The fiducial markers for AR applications provide robust tracking capabilities. However, it is more desirable to use natural features in the scene to estimate 6 DoF pose of the camera. The Studierstube Team at Graz University of Technology have demonstrated robust Natural Feature Tracking (NFT) for known planar surfaces. They use modified versions (for hand held devices) of SIFT and FERNS to compute and match features in the scene.
More details can be found in the following paper:
Daniel Wagner, Gerhard Reitmayr, Alessandro Mulloni, Tom Drummond and Dieter Schmalstieg, "Pose Tracking from Natural Features on Mobile Phones", ISMAR 2008
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment