A parametric approach for the geocoding of airborne visible/infrared imaging spectrometer (AVIRIS) data in rugged terrain

Abstract

A geocoding procedure for remotely sensed data of airborne systems in rugged terrain is affected by several factors: buffeting of the aircraft by turbulence, variations in ground speed, changes in altitude, attitude variations, and surface topography. The current investigation was carried from Central Switzerland, using two Airborne Visible/Infrared Imaging Spectrometer (AVIRIS) scenes out of NASA's Multi Aircraft Campaign (MAC) in Europe (1991). The first scene was acquired over flat and hilly terrain and the second over a mountainous area with steep slopes. The parametric approach reconstructs for every pixel the observation geometry based on the flight line, aircraft attitude, and surface topography. Aircraft navigation data, instrument engineering data, information from a conical radar tracking system, and a digital elevation model are used. To prevent changes to the radiometric characteristics, the original value is selected by an improved extraction algorithm, thereby eliminating the need to interpolate the values. The results are visually checked for correspondence at locations with different slope and aspect angles by overlaying scanned forest maps and digitized shorelines of the Swiss Topographical Map. This validation is completed by a quantization of the spatial deviation at selected points and a comparison with an improved nonparametric approach. In general, the results show very good correspondence with the maps. The algorithm reported in this article is a necessary base apply georadiometric correction methods for slope aspect-dependent illumination corrections, atmospheric corrections, and to use AVIRIS data within single-pixel based information systems for land use classification.