Cell] [![DataSize](https://img.shields.io/badge/DataSize-990066)|36,543] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|CatBoost, Light Gradient Boosting Method (LGBM), and eXtreme Gradient Boosting (XGBoost)] [![Metrics](https://img.shields.io/badge/Metrics-660066)|Claiming average detection and classification accuracies of 99.996% and 99.745%, respectively.] [Performance assessment of selective machine learning techniques for improved PV array fault diagnosis](https://www.sciencedirect.com/science/article/pii/S2352467721001454) **[Report|]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|International Journal of Applied Earth Observation and Geoinformation，Q2] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|IR] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|10] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|CNN+Classifier(SVM)] [![Metrics](https://img.shields.io/badge/Metrics-660066)|Claimed to achieve 99.49% on a self-built dataset.] [Developing a deep learning-based layer-3 solution for thermal infrared large-scale photovoltaic module inspection from orthorectified big UAV imagery data](https://www.sciencedirect.com/science/article/pii/S0303243421003597) **[Report|LightGBM is better than CatBoost, XGBoost, and Random Forest.]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|MDPI Electronics，Q3] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|array] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|Neighboring Pixel-Based Virtual Imaging (NPBVI) Technology] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Neighboring-Pixel-Based Maximum Power Point Tracking Algorithm for Partially Shaded Photovoltaic (PV) Systems](https://www.mdpi.com/2079-9292/11/3/359) **[Report|Propose a Maximum Power Point Tracking (MPPT) method. Uncertainty levels increase during high-resolution image testing, possibly due to small image details considered when capturing images with a high-resolution lens.]**

- [![文献类型](https://img.shields.io/badge/Conference-8A2BE2)|International Conference on Image Processing and Capsule Networks] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|EL] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|Image Denoising] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|"Optical or mechanical defects, such as microcracks, the size of cracks, and side effects caused by electrical or electromechanical interference during the image acquisition process.] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|Detecting defects through semantic segmentation.] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Defect Analysis of Electroluminescence Images of PV CELL](https://link.springer.com/chapter/10.1007/978-3-030-84760-9_57) **[Report|Although the quantity of microcracks is not directly related to the efficiency of PV cells, the size, width, and location of these cracks, along with the operational lifespan of the solar panel components, significantly impact the measured output power.]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|Remote Sensing，Q2] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|IR] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [RID—Roof Information Dataset for Computer Vision-Based Photovoltaic Potential Assessment](https://www.mdpi.com/2072-4292/14/10/2299) **[Report|The introduction of two new multi-class datasets is aimed at roof segmentation and semantic segmentation of roof structures, with a focus on extracting roof information from aerial images, particularly for applications in photovoltaic potential analysis. While deep learning-based photovoltaic potential analysis using 2D aerial images incorporates structural information, it lacks slope and shadow details. Therefore, it is necessary to investigate the learning of these details from aerial images or to merge 2D and 3D data to further enhance potential estimation accuracy. In addition to photovoltaic potential analysis, rich 3D models can also drive other research and applications, such as urban noise diffusion modeling, roof insulation assessment, or building energy demand estimation.]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|Energies，Q4] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|IR] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|Aerial Infrared Thermography (aIRT)] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Automatic Inspection of Photovoltaic Power Plants Using Aerial Infrared Thermography: A Review](https://www.mdpi.com/1996-1073/15/6/2055) **[Report|"Overview of the application of the Aviation Infrared Thermography (aIRT) framework in automating various tasks within photovoltaic power plants: fault detection and classification, optimization of flight path planning, simplification of on-site fault localization through PV station orthophoto projection, detection of module contamination, and differentiation from actual fault areas on PV modules.]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|Procedia Computer Science] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|IR] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|Panel] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|Classification + Object Detection] [![Metrics](https://img.shields.io/badge/Metrics-660066)|The claimed ResNet-50 transfer learning model achieved the highest F1 score of 85.37%, while Faster R-CNN attained a 67% Mean Average Precision (MAP).] [Solar panel hotspot localization and fault classification using deep learning approach](https://www.sciencedirect.com/science/article/pii/S1877050922008225) **[Report|Implementing Classification and Object Detection Using ResNet-50 and Faster R-CNN, Respectively]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|International Journal of Photoenergy，Q4] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|RGB+IR] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|hotspot] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Infrared Thermal Images of Solar PV Panels for Fault Identification Using Image Processing Technique](https://www.hindawi.com/journals/ijp/2022/6427076/) **[Report|The majority of faults in solar panels are documented as hotspots, primarily due to an increase in internal resistance.]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|Progress in Energy] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|IR] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Outdoor luminescence imaging of field-deployed PV modules](https://raw.githubusercontent.com/574168985/awesome_PV_faults_diagnosis/main/[https:/iopscience.iop.org/journal/2516-1083](https:/iopscience.iop.org/article/10.1088/2516-1083/ac9a33/meta) **[Report|Outdoor Electroluminescence]**

