Physics Informed Neural Networks application for seismic data interpolation

What is it about?

We propose to interpolate seismic data by utilizing physics informed neural networks (PINNs). Two feed-forward neural networks are jointly trained using the local plane wave differential equation as well as the available data as two terms in the objective function. The main one interpolates the seismic data, while the second smaller auxiliary network simultaneously estimates the local slope attribute of the events. Results on synthetic and field data validate the effectiveness of the proposed method in handling highly aliased data and data with large gaps.

Featured Image

Photo by Uriel SC on Unsplash

Photo by Uriel SC on Unsplash