Computer Vision Navigation Based on Fiducial Markers of Opportunity

M.A. Lakhani, J. Nielsen, G. Lachapelle

Abstract:	Visual Fiducial Markers (FM) have been extensively used for localizing in both indoor and outdoor environments. Such markers are typically comprised of known patterns and mounted in locations known to the handheld navigation device (HND). The Computer Vision (CV) algorithm applied to the camera output of the HND isolates, characterizes and identifies the FM’s. In this paper, the application of the FM is extended and generalized such that arbitrary FM’s of opportunity in unknown locations can be used for navigation. The FM’s may consist of simple patterns known only to be a rectangle or other geometric shape of generally unknown dimensions. The other necessary but rather benign assumption is that the FM is stationary. Theoretical and experimental results will be given in this paper that will demonstrate that any known or assumed attribute of an observed FM can provide information of practical significance in the context of the navigation objective. That is, the CV processing of the observed FM’s will result in constraints that can be directly applied to the navigation estimate resulting in lower uncertainty of the location of the HND. Furthermore, as will be shown, the CV processing required can be accomplished in real time with a modest processor. Index Terms— Computer Vision (CV), fiducial, Degrees of Freedom (DOF), Simultaneous Localization and Mapping (SLAM), handheld navigation device (HND), Field of View (FOV), feature points (fp).
Published in:	Proceedings of the ION 2013 Pacific PNT Meeting April 23 - 25, 2013 Marriott Waikiki Beach Resort & Spa Honolulu, Hawaii
Pages:	352 - 361
Cite this article:	Lakhani, M.A., Nielsen, J., Lachapelle, G., "Computer Vision Navigation Based on Fiducial Markers of Opportunity," Proceedings of the ION 2013 Pacific PNT Meeting, Honolulu, Hawaii, April 2013, pp. 352-361.
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In