## 【2021】IMAGE BASED METHODS

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|Energy Conversion and Management，Q1] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|IR] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|Including images obtained through unmanned aerial vehicles (drones) and ground operators.] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|The various preprocessing strategies include pixel standardization and normalization, grayscale conversion, thresholding, discrete wavelet transformation, as well as Sobel Feldman and box blur filtering.] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|10] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|Classification tasks: 1000, Segmentation tasks: 200] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|Claimed to achieve 100 on a self-built dataset.] [Deep learning strategies for automatic fault diagnosis in photovoltaic systems by thermographic images](https://www.sciencedirect.com/science/article/pii/S019689042100491X) **[Report|Introducing data augmentation techniques and Dropout layers to reduce overfitting, but based on the results, these techniques did not yield substantial improvement, and in some cases, may even lead to a decrease in accuracy by a few percentage points.]**

- [![文献类型](https://img.shields.io/badge/Conference-8A2BE2)|Fault Detection in aerial images of photovoltaic modules based on Deep learning] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|UAV] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|Scars, layering, discoloration, glass shattering, intact panels, and snail traces.] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|module] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|Convolutional Neural Network (CNN)] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Fault Detection in aerial images of photovoltaic modules based on Deep learning](https://iopscience.iop.org/article/10.1088/1757-899X/1012/1/012030/meta) **[Report|Apply CNN]**

- [![文献类型](https://img.shields.io/badge/Conference-8A2BE2)|International Conference on Environment and Electrical Engineering (EEEIC)] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|IR] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|UAV] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|CNN] [![Metrics](https://img.shields.io/badge/Metrics-660066)|Claimed to achieve 100 on a self-built dataset.] [Thermal anomalies detection in a photovoltaic plant using artificial intelligence: Italy case studies](https://ieeexplore.ieee.org/abstract/document/9584494) **[Report|The images are typically inspected individually by qualified operators. The application of artificial intelligence technology enables even non-specialized users to detect overheating, thereby reducing the time required for system checks.]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|Renewable Energy，Q1] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|IR] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|In order to address the imbalance in the dataset, the study employed both undersampling and oversampling methods. By utilizing data augmentation in the oversampling approach, the model's generalization performance improved by approximately 6% compared to undersampling.] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|In the case of certain photovoltaic (PV) faults, such as cracking, contamination, vegetation, and shading, there is high variability both within and between categories. This makes accurate classification of these faults challenging in unbalanced datasets with a small number of samples.] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Automatic fault classification in photovoltaic modules using Convolutional Neural Networks](https://www.sciencedirect.com/science/article/pii/S0960148121010752) **[Report|The challenges and efficiency of combining infrared thermography and machine learning methods for the classification of different defect categories in PV modules were assessed. For low-variability fault categories, reducing the fault categories to eight with lower internal and inter-category variability significantly impacted the performance of the automated classifier, achieving an accuracy of 78.85%. Regarding the ease of classification for specific fault categories, the study identified certain fault categories, such as diode combination and diode faults, which could result in up to a 66% reduction in PV system power with very specific patterns. Therefore, these types of faults can be easily classified with an accuracy exceeding 90% using the proposed model.]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|Processes，Q3] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|Considering the intermittent nature of photovoltaic power generation, we constructed a corresponding discontinuous regression model based on the XGBoost algorithm, and optimized it using the results of k-means clustering.] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Intelligent and Data-Driven Fault Detection of Photovoltaic Plants](https://www.mdpi.com/2227-9717/9/10/1711) **[Report|An unsupervised method that does not rely on labeled faulty data.]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|IEEE Journal of Photovoltaics，Q3] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|IR] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|overheated area] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|"Firstly, multiple large-scale images are generated from thermal imaging maps containing overheated regions to accurately detect targets in the overheated areas. Subsequently, regions of interest are extracted from these images to delineate areas where overheating may be present. Finally, a deep collaborative learning model is employed to identify the types and locations of overheated areas within these regions.] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Automated Overheated Region Object Detection of Photovoltaic Module With Thermography Image](https://ieeexplore.ieee.org/abstract/document/9314927) **[Report|Identifying specific types and accurate locations of different potential overheating areas targets in the thermal imaging of individual photovoltaic modules.]**

