Segmentation of cell-level anomalies in electroluminescence images of photovoltaic modules
dc.contributor.author | Otamendi, Urtzi | |
dc.contributor.author | Martinez, Iñigo | |
dc.contributor.author | Quartulli, Marco | |
dc.contributor.author | Olaizola, Igor G. | |
dc.contributor.author | Viles, Elisabeth | |
dc.contributor.author | Cambarau, Werther | |
dc.contributor.institution | SISTEMAS FOTOVOLTAICOS | |
dc.date.issued | 2021-05-15 | |
dc.description | Publisher Copyright: © 2021 International Solar Energy Society | |
dc.description.abstract | In the operation & maintenance (O&M) of photovoltaic (PV) plants, the early identification of failures has become crucial to maintain productivity and prolong components’ life. Of all defects, cell-level anomalies can lead to serious failures and may affect surrounding PV modules in the long run. These fine defects are usually captured with high spatial resolution electroluminescence (EL) imaging. The difficulty of acquiring such images has limited the availability of data. For this work, multiple data resources and augmentation techniques have been used to surpass this limitation. Current state-of-the-art detection methods extract barely low-level information from individual PV cell images, and their performance is conditioned by the available training data. In this article, we propose an end-to-end deep learning pipeline that detects, locates and segments cell-level anomalies from entire photovoltaic modules via EL images. The proposed modular pipeline combines three deep learning techniques: 1. object detection (modified Faster-RNN), 2. image classification (EfficientNet) and 3. weakly supervised segmentation (autoencoder). The modular nature of the pipeline allows to upgrade the deep learning models to the further improvements in the state-of-the-art and also extend the pipeline towards new functionalities. | en |
dc.description.sponsorship | This publication resulted (in part) from the PROMISE (Advances in Photovoltaic Solar Energy Operation and Maintenance Research) project (KK2019/00088), which is financed by the ELKARTEK program of the Basque Government and is a collaborative project between Tecnalia, Tekniker, Vicomtech, UPV-ISG, UPV-TIM and Koniker. W. Cambarau would like to acknowledge I.Aizpurua, J.M. Calama and I.Arrizabalaga for contributing to the generation of the TecnaliaPR dataset. | |
dc.description.status | Peer reviewed | |
dc.format.extent | 13 | |
dc.identifier.citation | Otamendi , U , Martinez , I , Quartulli , M , Olaizola , I G , Viles , E & Cambarau , W 2021 , ' Segmentation of cell-level anomalies in electroluminescence images of photovoltaic modules ' , Solar Energy , vol. 220 , pp. 914-926 . https://doi.org/10.1016/j.solener.2021.03.058 | |
dc.identifier.doi | 10.1016/j.solener.2021.03.058 | |
dc.identifier.issn | 0038-092X | |
dc.identifier.url | http://www.scopus.com/inward/record.url?scp=85104444224&partnerID=8YFLogxK | |
dc.language.iso | eng | |
dc.relation.ispartof | Solar Energy | |
dc.relation.projectID | UPV-TIM | |
dc.relation.projectID | Eusko Jaurlaritza | |
dc.rights | info:eu-repo/semantics/openAccess | |
dc.subject.keywords | Anomaly detection | |
dc.subject.keywords | Deep autoencoder | |
dc.subject.keywords | Deep learning | |
dc.subject.keywords | Electroluminescence images | |
dc.subject.keywords | Photovoltaic modules | |
dc.subject.keywords | Weakly supervised segmentation | |
dc.subject.keywords | Renewable Energy, Sustainability and the Environment | |
dc.subject.keywords | General Materials Science | |
dc.subject.keywords | SDG 7 - Affordable and Clean Energy | |
dc.title | Segmentation of cell-level anomalies in electroluminescence images of photovoltaic modules | en |
dc.type | journal article |