Advertisement

A Generative Model for Online Depth Fusion

  • Oliver J. Woodford
  • George Vogiatzis
Part of the Lecture Notes in Computer Science book series (LNCS, volume 7576)

Abstract

We present a probabilistic, online, depth map fusion framework, whose generative model for the sensor measurement process accurately incorporates both long-range visibility constraints and a spatially varying, probabilistic outlier model. In addition, we propose an inference algorithm that updates the state variables of this model in linear time each frame. Our detailed evaluation compares our approach against several others, demonstrating and explaining the improvements that this model offers, as well as highlighting a problem with all current methods: systemic bias.

Keywords

Markov Random Field Visible Surface Normalize Cross Correlation Fusion Frame Static 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.

References

  1. 1.
    Szeliski, R.: A multi-view approach to motion and stereo. In: Proceedings of CVPR (1999)Google Scholar
  2. 2.
    Kolmogorov, V., Zabih, R.: Computing visual correspondence with occlusions using graph cuts. In: Proceedings of ICCV (2001)Google Scholar
  3. 3.
    Merrell, P., Akbarzadeh, A., Wang, L., Mordohai, P., Frahm, J.M., Yang, R., Nistér, D., Pollefeys, M.: Real-time visibility-based fusion of depth maps. In: Proceedings of ICCV (2007)Google Scholar
  4. 4.
    Seitz, M.S., Dyer, C.R.: Photorealistic scene reconstruction by voxel coloring. In: Proceedings of CVPR (1997)Google Scholar
  5. 5.
    Hernández, C., Vogiatzis, G., Cipolla, R.: Probabilistic visibility for multi-view stereo. In: Proceedings of CVPR (2007)Google Scholar
  6. 6.
    Liu, S., Cooper, D.B.: A complete statistical inverse ray tracing approach to multi-view stereo. In: Proceedings of CVPR, pp. 913–920 (2011)Google Scholar
  7. 7.
    Elfes, A., Matthies, L.: Sensor integration for robot navigation: Combining sonar and stereo range data in a grid-based representation. In: Proceedings of IEEE Conference on Decision and Control (1987)Google Scholar
  8. 8.
    Konolige, K.: Improved occupancy grids for map building. Autonomous Robots 4, 351–367 (1997)CrossRefGoogle Scholar
  9. 9.
    Collins, T., Collins, J.J., Ryan, C.: Occupancy grid mapping: An empirical evaluation. In: Proceedings of the Mediterranean Conference on Control Automation (2007)Google Scholar
  10. 10.
    Guan, L., Franco, J.S., Pollefeys, M.: 3D object reconstruction with heterogeneous sensor data. In: Proceedings of 3DPVT (2008)Google Scholar
  11. 11.
    Kim, Y.M., Theobalt, C., Diebel, J., Kosecka, J., Miscusik, B., Thrun, S.: Multi-view image and ToF sensor fusion for dense 3D reconstruction. In: Proceedings of ICCV Workshops (2009)Google Scholar
  12. 12.
    Pathak, K., Birk, A., Poppinga, J., Schwertfeger, S.: 3D forward sensor modeling and application to grid based sensor fusion. In: Proceedings of IROS (2007)Google Scholar
  13. 13.
    Thrun, S.: Learning occupancy grids with forward models. Autonomous Robots 15, 111–127 (2001)CrossRefGoogle Scholar
  14. 14.
    Curless, B., Levoy, M.: A volumetric method for building complex models from range images. In: Proceedings of SIGGRAPH (1996)Google Scholar
  15. 15.
    Newcombe, R.A., Izadi, S., Hiliges, O., Molyneaux, D., Kim, D., Davison, A.J., Kohli, P., Shotton, J., Hodges, S., Fitzgibbon, A.: KinectFusion: Real-time dense surface mapping and tracking. In: Proceedings of ISMAR (2011)Google Scholar
  16. 16.
    Zach, C., Pock, T., Bischof, H.: A globally optimal algorithm for robust TV-L \(^\textrm{1}\) range image integration. In: Proceedings of ICCV (2007)Google Scholar
  17. 17.
    Vogiatzis, G., Hernández, C.: Video-based, real-time multi view stereo. Image and Vision Computing 29(7), 434–441 (2011)CrossRefGoogle Scholar
  18. 18.
    Bishop, C.M.: Pattern Recognition and Machine Learning. Springer (2006)Google Scholar
  19. 19.
    Minka, T.: A family of algorithms for approximate Bayesian inference. PhD thesis. MIT (2001)Google Scholar
  20. 20.
    Point Cloud Library, http://pointclouds.org

Copyright information

© Springer-Verlag Berlin Heidelberg 2012

Authors and Affiliations

  • Oliver J. Woodford
    • 1
  • George Vogiatzis
    • 2
  1. 1.Toshiba Research Europe Ltd.CambridgeUK
  2. 2.Aston UniversityBirminghamUK

Personalised recommendations