REDUCING ACCUMULATED ERRORS IN EGO-MOTION ESTIMATION USING LOCAL BUNDLE ADJUSTMENT

Akihiro Sugimoto, Tomohiko Ikeda

Abstract

Incremental motion estimation methods involve a problem that estimation accuracy gradually becomes worse as the motion trajectory becomes longer and longer. This is due to accumulation of estimation errors incurred in each estimation step. To keep estimation accuracy stable even for a long trajectory, we propose to locally apply the bundle adjustment to each estimated motion so that the modified estimation becomes geometrically consistent with time-series frames acquired so far. To demonstrate the effectiveness of this approach, we employ an ego-motion estimation method using the binocular fixation control, and show that (i) our modification of estimation is statistically significant; (ii) in order to reduce estimation errors most effectively, three frames are optimal for applying the bundle adjustment; (iii) the proposed method is effective in the real situation, demonstrating drastic improvement of accuracy in estimation for a long motion trajectory.

References

  1. H. Aoki, B. Schiele and A. Pentland (2000): Realtime Personal Positioning System for Wearable Computers, Vision and Modeling Technical Report, TR-520, Media Lab. MIT.
  2. B. Clarkson, K. Mase and A. Pentland (2000): Recognizing User's Context from Wearable Sensors: Baseline System, Vision and Modeling Technical Report, Vismod TR-519, Media Lab. MIT.
  3. A. J. Davison, W. W. Mayol and D. W. Murray (2003): Real-Time Localization and Mapping with Wearable Active Vision, Proc. of ISMAR, pp. 18-27.
  4. A. J. Davison and D. W. Murray (1998): Mobile Robot Localization using Active Vision, Proc. of ECCV, Vol. 2, pp. 809-825.
  5. G. N. DeSouza and A. C. Kak (2002): Vision for Mobile Robot Navigation: A Survey, IEEE Trans. on PAMI, Vol. 24, No. 2, pp. 237-267.
  6. M. W. M. G. Dissanayake, P. Newman, S. Clark, H. F. Durrant-Whyte and M. Csorba (2001): A Solution to the Simultaneous Localization and Map Building (SLAM) Problem, IEEE Trans. on RA, Vol. 17, No. 3, pp. 229-241.
  7. N. Gonc¸alves and H. Araújo (2002): Estimation of 3D Motion from Stereo Images, Proc. of ICPR, Vol. I, pp. 335-338.
  8. J. E. Guivant and E. Nebot (2001): Optimization of the Simultaneous Localization and Map-Building Algorithm for Real-Time Implementation, IEEE Trans. on RA, Vol. 17, No. 3, pp. 242-257.
  9. R. Hartley and A. Zisserman (2000): Multiple View Geometry in Computer Vision, Cambridge Univ. Press.
  10. B. D. Lucas and T. Kanade (1981): An Iterative Image Registration Technique with an Application to Stereo Vision, Proc. of IJCAI, pp. 674-679.
  11. N. Molton and M. Brady (2003): Practical Structure and Motion from Stereo When motion is Unconstrained, Int. J. of Computer Vision, Vol. 39, No. 1, pp. 5-23.
  12. T. Nakagawa, T. Okatani and K. Deguchi (2004): Active Construction of 3D Map in Robot by Combining Motion and Perceived Images, Proc. of ACCV, vol. 1, pp. 563-568.
  13. A. Sugimoto, W. Nagatomo and T. Matsuyama (2004): Estimating Ego Motion by Fixation Control of Mounted Active Cameras, Proc. of ACCV, pp. 67-72.
  14. A. Sugimoto and T. Ikeda (2004): Diverging Viewing-Lines in Binocular Vision: A Method for Estimating Ego Motion by Mounted Active Cameras, Proc. of the 5th Workshop on Omnidirectional Vision, Camera Networks and Non-classical Cameras, pp. 67-78.
  15. Y. Sumi, Y. Ishiyama and F. Tomita (2004): 3D Localization of Moving Free-Form Objects in Cluttered Environments, Proc. of ACCV, vol. 1, pp. 43-48.
  16. S. Thrun (2002): Robotic Mapping: A Survey, Exploring Artificial Intelligence in the New Millennium, Morgan Kaufmann.
  17. C. Tomasi and T. Kanade (1991): Detection and Tracking of Point Features, CMU Technical Report, CMU-CS91-132.
  18. B. Triggs, P. McLauchlan, R. Hartley and A. Fitzgibbon (2000): Bundle Adjustment - A Modern Synthesis, Vision Algorithms: Theory and Practice (B. Triggs, A. Zisserman and R. Szeliski eds.) LNCS, Vol. 1883, pp. 298-372, Springer.
  19. B. L. Welch (1938): The Significance of the Difference between Two Means when the Population Variances are Unequal, Biometrika, Vol. 29, pp. 350-362.
  20. M. Werman, S. Banerjee, S. Dutta Roy and M. Qiu (1999): Robot Localization Using Uncalibrated Camera Invariants, Proc. of CVPR, Vol. II, pp. 353-359.
  21. Z. Zhang and Y. Shan (2001): Incremental Motion Estimation through Local Bundle Adjustment, Technical Report MSR-TR-01-54, Microsoft Research.
  22. Z. Zhang and Y. Shan (2003): Incremental Motion Estimation through Modified Bundle Adjustment, Proc. ICIP, Vol.II, pp.343-346.
Download


Paper Citation


in Harvard Style

Sugimoto A. and Ikeda T. (2006). REDUCING ACCUMULATED ERRORS IN EGO-MOTION ESTIMATION USING LOCAL BUNDLE ADJUSTMENT . In Proceedings of the Third International Conference on Informatics in Control, Automation and Robotics - Volume 2: ICINCO, ISBN 978-972-8865-60-3, pages 418-425. DOI: 10.5220/0001213904180425


in Bibtex Style

@conference{icinco06,
author={Akihiro Sugimoto and Tomohiko Ikeda},
title={REDUCING ACCUMULATED ERRORS IN EGO-MOTION ESTIMATION USING LOCAL BUNDLE ADJUSTMENT},
booktitle={Proceedings of the Third International Conference on Informatics in Control, Automation and Robotics - Volume 2: ICINCO,},
year={2006},
pages={418-425},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0001213904180425},
isbn={978-972-8865-60-3},
}


in EndNote Style

TY - CONF
JO - Proceedings of the Third International Conference on Informatics in Control, Automation and Robotics - Volume 2: ICINCO,
TI - REDUCING ACCUMULATED ERRORS IN EGO-MOTION ESTIMATION USING LOCAL BUNDLE ADJUSTMENT
SN - 978-972-8865-60-3
AU - Sugimoto A.
AU - Ikeda T.
PY - 2006
SP - 418
EP - 425
DO - 10.5220/0001213904180425