Adaptive deconvolution of seismic signals-performance, computational analysis, parallelism

The deconvolution problem in the face of incomplete model knowledge is addressed. The proposed solution involves the use of the Lainiotis adaptive estimation algorithm in conjunction with two classes of conventional deconvolution algorithms. The overall performance is investigated through extensive simulation. The computational requirements of the adaptive algorithms are studied in detail. Results indicate that the use of the adaptive algorithms can be highly beneficial with regard to performance, especially when little or no information about the model is available. The computational overhead is low, especially if a parallel implementation is possible.<
>