Anmerkungen:
Nach Informationen von SSRN wurde die ursprüngliche Fassung des Dokuments July 13, 2023 erstellt
Beschreibung:
This paper presents a novel approach for defect detection in solar photovoltaic (PV) panels using electroluminescence (EL) imaging and advanced machine vision techniques. The proposed method involves two stages: feature extraction and image classification. Principal component analysis (PCA) is applied for image preprocessing, followed by the use of Convolutional Neural Networks (CNN) and ResNet51 for feature extraction. The extracted features are combined and classified using random forest and decision tree classifiers. Experimental results on the ELPV dataset show that the fusion of extracted features with the random forest classifier achieves the best performance. This approach demonstrates the effectiveness of deep learning and decision tree-based classifiers in accurately detecting defects in PV panels, offering a promising solution to enhance defect detection and ensure reliable operation of PV installations. The findings contribute valuable insights to researchers and practitioners in the field of solar energy and PV panel defect detection