We propose a fast method to calculate local wavefront attributes for 3D prestack seismic data. First step is to compute attributes on a coarse regular or irregular grid in time and space using conventional approaches. Second step is very fast and efficient inpainting of the attributes in remaining locations by artificial intelligence utilizing a specially trained deep neural network. The method incorporates multi-parameter attributes using a special colouring scheme and allows estimation of multiple attributes simultaneously during one run. We demonstrate that inpainting of local wavefront attributes for nonlinear beamforming can greatly speed up prestack enhancement of 3D seismic data. Other applications such as velocity analysis or seismic tomography can be implemented using a similar approach.
Presentation Date: Monday, September 16, 2019
Session Start Time: 1:50 PM
Presentation Time: 2:40 PM
Location: 221D
Presentation Type: Oral
