An intelligent defect detection method in solar cells via IDATLGAN

Abstract

Intelligent defect detection is a key section of the PDCA comprehensive quality management in the solar cells production. deep learning-based models for its powerful capacity of extracting features has successfully applied to defect detection tasks. Nevertheless, the deep learning-based models is susceptible to these problems such as insufficient labelled data and class imbalance, which deteriorates the performance of the deep models. What's worse, the new unknown defects without labels in the target domain tend to be recognized as the unknown defects in the source domain. To solve these problems, a novel method named IDATLGAN is proposed. Experimental results show that the proposed model outperforms the DCNN model in feature extraction ability and fulfil the transfer tasks even when faced with new unknown defects without labels.