A Physical-optics Model for Predicting VHF/UHF-band HH-pol Backscatter from Coniferous Forests in Undulating Terrain
Conference contribution, 2007
A physical optics based model for predicting backscatter from coniferous forest in the VHF- and UHF frequency bands is presented. It models the forest as a collection of stems standing on a well characterized undulating ground surface. The model is valid for SAR systems transmitting and receiving horizontal polarization (HH) and is useful for assessing the importance of ground topography in SAR images.
SAR data from the airborne systems CARABAS-II (22-82 MHz) and LORA (230-470 MHz) and forest characterization on individual tree level were used to validate the model. It was found that the model was able to predict much of the backscatter variations due to surface topography. At best, coefficients of determination (R2) of 0.44 and 0.65 for the CARABAS and LORA case were found. In comparison it should be noted that R2 drops to about 0.1 if a flat ground is assumed. These results indicate that it is essential to consider ground topography when predicting HH-polarized SAR images in these frequency bands.
The model failed to predict the absolute values of the backscatter intensity. Probable explanations are the lack of attenuation effects in the model and/or the value chosen for the dielectric constant of the stems.
In the near future efforts will be placed into the task of extending the model to include more polarizations, especially HV. This includes e.g. adding the backscatter from branches and leaves/needles, accounting for attenuation effects and including random small scale variations in the surface description.
Synthetic aperture radar