A Flexible Framework for Constrained Dix Inversion using Alternating Direction Method ofMultipliers

SUMMARY

We present a novel constrained Dix inversion method that leverages proximal operators within the Alternating Direction Method of Multipliers (ADMM) framework.

Addressing problems inherent to Dix inversion, our method robustly manages various noise types in picked RMS velocity functions, including spikes and outliers in measurements that can be distributed in both the space and time domains.

Moreover, the inverted model can be adaptively guided towards either blocky or smooth interval velocity fields, making it well-suited for large-scale three-dimensional projects.

The effectiveness of the proposed method is demonstrated through comprehensive validation on both synthetic and real data examples.