One of the most problematic areas where unexploded ordnance (UXO) can be situated is in an underwater environment, buried in sediment. Since burial frequently occurs in silt or mud, it is of great importance to have a capability for the detection of UXO in such buried conditions. This project considers detection of objects buried in mud using a side-looking, low frequency (LF) sonar system mounted on a boat. The main advantage of such a system is that it has a wider area coverage compared to downward-looking systems, enabling efficient surveying. The detection of buried UXO with an LF, side-looking sonar system, however, is a challenging problem. This is caused primarily by high reverberation and possible clutter. In addition, the amplitude of the target echo of an object is reduced by burial. For these conditions, tools need to be developed for the detection of UXO and other targets of interest and for the subsequent discrimination between targets and clutter contacts. It is critical that the processing and analysis techniques employed to achieve the task are evaluated on data acquired in operationally relevant conditions.

The objective of this project is to develop advanced processing techniques for improving detection and classification. Techniques include LF, synthetic aperture sonar and concepts originating from seismic surveying processing. With these advanced processing techniques, a set of sonar images, derived from data acquired in an operationally relevant environment, will be constructed. This sonar image gallery can subsequently be used to derive robust classification features.

Technical Approach

Data acquired with an experimental side-looking sonar system, as part of a series of experiments funded by the Netherlands Ministry of Defense, will be input for this project. The experimental sonar system comprises both a horizontal and a vertical receiver array, enabling the signal-to-reverberation ratio to be improved by the suppression of multipath reverberation and to synthetically increase the aperture by synthetic aperture sonar (SAS) processing, and covers a bandwidth between 1 and 30 kHz. The data are acquired in the MUD2009 and MUD2011 experiments conducted under operationally relevant conditions in an estuary (Haringvliet) in the Netherlands. The water depth ranges between 8 and 15 m. A set of test objects including sand-filled cylinders, boulders, 155 mm grenades, plastic targets, minelike targets, and Mk82 and Mk84 bombs—both proud and buried—was deployed (only cylinders and boulders in MUD2009). The MUD2009 data have been analyzed using SAS and sidescan images derived from both the horizontal and vertical array data.

The scope of the current project is to investigate more advanced processing techniques to optimally extract the available information on the targets that is captured in the data. This includes processing and image enhancement techniques for deriving multi-aspect acoustic color images (also called "acoustic templates" or "acoustic fingerprints" by different authors) and information on the height of targets. This project will focus on the analysis, leveraging the data collected during the MUD2009 and MUD2011 experiments, which are funded by the Netherlands Ministry of Defense.


The project will provide information on the acoustic response of targets in operationally relevant conditions. This is important for the development of robust algorithms for the detection and classification of buried UXO since these data are complementary to data acquired in controlled conditions such as, for example, the Office of Naval Research- (ONR) and SERDP-funded PONDEX trials carried out at the Naval Surface Warfare Center Panama City Division (NSWC PCD). The MUD2009 and MUD2011 data can therefore provide an opportunity for evaluating the detection and classification performance of multi-aspect acoustic color templates derived either from controlled experiments or from high-fidelity simulations. Tests in operationally relevant environments of this kind are critical for the development of an operational capability for the detection of buried UXO in muddy sediments and will therefore benefit Army munition remediation efforts. The muddy estuarine environments of the Netherlands represent one class of operationally relevant conditions. Furthermore, the processing techniques in the analysis of the MUD data could complement techniques investigated in already existing U.S. Department of Defense-sponsored efforts, a further major benefit of the work presented in this project. (Anticipated Project Completion - 2013)