Inference of a generalised texture for a compound – Gaussian clutter
In the context of a stochastic framework based on Jakeman's random walk model, Field and Tough demonstrated how the radar cross-section could be inferred from the intensity-weighted fluctuations of the phase (coherent data). With regard to the compound representation of the scattered amplitude, this property holds for an arbitrary texture. Extending previous work pertaining to the more specific K-distributed case (where the texture is Gamma distributed), the authors discuss the error arising during this inference process for a broader range of texture distributions. For three different texture models the authors then derive a condition, on the number of samples over which the phase fluctuations should be averaged, to optimise the extraction of the cross-section. Simulated data assert the viability of their findings. The practical implications of this technique for radar clutters are then discussed.
