Geometric Restoration of Satellite Image Data.
Abstract
We describe an approach for projecting satellite image data onto arbitrary map coordinates. This transformation removes distortions caused by the basic geometry of cross-track scanners and by spacecraft pitch, roll, and yaw. Geometrically restored images are useful for morphological analyses and correlative studies involving concurrent data from multiple sensors. It is possible, for example, to present AIRS passes from distinct receivers, all-sky camera images, radar data, and other relevant spatial measurements on one map in geographic coordinates. Out approach is rigorous in that arbitrary satellite motions are accounted for and exact principles of geometry are used throughout. The floating-point intensive geometry calculations are performed on an arbitrary (adjustable) reference grid, which can be coarse for quick-look purposes, or fine for highest accuracy. Concurrent images for distinct wavelengths can be efficiently projected without redundant geometry calculations. An IBM PC/AT compatible implementation for Polar BEAR AIRS image data is described. Results are presented for two data sets representing moderate and severe platform motions.
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 15, 1987
- Accession Number
- ADA190462
Entities
People
- R. D. Lucas
- Robert E. Robins
Organizations
- Northwest Research Associates