Utilization of muography and machine learning in underground exploration

Abstract

MUOGRAPHY, a novel technique of non-destructive monitoring for gigantic structure, is expected to apply to underground exploration. In this article, the possibility to incorporate machine learning into this interesting and challenging technical topic is studied. After looking at overview of both of physics in muography and deep learning briefly, to realize precise tomography and make observation time shorter as possible, the promising methods such as super resolution, GAN, and LSTM are reviewed. As a practical activity, the abandoned mine exploration using muography is reported.