LAUSR

A novel method for the estimation of a forest vertical structure using Polarimetric Synthetic Aperture Radar (PolSAR) data only without the need to interferometry data is proposed in the present paper. Electromagnetic (EM) simulation is used to develop the proposed method, where the SAR pulse is simulated as a plane wave incident in the direction of the side looking angle of the SAR. For this purpose, the forest canopy layer is modeled as clouds of randomly oriented thin straight dipoles which are randomly distributed within an inclined prism volume, whereas the forest soil surface is modeled as a random rough surface. The proposed method aims to estimate the average height of the canopy layer above the soil surface, the canopy layer thickness, and the roughness of the forest ground surface. The proposed method is based on the Radar Vegetation Index (RVI) and the normalized Radar Cross Section (RCS) calculated from the PolSAR data and their relevance to the parameters of the forest vertical structure. Some examples are presented to demonstrate the capability of the proposed method using some PolSAR images obtained through EM simulation of the scattering from forest regions and by applying the theorem of SAR target composition with the Multiple Component Scattering Model (MCSM). The phase differences between the components of scattering obtained from the solution of the SAR target decomposition problem are used in the estimation process. The accuracy of the proposed method is assessed by calculating the percentage error of the estimated vertical structure and ground roughness for each resolution cell of the simulated forest region. It is shown that the percentage errors of the estimated parameters are very low, which reflects the accuracy and efficiency of the proposed method.