Segmentation of cell-level anomalies in electroluminescence images of photovoltaic modules

• Published in: Solar energy Vol. 220; pp. 914 - 926
• Main Authors: Otamendi, Urtzi, Martinez, Iñigo, Quartulli, Marco, Olaizola, Igor G., Viles, Elisabeth, Cambarau, Werther
• Format: Journal Article
• Language: English
• Published: New York Elsevier Ltd 15.05.2021; Pergamon Press Inc
• Subjects: Anomalies; Anomaly detection; Deep autoencoder; Deep learning; Defects; Electroluminescence; Electroluminescence images; Image acquisition; Image classification; Image processing; Image segmentation; Information processing; Machine learning; Modules; Object recognition; Photovoltaic cells; Photovoltaic modules; Photovoltaics; Pipelines; Solar energy; Spatial discrimination; Spatial resolution; Weakly supervised segmentation; Deep learning; Weakly supervised segmentation; Electroluminescence images; Photovoltaic modules; Anomaly detection; Deep autoencoder
• ISSN: 0038-092X; 1471-1257
• DOI: 10.1016/j.solener.2021.03.058

[Display omitted] •A novel cell-level anomaly segmentation pipeline for solar panels is proposed.•Several cutting-edge deep learning techniques are used to achieve robust performance.•Electroluminescence images reveal the most slight and subtle anomalies.•Multiple data sources are combined and augmented to create a diverse dataset.•Weakly supervised model is used to infer high-level information from low-level knowledge. 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.