- [![文献类型](https://img.shields.io/badge/Conference-8A2BE2)|Proceedings of 2021 International Conference on Autonomous Unmanned Systems (ICAUS 2021)] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|Extend the dataset through Mosaic data augmentation method.] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|Selecting ResNet50 as the feature extraction network is aimed at improving efficiency and reducing computational overhead. By analyzing the convolutional features, an Attention algorithm is introduced to expand the model's receptive field. Additionally, the FPT structure is employed to enhance the fusion of feature information at the same position across different layers and at different positions within the same layer. A bipartite graph matching algorithm, based on minimizing a loss function, is designed to emphasize the importance of ground truth box distribution. Furthermore, the loss function is modified to alleviate the suppression of non-extreme predicted bounding boxes.] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [An Optimized One-Stage Detector for Autonomous PV Defect Detection](https://link.springer.com/chapter/10.1007/978-981-16-9492-9_168) **[Report|Optimizing the feature fusion module and loss function of the network aims to improve the poor performance of traditional single-stage object detection models on photovoltaic defects in infrared images.]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|SENSORS，Q3] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|IR] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|Classify PV panel defects into five categories, including bird droppings, single defects, splicing, horizontally arranged strings, and blocks.] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|panel] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|isolated convolution neural model (ICNM)] [![Metrics](https://img.shields.io/badge/Metrics-660066)|Claims to achieve 97.62%] [Photovoltaic Panels Classification Using Isolated and Transfer Learned Deep Neural Models Using Infrared Thermographic Images](https://www.mdpi.com/1424-8220/21/16/5668) **[Report|Applied Isolated Convolutional Neural Network (ICNN)]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|Energy，Q1] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|EL] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|The combination of data augmentation, appropriate model size, L2 weight regularization, batch normalization, and dropout, along with data augmentation techniques such as rotation, flipping, cropping, contrast adjustment, and blur processing, contributes to an improvement in model accuracy of around 6.5%.] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|Cell] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|Utilizing Lightweight Convolutional Neural Networks (CNN)] [![Metrics](https://img.shields.io/badge/Metrics-660066)|Claims 93.02%] [CNN based automatic detection of photovoltaic cell defects in electroluminescence images](https://www-sciencedirect-com.ezproxy.universite-paris-saclay.fr/science/article/pii/S0360544219320146) **[Report|State-of-the-art (SOTA) with Lightweight Convolutional Neural Networks (CNNs) on the public dataset [elpv-dataset]]**

