This paper demonstrates several improvements on implementing a fast monocular visual slam system (MonoSLAM) to navigate indoor aerial vehicle. These improvements include designing for the framework of the navigation system, redesigning of landmark patch matching method and giving new rules updating the image patch to overcome some particularly bad situation when landmarks are insufficient. We demonstrate the usefulness of these improvement in the SLAM system when guiding for a quadrotor in our lab. With those improvements, the SLAM system is capable not only for indoor navigation but also some outdoor environment with stable illumination. Several experiments are made to vindicate the effectiveness of the system.
Indoor navigation for aerial vehicle using monocular visual SLAM
01.08.2014
313309 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Autonomous aerial navigation using monocular visual‐inertial fusion
British Library Online Contents | 2018
|Monocular Trail Detection and Tracking Aided by Visual SLAM for Small Unmanned Aerial Vehicles
ArXiv | 2018
|