Waveform Inversion with Data Correlation

A common way to obtain a suitable velocity model for migration is migration velocity analysis. These methods somehow try to focus the migrated image in the depth-domain. Recently, a method was proposed to perform the velocity analysis in the data-domain. And first tests using the convolutional model and NMO velocities proved successful. In this paper we want to incorporate this method in a waveform inversion scheme. Classical waveform inversion tries to fit the data in a least-squares sence, and it has been shown that it is difficult to get an update of the background velocity this way. By combining classical waveform inversion with the data-correlation method we hope to be able to invert for the background model as well.