Publication: A Multi-Layer Deep Learning System for Fault Detection Solar Cell Electroluminescence Images
Date
2024
Authors
Almashhadani R.
Hock G.C.
Hani Nordin F.
Abdulrazzak H.N.
Ali Abbas Z.
Journal Title
Journal ISSN
Volume Title
Publisher
IEEE Computer Society
Abstract
An automatic solar defect detection with a classification system was proposed using deep learning. This paper focuses on solar defects in photovoltaic systems identified through electroluminescence (EL) images. Convolutional Neural Networks (CNNS), Artificial Neural Networks (ANN), Support Vector Machine (SVM), and a pre-trained VGG16 network for feature extraction were compared in this paper. They adopted two classification strategies: binary classification (defective or non-defective) and multi-class classification the class names are 0%, 33%, 67%, and 100% (here % represents the percentage of defectiveness), which represents the defect likelihood. The effectiveness of the model was evaluated using various metrics, including the recall, precision, accuracy, and F1-score for two and four classes and obtained on, supported by confusion matrices. VGG-16 model outperformed other models and achieved an accuracy of 89% for 2-classes and 93% for 4-classes. ? 2024 IEEE.
Description
Keywords
Convolutional neural networks , Deep neural networks , Multilayer neural networks , Classification system , Convolutional neural network , Defect detection , Electroluminescence images , Faults detection , Multi-layers , Neural-networks , Photovoltaic systems , Photovoltaics , VGG16 , Support vector machines