The inversion of 3D multichannel seismic data using wave-equation techniques is generally a cost prohibitive solution. The main reason for the high cost is the size of the prestack data and the time-variability of wave equation processes. The breakthrough with ICO is that the modeling operator, AMO, is very compact and therefore considerably cheaper to apply than other wave equation operators such as DMO or migration. Also, ICO is velocity independent since it is applied to normal-moveout corrected data. The iterative solution is therefore less sensitive to the velocity field such as the case for least-squares migration.

The following sections discuss several implementation tricks and robust techniques from inverse theory that enable cost effective implementation of ICO on 3D data.

- Log-stretch Fourier transform
- Iterative solution for the pseudo-inverse
- Diagonal weighting preconditioning
- Data-space preconditioning
- Model-space preconditioning
- Data and Model preconditioning
- Regularization of the inversion

1/18/2001