Defining and modeling the limits of high-resolution underwater acoustic imaging

In underwater warfare, detection of sea mines is a challenging discipline. Within the field of sea mine hunting, the detection of fully buried sea mines has always been the most difficult. In fact all objects located under water near to interfaces with high acoustic and or magnetic impedance differences and roughness are hard to detect. Today most up to date modern sea mines have a non magnetic hull and a minimum of magnetic materials on board. They preferably hide near interfaces that hinder detection. The shape often makes the modern sea mine stealth for acoustic detection instrumentation. The electronics on board of the mine also becomes increasingly smarter, smaller, lower power and more complex. Up till the time of writing this thesis, despite all the promising publications, there is no known instrument or combination of instruments capable of detecting buried sea mines within a reasonable timeframe. During the last decennia a number of instruments, methods and platforms have been thoroughly tested and evaluated. This research aimed at evaluating the applicability of very high resolution reflection seismic systems in buried object detection. The thesis basically covers 3 fields: field experiments (1), acoustic modeling and migration (2) and the evaluation of an existing acoustic 3D model and construction of a new alternative acoustic model based upon computer gaming technology (3). (1) The tests described cover detection of objects near interfaces. With interfaces is meant: air-water and seafloor- water interface. Some of the tests were performed by NATO partners, some tests were done in cooperation with NATO partners, and some tests were performed by the Belgian Navy. (2) Triggered by the failed processing of some field tests, 3D acoustic models were built and used for the evaluation of the processing algorithms. The research tried to interconnect field tests to existing mathematical models and focusing algorithms. The limitations of the migration algorithms and the implementations on the recording conditions were identified. (3) Trying to push the modeling to a higher level, the NURC model BORIS 3D was used and found to be inapplicable for the testing of migration algorithms. Eventually a completely new acoustic model enabling the accurate and reliable reproduction of acoustic signals and shadows in a full 3D environment with a velocity gradient and objects was developed. In short: the model handles very complex 3D environments. Unlike many other 3D models, the signals produced by this new model can also be used to test 3D migration algorithms and does not only serve as an estimate of sonar performance. The thesis had also some ‘spin-out’ results in the field of sediment dynamics. A new 4D sand dynamic model was developed to generate sand ripple fields where not only the pattern and shape of the ripples were more realistic but also parameters such as acoustic impedance and grain size distribution were made available. Based upon these results and existing swath bathymetry maps new algorithms for GIS applications to detect and vectorize sand crests are proposed