AI helps you reading Science

AI generates interpretation videos

AI extracts and analyses the key points of the paper to generate videos automatically


pub
Go Generating

AI Traceability

AI parses the academic lineage of this thesis


Master Reading Tree
Generate MRT

AI Insight

AI extracts a summary of this paper


Weibo:
We have proposed a method of projective reconstruction from multiple uncalibrated images

A Factorization Based Algorithm for Multi-Image Projective Structure and Motion

ECCV (2), pp.709-720, (1996)

Cited: 828|Views172
EI

Abstract

We propose a method for the recovery of projective shape and motion from multiple images of a scene by the factorization of a matrix containing the images of all points in all views. This factorization is only possible when the image points are correctly scaled. The major technical contribution of this paper is a practical method for the ...More

Code:

Data:

0
Introduction
  • In the last few years, the geometric and algebraic relations between uncalibrated views have found lively interest in the computer vision community.
  • Hartley [HGC92] derives from the fundamental matrix 2 projection matrices, equal to the true ones up to an unknown projective transformation
  • These are used to perform reconstruction by triangulation[HS94].
  • Projective reconstruction consists essentially of rescaling the image coordinates in order to place the joint image vector in a certain 4-dimensional subspace of the joint image space called the joint image
  • This subspace is characterized by the multi-linear matching constraints between the views.
  • Throughout this paper the authors will restrict attention to the case of bilinear matching constraints (fundamental ? This work has been done in the context of the MOVI project which belongs to CNRS, INPG, INRIA and UJF
Highlights
  • In the last few years, the geometric and algebraic relations between uncalibrated views have found lively interest in the computer vision community
  • The information one needs to do so is entirely contained in the fundamental matrix, which represents the epipolar geometry of the 2 views
  • Projective reconstruction consists essentially of rescaling the image coordinates in order to place the joint image vector in a certain 4-dimensional subspace of the joint image space called the joint image
  • We have proposed a method of projective reconstruction from multiple uncalibrated images
  • We have proposed a very simple way to obtain the individual scale factors, using only fundamental matrices and epipoles estimated from the image data
  • The results show that it is essential to work with normalized image coordinates
Methods
  • Experiments with Simulated Images

    The authors conducted a large number of experiments with simulated images to quantify the performance of the algorithm.
  • The simulations used three different configurations : lateral movement of a camera, movement towards the scene, and a circular movement around the scene.
  • In configuration 2, the depths of points lying on the line joining the projection centers can not be recovered.
  • Reconstruction of points lying close to this line is extremely difficult, as was confirmed by the experiments, which resulted in quite inaccurate reconstructions for this configuration.
  • The overall trajectory of the camera formed a quarter circle, centered on the scene.
  • The distance between the camera and the center of the sphere was 200
Results
  • Evaluation with Real Images

    The algorithm has been tested on several sequences of real images.
  • 38 points were tracked over the whole sequence, but only extracted with 1 pixel accuracy.
  • To estimate the quality of the projective reconstruction, the authors aligned it with an approximate Euclidean model of the scene obtained from calibrated views.
  • The bumpiness on the left side of the roof is due to the fact that the roof stands out slightly from the house’s front wall, causing occlusion in the last view of the edge point between roof and wall
Conclusion
  • Discussion and Further

