Utilization of Coherent Forest Scattering Model in Polarimetric SAR Measurement Interpretation

In this study we investigate pine forest backscattering in L-band by using a coherent field scattering model. We generate multilook data by a novel method for a realistic Scots Pine cylinder model and take a closer look at the probability density functions of the scattering. Multilook data are generated by rotating the tree model randomly around its vertical axis. We show that coherent field scattering model generates speckle and a realistic data distribution similar to real SAR measurement. The PDF is very close to multidimensional Gaussian distribution and, therefore, single averaged covariance matrix of the multilook data describes well the whole ensemble. We propose averaged covariance matrix formalism to be used for study also model output. Usage of covariance matrix formalism allows us to use descriptors like target entropy and alpha, which are commonly used to analyse SAR images, this helps also comparison between the model output and SAR image. We show that entropy and alpha values generated by our method for Scots Pine forest agree well with values measured for real forest with similar age and size.