Research Trends and Case Studies of Deep Learning Applications in Geo-electric and Electromagnetic Surveys

Abstract

Technological innovations within the context of electrical and electromagnetic (EM) surveys have allowed for a rapid, efficient, and easier acquisition of a high quantity of data. Such innovations have been integral in mineral exploration and groundwater surveys. On the other hand, conventional inversion of electrical or EM survey data is computationally time-consuming and expensive. To circumvent the limitations of conventional inversion, the implementation of deep learning (DL) using improved neural networks has garnered substantial attention. In this study, we review various DL methods that can be used as substitutes for traditional inversion methods. Specifically, we investigate cases highlighting the successful implementation of DL to electrical or EM surveys and also comprehensively examine the advantages and disadvantages of such an application of DL.