Iterative Separation of Blended Marine Data: Discussion On the Coherence-pass Filter

In conventional marine acquisition surveys the time intervals between the firing of successive sources are large enough to avoid interference in time. To obtain an efficient survey, the spatial source sampling is therefore often (too) large. However, much attention has been drawn recently to blended acquisition designs, where sources are shot in an overlapping fashion. Waiving the constraint of no overlap can potentially lead to significantly improved quality or economics since more sources can be utilized in a given time frame. Deblending is the procedure of recovering data as if they were acquired in the conventional, unblended way. A simple least-squares procedure however, does not remove the interference due to other sources, or blending noise. Fortunately, the character of this noise is different in different domains, e.g., it is coherent in the common source domain, but incoherent in the common receiver domain. Hence, a proper coherence-pass filter should be able to discriminate between signal and blending noise. Furthermore, such a filter can be integrated into a regularized inversion scheme, where the separation is performed in an iterative way. Three types of such coherence-pass filters, f - k, f - kr - ks, and t - p, are presented here as part of a steepest-descent type of algorithm. When applied to a numerically blended field dataset, the t - p filter outperforms the other two.