Pre-stack imaging of 3D seismic reflection data is at present a major challenge of seismic data processing. The main difficulty arises because of the limited power of today's computers. Manipulating volumes of 3D multi-fold seismic data and, especially, executing the large number of mathematical operations required for 3D migration is presently a very complicated and time consuming task. More than that, 3D seismic reflection surveys are characterized by very irregular acquisition geometries. This requires special migration algorithms and even new thinking.
In this thesis we present a new approach to imaging 3D pre-stack seismic reflection data. It is a practical alternative to the full 3D pre-stack depth migration and takes special care of irregular acquisition patterns. The objective of the method is three fold: (a) to perform migration before stack, (b) to correct for three-dimensional seismic effects, (c) to provide a three-dimensional velocity model and a depth image.
A 3D processing scheme is presented here. It is composed of three procedures which are different approximations to the full 3D pre-stack migration operation. The interaction between those schemes is simple, and each procedure serves as a building block for the next module. That is due to the special derivation of the algorithms in the frequency domain. The three modules that compose this work are: (1) Reorganization of multi-fold data on a regular pre-stack grid, based on Dip-Moveout and the inverse transform (DMO$\sp{-1}$). (2) 3D pre-stack time migration, using the PSI technique. (3) Two-pass 3D pre-stack depth migration, performed by splitting the full 3D calculation into a succession of 2D operations.
| Identifer | oai:union.ndltd.org:RICE/oai:scholarship.rice.edu:1911/16713 |
| Date | January 1994 |
| Creators | Canning, Anat |
| Contributors | Gardner, G. H. F. |
| Source Sets | Rice University |
| Language | English |
| Detected Language | English |
| Type | Thesis, Text |
| Format | 192 p., application/pdf |
Page generated in 0.0133 seconds