Projection-independent earth-solar-sensor geometry for surface reflectance correction

We present a novel method for calculating Earth-solar-sensor geometry for remote sensing data in an arbitrary map projection. The geometry algorithm uses a recursive bisection and bilinear interpolation procedure to generate geodetic coordinates, accurate astronomical calculation to determine satellite position as a function of time, and a sampling-gridding-interpolation procedure to produce per-pixel view and illumination geometry. The algorithm is implemented using open-source numeric and geospatial components. We also discuss the use of the algorithm in production-scale NBAR calculations for UTM Landsat data.