Stack algorithm for recursive deconvolution of Bernoulli-Gaussian processes (seismic exploration)

The deconvolution of pulse trains modeled as Bernoulli-Gaussian processes is addressed. The detector presented implements a stack Viterbi algorithm with a parametric level of suboptimality. Simulation results are satisfactory, even in the difficult case of a poor spectral content of the wavelet. The recursive and parallel structure of the method allows fast data processing on modern architectures.<
>