Differencing Phases of Volume and Double Scattering Components to Improve Tree Height Estimate

Abstract

A polarimetric interferometric synthetic aperture radar (PolInSAR) approach can estimate tree heights, but underestimation happens. There is a need to mitigate the underestimating. With the Freeman-Durden polarimetric SAR (PolSAR) decomposition algorithm, one identifies the volume, single, and double scattering components from PolSAR data. Here, the PolSAR decomposition and PolInSAR techniques are combined to estimate a tree height. The decomposed volume scattering phase center replaces the canopy phase center and the double scattering phase center the ground phase center. The tree height is then calculated using the phase difference of the two phase centers and effective vertical wavenumber. Using simulated PolSAR datasets, the algorithm combining the decomposition and PolInSAR approaches output tree heights with an average of 8.6 m and one standard deviation of 2.1 m. The mean and one standard deviation of tree heights are 5.1 m and 2.4 m after the PolInSAR technique. Thus, an improvement of 3.5 m in the mean height is achieved.