Classification of grassland types by means of multi-seasonal TerraSAR-X and RADARSAT-2 imagery

Abstract

The management and protection of grassland biodiversity is of utmost importance as they play a key role in the carbon and hydrological cycle. Therefore, the analysis of their dynamics is of great value given the current ongoing intensification of agricultural land use. To this aim, in this paper we present a novel approach for monitoring grassland dynamics based on polarimetric high-resolution SAR imagery, which is i) capable of handling either dual- or quad-polarization multi-temporal data and ii) supports targeted classification. Results based on dualpol TerraSAR-X as well as dual- and quadpol Radarsat-2 data acquired over a test area in Bavaria (Germany) in 2011 are extremely promising and confirm the effectiveness of the proposed approach. © 2014 IEEE.