Efficient deep feature extraction and classification for identifying defective photovoltaic module cells in Electroluminescence images

作者：

Highlights：

• Efficient solar cell Electroluminescence image classification methods are proposed.

• A novel fast-learning lightweight convolutional neural network model is proposed.

• Faster feature extraction performed using pre-trained Deep Neural Networks.

• State-of-art results achieved using feature selection and machine learning methods.

• 90.57% and 94.52% classification accuracy obtained for 4 and 2 class datasets.

摘要

•Efficient solar cell Electroluminescence image classification methods are proposed.•A novel fast-learning lightweight convolutional neural network model is proposed.•Faster feature extraction performed using pre-trained Deep Neural Networks.•State-of-art results achieved using feature selection and machine learning methods.•90.57% and 94.52% classification accuracy obtained for 4 and 2 class datasets.

论文关键词：Electroluminescence imaging,Defect detection,Feature extraction,Feature selection,Deep features,Deep learning

论文评审过程：Received 13 August 2020, Revised 5 January 2021, Accepted 28 February 2021, Available online 3 March 2021, Version of Record 18 March 2021.

论文官网地址：https://doi.org/10.1016/j.eswa.2021.114810