Satellite Hyperspectral Imaging Simulation

Abstract

Simulation of generic pushbroom satellite hyperspectral sensors have been performed to evaluate the potential performance and validation techniques for satellite systems such as COIS(NEMO), Warfighter-1(OrbView-4) and Hyperion(EO-1). The simulations start with a generation of synthetic scenes from material maps of studied terrain. Scene-reflected radiance is corrected for atmospheric effects and convolved with sensor spectral response using MODTRAN 4 radiance and transmissions calculations. Scene images are further convolved with point spread functions derived from Optical Transfer Functions (OTF's) of the sensor system. Photon noise and etectorr/electronics noise are added to the simulated images, which are also finally quantized to the sensor bit resolution. Studied scenes include bridges and straight roads used for evaluation of sensor spatial resolution, as well as fields of minerals, vegetation and manmade materials used for evaluation of sensor radiometric response and sensitivity. The scenes are simulated with various seasons and weather conditions. Signal-to-noise ratios and expected performance are estimated for typical satellite system specifications and are discussed for all the scenes.