Effect of polarization and frequency using GlobeSAR data on vegetation discrimination

Abstract

The objective of this study is using the multipolarization and multifrequence airborne SAR data collected through the GlobeSAR Mission on November 20 and 21, 1993 to study the relationship between the radar backscatter intensities and the vegetation structure. The radar backscatter intensity visualization model (RBIVM) was applied to demonstrate the diversity of the radar signals from the agricultural crops and other economic plants. The effects of polarization and frequency on the radar return of agricultural features are also studied. Results show that the polarization is an important factor, as only VV polarization can distinguish euryale ferox from fish ponds, and radar backscatter of rice fields changed dramatically with the polarization. The RBIVM of vegetation shows that the radar backscatter of the vegetation changes over a wide range. This paper also presents the results of vegetation discrimination, which is a part of the GlobeSAR Land Use Mapping Project. Ten types of land cover were classified. They are water bodies, euryale ferox, harvested rice, rice, rice with wet soil, banana trees, orange trees, grassland, bare soil and sugar cane.