Object Identification and Pose Estimation for Automatic Manipulation

  • Benjamin Hohnhaeuser
  • Guenter Hommel
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 1998)


In this paper we present a framew rk to recognize objects and to determine their pose from a set of bjects in a scene for automatic manipulation (bin picking) using pixel-synchronous range and intensity images. The approach uses three-dimensional bject models. The object identification and pose estimation process is structured into three stages. The first stage is the feature collection stage, where the feature detection is performed in an area of interest followed by the hypothesis generation which tries to form hypotheses from consistent features. The last stage, the hypothesis verification, tries to evaluate the hypotheses by comparing the measured range data to the predicted range data from hypothesis and the model.


Intensity Image Range Image Feature Extraction Process Range Information Cluttered Scene 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    R. C. Bolles, P. Horaud: 3dpo:A threedimensional part orientation system. J. of Robotics Research 5 (1986) 3–26.53CrossRefGoogle Scholar
  2. 2.
    O. D. Faugeras, M. Hebert: The representation, recognition and location of 3-d objects. J. of Robotics Research 5 (1986) 27–52.53Google Scholar
  3. 3.
    W. E. L. Grimson: Object Recognition by Computer. MIT Press, (1990). 53Google Scholar
  4. 4.
    K. Hara, H. Zha, T. Hasegawa: Regularization-based 3-d bject modeling from multiple range images. In: IEEE-Proceed. of the Internat. Conf. on Pattern Recog-nition ICPR’ 98, Queensland, Australia, (August 1998). 53Google Scholar
  5. 5.
    A. E. Johnson: Spin-images: a representation for 3-D surface matching. PhD thesis, Carnegie Mellon University, Pittsburgh, Pennsylvania, (August 1997). 53Google Scholar
  6. 6.
    A. C. Kak, J. L. Edwards: Experimental state of the art in 3d bject recognition and localization using range data. In: Proceed. of Workshop on Vision for Robots in IROS’95 Conference, Pittsburgh, Pennsylvania, (1995). 53Google Scholar
  7. 7.
    B. Krebs: Probabilistische Erkennung von 3d Freiformobjekten mit Bayesschen Netzen. PhD thesis, TU Braunschweig, (October 1999). 53Google Scholar
  8. 8.
    Y. Lamdan, H.J. Wolfson: Geometric hashing: a general and efficient model-based recognition scheme. In: IEEE-Proceed. of the Internat. Conf. on Computer Vision Tarpon Springs, Florida, (1988) 238–249.53Google Scholar
  9. 9.
    F. Pipitone, W. Adams: Tripod perators for recognizing bjects in range images; rapid recognition f library objects. In: IEEE-Proceed. of the Internat. Conf. on Robotics and Automation, ICRA92, Nice, France, (1992) 1596–1601.53Google Scholar
  10. 10.
    K. Rahardja, A. Kosaka: Vision-based bin-picking:Recognition and localization of multiple complex bjects using simple visual cues. In: Proceed. of the IEEE/RSJ Internat. Conf. on Intelligent Robotics and Systems, IROS’ 96, Osaka, Japan, (1996). 53Google Scholar
  11. 11.
    W. Schroeder, E. Forgber, G. Roeh: Laser range camera application. Technical report, 1999. 52, 55Google Scholar
  12. 12.
    T. Stahs: Objekterkennung mit einem aktiven 3D-Rob tersens rsystem. PhD thesis, TU Braunschweig, (June 1994). 53Google Scholar
  13. 13.
    H. Zha, S. Tahira, T. Hasegawa: Multi-resolution surface descripion f 3-d objects by shape-adaptive triangular meshes. In: IEEE-Proc. of the Internat. Conf. on Pattern Recognition, ICPR’ 98, Queensland, Australia, (August 1998). 53Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2001

Authors and Affiliations

  • Benjamin Hohnhaeuser
    • 1
  • Guenter Hommel
    • 1
  1. 1.Institut f.Technische InformatikTechnische Universität BerlinBerlinGermany

Personalised recommendations