CONCLUSIONS

When the wavelet is known, the ICP decon completely retrieves that components of the reflectivity series that share the same spectral region with the wavelet. In addition, the ICP algorithm keeps the same amplitude level in the whole frequency domain, resulting in a recovered time series that has the same sharpness as the original reflectivity sequence. When the wavelet is unknown the final result will depend on the choice of the initial wavelet, and when the minimum-phase equivalent of the trace is chosen to start the process, the ICP decon corresponds to a slightly sharper version of the predictive deconvolution.

Although a mathematical proof to assure convergence is still required, the method reached convergence in all the tests performed with the use of different choices for the starting wavelet.