Using Multilook Averaging for Coherently Modeled Scattering From a Pine Tree at L-Band

In this letter, we study scattering from a scots pine tree by using a realistic cylinder model of the tree and the coherent electric field scattering model. We study the directional variability of bistatic and monostatic scattering at L-band and show that, due to large variations in results, some averaging technique should be used to describe and interpret the model output efficiently. We propose that the needed averaging can be done by multilooking and that multilook data could be easily generated by rotating the tree model randomly around its vertical axis. We show that the resulting scattering data obey generally the multidimensional Gaussian distribution or the more general K-distribution, in a way similar to synthetic aperture radar (SAR) image pixels, and can therefore be represented by a single averaged covariance matrix. The trunk-ground reflection to the backscattering direction adds to the model output non-Gaussian behavior, which can be treated as texture. Covariance matrix formalism allows us to use descriptors which are commonly used to analyze SAR images, like target entropy and alpha angle. The method helps interpretation and comparison between the model output and SAR image.