- [![文献类型](https://img.shields.io/badge/Conference-8A2BE2)|Chinese Automation Congress (CAC)] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|IR] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|UAV] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|pannel] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|Segmentation first, followed by object detection] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Deep Learning-based Method for PV Panels Segmentation and Defects Detection with Infrared Images](https://ieeexplore.ieee.org/abstract/document/9728350) **[Report|Establishing a workflow for solar photovoltaic array segmentation and module-level defect detection based on infrared imagery captured by unmanned aerial vehicles (UAVs). In this framework, segmentation precedes identification (object detection, not classification). The U2-Net is employed for precise segmentation of photovoltaic panels, while YOLOv4 is trained to recognize fault information in the photovoltaic modules]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|Energies，Q4] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [A Review on Artificial Intelligence Applications for Grid-Connected Solar Photovoltaic Systems](https://www.mdpi.com/1996-1073/14/15/4690) **[Report|A review has been conducted on the observation results of utilizing different artificial intelligence technologies at the system level in various purposes along the solar photovoltaic value chain. In optimized scenarios, evolutionary algorithms such as Particle Swarm Optimization (PSO) and Genetic Algorithms (GA) are widely employed. In the case of time series data, such as irradiance or power prediction, artificial neural networks have been successfully applied. Additionally, inference models, data-driven algorithms, and learning methods widely used in control and maintenance are considered more specifically applicable to certain conditions, problems, or datasets, lacking universality in the same aspects. However, despite the significant successes reported in most papers, it must be noted that this success is based on tuning specific parameters of particular models while maintaining default parameters of other models. Furthermore, some models proposed based on input data may not be universally applicable. Currently, there seems to be an issue of data inconsistency in the industry, but with the emergence of Internet of Things solutions, the deployment of a large number of sensors, the provision of video streams by drones in maintenance, and the advent of natural language processing technology, the problem of data scarcity may disappear.]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|Applied Sciences-Basel，Q4] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|RGB] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|Cell] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Evaluation of Artificial Intelligence-Based Models for Classifying Defective Photovoltaic Cells](https://www.mdpi.com/2076-3417/11/9/4226) **[Report|A solar photovoltaic cell defect classifier using five different classification models (KNN, SVM, RF, MLP, and CNN).]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|Applied Sciences-Basel，Q4] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|IR+RGB] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|UAV] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|Grayscale conversion, filtering, three-dimensional temperature representation, probability density function, and cumulative distribution function analysis.] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|Cell] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|A four-step image analysis process was proposed. In the first step, the raw infrared image undergoes filtering and is transformed into a grayscale image. Then, in the second step, the image is converted into a 3D color representation to amplify temperature differences. The third step involves establishing a histogram of the thermal temperature distribution to identify the percentage of high-temperature regions. The final step constructs an accumulated chart of the temperature distribution to determine the potential number of batteries with anomalies or defects.] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Using UAV to Detect Solar Module Fault Conditions of a Solar Power Farm with IR and Visual Image Analysis](https://www.mdpi.com/2076-3417/11/4/1835) **[Report|The recognition of images can enhance the clarity of infrared images.]**

- [![文献类型](https://img.shields.io/badge/Conference-8A2BE2)|IOP Conference Series: Earth and Environmental Science] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|IR] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|Utilize masking to exclude irrelevant background regions in the original image, and annotate the image data using labelme software.] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|hotspot] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|Cell] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|SegNet] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Efficient Region Segmentation of PV Module in Infrared Imagery using Segnet](https://iopscience.iop.org/article/10.1088/1755-1315/793/1/012018/meta) **[Report|To enhance the accuracy of hotspot detection, it is crucial to perform effective region segmentation on the infrared images of photovoltaic modules.]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|Journal of Mechanical Engineering Research and Developments] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Defect detection by automatic control in the photovoltaic panel manufacturing process](https://jmerd.net/Paper/Vol.44,No.5(2021)/258-262.pdf) **[Report|Based on Matlab]**

- [![文献类型](https://img.shields.io/badge/毕设-8A2BE2)| Campus Tecnológico Local San Carlos] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|IR+RGB] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|UAV] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|Partial shadow, pollution, and electrical faults] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Comparison of the effectiveness among different photovoltaic fault detection techniques](https://repositoriotec.tec.ac.cr/handle/2238/14050) **[Report|The effectiveness of different fault detection methods: a quantitative comparison of three fault detection technologies in photovoltaic systems, namely a) Visual Imaging (VI), b) Infrared Thermography (IRT), and c) Electrical Analysis (EA). The results show that there is no significant difference in the total number of faults detected by the three technologies. However, VI performs best in detecting contamination while showing the poorest performance in detecting electrical faults. Additionally, partial shading is the most easily detected fault among all technologies. Therefore, for comprehensive fault detection, VI should be combined with either IRT or EA. Fusion-based methods can be considered for maximizing fault detection capabilities.]**

- [![文献类型](https://img.shields.io/badge/Conference-8A2BE2)|Annual Conference of Industrial Electronics Society] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|local shadows and low irradiance] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|The effectiveness of this method in detecting faults under various atmospheric and local shadow conditions has been verified through simulation using MATLAB/Simulink] [Power based Fault Detection Method for PV Arrays](https://ieeexplore.ieee.org/abstract/document/9589163) **[Report|Faults in Solar Photovoltaic Systems under Uniform and Partial Shade Conditions]**

