Transient electromagnetic deep learning inversion based on Pseudo Wave Field Data

Deep learning inversion techniques have been widely used in the field of geophysical data interpretation, as they can efficiently process complex data and extract useful information. However, there are some challenges in directly applying deep learning methods in transient electromagnetic (TEM) data processing. Due to the insensitivity of TEM to small-sized or large depth anomalous bodies, the imaging accuracy and resolution are poor during deep learning inversion. To address this issue, this study proposes using Pseudo Wave Field Data (PSW) obtained through TEM wavefield transformation for deep learning inversion. The simulation results show that PSW data has higher sensitivity in identifying underground anomalous electrical changes than TEM data, and deep learning inversion based on PSW can significantly improve TEM imaging accuracy. This study indicates that PSW data can be used for TEM deep learning inversion, providing a new approach for the combination of electromagnetic data processing and deep learning techniques.