Synthetic-Aperture Radar Imaging of Forests in the VHF and UHF Bands: Electromagnetic Models and Data Analysis
Doktorsavhandling, 2007
Synthetic-aperture radar (SAR) is a technique which can image the Earth’s surface, from a spaceborne or airborne platform, irrespective of daylight or cloud cover. During the last decades there has been an increasing interest in using SAR for estimation of forest parameters. It has been shown that the radar image-intensity over forested areas initially increases with increasing biomass up to a certain point referred to as the saturation point. The latter is dependent on forest type but also the frequency of the SAR system and the corresponding biomass level is higher for lower frequencies compared to higher frequencies. Spaceborne systems are the only realistic choice for global forest monitoring but do not currently operate below 1 GHz and saturate at relative low biomasses. Therefore, many studies have addressed the possibility of using frequencies below 1 GHz for future biomass missions.
In this thesis, some of the problems and possibilities of using a SAR for forest imaging in the VHF (30-300 MHz) and UHF (300-1000 MHz) bands are studied through the use of electromagnetic models and by observations made by airborne SAR systems which today operate in these frequency bands. A method to co-register and combine multiple VHF SAR images in order to improve spatial and radiometric resolution is proposed and it is found that it is possible to enhance resolution to a degree where measurements on individual trees are possible. By studying these high resolution SAR images it is also found that previously existing models of forest scattering in VHF and UHF bands sometimes must be extended in order to correctly account for the effects of ground topography. A new model based on physical optics and with a generalized ground-parameterization is proposed and validated against observations done with airborne SAR systems. Furthermore, a model for dependency of stem number density on P-band (440 MHz) scattering is proposed and compared with observations. Finally, some problems associated with a P-band SAR satellite for forest monitoring, e.g. degradation of spatial resolution due to ionospheric scintillations and polarization-channel cross-talk due to Faraday rotation, are assessed by model simulations. It is found that these problems are likely to be surmountable, at least for equatorial regions.
image formation
biomass mapping
Synthetic-aperture radar (SAR)
ionospheric propagation
electromagnetic models
forest monitoring
remote sensing
EB-salen, Hörsalsvägen 11, Chalmers, Göteborg.
Opponent: Professor Mahata Moghaddam, The Electical Engineering and Computer Science (EECS) Department, University of Michigan, USA