Dense 3D Reconstruction in Multi-camera Systems

  • Soonmin BaeEmail author
Conference paper
Part of the Lecture Notes in Electrical Engineering book series (LNEE, volume 306)


The 3D information is useful in the security services and media production markets. However, it is difficult to achieve a reliable and dense 3D reconstruction from images. In this paper, we introduce a new approach to generating a dense 3D map from stereo images. We estimate the epipolar geometry and find matches using the geometric model. Given matches, we triangulate the 3D points and propagate the 3D information to neighboring pixels. For propagation, we use segmentation for a plane fitting and non-homogeneous optimization for an edge-preserving smoothing.


3D reconstruction Stereo 


  1. 1.
    Bleyer M, Gelautz M (2005) Graph-based surface reconstruction from stereo pairs using image segmentation. Proc SPIE 5665:288–299Google Scholar
  2. 2.
    Comaniciu D, Meer P (2002) Mean shift: a robust approach toward feature space analysis. IEEE Trans Pattern Anal Mach Intell 24:603–619CrossRefGoogle Scholar
  3. 3.
    Hartley R, Zisserman A (2000) Multiple view geometry in computer vision. Cambridge Press, CambridgezbMATHGoogle Scholar
  4. 4.
    Hong L, Chen G (2004) Segment-based stereo matching using graph cuts. In: IEEE Computer Society conference on computer vision and pattern recognitionGoogle Scholar
  5. 5.
    Levin A et al (2004) Colorization using optimization. ACM Transactions on Graphics, In: Proceedings of ACM SIGGRAPH conference 23(3):689–694Google Scholar
  6. 6.
    Scharstein C, Pal C (2007) Learning conditional random fields for stereo. In: IEEE Computer Society conference on computer vision and pattern recognitionGoogle Scholar
  7. 7.
    Strecha C et al (2003) Dense matching of multiple wide-baseline views. In: IEEE international conference on computer vision Google Scholar

Copyright information

© Springer International Publishing Switzerland 2014

Authors and Affiliations

  1. 1.Intelligent Control Technology Team, Advanced Technology CenterSamsung TechwinSeongnamSouth Korea

Personalised recommendations