    Work

    In this paper, the authors have proposed a method of projective reconstruction from multiple uncalibrated images.
  • The method is very elegant, recovering shape and motion by factorization of one matrix, containing all image points of all views.
  • This factorization is only possible when the image points are correctly scaled.
  • The authors have proposed a very simple way to obtain the individual scale factors, using only fundamental matrices and epipoles estimated from the image data.
  • Data for this research were partially provided by the Calibrated Imaging Laboratory at Carnegie Mellon University, supported by ARPA, NSF, and NASA
Funding
  • This work was partially supported by INRIA France and E.C. projects HCM and SECOND
Reference
  • [Fau92] O. Faugeras. What can be seen in three dimensions with an uncalibrated stereo rig? In G. Sandini, editor, Proceedings of the 2nd European Conference on Computer Vision, Santa Margherita Ligure, Italy, pages 563–578. Springer-Verlag, May 1992.
    Google ScholarLocate open access versionFindings
  • [FM95] O. Faugeras and B. Mourrain. On the geometry and algebra of the point and line correspondences between n images. In Proceedings of the 5th International Conference on Computer Vision, Cambridge, Massachusetts, USA, pages 951–956, June 1995.
    Google ScholarLocate open access versionFindings
  • [Har93] R.I. Hartley. Euclidean reconstruction from uncalibrated views. In Proceeding of the DARPA–ESPRIT workshop on Applications of Invariants in Computer Vision, Azores, Portugal, pages 187–202, October 1993.
    Google ScholarLocate open access versionFindings
  • [Har95] R. Hartley. In defence of the 8-point algorithm. In Proceedings of the 5th International Conference on Computer Vision, Cambridge, Massachusetts, USA, pages 1064–1070, June 1995.
    Google ScholarLocate open access versionFindings
  • [HGC92] R. Hartley, R. Gupta, and T. Chang. Stereo from uncalibrated cameras. In Proceedings of the Conference on Computer Vision and Pattern Recognition, Urbana-Champaign, Illinois, USA, pages 761–764, 1992.
    Google ScholarLocate open access versionFindings
  • [HS94] R. Hartley and P. Sturm. Triangulation. In Proceedings of ARPA Image Understanding Workshop, Monterey, California, pages 957–966, November 1994.
    Google ScholarLocate open access versionFindings
  • [MM95] P.F. McLauchlan and D.W. Murray. A unifying framework for structure and motion recovery from image sequences. In Proceedings of the 5th International Conference on Computer Vision, Cambridge, Massachusetts, USA, pages 314–320, 1995.
    Google ScholarLocate open access versionFindings
  • [MVQ93] R. Mohr, F. Veillon, and L. Quan. Relative 3D reconstruction using multiple uncalibrated images. In Proceedings of the Conference on Computer Vision and Pattern Recognition, New York, USA, pages 543– 548, June 1993.
    Google ScholarLocate open access versionFindings
  • [PK94] C. J. Poelman and T. Kanade. A paraperspective factorization method for shape and motion recovery. In J.O. Eklundh, editor, Proceedings of the 3rd European Conference on Computer Vision, Stockholm, Sweden, pages 97–108, May 1994.
    Google ScholarLocate open access versionFindings
  • [Sha94] A. Shashua. Projective structure from uncalibrated images: Structure from motion and recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence, 16(8):778–790, August 1994.
    Google ScholarLocate open access versionFindings
  • [Sha95] A. Shashua. Algebraic functions for recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence, 17(8):779–789, August 1995.
    Google ScholarLocate open access versionFindings
  • [TK92] C. Tomasi and T. Kanade. Shape and motion from image streams under orthography: A factorization method. International Journal of Computer Vision, 9(2):137–154, 1992.
    Google ScholarLocate open access versionFindings
  • [Tri95a] B. Triggs. The geometry of projective reconstruction i: Matching constraints and the joint image. International Journal of Computer Vision, 1995. submitted.
    Google ScholarLocate open access versionFindings
  • [Tri95b] B. Triggs. Matching constraints and the joint image. In Eric Grimson, editor, Proceedings of the 5th International Conference on Computer Vision, Cambridge, Massachusetts, USA, pages 338–343. IEEE, IEEE Computer Society Press, June 1995.
    Google ScholarLocate open access versionFindings
0
Your rating :

No Ratings

Tags
Comments
avatar
数据免责声明
页面数据均来自互联网公开来源、合作出版商和通过AI技术自动分析结果,我们不对页面数据的有效性、准确性、正确性、可靠性、完整性和及时性做出任何承诺和保证。若有疑问,可以通过电子邮件方式联系我们:report@aminer.cn