An efficient subband decomposition based on the Hilbert transform for high-resolution spectral estimation

This paper deals with high-resolution frequency estimation for narrow-band plane waves. We propose an approach based on subband decomposition in the spectral domain to improve the performance of high-resolution analysis. This decomposition is based on the Hilbert transform for one and two-dimensional signals. This improvement is tested on ESPRIT and MEMP techniques. We first present the subband decomposition based on the Hilbert transform (SDBHT) for one-dimensional (1D) signals. Secondly the SDBHT method is extended to the two-dimensional (2D) case. Finally the advantages of such a method are illustrated with simulation examples.