conclusion

I have shown that the multiple reflections may be attenuated in the prestack domain using the Hyperbolic Radon Transform. Different inverse problems have been solved to obtain the velocity panels showing different properties: (1) the l² and l¹ norm produce comparable results in the velocity space and for the reconstructed data, (2) the l¹ norm with l¹ regularization shows spikier results in the velocity domain but converge much more slowly than the other methods, and (3) the stacked sections of the multiple-free data are very similar for the different inverse problems. The good quality of the data explains the small discrepancy between the l² and the l¹ norm; more noisy gathers would lead to different conclusions.