Addressing uncertainty in seismic imaging by using deep learning surrogate model for reverse time migration

Resumo

In seismic exploration, many decisions are based on interpretations of seismic images, which are affected by the
presence of multiple sources of uncertainty. Uncertainties exist in data measurements, source positioning, and subsurface
geophysical properties. Reverse time migration (RTM) is a high-resolution depth migration approach useful for extracting
information such as reservoir localization and boundaries. RTM, however, is time-consuming and data-intensive as it
requires computing twice the wave equation to generate and store an imaging condition. RTM, when embedded in an
uncertainty quantification algorithm (like the Monte Carlo method), shows a many-fold increase in its complexity and
cost due to the high input-output dimensionality with computationally intensive requirements’. Hence, one of the main
challenges facing uncertainty quantification in seismic imaging is reducing the computational cost of the analysis. In this
work, we propose an encoder-decoder deep learning surrogate model for RTM under uncertainty. Inputs are an ensemble
of velocity fields, expressing the uncertainty, and outputs the seismic images. We show by numerical experimentation
that the surrogate model can reproduce the seismic images accurately, and, more importantly, the uncertainty propagation
from the input velocity fields to the image ensemble.