CDR Grady H. Tuell
National Oceanic and Atmospheric Administration, National Ocean Service, National Geodetic Survey

Analysts working with high-dimensional remotely sensed images (multispectral and hyperspectral) often apply atmospheric correction procedures as part of the overall process of converting sensor radiance data to ground reflectance estimates. This correction facilitates matching pixel signatures against libraries of known reflectance spectra. But the application of these corrections may significantly alter the spectral contrast of adjacent pixels and thereby alter the geometric resolution in any spectral channel. Therefore, channel-by-channel estimation of the geometric performed both before and after the correction provides one possible metric for studying the overall effect on the image cube. In this experiment, an Optical Transfer Function (OTF) based methodology is used to estimate the resolution in several channels of data from an imaging spectrometer. These estimates are then repeated after the application of a simple atmospheric removal procedure. The 2 sets of resolution estimates are examined as a basis for discussing the net effect of the procedure on image resolution.