Multi-resolution Mapping Using Surface, Descent and Orbital Images

Our objective is to produce high-accuracy maps of the terrain elevation at landing sites on planetary bod- ies through the use of all available image data. We use images on the planetary surface from landers and rovers, images captured during the lander descent to the surface, and orbital images. Three new capabil- ities have been developed. First, we generate eleva- tion maps from descent images using structure-from- motion techniques. These maps are useful for rover navigation and provide a link between the orbital im- ages and surface images. We have developed a method- ology for performing rover localization using bundle adjustment that uses tie points between the rover and descent images to determine both the camera and the tie point locations. Finally, a new method to perform registration between orbital images and descent images has been developed that locates the landing position in the orbital imagery and allows integration of the entire data set. These technologies are important for per- forming rover navigation in future space missions and the maps provide a tool for coordinating rovers in a robotic colony.