Succesive estimation of the asteroid shape and probe motion using sequential images

Abstract

In order for a probe to safely land on an asteroid, it is necessary to select the landing point and navigate the probe based on information of the asteroid shape and the probe motion. The procedures for extracting such information from images taken by the probe is called global mapping. In the Hayabusa project of JAXA, the global mapping was conducted with much manual manipulation. As the manual global mapping is too costly, automating the procedures is strongly desired in the future projects. In this paper, we propose a new approach for the automation of global mapping by using the technique of image recognition and robotics. Throughout the experiment using the dataset provided by JAXA, we checked the validity of the method which we proposed.