3D visualization of aurora from optional viewpoint at optional time

Abstract

Three-dimensional analysis of the aurora is significant because the shape of aurora depends on solar wind which influences electric equipment such as satellites. Our research group set two fish-eye cameras in Alaska, U.S.A and reconstructed the Aurora's shape from a pair of stereo images [Fujii et al. 2014]. However, the method using the feature-based matching cannot detect dense enough feature points accurately since they are hard to detect from the aurora image whose most parts are with low contrast. In this paper, we achieved both increasing the detected feature points and improving accuracy. Applying this method, the 3D shape of aurora from optional view point at optional time can be visualized.