Studies into the detection of buried objects (particularly optical fibres) in saturated sediment. Part 4: experimental investigations into the acoustic detection of objects buried in saturated sediment

Abstract

This report is the fourth in a series of five, designed to investigate the detection of
targets buried in saturated sediment, primarily through acoustical or acoustics-related
methods. Although steel targets are included for comparison, the major interest is in
targets (polyethylene cylinders and optical fibres) which have a poor acoustic
impedance mismatch with the host sediment. This particular report aims to review the
signal processing requirements of the system and to investigate the relative
performance of a number of different detection algorithms.
In order to do this, some basic signal processing concepts are presented and a number
of different approaches to detection processing were discussed, noting that the
detection system should be optimised to the class of object being sought. To this end,
the scattering characteristics of spheres and cylinders are calculated. These were used
to aid in the selection of an optimal frequency range for later incorporation into the
detection algorithms.
The study examines the use of waveform dependent filtering, processing in a number
of ways (forms of optimal filter and Synthetic Aperture Sonar, SAS) the same
scattered data obtained with incident FM pulse compression waveforms. (There is a
preliminary study which examines a simple scattered power measurement, using
pulsed AM waveforms, but this can only be seen as rudimentary preparatory work
because: (i) it cannot be compared with the waveform dependent filtering methods
because it uses a different incident waveform; and (ii) the designers of a realistic
system would almost certainly consider some form of matched filter, at the very least
an attempt would be made to match the filter input bandwidth to that of the signal).
The algorithms presented in this report are applied in an experiment and prove to be
very successful in detecting objects buried at depths of between 25 and 30 cm in the
saturated sediment of the test tank. In every case, either 60 or 300 sample points were
measured over a series of planes extending vertically into the sediment. It was found
that with 60 points (having a sample spacing of 5 cm) the resolution was not high
enough to provide conclusive detection results. Conversely, with 300 points (having a
sample spacing of around 2 cm) the buried objects could readily be detected.
xi
Simple matched filtering is shown to be useful in an environment dominated by noise.
However, the optimal filter is shown to be more successful in dealing with the
cluttered seabed environment. Target optimisation techniques had mixed success.
When the target scattering responses (for both rigid and the elastic scattering) were
incorporated into the filters, qualitative improvements in target localisation were
observed. However, these were not accompanied by an increase in the average values
of the signal-to-noise ratio.
The investigation into Synthetic Aperture techniques was limited by the positional
error in the laboratory apparatus and the small number of measurement positions used
to form the synthetic aperture. As a result of these experimental limitations, no
significant performance improvement was actually observed.
This series of reports is written in support of the article “The detection by sonar of
difficult targets (including centimetre-scale plastic objects and optical fibres) buried
in saturated sediment” by T G Leighton and R C P Evans, written for a Special Issue
of Applied Acoustics which contains articles on the topic of the detection of objects
buried in marine sediment. Further support material can be found at
http://www.isvr.soton.ac.uk/FDAG/uaua/target_in_sand.HTM.

More information

Identifiers

ORCID for R.C.P. Evans: orcid.org/0000-0002-1649-8750

Catalogue record

Export record