Image Space

We now apply adaptive subtraction after migration. We perform split-step wavefield downward-continuation migration to go from the data to the image space. Both the original seismic data and the multiples model were migrated with the same algorithm. Because the velocity model is still an unknown at this stage of the processing, we use a simple, vertical-gradient velocity function. We use the same velocity model for the original data set and the multiples model. Since this particular data set corresponds to an OBC acquisition, both the original data set and the multiples model were re-datumed in order to have both sources and receivers at the same depth level before the migration.

Figure  shows one angle-domain common-image gather, following the same presentation scheme for the results in the data space. The figure shows, from top to bottom: A) the migrated multiples model; B) the result of adaptive subtraction in the image space, that is the primaries; C) the filter obtained with the migrated multiples model for performing the subtraction; and D) the migrated data set.

We first observe that both the migration of the entire data set and the migration of the multiples (panels A and D) have a residual curvature in the angle gathers because we have the correct velocity model. However, this is not an obstacle to performing the multiples subtraction in this domain, since both panels present the same residual moveout. After estimating the filter for performing the subtraction (panel C), we were able to eliminate almost all the multiples present in our multiples model.

 

ispace_4panel
Figure 4 Image space multiple removal. From top to bottom: A) Migrated multiple model; B) Primaries; C) Filter; D) Migrated data.