Tracking of industrial objects by using CAD models
H. Wuest, and D. Stricker. Journal of Virtual Reality and Broadcasting, (August 2007)ISSN 1860-2037.
Abstract
In this paper we present a model-based approach for real-time camera pose estimation in industrial scenarios. The line model which is used for tracking is generated by rendering a polygonal model and extracting contours out of the rendered scene. By un-projecting a point on the contour with the depth value stored in the z-buffer, the 3D coordinates of the contour can be calculated. For establishing 2D/3D correspondences the 3D control points on the contour are projected into the image and a perpendicular search for gradient maxima for every point on the contour is performed. Multiple hypotheses of 2D image points corresponding to a 3D control point make the pose estimation robust against ambiguous edges in the image.
%0 Journal Article
%1 ws07
%A Wuest, Harald
%A Stricker, Didier
%D 2007
%E Herder, Jens
%J Journal of Virtual Reality and Broadcasting
%K 4(2007)1 4.2007 AR Augmented_Reality CAD_Model DiPP Digital_Peer_Publishing_Initiative GI2006 JVRB Journal_of_Virtual_Reality_and_Broadcasting Line_Tracking Open_Access Peer_Reviewed Tracking [WS07]
%N 1
%T Tracking of industrial objects by using CAD models
%U urn:nbn:de:0009-6-11595
%V 4
%X In this paper we present a model-based approach for real-time camera pose estimation in industrial scenarios. The line model which is used for tracking is generated by rendering a polygonal model and extracting contours out of the rendered scene. By un-projecting a point on the contour with the depth value stored in the z-buffer, the 3D coordinates of the contour can be calculated. For establishing 2D/3D correspondences the 3D control points on the contour are projected into the image and a perpendicular search for gradient maxima for every point on the contour is performed. Multiple hypotheses of 2D image points corresponding to a 3D control point make the pose estimation robust against ambiguous edges in the image.
@article{ws07,
abstract = {In this paper we present a model-based approach for real-time camera pose estimation in industrial scenarios. The line model which is used for tracking is generated by rendering a polygonal model and extracting contours out of the rendered scene. By un-projecting a point on the contour with the depth value stored in the z-buffer, the 3D coordinates of the contour can be calculated. For establishing 2D/3D correspondences the 3D control points on the contour are projected into the image and a perpendicular search for gradient maxima for every point on the contour is performed. Multiple hypotheses of 2D image points corresponding to a 3D control point make the pose estimation robust against ambiguous edges in the image.},
added-at = {2007-10-24T13:04:49.000+0200},
author = {Wuest, Harald and Stricker, Didier},
biburl = {https://www.bibsonomy.org/bibtex/2b0b766b77b41fbfc32e31a7c3e3e7ff0/jvrb_regulski},
editor = {Herder, Jens},
interhash = {344cc0b6660dbdf058b55b9a37420104},
intrahash = {b0b766b77b41fbfc32e31a7c3e3e7ff0},
journal = {Journal of Virtual Reality and Broadcasting},
keywords = {4(2007)1 4.2007 AR Augmented_Reality CAD_Model DiPP Digital_Peer_Publishing_Initiative GI2006 JVRB Journal_of_Virtual_Reality_and_Broadcasting Line_Tracking Open_Access Peer_Reviewed Tracking [WS07]},
month = aug,
note = {ISSN 1860-2037},
number = 1,
timestamp = {2007-10-24T13:04:49.000+0200},
title = {Tracking of industrial objects by using CAD models},
url = {urn:nbn:de:0009-6-11595},
volume = 4,
year = 2007
}