U-net-based pseudoseismic imaging for the short-offset transient electromagnetic method

Short-offset transient electromagnetic (SOTEM) is an innovation of the artificial-source electromagnetic method, and detection is conducted in the near-source area using an electromagnetic formation wave so that the signal has strong strength and large bandwidth. Inspired by the parallelism of the propagation of seismic body waves and EM formation waves, this paper intends to implement pseudoseismic imaging (PSI) on SOTEM data to give play to the high resolution of SOTEM. Traditionally, to perform the wave-field transform (WFT), a set of ill-posed linear equations needs to be solved, and the solving process has strong ambiguity and instability. This paper presents a new method for performing WFT based on U-Net, and a nonlinear mapping from the EM diffusion field to the pseudoseismic wave field under the same velocity model is established by training the network. To eliminate the error that might result when the time‒depth transform is conducted for the wave-field record, another U-Net is used to transform the wave-field record into a depth-offset profile to accomplish the whole pseudoseismic imaging process for SOTEM. Simulation and measured data are used to validate the effect of the networks, and the results indicate that this approach is highly feasible, thus providing a new strategy for using SOTEM to detect sedimentary strata.