Next: Relation between the fundamental Up: Multi view geometry Previous: Multi view geometry   Contents

## Two view geometry

In this section the following question will be addressed: Given an image point in one image, does this restrict the position of the corresponding image point in another image? It turns out that it does and that this relationship can be obtained from the calibration or even from a set of prior point correspondences.

Although the exact position of the scene point is not known, it is bound to be on the line of sight of the corresponding image point . This line can be projected in another image and the corresponding point is bound to be on this projected line . This is illustrated in Figure 3.5.

In fact all the points on the plane defined by the two projection centers and have their image on . Similarly, all these points are projected on a line in the first image. and are said to be in epipolar correspondence (i.e. the corresponding point of every point on is located on , and vice versa).

Every plane passing through both centers of projection and results in such a set of corresponding epipolar lines, as can be seen in Figure 3.6. All these lines pass through two specific points and . These points are called the epipoles, and they are the projection of the center of projection in the opposite image.

This epipolar geometry can also be expressed mathematically. The fact that a point is on a line can be expressed as . The line passing trough and the epipole is

 (C23)

with the antisymmetric matrix representing the vectorial product with .

From (3.9) the plane corresponding to is easily obtained as and similarly . Combining these equations gives:

 (C24)

with indicating the Moore-Penrose pseudo-inverse. The notation is inspired by equation (2.7). Substituting (3.23) in (3.24) results in

Defining , we obtain
 (C25)

and thus,
 (C26)

This matrix is called the fundamental matrix. These concepts were introduced by Faugeras [31] and Hartley [46]. Since then many people have studied the properties of this matrix (e.g. [82,83]) and a lot of effort has been put in robustly obtaining this matrix from a pair of uncalibrated images [153,154,177].

Having the calibration, can be computed and a constraint is obtained for corresponding points. When the calibration is not known equation (3.26) can be used to compute the fundamental matrix . Every pair of corresponding points gives one constraint on . Since is a matrix which is only determined up to scale, it has unknowns. Therefore 8 pairs of corresponding points are sufficient to compute with a linear algorithm.

Note from (3.25) that , because . Thus, . This is an additional constraint on and therefore 7 point correspondences are sufficient to compute through a nonlinear algorithm. In Section 4.2 the robust computation of the fundamental matrix from images will be discussed in more detail.

Subsections

Next: Relation between the fundamental Up: Multi view geometry Previous: Multi view geometry   Contents
Marc Pollefeys 2002-11-22