Improving the classification of high relief RADARSAT imagery through the fusion of ascending and descending look angle images
Date
2001
Authors
Sykes, Richard John
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
The interpretation of Synthetic Aperture RADAR (SAR) imagery is impacted by viewing direction and the visual effects of foreshortening, shadowing, and layover. In areas of high relief, these effects of topography are amplified to the extent that classifications of opposite side imagery of the same area can vary significantly. Traditionally, to compensate for these distortions, digital elevation models (DEMs) have been employed to either ortho correct the imagery, or to create synthetic SAR images.
This paper explores a data fusion approach to SAR interpretation. Opposite side (ascending/descending) RADARSAT imagery are fused to create a single image combining the interpretable features contained in both input images. Each member of the image pair is geocorrected using British Columbia Terrain Resource Information Management (TRIM) planimetric data, fused, and the resultant images then compared to the originals using TRIM DEM data.