Conclusions

We have developed an algorithm for elastic seismic inversion in the k-$\omega$ domain. domain. It uses diffraction tomography techniques and DFT interpolation to map the data space into the model space. The linearized k-$\omega$ approach allows to analyze the inverse problem and to control the indeterminations due to the geometry of acquisition (finite bandwidth and cable length). The algorithm is fast and efficient.

Tests on synthetic data have shown that, even in this ideal case, the ill conditioning of the inverse problem does not allow the estimation of all the elastic parameters. The inversion of P-P reflections produces an accurate estimate of P-impedance perturbation only. S-impedance perturbation can be recovered with multicomponent data.
Our future work will address
- inclusion of the algorithm in a target-oriented inversion scheme, adding the treatment of stratigraphic overburdens;
- application to real data;
- extension to 3D media.