Regresar

Landmark-based virtual path estimation for assisted UAV FPV tele-operation with augmented reality

Abstract:

In this paper we proposed an Assisted UAV Tele-Operation System, specifically for FPV navigation based on Artificial Landmarks in obstacle free environments. The system estimates the optimal path through landmarks and traces an artificial route to be followed. Path recognition uses color space and morphological transformation such as eroding and dilating to reduce noise due to different lighting environments. Once path is recognized ORB detector is used for getting a set of the most representative pixels coordinates, this is done for each ROI (Region of Interest) in the camera image. Later, the median of each pixel coordinate in the specific ROI is considered for interpolation needed to trace the route. Parrot’s drone Bebop 2 was used for the purpose of this study as it has a fisheye lens camera that allows us to face downwards to detect the landmarks.