Learn More
— Real-time approach for monocular visual simultaneous localization and mapping (SLAM) within a large-scale environment is proposed. From a monocular video sequence, the proposed method continuously computes the current 6-DOF camera pose and 3D landmarks position. The proposed method successfully builds consistent maps from challenging outdoor sequences(More)
CONTRIBUTION We estimate precise 6-dof camera pose from monoc-ular video given a SfM scene reconstruction computed offline. Our key contributions are: • A new framework for accurate video-rate local-ization that consistently maintains low latency. • An efficient 2d-3d direct feature matching approach interleaved with 2d keypoint tracking. Our method scales(More)
— This paper proposes a vision-based obstacle avoidance strategy in a dynamic environment for a fixed-wing unmanned aerial vehicle (UAV). In order to apply a nonlinear model predictive control (NMPC) framework to image-based visual servoing (IBVS), a dynamic model from UAV control input to image features is derived. From this dynamics, a visual(More)
This paper considers onboard control of a small-sized quadrotor using a strapdown embedded optical flow sensor which is conventionally used for desktop mice. The vehicle considered in this paper can carry only few dozen grams of payload, therefore conventional camera-based optical flow methods are not applicable. We present hovering control of the(More)