A spectral analysis algorithm for the estimation of sea SAR image fractal dimension

This paper proposes a spectral analysis algorithm for sea SAR image fractal dimension estimation. The power spectral density of the sea SAR image is obtained through the periodogram approach and by decomposing the spectrum into sub-bands according to a dyadic scale. The fractal dimension of the image is obtained by a linear regression of the absolute maxima of each sub-band. This method can be applied globally or locally to have a single value or a map of the image fractal dimension. Examples on simulated and real data are reported.