High-resolution seismic velocity inversion and imaging of the shallow-subsurface using a combined jmi/fwi workflow

Abstract

Full waveform inversion (FWI) is a popular high-resolution velocity estimation technique which is distinctively hindered when applied to the shallow-subsurface due to the high level of noise, complex structures,\elastic effects as well as the often lacking low frequencies in the dataset. In this abstract we propose a to use a data-driven, reflection based inversion technique (Joint-Migration Inversion (JMI)) to create a starting model for FWI which pushes the inversion out of a local minimum to recover previously hidden structures. The research is applied to a 2D acoustic model and compares the results of using three different types of starting model, including one generated using JMI.