A tomographic approach to multiarray ocean surveillance

One of the most widely applicable areas of signal processing in the last few years has been tomography, or image reconstruction. This area consists of a set of inverse problems characterized by the common trait that the desired quantities are the values within a field which can only be observed by looking through the field. This paper explores the applicability of tomographic ideas to the problems of ocean surveillance. Succinctly stated, beamformer outputs look like integrals of the acoustic field along the paths which include many sources of acoustic energy. By observing many beams from several arrays, the acoustic field in the ocean might be reconstructable. The basic ideas of tomography are presented, with a description of how available measurements in sonar are appropriate for the technique. The ideas are applied to two germane problems: noise mapping and localization.