Low frequency synthetic aperture sonar (SAS) is a key technique for the detection and characterization of unexploded ordnance (UXO) in the underwater environment. For buried UXO, however, it is rather difficult to determine the range and depth up to which the detection of UXO is feasible. This critically depends on the detection system properties, environmental properties such as water depth and sediment type, target properties, and the actual burial depth up to which UXO needs to be detected. For the evaluation of the effectiveness of UXO detection surveys, it is therefore not only relevant to analyze the acquired data to determine the presence or absence of UXO, but also to assess which area effectively has been covered by the survey, and up to which depth effective coverage has been obtained.
The objective of this project was to develop an approach for coverage assessment, and to evaluate the applicability of the coverage assessment approach. The development and evaluation of the coverage assessment methodology is supported by using data examples acquired with the Netherlands Organisation for Applied Scientific Research’s low-frequency synthetic aperture sonar (LF-SAS) system.
The coverage assessment approach relied on high-fidelity target-in-environment response (TIER) simulations of UXO objects of interest, which were obtained using Finite Element Modelling simulations. They were subsequently merged with measured reverberation data. Using augmented targets, it became possible to assess whether a target of interest would be detectable, if present. The detectability was assessed in SAS images using an intensity threshold detector, i.e. by considering the Signal Excess for different target depths, target orientations, and target ranges. The coverage assessment for more sophisticated detectors would be supported as well by the developed high-fidelity TIER simulation approach.
The coverage assessment results show that the maximum range for detecting buried objects was limited by the critical angle of incidence for a side looking LF-SAS system. It is difficult to obtain information on sediment sound speed using an LF-SAS system, which primarily senses back-scattering properties of the seabed. It is therefore important to collect this information on seabed properties by alternative means, e.g. grab samples and/or gravity cores. There are well-established relations between grain size and sediment sound speed.
The coverage assessment methodology critically relies on the calibration of the measurement system, processing, and TIER simulation chain. Hardware characteristics, such as beam patterns, need to be established by conducting basin measurements, and scale factors introduced in data acquisition, processing, and simulations need to be taken into account. Errors in the calibration will result in biased coverage assessment results.
The coverage assessment approach has been successfully demonstrated on experimental data acquired by the LF-SAS system near Gdynia, Poland. Simulation results for the signal-to-reverberation ratio (SRR) for a scientific cylindrical target agree well with measurements. Single-view coverage assessment results are in agreement with range and depth trends in SRR observed in experimental data.
Finally, coverage assessment results can be generated during an UXO detection survey, provided that the simulated TIER SAS images are generated prior to the survey. Timely availability of coverage assessment results helps to improve the quality of UXO surveys. Furthermore, averaged performance curves, such as P(y) curves, are useful input to the planning of future UXO surveys. This study therefore demonstrated that high-quality coverage assessment results can be efficiently obtained to aid UXO survey planning, execution, and evaluation.