## 3-D Reconstruction of Urban Scenes from Sequences of Images (1995)

Venue: | Automatic Extraction of Man-Made Objects from Aerial and Space Images. Birkhauser |

Citations: | 42 - 1 self |

### BibTeX

@TECHREPORT{Faugeras953-dreconstruction,

author = {Olivier Faugeras and Olivier Faugeras and Stéphane Laveau and Stephane Laveau and Luc Robert and Luc Robert and Gabriella Csurka and Gabriella Csurka and Cyril Zeller and Cyril Zeller and Projet Robotvis},

title = {3-D Reconstruction of Urban Scenes from Sequences of Images},

institution = {Automatic Extraction of Man-Made Objects from Aerial and Space Images. Birkhauser},

year = {1995}

}

### Years of Citing Articles

### OpenURL

### Abstract

In this paper, we address the problem of the recovery of the Euclidean geometry of a scene from a sequence of images without any prior knowledge either about the parameters of the cameras, or about the motion of the camera(s). We do not require any knowledge of the absolute coordinates of some control points in the scene to achieve this goal. Using various computer vision tools, we establish correspondences between images and recover the epipolar geometry of the set of images, from which we show how to compute the complete set of perspective projection matrices for each camera position. These being known, we proceed to reconstruct the scene. This reconstruction is defined up to an unknown projective transformation (i.e. is parameterized with 15 arbitrary parameters). Next we show how to go from this reconstruction to a more constrained class of reconstructions, defined up to an unknown affine transformation (i.e. parameterized with 12 arbitrary parameters) by exploiting known geometr...

### Citations

1731 | A Combined Corner and Edge Detector
- Harris, Stephens
- 1988
(Show Context)
Citation Context ...tween images The algorithm used to compute the projection matrices needs correspondences and a few epipoles in order to work. We first obtain feature points using very simple corner detectors (we use =-=[19]-=-, but other possibilities are [23, 35, 18, 15]) and we refine their position using a model-based approach [3]. In the S-situation, we then establish correspondences between the corners using grey-leve... |

1262 |
Three-Dimensional Computer Vision — A Geometric Viewpoint
- Faugeras
- 1993
(Show Context)
Citation Context ...an be recovered. Since projective geometry is now widely used in computer vision and has been in computer graphics since the early days, we refer the interested reader to the corresponding literature =-=[39, 34, 9]-=-. The estimation of the perspective projection matrices is done in three steps: 1. We estimate the N \Gamma 1 fundamental matrices defining the epipolar geometry between consecutive images, 2. these a... |

502 |
What can be seen in three dimensions with an uncalibrated stereo rig
- Faugeras
- 1992
(Show Context)
Citation Context ...try Recently, it has been discovered that the full calibration of the cameras (intrinsic and extrinsic parameters) is not needed to obtain a useful reconstruction of a scene viewed by a stereo system =-=[8, 21]-=-. This theory make use of epipolar geometry which can be retrieved easily from point correspondences in pair of images. Since these first attempts at an uncalibrated stereovision, a lot of work has be... |

361 | Camera self-calibration: theory and experiments
- Faugeras, Luong, et al.
- 1992
(Show Context)
Citation Context ...scene. Nonetheless, it does require, in the version presented in this article, some knowledge about the 3-D geometry of the scene such as parallel lines and angles. But we have shown in previous work =-=[33, 28, 12]-=- that even this assumption is not necessary. We use it here because there is such a rich geometric information in images of urban scenes and because the system we are developing is partially interacti... |

236 | Euclidean reconstruction from uncalibrated views
- Hartley
(Show Context)
Citation Context ...d easily from point correspondences in pair of images. Since these first attempts at an uncalibrated stereovision, a lot of work has been done on the estimation of the epipolar geometry of two images =-=[29, 26, 32, 31, 30, 22, 20, 36, 4]-=-. Robust programs which work automatically are now publicly available. We will consider this problem as solved for the rest of this article; the interested reader is referred to the bibliography. We w... |

211 |
A fast operator for detection and precise location of distinct points, corners and centres of circular features
- Förstner, Gülch
- 1987
(Show Context)
Citation Context ... compute the projection matrices needs correspondences and a few epipoles in order to work. We first obtain feature points using very simple corner detectors (we use [19], but other possibilities are =-=[23, 35, 18, 15]-=-) and we refine their position using a model-based approach [3]. In the S-situation, we then establish correspondences between the corners using grey-level correlation between neighboring regions of t... |

