The overall objective of this project was to work towards resolving issues that affect sonar detection and classification of underwater unexploded ordnance (UXO). A central hypothesis is that the environment within which UXO must be detected and classified significantly alters the acoustic response of UXO and the environment must be taken into account in order to develop robust detection and classification strategies. The primary aim was to identify and understand target and environmental factors affecting sonar performance against proud, partially buried, and buried targets and to identify robust signal features or image characteristics unique to a given UXO.
Acoustic scattering from a set of targets was collected under well-controlled and measured environmental conditions in the test pond facility at the Naval Surface Warfare Center, Panama City Division (NSWC PCD) in Panama City, Florida. The data form a database of target signatures and offer ground truth for model predictions of the acoustic scattering. The hybrid model used in this research combines a 2D finite-element (FE) model for a target response with 3D wave propagation via a Sommerfeld-Helmholtz integral. The FE model exploits the axisymmetry of the target, and each UXO has a unique FE mesh. Once the FE models are validated, environmental conditions used within the models can be altered to study changes in the acoustic scattering and to simulate environments where collection of acoustic scattering data may be infeasible due to either cost or location. Both data and model results then can be used to test current classification algorithms and provide insight into improvements of these algorithms or guidance for development of new classification schemes.
The experimental data and model predictions were generated in a manner suitable for synthetic aperture sonar (SAS) processing. SAS images of individual targets and collections of targets have been constructed. In addition, a novel SAS filtering algorithm that allows one to isolate the acoustic scattering associated with a specific target has been applied to the data. SAS filtering was able to isolate target signatures for adjacent targets separated by approximately 1.5 meters. The isolated target signatures are then amenable to additional signal processing techniques. Acoustic templates (also called acoustic color plots) have been generated from isolated signals. An acoustic template is a colored representation of the (absolute) target strength (TS) as a function of frequency and aspect angle. Finally, data and model predictions have been fed into a binary classifier available in the mine countermeasure community. This classifier identifies features within the data to discriminate whether a signature is UXO-like or not UXO-like.
This research provides acoustic data on a collection of underwater targets under various environmental conditions. The data provides the ground truth in the construction of models, which then can be used to simulate the scattering responses of targets in other environments. The models can provide cost savings by reducing the number and/or duration of field tests. By providing a capability to simulate sonar performance, the Department of Defense can make informed decisions on the relative merits of existing sonar systems and on proposed modifications to these systems for underwater UXO management. Other researchers also can leverage the data and analysis techniques.