抄録
In the JAROS/NASA JERS-1 Simulation Project, the NASA/JPL DC-8 SAR (AIRSAR) was flown over six test sites in the U.S.A. The AIRSAR operates at three frequencies (L, C, P-band) and four polarization realizations (HH, HV, VH, VV) and also works as a polarimetric SAR.
This paper describes feasibility of vegetation analysis with SAR data derived from AIRSAR data analysis. Ten targets with vegetation and without vegetation cover are selected in the calibrated data, Then radar backscatter intensity at three frequencies and like and cross polarizations and polarimetry of these targets are investigated.
Radar backcatter at P and L-band is much affected by vegetation. Vegetation indicates a smaller difference of radar backscatter intensity between like and cross polarizations. The difference at C-band is small for all targets with greater than 20-percent vegetation cover. The differences at P and L-band decrease with an increase in vegetation density and height. Using these characteristics, vegetation volume classification was made for three test sites. Polarimetric signatures give us an additional information about scatterers. They allow us to discriminate corn from short grass, which was impossible with only like and cross polarizations of P, L, C-band. Multifrequency polarimetric SAR can be a more useful tool for vegetation analysis than single or multi polarization and multifrequency SAR.
