.png){kind=icon}
.png){kind=icon}
Return
Imaging based fault diagnosis of photovoltaic modules: a systematic review
.png){kind=icon}
PRE
AI
M
A
K
DOI：10.1016/j.solener.2026.114679.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
This review provides a comprehensive synthesis of imaging-based photovoltaic (PV) fault detection methods, with particular emphasis on infrared thermography (IRT), electroluminescence (EL), and RGB visual imaging. The review systematically analyzes recent advances in imaging-based fault diagnosis, including data acquisition strategies, fault taxonomies, and machine learning based techniques applied to PV module inspection. This study integrates macro-level bibliometric analysis with micro-level technical assessment of the 2021–2024 literature, using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. The macro- and micro-level analyses enable both the identification of research trends and the detailed comparison of methodologies, respectively. The analysis reveals that the IRT imaging has been predominantly used for PV plant monitoring and fault detection. EL imaging can accurately identify cellular defects, microcracks, finger interruptions, and material degradation. However, EL imaging poses significant challenges for the field inspection of PV plants. RGB-based imaging can easily detect surface soiling, physical damage, shading, and debris. This review highlights the prevalence and impact of PV faults on energy production and emphasizes the importance of early detection. Furthermore, current challenges and future research directions are identified. These include multimodal data fusion, UAV-based inspection, edge intelligence, and emerging deep learning techniques, and highlight their potential to address current limitations in large-scale PV monitoring. The review aims to support the development of reliable, scalable imaging-based PV fault-detection systems.
Keywords:
infrared thermography
electroluminescence
RGB imaging
photovoltaic fault detection
machine learning
Journal
IF:
6.6
Papers:
1.4W
Citations:
6.2W
Organization
Cited Papers