Tomographic Image Reconstruction of MCM Targets Using Synthetic Dolphin Signals

摘要

Synthetic aperture sonar/radar (SAS/SAR) imaging and back projection (BP)acoustic tomography are identical processes if echoes are appropriately filtered. This equivalence has several important implications. First, the point spread function of a BP image is equivalent to the range, cross-range ambiguity function (RCAF) of the SA system. Any process that makes the RCAF more like a two-dimensional impulse will improve image quality. The rotated wavelet transform, for example, yields an impulse-like RCAF even for sparse angular sampling. The wave-forms and beam patterns for rotated wavelet imaging resemble those used by an echolocating dolphin. Tradeoffs between system bandwidth and angular observation interval can be predicted from the RCAF. The SAS/BP equivalence also implies that part of a BP image can be constructed without having to construct the whole image. Another implication is that a high-resolution BP 'feature image' can be constructed. A feature image is a generalization of conventional reflectivity vs. position. Reflectivity is replaced by a feature that represents surface roughness, for example. Synthetic aperture processing can also be used to construct a three-dimensional (3-D) BP image when the sonar is above the plane of target rotation. Applications include clutter/speckle suppression via feature images, better focusing via a composite 2-D image that combines a set of range-azimuth images from different evaluations, and a 3-D representation of an object's surface (as opposed to its reflectivity), thus permitting a sonar system to synthesize a 3-D vision-like target image. These applications are illustrated with examples derived using sonar backscatter from a very shallow water (VSW) mine simulator ensonified with synthetic broadband bottlenose dolphin echolocation clicks.