Visual Navigation for UAVs Landing on Accessory Building Floor

Abstract

Landing is a critical step in most real world UAV applications, especially in delivery. A package delivery is successful only when a landing or a low-altitude drop-off (pseudo-landing) is completed. For precision landing requirement, vision-based navigation techniques are of high potential to be reliable and accurate. In this paper, we present a research work on autonomous visual navigation for precision landing on accessory building floor. We incorporate some state-of-the-art vision-based methods, develop other functional components to present an employable autonomous navigation system for precision landing near buildings. Initial experiments in a real world scenario show an encouraging results with high success rate of performing precision landing.