An integrated approach for segmentation and representation of range images

  • Olga R. P. Bellon
  • Clesio L. Tozzi
Poster Session A: Color & Texture, Enhancement, Image Analysis & Pattern Recognition, Segmentation
Part of the Lecture Notes in Computer Science book series (LNCS, volume 1310)


This paper presents an integrated approach for segmentation and representation of objects in range images. The segmentation is based on the association of edge detection with clustering techniques, and it produces a set of labeled regions plus the coefficients of the plane fitted to each region. With this information, the reconstruction error for the whole image is estimated, and the initial number of regions supplied to the clustering algorithm can be updated, based on this error. Once the image reconstruction error is smaller than the desired, the next step is to create a polyhedral representation to the surfaces of the image. For each segmented region, the representation process yields its 3D vertices' coordinates, ordered to form a polygon. The final representation is robust enough to guarantee that there are no cracks on the reconstructed surfaces. The main contributions of this work are: (1) A new approach for determining the optimal number of image regions; (2) A suitable representation that can be applied to both polyhedral and non-polyhedral objects.


  1. [1]
    P. Besl and R. Jain, 1988. “Segmentation through variable-order surface fitting”. IEEE T-PAMI, Vol. 10, No.2, pp. 167–192.Google Scholar
  2. [2]
    S. Bhandarkar and A. Siebert, 1992. “Integrating edge and surface information for range image segmentation”. Pattern Recognition, Vo1.25, No.9, pp.947–962.Google Scholar
  3. [3]
    J. Chen, S. Castan, 1986. “An optimal linear operator for edge detection”. Proc. of CVPR'86, Miami.Google Scholar
  4. [4]
    R. Dubes and R.C. Jain, 1976. “Clustering techniques: the user's dilemma”. Pattern Recognition Letters, Vol.8, pp.247–260.Google Scholar
  5. [5]
    O.D. Faugeras and M. Herbert, 1987. “The representation, recognition, and positioning of 3D shapes from range data”. Techniques for 3D Machine Perception, Ed. North Holland, Netherlands, pp. 13–52.Google Scholar
  6. [6]
    R.C. Gonzalez, R.E. Woods, 1992. Digital Image Processing. Addison-Wesley.Google Scholar
  7. [7]
    J.F. Haddon, 1988. “Generalized threshold selection for edge detection”, Pattern Recognition, Vol.3, pp. 195–203.Google Scholar
  8. [8]
    R. Hoffman and A. Jain, 1987. “Segmentation and classification of range images”. IEEE T-PAMI, Vo1.9, No.5, pp.608–620.Google Scholar
  9. [9]
    R. Krishnapuram and A Munshi, 1991. “Cluster-based segmentation of range images using differential-geometric features”. Optical Engineering, Vo1.30, No.10, pp.1468–1478.Google Scholar
  10. [10]
    B. Noble and J.W. Daniel, 1988. Applied Linear Algebra. Prentice-Hall Int.Google Scholar
  11. [11]
    F.P. Preparata and M.I. Shamos, 1985. Computational Geometry: An Introduction. Springer-Verlag, NY-USA, 1985.Google Scholar
  12. [12]
    N. Yokoya and M. Levine, 1989. “Range image segmentation based on differential geometry: a hybrid approach”. IEEE T-PAMi, Vol.11, No.5, pp.643–649.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1997

Authors and Affiliations

  • Olga R. P. Bellon
    • 1
  • Clesio L. Tozzi
    • 2
  1. 1.UFPR-Depto. de InformáticaCuritiba-PRBrazil
  2. 2.UNICAMP-FEEC-DCACampinas-SPBrazil

Personalised recommendations