A full waveform inversion scheme for automated salt velocity model building

F. S. Dzulkefli, Mahesh Kalita, K. Xin, Tariq Ali Alkhalifah, A. R. Ghazali

Research output: Chapter in Book/Report/Conference proceedingConference contribution


In area with the presence of complex large salt bodies results in a much more complicated and highly nonlinear inversion problem where we have multiple local minima with the possibility of having an ill-pose FWI problem. Current method typically pick the top salt, flood it and pick the bottom salt and include the manually developed salt in the initial velocity model. It requires a clear contrast boundary between the salt and the background sediments. This is not only time consuming but the manual interpretation is vulnerable to errors especially the bottom of the salt. An automatic velocity model building is a better alternative to manual interpretation and horizon picking. By applying FWI in two stages of model regularization namely, FWI+TV and flooding, we successfully implemented an automated velocity model building for salt body on SEAM 3D salt model. FWI+TV act as a penalty function to control the variation in the model while preserving the edges of the salt body and flooding smear the high velocity below the top of the salt across the region where there is a drop in the velocity with depth. The proposed method is more robust and less time consuming compared to the standard 'migrate-pick-flood' approach.
Original languageEnglish (US)
Title of host publicationAPGCE 2019
PublisherEuropean Association of Geoscientists & Engineers
StatePublished - Jun 14 2020


Dive into the research topics of 'A full waveform inversion scheme for automated salt velocity model building'. Together they form a unique fingerprint.

Cite this