148 | On the geometry and algebra of the point and line correspondences between N images
- Faugeras, Mourrain
- 1995
(Show Context)
Citation Context ...ers). This leads to O(N) parameters for the cameras and since the fundamental matrices are represented by O(N 2 ) parameters, there exist constraints between them. These constraints can be enumerated =-=[13, 14]-=-, but they are difficult to use and we prefer the more compact representation of the projective camera matrices. 1.2 Perspective Projection Matrices A point M in space projects to a point m in the ima... |

50 | Characterizing the uncertainty of the fundamental matrix
- Csurka, Zeller, et al.
- 1997
(Show Context)
Citation Context ...equence. RR n2572 10 Olivier Faugeras, Stephane Laveau, Luc Robert and Gabriella Csurka, Cyril Zeller the parameter vector of the fundamental matrix. The details of those computations can be found in =-=[7]-=-. 2.4 Recovering the geometry of the N cameras Up to this stage in the processing, we have estimated the fundamental matrices of consecutive pairs of images as well as obtained a number of point corre... |

48 | Stratification of 3-d vision: projective, affine, and metric representations
- Faugeras
- 1995
(Show Context)
Citation Context ... estimate the affine geometry of the scene, i.e. those properties of the scene which are invariant under the action of 3-D affine transformations. This can be achieved in several ways as described in =-=[10]-=- depending upon whether or not one can control the motion of the sensors. In the application described in this article we only use a priori information about the scene such as parallel lines or known ... |

37 | Epipole and fundamental matrix estimation using virtual parallax
- Boufama, Mohr
- 1995
(Show Context)
Citation Context ...d easily from point correspondences in pair of images. Since these first attempts at an uncalibrated stereovision, a lot of work has been done on the estimation of the epipolar geometry of two images =-=[29, 26, 32, 31, 30, 22, 20, 36, 4]-=-. Robust programs which work automatically are now publicly available. We will consider this problem as solved for the rest of this article; the interested reader is referred to the bibliography. We w... |

25 | Invariant linear methods in photogrammetry and model-matching - Barrett, Brill, et al. - 1992 |

20 |
Multivariate Calibration
- Brown
- 1982
(Show Context)
Citation Context ...ges, 2. these are then used to obtain a first estimate of the perspective projection matrices, 3. which is then refined using possibly several methods, one of them, being the famous bundle adjustment =-=[5, 6, 16, 17]-=-. Once this estimation has been completed, the 3-D scene can be reconstructed up to an arbitrary projective transformation. To be complete, let us mention that we track the uncertainty at all levels, ... |

15 |
Representing three-dimensional data as a collection of images and fundamental matrices for image synthesis
- FAUGERAS, LAVEAU
- 1994
(Show Context)
Citation Context ...computed. We can even project it onto an arbitrary virtual camera: this way we can produce new views of the scene. This process, usually called view transfer [2, 41], has been used for image synthesis=-=[11, 25]-=-. However, for simple reasons, the projective reconstruction may not be sufficient: for instance withinsRealise, the generated building models are used for realistic rendering and virtual walk-through... |

14 |
A solution to the general problem of multiple station analytical stereotriangulation
- Brown
- 1958
(Show Context)
Citation Context ...ges, 2. these are then used to obtain a first estimate of the perspective projection matrices, 3. which is then refined using possibly several methods, one of them, being the famous bundle adjustment =-=[5, 6, 16, 17]-=-. Once this estimation has been completed, the 3-D scene can be reconstructed up to an arbitrary projective transformation. To be complete, let us mention that we track the uncertainty at all levels, ... |

12 |
System Calibration Through Self-Calibration
- Gruen, Beyer
- 2001
(Show Context)
Citation Context ...ges, 2. these are then used to obtain a first estimate of the perspective projection matrices, 3. which is then refined using possibly several methods, one of them, being the famous bundle adjustment =-=[5, 6, 16, 17]-=-. Once this estimation has been completed, the 3-D scene can be reconstructed up to an arbitrary projective transformation. To be complete, let us mention that we track the uncertainty at all levels, ... |

9 |
Corner characterization by differential geometry techniques
- Guiducci
- 1988
(Show Context)
Citation Context ... compute the projection matrices needs correspondences and a few epipoles in order to work. We first obtain feature points using very simple corner detectors (we use [19], but other possibilities are =-=[23, 35, 18, 15]-=-) and we refine their position using a model-based approach [3]. In the S-situation, we then establish correspondences between the corners using grey-level correlation between neighboring regions of t... |

3 |
reliability and statistics in close range photogrammetry
- Accuracy
- 1978
(Show Context)
Citation Context |