Deep learning based model for Defect Detection of Mono-Crystalline-Si Solar PV Module Cells in Electroluminescence Images Using Data Augmentation

With the large-scale application of renewable energy, the solar energy has drawn great attention. Due to the huge capacity of solar energy installed during the past decades, this paper proposes a method for automatic classification of defect in EL image of mono-crystalline-Si PV module cell, which is helpful for the intelligent operation and maintenance of photovoltaic power station. This paper presents a method for data augmentation based on the geometric morphometric characteristic first and then designs a CNN based model for defect classification in EL image by extracting those features. The performance of model is evaluated with various methods and it is compared with the existing method. The result of comparison demonstrates the efficiency and excellence of model for classification.