PV Cell Defects Classification in Electroluminescence Images using Gradient Histogram (HOG)

• Published in: IEEE Symposium on Wireless Technology and Applications, ISWTA
• Main Author: Yani A.N.A.; Shariff K.K.M.; Othman Z.; Sulaiman S.I.; Zakaria N.A.Z.; Yassin A.I.
• Format: Conference paper
• Language: English
• Published: IEEE Computer Society 2023
• Online Access: https://www.scopus.com/inward/record.uri?eid=2-s2.0-85174315103&doi=10.1109%2fISWTA58588.2023.10249747&partnerID=40&md5=3157db69cc85b3b41b868599d571e849
• ISSN: 23247843
• DOI: 10.1109/ISWTA58588.2023.10249747

This paper presents an automated approach for the inspection of photovoltaic (PV) cells using electroluminescence (EL) imaging. Histogram of Gradient (HoG) features are extracted from EL images and then reduced using Principal Component Analysis (PCA). The reduced features are then classified using a Support Vector Machine (SVM) with a linear kernel. The performance of the model is evaluated using a publicly available dataset of EL images. Results show that the proposed method is able to classify EL images of PV cells with an accuracy of 100% for PV cells of size 4 and 16. The evaluation of the model is faster and more efficient when PCA is used. The proposed method is a promising tool for automating the process of EL image inspection. © 2023 IEEE.