Statistical characterization of scattering signals from vegetation at frequency above X-band by numerical simulations

Abstract

Since the wave scattering from a vegetated canopy is due to scattering from different vegetation components (leaves, branches, stems, etc.) within the volume enclosed by the transmitting and receiving antenna beams, it is necessary to understand the role played by each component on the total radar return from the whole canopy. To meet the objective of this study, dielectric non-spherical scatters are used to model the vegetation components. Circular disks are used to model deciduous leaves, needles to model coniferous leaves, and finite cylinders to model branches or trunk scatterers. The radar return from disc or needle has been evaluated by applying the Generalized Rayleigh Gans approximation for a disc or needle scatterers. As for the signal return from a finite cylinder, it has been calculated by using the inner field inside similar cylinder of infinite length. Then the scattered field from the canopy is formulated in terms of the scattered field from each forest component through the first-order radiative transfer theory. By means of computer simulation, the samples of signal returns from such vegetation components are obtained and the signal probability density is estimated from these samples. The goodness of fit procedure is then applied to fit the simulated data to known statistical models.