- [![文献类型](https://img.shields.io/badge/Conference-8A2BE2)|Fifteenth International Conference on Quality Control by Artificial Vision] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|IR] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|Using methods based on different features, the prediction results from a subtraction network are employed with a voting strategy to identify abnormal battery panels in the input image.] [![Metrics](https://img.shields.io/badge/Metrics-660066)|They constructed two datasets to assess their approach: one composed of manually extracted images of solar panels, forming a clear dataset, and another consisting of automatically extracted data with added noise. The method achieved over 90% classification accuracy on both datasets.] [Identify solar panel defects by using differences between solar panels](https://www.spiedigitallibrary.org/conference-proceedings-of-spie/11794/1179415/Identify-solar-panel-defects-by-using-differences-between-solar-panels/10.1117/12.2586911.short#_=_) **[Report|The identification of anomalies becomes challenging due to the potential similarity in temperature characteristics between normal and defective solar panels in a single panel image]**

- [![文献类型](https://img.shields.io/badge/arxiv-8A2BE2)|arxiv] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|3200 training samples and 800 testing samples (200 for each category)] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Graph neural network-based fault diagnosis: a review](https://arxiv.org/abs/2111.08185) **[Report|Reviewed fault diagnosis (FD) based on Graph Neural Networks (GNN), highlighting the effective representation of data from various application domains through a graph structure. In comparison to traditional fault diagnosis methods, this representation achieves superior performance.]**

## 【2023】FUSION BASED METHODS

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|Energies，Q4] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|Process signals using SDP to generate images with specific snowflake-like features] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|Defective welding, cracks, and failure of bypass diodes] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|Cell] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|Combining Symmetric Dot Patterns (SDP) with AlexNet] [![Metrics](https://img.shields.io/badge/Metrics-660066)|Claimed 99.8%] [Fault Diagnosis for PV Modules Based on AlexNet and Symmetrized Dot Pattern](https://www.mdpi.com/1996-1073/16/22/7563) **[Report|The detection performance of five algorithms was compared, namely three machine learning algorithms: SDP+HOG+SVM, SDP+HOG+ENN, and SDP+HOG+BPNN, as well as two deep learning algorithms: SDP+CNN and SDP+AlexNet. This paper transforms subtle changes in signals into more pronounced variations using snowflake diagrams. These diagrams are then fed into AlexNet, a fast-training and testing deep learning model, for fault diagnosis.]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|Applied Sciences-Basel，Q4] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|IR+RGB] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Infrared and Visible Image Fusion: Methods, Datasets, Applications, and Prospects](https://www.mdpi.com/2076-3417/13/19/10891) **[Report|Overview of Infrared and Visible Light Image Fusion, Including a Classic Fusion Article from 2003 in the PV Direction[Fault diagnosis for photovoltaic array with the technique of infrared/visible image fusion](https://www.spiedigitallibrary.org/conference-proceedings-of-spie/5286/0000/Fault-diagnosis-for-photovoltaic-array-with-the-technique-of-infrared/10.1117/12.539825.short)]**

## 【2022】FUSION BASED METHODS

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|MODERN ELECTRIC POWER] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|DC Series Arc Fault] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|system] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [A Multi-feature Fusion-Based Method to Recognize Series DC Arc Fault in Photovoltaic System](http://xddl.ncepujournal.com/en/article/Y2022/I5/529) **[Report|A feature space for fault arcs was constructed by extracting features in the time domain, frequency domain, and energy entropy]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|Energies，Q4] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|Using SDP to process signals and generate images with specific snowflake-like features.] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|Poor Welding, Breakage, and Bypass Diode Failure] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|module] [![DataSize](https://img.shields.io/badge/DataSize-990066)|3200] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|Integrating Symmetry-Differentiated Patterns (SDP) with Convolutional Neural Networks (CNN).] [![Metrics](https://img.shields.io/badge/Metrics-660066)|The claim of achieving 99.88% on a self-built dataset.] [A Fault Detection Method Based on CNN and Symmetrized Dot Pattern for PV Modules](https://www.mdpi.com/1996-1073/15/17/6449) **[Report|]**

## 【2021&Before】FUSION BASED METHODS

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|Journal of Materials Science Research] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|IR+RGB] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|Scratches and Bypasses in Polycrystalline Photovoltaic Solar Cells] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Fusion of Thermal and Visible Acquisitions for Evaluating Production-borne Scratches and Shunts in Photo-Voltaic PV Cells](https://www.ccsenet.org/journal/index.php/jmsr/article/view/20250) **[Report|Fusion of Infrared and Visible Light Images]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|IEEE Transactions on Industrial Informatics，Q1] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|IR+EL] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|Scratches, broken cables, surface impurities, hidden cracks.] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|Cell] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Electromagnetic Induction Heating and Image Fusion of Silicon Photovoltaic Cell Electrothermography and Electroluminescence](https://ieeexplore.ieee.org/abstract/document/8736340) **[Report|A fusion rule based on the L1 norm was proposed for merging sparse vectors of ET (Electrothermal) and EL (Electroluminescent) images, enabling the integration and complementarity of data from two different wavelengths in detection.]**

