Modeling and Correcting Spatial Non-uniformity of the Apex Pushbroom Imaging Spectrometer

The Airborne Prism Experiment (APEX) is currently being built as a demonstrator and calibrator for potential future European spaceborne imaging spectrometers (Itten et al., 1997, Schaepman et al., 2003). It allows the retrieval of traditional and new products for vegetation, geological, limnological and atmospheric sciences. The uniformity of pushbroom imaging spectrometers is affected by spectral and spatial misregistration within each detector array as well as by coregistration errors between the VNIR and the SWIR spectral range (Mouroulis et al., 2000). Correction of such persistent effects may only be done by dedicated interpolation methods. However, any interpolation produces spectra which may differ from real uniform measurements. The impact of interpolation techniques on the uniformity is tested by evaluating their impact on the radiometric measurements. It has to be evaluated, which correction strategies can increase the overall uniformity of the system while maintaining the radiometric accuracy. An extensive library of measured surface reflectance spectra as well as samples of AVIRIS imaging spectrometer data are used for these analyses. The original spectra are systematically convolved to distorted positions, wherefrom correction methods are applied to retrieve the original values. The radiometric residual error can be quantified taking the expected APEX pixel misregistration error into account in the sensor model. Potential improvements can then be modeled by interpolations to uniform positions.