Reducing ERTM artifact arising from incorrectly imaged mode conversions in the dip-angle domain

Abstract

Summary A straightforward and naive implementation of elastic reverse time migration (ERTM) might produce images that are not very useful mainly because they are filled with noises arising from incorrectly imaged mode conversions, cross-talk, and back-scattered energy. Therefore, we present a ERTM workflow that overcomes these problems. We first employ vector-based normalized cross-correlation imaging conditions (VBNICs) for PP and PS reflections to avoid cross-talk and polarity reversal. Based on Poynting-vector imaging condition, we implement angle filtering to reduce low-wavenumber and high-amplitude noise. Then, we efficiently extract dip-angle gathers by using Poynting vectors and finally utilize them to attenuate the artifact caused by non-physical wave-mode conversions because their slopes differ from the migration dip. Numerical examples demonstrate that the proposed ERTM workflow can produce images with less noise and higher quality.