Future work - Conclusion

We presented a study on the estimation of the anisotropic velocity parameters based on the residual moveout curves in anisotropic angle-domain common image gathers.

We first performed a brief theoretical study of the estimation problem by formulating it mathematically as an inversion problem. We showed that provided that the range of the aperture angles is not too small, the problem is well posed with relatively small condition numbers, whatever the sampling of the aperture-angle axis. The theory also highlighted the well-known fact that the estimation of the horizontal velocity is poor when the range of the aperture angles is small and that one should be careful when trying to estimate anisotropic parameters when the range of aperture angle is large (significant noise propagation).

In a second part, we computed anisotropic velocity spectra from synthetic angle-domain common-image gathers. This study showed that velocity spectra can introduce some bias in the estimation of the anisotropic parameters and that there are large trade-offs between the different anisotropic parameters. However, picking the set of anisotropic parameters that correspond to the maximum semblance, proved to overall improve the anisotropic parameters' estimation.

Our future work will consists in the development of the two following estimation techniques:

• The computation of the velocity spectra and the iterative estimation of the anisotropic parameters have to be explored: is there an optimal search path to determine the anisotropic parameters (successive estimation of the different parameters or estimation of several parameters at the same time)? This procedure will also have to be tested on real data estimation problem, for which some continuity constraints will be key.
• The actual measure of the residual moveout (using an automatic technique such as the one presented by (8)) and its inversion is a way of estimating the anisotropic parameters that has to be explored and that could be very successful. It has the advantage of better constraining the estimation of the anisotropic parameters to the data, in comparison with the method developed by (1).