A lightweight network for photovoltaic cell defect detection in electroluminescence images based on neural architecture search and knowledge distillation

TL;DR

This paper introduces a lightweight neural network for PV cell defect detection in electroluminescence images, utilizing neural architecture search and knowledge distillation to achieve high accuracy with fewer parameters, suitable for industrial deployment.

Contribution

The paper pioneers the use of neural architecture search combined with knowledge distillation to design a lightweight, high-performance model for PV defect detection.

Findings

Achieved 91.74% accuracy on PV defect dataset.

Model has only 1.85 million parameters.

Outperforms existing methods in accuracy and efficiency.

Abstract

Nowadays, the rapid development of photovoltaic(PV) power stations requires increasingly reliable maintenance and fault diagnosis of PV modules in the field. Due to the effectiveness, convolutional neural network (CNN) has been widely used in the existing automatic defect detection of PV cells. However, the parameters of these CNN-based models are very large, which require stringent hardware resources and it is difficult to be applied in actual industrial projects. To solve these problems, we propose a novel lightweight high-performance model for automatic defect detection of PV cells in electroluminescence(EL) images based on neural architecture search and knowledge distillation. To auto-design an effective lightweight model, we introduce neural architecture search to the field of PV cell defect classification for the first time. Since the defect can be any size, we design a proper search structure of network to better exploit the multi-scale characteristic. To improve the overall performance of the searched lightweight model, we further transfer the knowledge learned by the existing pre-trained large-scale model based on knowledge distillation. Different kinds of knowledge are exploited and transferred, including attention information, feature information, logit information and task-oriented information. Experiments have demonstrated that the proposed model achieves the state-of-the-art performance on the public PV cell dataset of EL images under online data augmentation with accuracy of 91.74% and the parameters of 1.85M. The proposed lightweight high-performance model can be easily deployed to the end devices of the actual industrial projects and retain the accuracy.