The present study focuses on automatic defects classification of PV cells in electroluminescence images. Two machine learning approaches, features extraction-based support vector machine (SVM) and convolutional neural network (CNN), are used for the solar cell defect classifications.
PV cell defects are classified by training a model with EL images using a radial-based kernel SVM. First, features are extracted from EL images of the cell using feature extraction techniques. Then, these features are fed to the SVM classifier.
(Tang et al., 2020) also based their research on ELPV dataset and classified PV cell EL images according to their defect types by creating synthetic data using Generative Adversarial Networks (GAN). CNN were trained from scratch and the classification accuracy between 81% and 84% is achieved.
A hybrid deep CNN architecture is proposed to achieve high classification performance in PV solar cell defects. The proposed method is based on the integration of residual connections into the inception network. Therefore, the advantages of both structures are combined and multi-scale and distinctive features can be extracted in the training.
Photovoltaic (PV) defects can be classified using various techniques such as infrared (IR) imaging, electroluminescence (EL), large-area laser beam induced current, and current–voltage characteristics [6, 7]. Recent advancements in EL imaging have made it possible to extract defect information hidden within the PV cell.
To classify the seven types of defects in a polycrystalline silicon PV cell, the proposed machine learning approaches are applied to the public dataset of solar cell EL images. The successful classification of these defects is a challenging task due to the background texture of the cells.
Using this dataset, (Deitsch et al., 2019) performed PV cell classification on the original dataset with 4-class (i.e. Non-defected, Possibly normal, Possibly defected and …
The present study is carried out for automatic defects classification of PV cells in electroluminescence images. Two machine learning approaches, features extraction-based support vector machine (SVM) and …
A hybrid deep CNN architecture is proposed to achieve high classification performance in PV solar cell defects. The proposed method is based on the integration of …
In this paper, we applied several deep learning networks such as AlexNet, SENet, ResNet18, ResNet34, ResNet50, ResNet101, ResNet152, GoogleNet (Inception V1), …
Photovoltaic (PV) system performance and reliability can be improved through the detection of defects in PV modules and the evaluation of their effects on system operation. …
In this work, we propose two classification pipelines that automatically solve the second task, i.e., to determine a per-cell defect likelihood that may lead to efficiency loss. The …
First, we proposed a CNN model that performs binary classification between good and defective solar cells. After that we proposed a multiclass classification model using …
The common defects observed on the photovoltaic cells during the manufacturing process include chipping, tree crack, micro-line, soldering, and short circuits.
A dataset has been created for detecting anomalies in photovoltaic cells on a large scale in [], this dataset consists of 10 categories, several detection models were …
The defect classification in PV cells has a key role in controlling the quality and output power of PV cells. The fast and accurate determination of the defect locations in PV module and cell is …
Data augmentation is the most common method to deal with image data scarcity by means of the introduction of slight ... is proposed for the efficient classification of PV cell …
Qualitative defect classification results in a PV module previously not seen by the deep regression network. The red shaded circles in the top right corner of each solar cell …
The present study is carried out for automatic defects classification of PV cells in electroluminescence images. Two machine learning approaches, features extraction-based …
Finally, the images of individual cells are inputted into a deep neural network classifier. Our leading model achieves an F1 score of 0.93 while processing an average of 240 images per …
A change in the operating conditions of the PV array indicates implicitly that a fault has occurred. This fault can be divided into three categories []: physical faults can be a …
Generally, solar cell defects can be divided into two broad defect categories: intrinsic and extrinsic defects. Figure 1 shows an example of a cell extracted from an EL image …
In this work, we propose two classification pipelines that automatically solve the second task, i.e., to determine a per-cell defect likelihood that may lead to efficiency loss. The …
This work introduces neural architecture search to the field of PV cell defect classification for the first time and proposes a novel lightweight high-performance model for automatic defect …
This work introduces neural architecture search to the field of PV cell defect classification for the first time and proposes a novel lightweight high-performance model for automatic defect …
In this paper, we propose a deep-learning-based defect detection method for photovoltaic cells, which addresses two technical challenges: (1) to propose a method for data …
Photovoltaic (PV) system performance and reliability can be improved through the detection of defects in PV modules and the evaluation of their effects on system operation. …
The second classification for solar PV defects is based on time characteristics: intermittent faults caused by external factors like shading and dust are temporary but reduce …