- [![文献类型](https://img.shields.io/badge/Conference-8A2BE2)|IEEE PES Asia-Pacific Power and Energy Engineering Conference (APPEEC)] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|The claim of achieving 99.49% on a self-built dataset.] [Research on Fault Detection of PV Array Based on Data Fusion and Fuzzy Mathematics](https://ieeexplore.ieee.org/abstract/document/5749018) **[Report|Multisensor Data Fusion]**

## 【2023】Review of Fault Detection Methods for Photovoltaic Modules

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|Energies，Q4] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Artificial Intelligence in Photovoltaic Fault Identification and Diagnosis: A Systematic Review](https://www.mdpi.com/1996-1073/16/21/7417) **[Report|The latest advancements in Artificial Intelligence (AI) technology for addressing issues in fault detection within Photovoltaic (PV) systems are discussed in this introduction. Combining traditional methods with AI approaches appears to yield better results. Additionally, integrated learning and stacked classifiers prove particularly effective in diagnosing faults. However, there are still some challenges that need to be overcome. A shortage of publicly available data for testing these methods is noted, highlighting the need for multimodal complementarity in data scarcity. In real-world environments, the application of certain complex algorithms may also pose difficulties. The research identifies that certain fault types receive more attention than others in the study.]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|International Journal of Low-Carbon Technologies，Q4] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Detection, location, and diagnosis of different faults in large solar PV system—a review](https://academic.oup.com/ijlct/article/doi/10.1093/ijlct/ctad018/7083000login=false) **[Report|The faults that may occur in solar photovoltaic systems are categorized into physical, environmental, and electrical faults, and further classified.]**

- [![文献类型](https://img.shields.io/badge/Conference-8A2BE2)|International Conference on Green Energy, Computing and Intelligent Technology (GEn-CITy 2023)] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Comparative review on recent electrical and non-electrical methods utilized for fault detection and diagnosis in PV systems](https://digital-library.theiet.org/content/conferences/10.1049/icp.2023.1780) **[Report|This article compares recent research findings on fault detection and diagnostic techniques for photovoltaic (PV) systems based on both electrical and non-electrical methods. It provides a detailed discussion on performance metrics such as detection accuracy, computation time, and the types of faults considered.]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|Energies，Q4] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|EL] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [A Review on Defect Detection of Electroluminescence-Based Photovoltaic Cell Surface Images Using Computer Vision](https://www.mdpi.com/1996-1073/16/10/4012) **[Report|Implementation of Photovoltaic Cell Automated Defect Detection Based on Electroluminescence (EL) Using Convolutional Neural Network (CNN) Architecture]**

## 【2022】Review of Fault Detection Methods for Photovoltaic Modules

- [![文献类型](https://img.shields.io/badge/Conference-8A2BE2)|IEEE Conference and Expo Transportation Electrification Asia-Pacific (ITEC Asia-Pacific)] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Review on Fault Characterization and Diagnosis Technique in Photovoltaic Systems](https://ieeexplore.ieee.org/abstract/document/9941956) **[Report|Different Fault Detection Methods: Visual Diagnostic Method: Can only diagnose faults that have already resulted in serious consequences. Imaging Solutions: Utilizes various types of camera sensors to identify hotspots or aging in photovoltaic modules. Electrical Detection Methods: Discusses offline and online techniques, with I-V curve measurement being a commonly used offline technique. Among online methods, Maximum Power Point Tracking (MPPT) is the simplest and most accurate diagnostic approach: Overall, MPPT-based diagnostic methods are the simplest and most accurate among all online methods and can be embedded in commercial inverters.]**

