Scene Reconstruction from Images
The CAIP’99 invited lecture overviews the recent development in the scene reconstruction in the context of the Prague research contribution to it by the Center for Machine Perception.
3D Scene Reconstruction from (uncalibrated) 2D images The second half of nineties witnessed a qualitative move from stereovision that remained for a long time in a one hundred old photogrammetric framework providing relation between two views only. The new impulse was the discovered trilinear and quadrilinear relation among views in projective geometry. Another impulse was the transition from Euclidean using calibrated cameras to projective reconstruction where uncalibrated cameras are sufficient. The observations of the scene provides extensive number (tenth of) images that yield qualitatively better results than before.
The topic will be overviewed from the perspective of own achievements: (a) interpolation from two images , (b) relation among interpolation, extrapolation and reconstruction of a full 3D model , (c) projective reconstruction from three uncalibrated images , (d) introduction of the oriented projective reconstruction 14, 12, (e) selection of an optimal set of reference images , (f) search for correspondences if observer just translates or rotates , (g) the attempt to generalise a dense sequence correspondence to more general cases , (h) autocalibration from uncalibrated views 8, 4.
Omni-directional Vision Epipolar geometry and egomotion estimation algorithm for central panoramic cameras was developed and presented 6, 3, 7. Design and image formation for newly defined central panoramic cameras have been studied . New approach to mobile robot localization has been proposed . The approach relies on a visual map comprising panoramic images which are represented in a rotationally invariant manner.
KeywordsMotion Estimation Panoramic Image Epipolar Geometry Projective Reconstruction Mobile Robot Localization
- 1.Václav Hlaváĉ, Aleš Leonardis, and Tomáš Werner. Automatic selection of reference views for image-based scene representations. In Proceedings of the European Conference in Computer Vision, volume 1 of Lecture Notes in Computer Science, No. 1064, pages 526–535, Heidelberg, Germany, April 1996. Springer.Google Scholar
- 2.Tomáš Pajdla. Robot localization using shift invariant representation of panoramic images. Technical Report K335-CMP-1998-170, FEE CTU, FEL ĈVUT, Karlovo náměstí 13, Praha, Czech Republic, November 1998.Google Scholar
- 3.Tomáš Pajdla, Tomáš Svoboda, and Václav Hlaváĉ. Robot motion estimation from panoramic images. In First International Workshop on Robot Perception for Autonomous Aerial Vehicles, June 1998. to appear.Google Scholar
- 4.Tomáš Pajdla and Martin Urban. Camera calibration from bundles of parallel lines. Research Report K335-CMP/99/180, Czech Technical University, Prague, Karlovo nam. 13, 121 35 Prague, Czech Republic, January 1999.Google Scholar
- 5.Tomáš Svoboda, Tomáš Pajdla, and Váaclav Hlaváĉ. Central panoramic cameras: Design and geometry. In Aleš Leonardis and Franc Solina, editors, Proceedings of Computer Vision Winter Workshop in Gozd Martuljek, Slovenia, pages 120–133, Ljubljana, Slovenia, February 1998. IEEE Slovenia Section, IEEE Slovenia Section.Google Scholar
- 6.Tomáš Svoboda, Tomáš Pajdla, and Váaclav Hlaváĉ. Epipolar geometry for panoramic cameras. In Hans Burkhardt and Neumann Bernd, editors, the fifth European Conference on Computer Vision, Freiburg, Germany, number 1406 in Lecture Notes in Computer Science, pages 218–232, Berlin, Germany, June 1998. Springer.Google Scholar
- 7.Tomáš Svoboda, Tomáš Pajdla, and Váaclav Hlaváĉ. Motion estimation using central panoramic cameras. In Stefan Hahn, editor, IEEE International Conference on Intelligent Vehicles, pages 335–340, Stuttgart, Germany, October 1998. Causal Productions.Google Scholar
- 8.Martin Urban, Tomá Pajdla, and Václav Hlaváĉ. Camera self-calibration from multiple views. Research Report CTU-CMP-1998-169, CMP FEE CTU, Karlovonám. 13, 121 35 Prague, Czech Republic, September 1998.Google Scholar
- 9.Jan Vydržel and Václav Hlaváĉ. Tracking correspondences in dense motion sequences. Research Report K335-1998-154, CMP FEE CTU, FEL ĈVUT, Karlovo náméstí 13, Praha, Czech Republic, January 1998.Google Scholar
- 10.T. Werner, R.D. Hersch, and V. Hlaváĉ. Rendering real-world objects using view interpolation. In Proc. 5th International Conf. on Computer Vision, pages 957–962, Boston, USA, June 1995. IEEE Computer Society Press.Google Scholar
- 11.T. Werner, T. Pajdla, and V. Hlaváĉ. Efficient rendering of projective model for image-based visualization. In Proceedings of the 14th International Conference on Pattern Recognition, Brisbane, Australia, pages 1705–1707, Los Alamitos, California, August 1998. International Association for Pattern Recognition, IEEE Computer Society.Google Scholar
- 12.Tomáš Werner. Image-Based Visualization of Real 3D Scenes. PhD thesis, Faculty of Eletrical Engineering, Czech Technical University, FEL ĈVUT, Karlovo náméstí, 13,_Praha, Czech Republic, 1998.Google Scholar
- 13.Tomáš Werner, Tomáš Pajdla, and Václav Hlaváĉ. Efficient 3-D scene visualization by image extrapolation. In Hans Burkhardt and Bernd Neumann, editors, Proc. 5th European Conf. Computer Vision, volume 2, pages 382–395, Berlin, Germany, June 1998. Springer Verlag.Google Scholar
- 14.Tomáš Werner, Tomáš Pajdla, and Václav Hlaváĉ. Oriented projective reconstruction. In M. Gengler, M. Prinz, and E. Schuster, editors, Pattern Recognition and Medical Computer Vision: 22-nd Workshop of the Austrian Association for Pattern Recognition (öAGM/IAPR), pages 245–254, Wien, Austria, May 14—15 1998. österreichische Computer Gesselschaft.Google Scholar