A unified analysis of radar interferometry and polarimetry for the estimation of forest parameters

Radar interferometry is primarily sensitive to the spatial distribution of the scatterers constituting vegetated terrain, while polarimetry is primarily sensitive to their shape and orientation. For forests, there are obvious qualitative relationships between the spatial distribution and orientation of certain typical features. Ground surfaces, which are spatially localized at the bottom of a forest are also horizontally oriented. Leaf-branch volumes, which are frequently concentrated more at the middle and top of the forest, are more randomly oriented. Different admixtures of ground, ground-volume, and direct volume returns induce distinctive signatures in both interferometric and polarimetric data. This paper explores the potential for unifying interferometric and polarimetric data by simultaneous analysis in order to estimate vegetation and surface characteristics. Estimation of parameters from a combined data set, unified with a physical scattering model, has the potential for being more accurate than estimations from interferometry or polarimetry independently. In this paper, “interferometry” will mean the acquisition and cross-correlation of complex signals at two different ends of a baseline, but with the same receive and transmit polarization at each end