## 【2021】Review of Fault Detection Methods for Photovoltaic Modules

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|RENEWABLE & SUSTAINABLE ENERGY REVIEWS，Q1] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Potential measurement techniques for photovoltaic module failure diagnosis: A review](https://www.sciencedirect.com/science/article/pii/S1364032121008108) **[Report|The main sources of degradation in PV modules include discoloration, delamination, corrosion, and cracking. These can be directly quantified through infrared (IR) thermal imaging and I-V curve measurements. Additionally, environmental parameters such as temperature, humidity, and ultraviolet radiation should also be considered.]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|RENEWABLE & SUSTAINABLE ENERGY REVIEWS，Q1] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Application of Artificial Neural Networks to photovoltaic fault detection and diagnosis: A review](https://www.sciencedirect.com/science/article/pii/S136403212030798X) **[Report|The application of ANN and hybrid ANN models in photovoltaic fault detection and diagnosis (FDD) is comprehensively reviewed.]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|Machines，Q3] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Review of Artificial Intelligence-Based Failure Detection and Diagnosis Methods for Solar Photovoltaic Systems](https://www.mdpi.com/2075-1702/9/12/328) **[Report|The main characteristics of artificial intelligence-based methods and the effectiveness in the application of photovoltaic fault detection and diagnosis.]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|IEEE Access，Q3] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Deep Learning-Based Fault Diagnosis of Photovoltaic Systems: A Comprehensive Review and Enhancement Prospects](https://ieeexplore.ieee.org/abstract/document/9530686) **[Report|Deep learning-based Fault Detection and Diagnosis (FDD) can be categorized into five types: FDD based on Convolutional Neural Networks (CNN), FDD based on Recurrent Neural Networks (RNN), FDD based on Stacked Autoencoder Networks (SAEN), FDD based on Deep Belief Networks (DBN), and FDD based on Deep Transfer Learning (DTL). The main advantages and disadvantages of each approach are highlighted.]**

- [![文献类型](https://img.shields.io/badge/Conference-8A2BE2)|International Conference on Applied System Innovation (ICASI)] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Fault identification and diagnosis methods for photovoltaic system: A review](https://ieeexplore.ieee.org/abstract/document/9568414) **[Report|The accuracy of measurements is correlated with the effectiveness of monitoring/fault detection systems; if photovoltaic inverters provide inaccurate measurement data, additional measurement devices such as sensors may be required.]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|Engineering Review] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [A comprehensive review of diagnosis technics for faults occurring in the photovoltaic generator](https://er.riteh.hr/index.php/ER/article/view/1714) **[Report|Summarized many forms of degradation, types of faults, and diagnostic techniques related to photovoltaic generators.]**

- [![文献类型](https://img.shields.io/badge/Journal-8A2BE2)|International Journal of Green Energy，Q4] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Fault diagnosis of visual faults in photovoltaic modules: A Review](https://www.tandfonline.com/doi/full/10.1080/15435075.2020.1825443) **[Report|Visible faults, electrical faults, and other faults.]**

## Photovoltaic Module Fault Monitoring Dataset

- [![文献类型](https://img.shields.io/badge/Dataset-8A2BE2)] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|IR] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|11 Types of Faults] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|module] [![DataSize](https://img.shields.io/badge/DataSize-990066)|10000+10000] [InfraredSolarModules|Infrared solar module dataset for anomaly detection] [![pdf](https://img.shields.io/badge/pdf-1A2BE2)](https://ai4earthscience.github.io/iclr-2020-workshop/papers/ai4earth22.pdf) [![github](https://img.shields.io/badge/github-5A2BE2)](https://github.com/RaptorMaps/InfraredSolarModules) [![kaggle](https://img.shields.io/badge/kaggle-9A2BE2)](https://www.kaggle.com/code/marcosgabriel/dataset-intro-infrared-solar-modules) **[Report|The dataset comprises authentic infrared images depicting various anomalies in solar fields.]**

- [![文献类型](https://img.shields.io/badge/Dataset-8A2BE2)|IEEE Transactions on Industrial Informatics，Q1] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|EL] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|Manufacturing] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|10] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|Cell] [![DataSize](https://img.shields.io/badge/DataSize-990066)|36,543] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|claimed 99.49] [PHOTOVOLTAIC CELL ANOMALY DETECTION DATASET] [![pdf](https://img.shields.io/badge/pdf-1A2BE2)](https://ieeexplore.ieee.org/abstract/document/9744494/algorithmstabFilter=dataset#algorithms) [![IEEE DataPort](https://img.shields.io/badge/IEEE_DataPort-992BE2)](https://ieee-dataport.org/documents/photovoltaic-cell-anomaly-detection-dataset) **[Report|"Provide public datasets. The first one is a public dataset that provides ground truth for the framework of abnormal detection in photovoltaic solar cells (object detection task).]** HIGHLIGHT

