Detection of the number of signals in super-resolution ocean surface current algorithm for OSMAR2000

Super-resolution ocean surface current algorithm based on MUltiple SIgnal Classification (MUSIC) is used for current mapping of Wuhan University´s ocean state measuring and analyzing radar (OSMAR2000), in which MUSIC is applied to estimate the bearings of the first-order sea echo signals. The premise of MUSIC processing is that the number of signals is known in advance. In fact, the number of signals is unknown, and needs to be estimated from received data. In the case of discrete targets, the problem can be perfectly solved using information theoretic criteria such as the Akaike Information Criterion (AIC) or the Minimum Description Length (MDL) criterion. However, these criteria are proved unsuccessful for sea echo signals both in simulation and in actual application, since sea surface is essentially a continuum. Therefore, an ad hoc method of detection of number of signals, a bit different from the existent methods, is developed for the current algorithm of OSMAR2000. The underlying idea is to determine the number of signals based on the variance of MUSIC spectra structure under different candidate number of signals