作者: Jonathan Kelly , Gaurav S. Sukhatme
DOI: 10.3182/20070903-3-FR-2921.00036
关键词:
摘要: Abstract Unmanned aerial vehicles normally rely on GPS to provide pose information for navigation. In this work, we examine stereo visual odometry (SVO) as an alternative estimation method situations in which unavailable. SVO is incremental procedure that determines ego-motion by identifying and tracking landmarks the environment, using cameras mounted on-board vehicle. We present experiments demonstrating how performance varies with camera pointing angle, a robotic helicopter platform. Our results show oblique angle produces better motion estimates than nadir view reliable navigation over distances of more 200 meters possible alone.