Generalizing and stabilizing reverse time migration deconvolution imaging

While for reverse time migration deconvolution imaging condition offers the strongest possibility for source designature and compensation of illumination, cross-correlation-based imaging conditions are the most widely used due to their high stability. We propose here a general theoretical frame and practical solution for stabilizing deconvolution-based reverse time migration. Our approach involves an optimization scheme regularized by a set of constraints. The proposed constraints insure both high-resolution and removal of low frequency migration noise arising in case of diving or reflected waves in the background model. We show the improvement obtained with our approach on synthetic and real datasets.