Model-based seafloor identification with parametric sonar data: Experimental results

Abstract

A method is proposed to estimate the geoacoustic parameters of the uppermost seafloor strata using normal-incidence backscattered acoustic data obtained with a parametric sonar in the 4–12 kHz frequency range. A model-based identification approach has been pursued. The time-domain backscattering model BORIS [Bergem et al., SACLANTCEN Memorandum SM-328 (1997)] has been selected as a forward model. A specific wavelet transform has been defined in order to obtain a significant measure of the discrepancy between data and model predictions in the wavelet-transformed domain. This discrepancy is minimized as a function of the bottom parameters. The technique has been tested on experimental data gathered in different locations in order to assess its perfomance on different bottom types. Identification results have been compared with independent ground truth collected by core analysis and video camera inspection. The comparison shows excellent agreement between acoustically determined parameters and ground truth. [This work has been partially supported by European Union, Contract No. MAS3-CT95-046, project ISACS.]

Calibration sphere for low-frequency parametric sonars

“…Another, larger class of applications exploits the properties of parametric sonar in backscattering, hence over a two-way path. These include acoustic scattering by the sea surface, 6,7 water column, 8 bottom as in the determination of geoacoustic properties 6,[9][10][11] and seafloor characterization, [12][13][14] and seabed vegetation. 15 Some other backscattering applications include sub-bottom profiling, 16 detection of objects on and in the seabed, [17][18][19][20][21] and marine archaeology.…”