Stewart A. L. Glegg
Marc Oliveri
Ctr. for Acoust. and Vib., Dept. of Ocean Eng., Florida Atlantic Univ., Boca Raton, FL 33431
The concept of acoustic imaging using ambient noise was introduced by Buckingham et al. [Nature 356, 327 (1992)] as a method for passively detecting objects in the ocean. It was suggested that two-dimensional pictorial images could be obtained using this approach, and it has been shown by Potter [J. Acoust. Soc. Am. 95, 21 (1994)] that at least 900 pixels are required to show the details of a spherical object in an underwater sound channel. This paper will describe an imaging technique that makes use of the cross correlation between an array of sensors to generate an image with over 1000 pixels. The difference between this and other approaches is that broadband signal processing is utilized, which optimizes the use of the available sound energy scattered by the object, and eliminates many of the problems associated with spatial aliasing inherent to narrow-band analysis. Consequently images can be obtained using a relatively small number of sensors. This method has been evaluated using a simple experiment in air, scaled to be equivalent to an ocean application, using loudspeakers to generate the ambient noise field. Results will be presented that show the images of various objects. [Work supported by ONR.]