Abstract: A permutation that optimizes correspondence between the seismic data and the simulated data is computed using a graph space optimal transport formulation-based misfit. The seismic data or simulated data are transformed into auxiliary data by applying a portion of time shifts computed from the optimal permutation before updating the structural model of the explored underground formation. The full-waveform inversion minimization of the distance between auxiliary data and the seismic data or simulated data to which partial time shifts have not been applied, may be embedded in a Kantorovich-Rubinstein norm.

Abstract: A DUnet engine produces a processed image of seismic data acquired over an underground formation. The DUnet engine includes: a contractive path that performs multilayer convolutions and contraction to extract a code from the seismic data input to the DUnet, an expansive path configured to perform multilayer convolutions and expansion of the code, using features provided by the contractive path through skip connections, and a model level that performs multilayer convolutions on outputs of the contractive path and expansive paths to produce the processed image and/or an image that is a difference between the processed image and the seismic data. A fraction of the seismic data may be selected for training the DUnet engine using an anchor method that automatically extends an initial seismic data subset, based on similarity measurements. A reweighting layer may further combine inputs received from layers of the DUnet model to preserve signal amplitude trend.

Abstract: A permutation that optimizes correspondence between the seismic data and the simulated data is computed using a graph space optimal transport formulation-based misfit. The seismic data or simulated data are transformed into auxiliary data by applying a portion of time shifts computed from the optimal permutation before updating the structural model of the explored underground formation. The full-waveform inversion minimization of the distance between auxiliary data and the seismic data or simulated data to which partial time shifts have not been applied, may be embedded in a Kantorovich-Rubinstein norm.