Polarity reversal correction for elastic reverse time migration

Polarity reversals in PS and SP images of elastic reverse time migration will cause destructive interference in the final stacked results, which may lead to an erroneous interpretation. We derive a polarity reversal correction based on an analysis of the polarity distribution of PS and SP images. The key aspect for a polarity reversal correction is the polarity distribution of the S-wave component. Polarity reversal correction methods can be performed according to the actual patterns of the polarity distribution in gathers. For the 2D isotropic case, a sign factor is introduced to represent the polarity distribution of the S-wave component; this sign factor can be computed using the energy flux density vector. We present a procedure to correct the polarity reversal for elastic reverse time migration in the common-shot domain. First, the sign factor is computed during the wavefield reconstruction for every imaging point. Then, the polarity reversal is corrected by multiplying the PS and SP images with the sign factor at every time step when an elastic imaging condition is applied. We also design a filter algorithm for the sign factor to improve its consistency along an event and thereby to diminish the impact of the inaccuracy of the energy flux density vector and to improve the imaging results. Numerical examples have shown that the polarity reversal correction procedure works and that the sign filter algorithm successfully eliminates the consequences of the inaccuracy of the energy flux density vector. The quality of the migration events is effectively improved.