3-D Depth migration: parallel processing and migration movies

Dmitri Gavrilov, Laurence R. Lines, Henry C. Bland, Tony Kocurko

The advent of 3-D depth migration has brought about many challenges and opportunities in the applications of computer science to seismic imaging. We shall examine two fields of computer science, which will impact the 3-D depth migration technology: parallel processing and visualization. Parallel computational applications will make 3-D migration calculations tractable while visualization will allow us to examine the input models, snapshots of computations, and the output images.

In our depth migrations we primarily examine two methods: Kirchhoff migration and reverse-time migration. Zhu and Lines (1998) have examined the computational advantages and disadvantages of these methods. The Kirchhoff algorithm is generally the faster of the two, but is slightly less accurate than reverse-time migration when velocity models are accurately defined since the Kirchhoff is based on ray theory approximations to the wave equation.