De-noising procedures for inverting underwater acoustic signals in applications of acoustical oceanography

Invited paper

Michael Taroudakis

University of Crete and FORTH

Tuesday 2 june, 2015, 16:00 - 16:20

0.8 Rome (118)

Abstract:
Applications of ocean acoustic tomography and sea-bed classification using measurements of acoustic signals emitted from known sources are based on the extraction of specific features of the signal and their subsequent exploitation by post processing. The extraction of these features is sometimes very difficult to be done in a reliable way, due to the fact that the measurements are made in the presence of noise, which in some cases is a severe handicap for the exploitation of the measured signals. Although several inversion methods have been proposed to by-pass the noise problem, the issue is open to further research and de-noising strategies are in the process of being studied. In this paper we focus on a specific inversion method based on the statistical characterization of the acoustic signal (Taroudakis et al. J. Acoust. Soc. Am Vol. 119, pp 1396-1405 (2006)) and compare a couple of alternative de-noising procedures in order that the statistical features of the signal extracted from the noisy measurements could lead to reliable inversions of the critical parameters of the sea-bed and or the water column. The results presented are based on synthetic signals produced using typical characteristics of tomograpahic sources and adding noise to a level appropriate for simulating realistic experiments in a shallow water environment.

ICS file for iCal / Outlook

[ Close ]