A Vision System for Landing an Unmanned Aerial Vehicle

W e present the design ami implementation of a real-time computer vision syslem for a rotor-craft unmanned aerial vehicle to land onto a known landing target. This vision system consists of customized software and off-the-shelf harc!ware which perform image processing, segmentation, feature point extraction, camera pan/tilt control, and mot ion estimation. W e introduce the design of a landing target which significantly simplifies the computer vision tasks such as corner detection and correspondence matching. Oustomized algorithms are deveiloped to allow f o r realt ime computation at a frame rate of 3OHz. Such algori thms include certain linear and nonlinear optimizat ion schemes f o r model-based camera pose estimation. W e present results f r o m a n actual flight test which show the vision-based state estimates are accurate to within 5 c m in each axis of t.ranslation and 5 degrees in each axis of rotation, making vision a viable sensor to be placed in the control loop of a hierarchical flight management system.