Affiliation:
1. ISME — Interuniversity Centre of Integrated Systems for the Marine Environment & DII — Dip. Ing. Informazione, University of Siena, Italy
2. ISME — c/o DSEA, Dept. Electrical Systems & Automation, University of Pisa, via Diotisalvi 2, 56100 Pisa, Italy
Abstract
An inversion technique is proposed for the determination of the geoacoustic and morphological properties of the uppermost sediment layer. The methodology is based on the use of the backscattering acoustic return when the source is a parametric instrument steered at normal incidence with respect to the seafloor. The peculiarity of the parametric sonar (i.e., its narrow beam and the absence of sidelobes) allows for discriminating between the scattering effects due to surface roughness and those due to volume perturbations. The inversion procedure is based on the minimization of a discrepancy measure between data and model predictions. Model predictions are obtained as time series realization of a stochastic process, modeling the backscattering process with the Kirchoff approximation for surface scattering and the small perturbation theory for volume scattering. The BoRIS code is used to generate the time series predictions. It is important to note that the model is stochastic, i.e., the model predicted time series with the same nominal parameters may differ from one realization to another. However, by use of wavelet transform of the signals involved, and measuring the data-model discrepancy in a generalized time-frequency domain, the stochasticity of the problem is greatly reduced. In particular, the wavelet transform is insensitive to different model realizations obtained with the same set of parameters, and sensitive to changes in the parameters. By appropriately weighing the discrepancy in the wavelet domain, and exploiting the properties of the parametric source, it is possible to separately recover the parameters influencing the surface backscattering (acoustic impedance and surface roughness) and those influencing the volume backscattering (P-wave attenuation and volume inhomogeneity), avoiding ambiguities and nonuniqueness problems. The approach proposed requires, however, a precise calibration of the parametric sonar, in terms of source level and beam pattern. Comparison of inversion results with independently measured ground truth at three different sites in the Mediterranean Sea are reported.
Publisher
World Scientific Pub Co Pte Lt
Subject
Applied Mathematics,Acoustics and Ultrasonics