Detection and classification of buried objects with an adaptive acoustic mine-hunting system

Abstract

Cost- and time-effective mine countermeasures have become high priority in today's U.S. Navy. Current systems lack adequate target classification/localization capabilities; and thus development of new and innovative technologies is essential for mine search operations in littoral environments. A unique system design is described that fuses sub-bottom seafloor imagery and signal classification algorithms. Seafloor and subbottom maps are produced by a compact 6 transmitter, 32 element receive array sonar system employing a FM upsweep transmit signal containing energy from 5 to 23 kHz. This system provides 4 to 8 cm spatial resolution, up to 2 m bottom penetration, and is ideally suited for detecting proud and buried mine-like targets. Image processing algorithms automatically detect and localize targets of interest. Targets are extracted and passed to biomimetic signal classification algorithms that map time-frequency patterns into object class declarations. The system and processing stages are presented and an experiment is described in which buried objects consisting of a concrete block, coral head, sand-filled aluminum spheres, sand-filled scuba tanks, 155 mm ordnance, and a mine-shape are successfully differentiated. These results are encouraging, and suggest that a hybrid system employing a conjunct seafloor image and biomimetic signal classification can rapidly and accurately detect and classify buried mine-like objects in the littorals.