3-D Reconstruction of Urban Scenes from Image Sequences

@article{Faugeras19983DRO,
  title={3-D Reconstruction of Urban Scenes from Image Sequences},
  author={Olivier D. Faugeras and Luc Robert and St{\'e}phane Laveau and Gabriela Csurka and Cyril Zeller and Cyrille Gauclin and Imad Zoghlami},
  journal={Comput. Vis. Image Underst.},
  year={1998},
  volume={69},
  pages={292-309}
}
In this paper, we address the problem of the recovery of a realistic textured model of a scene from a sequence of images, without any prior knowledge either about the parameters of the cameras or about their motion. We do not require any knowledge of the absolute coordinates of some control points in the scene to achieve this goal. First, using various computer vision tools, we establish correspondences between the images and recover the epipolar geometry, from which we show how to compute the… 

Architectural Scenes Reconstruction from Uncalibrated Photos and Map Based Model Knowledge

TLDR
This paper considers the problem of reconstructing architectural scenes from multiple photographs taken from arbitrarily viewpoints, and shows how simple constraints can be used to calibrate the cameras, to recover the projection matrices for each viewpoint, and to obtain 3D models by using triangulation.

Interactive 3D modeling from multiple images using scene regularities

TLDR
This paper presents two interactive (semi-automated) systems for recovering 3D models of large-scale environments from multiple images, and specifies some geometric knowledge ahead of time to guide the matching and reconstruction algorithms.

Calibration, pose and novel views from single images of constrained scenes

Tracking of hidden parts in video sequences and other problems raised by the 3D reconstruction of the urban environment

TLDR
A new technique of layer segmentation which extracts regions of similar parametric motion in a video sequence based on temporal constraints defined and optimized over all images simultaneously and not successively, without any a priori on the observed scene.

3D models of architectural scenes from uncalibrated images and vanishing points

  • R. CipollaD. Robertson
  • Computer Science
    Proceedings 10th International Conference on Image Analysis and Processing
  • 1999
TLDR
A simple, geometrically intuitive method which exploits the strong rigidity constraints of parallelism and orthogonality present in indoor and outdoor architectural scenes to recover 3D models from uncalibrated images from arbitrary viewpoints is proposed.

AUTOMATIC IMAGE TO MODEL ALIGNMENT FOR PHOTO-REALISTIC URBAN MODEL RECONSTRUCTION

TLDR
A hybrid approach in which images of an urban scene are automatically aligned with a base geometry of the scene to determine model-relative external camera parameters, which is an improvement over previous methods because it does not require user assistance for camera calibration.

Affine reconstruction from a plane and two parallel lines

TLDR
It is proposed that if one plane and a pair of parallel lines are presented in the scene, the affine reconstruction can be done linearly from two images taken by a translating camera.

PhotoBuilder-3D models of architectural scenes from uncalibrated images

TLDR
A simple, geometrically intuitive method is proposed which exploits the strong rigidity constraints of parallelism and orthogonality present in indoor and outdoor architectural scenes to recover 3D models from uncalibrated images of architectural scenes.

Efficient Fourier-Based Approach for Detecting Orientations and Occlusions in Epipolar Plane Images for 3D Scene Modeling

TLDR
A novel panoramic EPI analysis method is proposed that combines the advantages of both PVIs and EPIs efficiently in three important steps: locus orientation detection in the Fourier frequency domain, motion boundary localization in the spatio-temporal domain, and occlusion/resolution recovery only at motion boundaries.
...

References

SHOWING 1-10 OF 55 REFERENCES

3-D scene representation as a collection of images

TLDR
This work addresses the problem of the prediction of new views of a given scene from existing weakly or fully calibrated views called reference views, and derives an algorithm from the constraints between the images to predict the new views.

Epipole and Fundamental Matrix Estimation Using the Virtual Parallax Property

TLDR
A novel method based on virtual parallax is proposed, instead of computing directely the 3 3 fundamental matrix, a homography with one epipole position is computed, and it is shown that this is equivalent to compute the fundamental matrix.

Physically-valid view synthesis by image interpolation

  • S. SeitzC. Dyer
  • Computer Science
    Proceedings IEEE Workshop on Representation of Visual Scenes (In Conjunction with ICCV'95)
  • 1995
TLDR
It is shown that two basis views are sufficient to predict the appearance of the scene within a specific range of new viewpoints and that generating this range of views is a theoretically well-posed problem, requiring neither knowledge of camera positions nor 3D scene reconstruction.

Camera Self-Calibration from Video Sequences: the Kruppa Equations Revisited

The use of some new techniques such as self-calibration and calibration from partial a priori information about the observed scene (angles, distances), together with the use of some classical

REALISE: reconstruction of REALity from Image SEquences

TLDR
The R&D work is been performed principally in the following areas of computer vision and computer graphics: structure from motion, recovery of geometries, Recovery of photometric and texture information, highly realistic rendering on the basis of empirically-based reflectance models, and the design and development of improved rendering processes together with a new VR system.

An optimization framework for efficient self-calibration and motion determination

  • Q. LuongO. Faugeras
  • Computer Science
    Proceedings of 12th International Conference on Pattern Recognition
  • 1994
TLDR
This paper proposes an alternative method to compute at the same time camera calibration and motion, which is robust and efficient, and allows also to take into account the important trinocular constraints.

Matching constraints and the joint image

  • B. Triggs
  • Mathematics
    Proceedings of IEEE International Conference on Computer Vision
  • 1995
TLDR
The geometry of multi image perspective projection and the matching constraints that this induces on image measurements are studied and their complex algebraic interdependency is captured by quadratic structural simplicity constraints on the Grassmannian.

Canonic Representations for the Geometries of Multiple Projective Views

We show how a special decomposition of general projection matrices, called canonic enables us to build geometric descriptions for a system of cameras which are invariant with respect to a given group

Epipole and fundamental matrix estimation using virtual parallax

  • B. BoufamaR. Mohr
  • Computer Science
    Proceedings of IEEE International Conference on Computer Vision
  • 1995
TLDR
Instead of computing directly the 3/spl times/3 fundamental matrix, a homography with one epipole position is computed, and it is shown that this is equivalent to computing the fundamental matrix.
...