Combined multi-temporal ERS-1/JERS-1 imagery for land-cover classification

Today all of Earth's land masses are repeatedly mapped with high spatial resolution SAR imagery available from spaceborne platforms (ERS-1/2, JERS-1, RADARSAT). These free-flying sensors are all equipped with single frequency and single polarization instruments (C-vv, L-hh, C-hh). While the capability of "all weather" image acquisition makes these systems very attractive for regional to global scale land-cover classification and monitoring, especially of vegetated, hence cloudy, regions, the lack of multi-frequency/polarization data diminishes the usefulness of either instrument taken alone. This paper describes how vegetation classification results could be improved through various time/sensor combinations using ERS-1 and JERS-1 SAR image data from the Race test site in the Upper Peninsula of Michigan.