Publication: Solar cells defects detection using deep learning approach based on yolov4 framework
Loading...
Date
2022-07-01
Authors
Omairah, Amran Khamis Mohammed
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Two most common defects encountered during manufacturing of crystalline silicon solar cells are microcrack and dark spot or dark region. The microcrack in particular is a major threat to module performance since it is responsible for most PV failures and other types of damage in the field. On the other hand, dark region in which one cell or part of the cell appears darker under UV illumination is mainly responsible for PV reduced efficiency, and eventually lost of performance. Therefore, one key challenge for solar cell manufacturers is to remove defective cells from further processing. Recently, few researchers have investigated deep learning as alternative approach for defect detection in solar cell manufacturing. The results are quite encouraging. This paper evaluates the convolutional neural network based on You Only Look Once (YOLO) algorithm for use in solar cell inspection. The state-of-the art YOLOv4 and Tiny-YOLOv4 are used in the evaluation from which the best model is proposed for online inspection and automation. Experiments results show that the multi-calss YOLOv4 model outperforms the single-class model combined with 98.78% mAP and detection speed of 62.89ms compared to an average of 96.89% mAP and prediction time of 104.73ms. The Tiny-YOLOv4 model optimized in this study achieves good results with averaged 91.92% and prediction time of 28.17ms.