Deep convolutional neural networks for SOFC core-shell recognition

The quantitative analysis of the microstructure and related electrical properties of anode substrate are necessary to optimize the fabrication conditions for the best performance of Solid Oxide Fuel Cell (SOFC). In SOFC Optical Microscopic (OM) images, core-shell microstructure can observed. The core-shell will affect the electrochemical performance of the SOFC, and reduce its service life to same degree. Highly accurate recognition of core-shell images of SOFC microstructure is of great significance. In this paper, a Convolutional Neural Network (CNN) based on SOFC OM images using supervised training is proposed. In our CNN model, we do not need the manual design of the feature space, seek an effective feature vector classifier or segment specific detection object and image patches, which are the main technological difficulties in the adoption of traditional image recognition methods. With the utilization of a simple data augmentation method and fast convergence speed, our method can achieve the high accuracy rate and outstanding recognition capability for core-shell images.