Electromagnetic scattering from multi-scale rough surfaces

It is shown that the wavelet correlation dimension is a very relevant quantity for the characterization of rough surfaces by remote sensing means. First, the concept of correlation length is generalized to surfaces with wide power spectrum. Second, it is demonstrated that, in the framework of the small-perturbation theory, the wavelet correlation dimension can be retrieved from a knowledge of the backscattered cross section for a discrete set of frequencies. Rigorous numerical experiments confirm these predictions, and in the last section an experimental scheme for a straightforward derivation of the wavelet correlation dimension is proposed.