Defect Classification in Photovoltaic Modules through Thermal Infrared Imaging using Machine Learning
Christopher Dunderdale, Warren Brettenny, Chantelle Clohessy, and 1 more author
2020
As the global energy demand continues to soar, solar energy has become an attractive and environmentally conscious method to meet this demand. This study examines the use of machine learning techniques for defect detection and classification in photovoltaic systems using thermal infrared images. A deep learning and feature-based approach is also investigated for the purpose of detecting and classifying defective photovoltaic modules. The VGG-16 and MobileNet deep learning models are shown to provide good performance for the classification of defects. The scale invariant feature transform (SIFT) descriptor, combined with the random forest and support vector machine classifier, is also used to discriminate between defective and non-defective photovoltaic modules in a South African setting. The successful implementation of this approach has significant potential for cost reduction in defect classification over currently available methods.