Peter T. Gough
Dept. of Elec. and Electron. Eng., Univ. of Canterbury, P. O. Box 4800, Christchurch, New Zealand
Michael P. Hayes
Industrial Res. Ltd., Christchurch, New Zealand
Most synthetic aperture (side scan) sonars, SAS, assume that the towfish travels along a straight track. Unfortunately, even a sub-wavelength sideways movement (sway) causes significant errors. For SAS to reach its potential, methods need to be found to monitor and correct these errors. Three techniques exist to estimate the unknown movements of the towfish. The first is to use precise navigation instruments, the second is to correct the image using point-spread-variant image enhancement---relying on dominant point reflectors in the object field. The third is the topic of the simulations reported in this paper. This technique makes one assumption: that in the absence of any dominant echoes, the only parameter common to all ranges is the movement of the towfish. Thus after all the range-gated echoes have been averaged in some fashion, any residual bias is due to the movement of the towfish. Here, a series of simulated trials to test this hypothesis is reported. It is shown this technique fails when a dominant reflector is introduced in the clutter-limited object field. The conclusion is that a combination of techniques two and three should enhance the quality of SAS image reconstruction under most realizable situations.