GeoCARB image navigation and registration performance

Abstract

The geoCARB sensor uses a 4-channel slit-scan infrared imaging spectrometer to measure the absorption spectra of sunlight reflected from the ground in narrow wavelength regions. The instrument, which is to be hosted on a geostationary communication satellite, is designed to provide continual monitoring of greenhouse gas over continental scales, several times per day, with a spatial resolution of a few kilometers. The paper discusses the image navigation and registration (INR) of the geoCARB optical footprints on to the earth’s surface. The instrument acquires data in a step and stare mode with 4.08 s stare time and 0.34s step time on 1016 footprints spaced by 2.7 km at nadir in the NS direction along the slit, which is stepped in 3 km EW increments. Knowledge of the instrument line of sight is obtained through use of a dual-head star tracker system (STS), high-precision optical encoders for the scan mirrors, a GPS receiver, and a highly stable common optical bench to which the instrument components, the scan mirror assembly, and the heads of the STS are kinematically mounted. While attitude disturbances due to jitter and solar array flex affect spatial resolution, we show that the effect on INR is negligible. GeoCARB performs a star sighting every 30 minutes to compensate for its diurnal alignment variation relative to the STS, enabling a 1 sigma INR accuracy of 0.38 and 0.51 km at nadir in the NS and EW directions, respectively. Coastline identification may be used to improve accuracy by 6%, while an additional 20% improvement is achievable through identification of systematic errors via extensive post-processing. The paper quantifies all error sources and describes how each of them affects overall INR accuracy.