No abstract

Ever since the first polarimetric SAR images were generated in the mid-19807s, there has been speculation and research concerning the use of the combined multichannel phase and amplitude data, and the spatial and statistical properties of these data, as an analytical tool for remote sensing of the Earth's surface. This has been especially true for terrestrial ecosystems because differences in the architecture of the vegetation canopy can give rise to different polarimetric scattering properties and spatial variations. The purpose of the study was to access the characteristics of the polarimetric radar clutter collected over various types of vegetation, and to evaluate the potential utility of various measures to classify vegetated terrain. This study utilized airborne radar data collected in October 1988 of pure and mixed stands of jack pine and aspen located in central Minnesota, and data collected in August 1989 of forested areas surrounding Durham, NC. The measured parameters are the polarimetric signature, a statistical distribution shape parameter (mean squared intensity divided by the variance of the intensity), and the texture autocorrelation function. Using a model for clutter that includes Rayleigh fading (speckle) as well as the inherent texture of the clutter, a method is presented for calculating the texture autocorrelation from single-look complex SAR imagery. This method is valid assuming the system impulse response is narrow relative to the peak in the texture autocorrelation. The various parameters from different clutter types were compared at a variety of frequency, polarization, and incidence angle combinations. Based on this comparison, it was found that different terrain types exhibit unique polarimetric, statistical, and spatial properties.

No abstract

No abstract