Google Map Aided Visual Navigation for UAVs in GPS-denied Environment

摘要

We propose a framework for Google Map aided UAV navigation in GPS-deniedenvironment. Geo-referenced navigation provides drift-free localization anddoes not require loop closures. The UAV position is initialized viacorrelation, which is simple and efficient. We then use optical flow to predictits position in subsequent frames. During pose tracking, we obtain inter-frametranslation either by motion field or homography decomposition, and we use HOGfeatures for registration on Google Map. We employ particle filter to conduct acoarse to fine search to localize the UAV. Offline test using aerial imagescollected by our quadrotor platform shows promising results as our approacheliminates the drift in dead-reckoning, and the small localization errorindicates the superiority of our approach as a supplement to GPS.