Towards Deep Computer Vision for In-Line Defect Detection in Polymer Electrolyte Membrane Fuel Cell Materials

Abstract

Polymer Electrolyte Membranes (PEM) fuel cells are a promising source of alternative energy. However, their production is limited by a lack of well-established methods for quality control of their constituent materials like the electrode and PEM during roll-to-roll manufacturing. One potential solution is the implementation of deep learning methods to detect unwanted defects through their detection in scanned images. We explore the detection of defects like scratches, pinholes, and scuffs in a sample dataset of PEM optical images using two deep learning algorithms: Patch Distribution Modeling (PaDiM) for unsupervised anomaly detection and Faster-RCNN for supervised object detection. Both methods achieve scores on performance metrics (ROC-AUC and PRO-AUC for PaDiM and AP for Faster-RCNN) that are comparable to their scores on benchmark datasets and show potential for localizing relevant defects of interest. Overall, deep learning methods show promise at detecting defects and has the potential to achieve real-time defect detection.