- [![文献类型](https://img.shields.io/badge/Dataset-8A2BE2)|mdpi data] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|RGB] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|Dynamic reconfiguration and maximum power point tracking.] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|Cell] [![DataSize](https://img.shields.io/badge/DataSize-990066)|5211] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|Utiliser une méthode basée sur l'imagerie pour estimer les caractéristiques électriques des panneaux solaires.] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [Dataset for Detecting the Electrical Behavior of Photovoltaic Panels from RGB Images] [![pdf](https://img.shields.io/badge/pdf-1A2BE2)](https://www.mdpi.com/2306-5729/7/6/82) [![Zenodo](https://img.shields.io/badge/Zenodo-999BE2)](https://ieee-dataport.org/documents/photovoltaic-cell-anomaly-detection-dataset)) **[Report|The dataset introduces an association between RGB images and photovoltaic panel electrical data under varying irradiance and shadow conditions. Additionally, the dataset provides supplementary weather data and other image features to support the training of estimation models. Specifically, this dataset is designed to facilitate the development of image-based models for estimating electrical data, serving as a substitute for extensive sensor arrays.]**

- [![文献类型](https://img.shields.io/badge/Dataset-8A2BE2)|Remote Sensing，Q2] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|IR] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|aerial images] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|] [![DataSize](https://img.shields.io/badge/DataSize-990066)|] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [RID—Roof Information Dataset for Computer Vision-Based Photovoltaic Potential Assessment] [![pdf](https://img.shields.io/badge/pdf-1A2BE2)](https://www.mdpi.com/2072-4292/14/10/2299) [![github](https://img.shields.io/badge/github-5A2BE2)](https://github.com/TUMFTM/RID) **[Report|Two new multiclass datasets have been introduced for roof segmentation and semantic segmentation of roof structures, with a focus on extracting roof information from aerial images, particularly in the context of photovoltaic potential analysis. While deep learning-based photovoltaic potential analysis using 2D aerial images incorporates structural information, it lacks slope and shadow information. Therefore, it is necessary to investigate learning these details from aerial images or combining 2D and 3D data to further improve potential estimation accuracy. In addition to photovoltaic potential analysis, rich 3D models can also drive other research and applications, such as urban noise diffusion modeling, roof insulation assessment, or estimation of building energy demands.]**

- [![文献类型](https://img.shields.io/badge/Dataset-8A2BE2)|35th EU PVSEC 2018] [![ImageType](https://img.shields.io/badge/ImageType-2b83e2)|EL] [![CollectionMethod](https://img.shields.io/badge/CollectionMethod-e28a2b)|lab] [![ImageProcessing](https://img.shields.io/badge/ImageProcessing-83e22b)|] [![FaultType](https://img.shields.io/badge/FaultType-ff0000)|Cracks, fracture interconnection, PID, and battery quality.] [![ApplicationLevel](https://img.shields.io/badge/ApplicationLevel-009999)|cell] [![DataSize](https://img.shields.io/badge/DataSize-990066)|2,426] [![Methodology](https://img.shields.io/badge/Methodology-6666cc)|] [![Metrics](https://img.shields.io/badge/Metrics-660066)|] [elpv-dataset] [![pdf](https://img.shields.io/badge/pdf-1A2BE2)](https://userarea.eupvsec.org/proceedings/35th-EU-PVSEC-2018/5CV.3.15/) [![github](https://img.shields.io/badge/github-5A2BE2)](https://github.com/zae-bayern/elpv-dataset) **[Report|The dataset, consisting of 2,426 solar cells extracted from high-resolution electroluminescence (EL) images, is employed for automated defect probability recognition. The collected images encompass both monocrystalline and polycrystalline solar modules, featuring solar cells with varying degrees of normal functionality and defects. Expert annotations categorize the solar cells in each image based on the likelihood of defects. These labeled images serve as valuable resources for developing computer vision and machine learning approaches to automatically detect various defects, such as cracks, interconnection fractures, Potential-Induced Degradation (PID), and overall cell quality. The dataset can also be utilized for predicting power efficiency losses.]**

awesome_PV_faults_diagnosis

awesome_PV_faults_diagnosis 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

awesome_PV_faults_diagnosis

awesome_PV_faults_diagnosis 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

Download README