Application of multiscale analysis to coherent pulsed sea clutter returns

The variability of sea clutter when viewed at low grazing angles has divided opinion upon the correct model for the amplitude statistics. If the scattering mechanism is non-stationary then a method of utilising existing stationary theory is to identify correlated regions. The continuous wavelet transform applied to coherent pulsed data demonstrates that scattering events from the sea surface can be discrete and occur in different forms. Results are presented to show the varying wave velocity spectrum as a function of time. This ultimately gives us second order statistics over a range of velocities which allows us to propose a target detection strategy at clutter velocities, based on the time frequency aspect ratio of the observed event. Measuring the persistence of maxima in the transform allows detection below the clutter level without setting an amplitude threshold.