Ellipsoidal Cube Maps for Accurate Rendering of Planetary-Scale Terrain Data

Abstract

Advances in sensor technology lead to a rapidly growing number of terrain data sets with very high spatial resolution. To allow reliable visual analysis of this data, terrain data for planetary objects needs to be rendered with accurate reproduction of every detail. This combination of very large scale and very fine detail is challenging for multiple reasons: the numerical accuracy of typical data types is not sufficient, simple spherical planet models fail to accurately represent the data, and distortions in map projections used for data storage lead to sampling problems. In this paper, we propose the Ellipsoidal Cube Map model to address these problems. We demonstrate the possibilities of the model using a simple renderer implementation that achieves interactive frame rates for a variety of data sets for Earth, Moon, and Mars.