Abstract
In this paper, we address the following problem: given a camera moving in an unknown environment, where the camera motion is known, we want to build a 3D map of the environment. The technique we propose is two-stage: (1) bootstrapping: 2D tokens are tracked using an order 1 dynamic model for their evolution; (2) steady-state: 3D fusion of 2D tokens is performed recursively, then the value predicted at time t for these 3D tokens is projected at time t + 1 onto the camera focal plane and replaces the dynamic model used in the bootstrapping mode. This allows for introducing 3D information into the 2D token tracker without prior knowledge of the environment. Uncertainties are taken into account through extended Kalman filtering. 2D and 3D token parameters are chosen such as to simplify the linearization process and to ensure numerical stability.
supported in part by Direction des Recherches, Etudes et Techniques grant # 87331
This is a preview of subscription content, log in via an institution to check access.
Preview
Unable to display preview. Download preview PDF.
References
[Ayache N., Faugeras O.D.], “HYPER: A New Approach for the Recognition and Positioning of Two-Dimensional Objects”, IEEE Trans. PAMI-8, n° 1, Jan. 1986.
[Ayache N. 89], “Vision Stéréoscopique et Perception Multisensorielle - Applications à la Robotique Mobile”, Thèse d’Etat, InterEditions editor, Paris, 1989.
[Bolles R.C., Cain R.A.], “Recognizing and Localizing Partially Visible Objects: the Local Feature Focus Method”, Int. Journal of Robotics Research, 1 (3), pp 57–82, 1982.
[Bolles R.C., Horaud R.], “3DPO: a Three-Dimensional Part Orientation System”, Int. Journal of Robotics Research, 5 (3), pp 3–26, 1986.
[Brooks R.A. -81], “Symbolic Reasoning among 3-D Models and 2-D Images”, Artificial Intelligence Journal, 8, pp 77–98.
[Crowley J.L., Stelmaszyk P., Discours C.], “Measuring Image Flow by Tracking Edge Lines”, Second ICCV,Dec 88.
[Deriche R.], “Using Canny’s Criteria to Derive an Optimal Edge Detector Recursively Implemented”, International Journal of Computer Vision, 1987.
[Grimson W. -88], “Object Recognition from Noisy Data in Cluttered Scenes”, Journées Thomson, Septembre 1988.
Grimson W., Lozano-Pérez T. -87, “Localizing Overlapping Parts by Searching the Interpretation Tree”, IEEE Trans. PAMI, 9 (4), pp 469–482, 1987.
[Herman M., Kanade T.], “The 3D MOSAIC Scene Understanding System: Incremental Reconstruction of 3D Scenes from Complex Images”, Technical Report CMU-CS84–102, Carnegie-Mellon University, Feb. 1984.
[Lamdan Y., Wolfson H.J.], “Geometric Hashing: a General and Efficient Model-Based Recognition Scheme”, New York University, Courant Institute Robotics Report n° 152, May 88.
[Lowe D.G.], “Three-Dimensional Object Recognition from Single Two-Dimensional Images”, Artificial Intelligence 31, 1987, 355–395.
[Maître H., Wu Y.], “Combination of Dynamic Programming and Autoregressive Model for Elastic Matching of Distorted Images”, Proc. 5th Scandinavian Conference on Image Analysis, Stockholm, June 2–5, 1987, 109–116.
[Roberts L.G. -65], “Machine Perception of Three Dimensional Solids”, in Tipett et al. Optical and Electro-Optical Information Processing,MIT Press, Cambridge, Mass.
[Thompson D.W., Mundy J.L.], “Three-Dimensional Model Matching from an Unconstrained Viewpoint”, Proc. Int. Conf. on Robotics and Automation, IEEE Computer Society Press, pp 208–220, 1987.
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 1993 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Jezouin, J.L., Ayache, N. (1993). Three-Dimensional Fusion from a Monocular Sequence of Images. In: Aggarwal, J.K. (eds) Multisensor Fusion for Computer Vision. NATO ASI Series, vol 99. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-662-02957-2_10
Download citation
DOI: https://doi.org/10.1007/978-3-662-02957-2_10
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-08135-4
Online ISBN: 978-3-662-02957-2
eBook Packages: Springer Book Archive