262x Filetype PDF File size 0.32 MB Source: www.cse.iitd.ac.in
Multiple Views Geometry
Subhashis Banerjee
Dept. Computer Science and Engineering
IIT Delhi
email: suban@cse.iitd.ac.in
January 12, 2008
1
1 Epipolar geometry
Fundamental geometric relationship between two perspective cameras:
epipole: isthepointofintersectionofthelinejoiningtheopticalcenters-thebaseline
- with the image plane. The epipole is the image in one camera of the optical
center of the other camera.
epipolar plane: is the plane defined by a 3D point and the optical centers. Or,
equivalently, by an image point and the optical centers.
epipolar line: is the line of intersection of the epipolar plane with the image plane.
It is the image in one camera of a ray through the optical center and the image
point in the other camera. All epipolar lines intersect at the epipole.
Epipolar geometry provides a fundamental constraint for the correspon-
dence problem
1.1 Epipolar geometry: uncalibrated case
² Given the two cameras 1 and 2, we have the camera equations:
˜ ˜
x =P Xandx =P X
1 1 2 2
² The optical center projects as
˜
PX=0
i
² Writing
˜
P =[P |−Pt]
i i i i
where P is 3×3 non-singular we have that t is the optical center.
i i
· t ¸
[P | −P t ] i =0
i i i 1
² The epipole e2 in the second image is the projection of the optical center of the
first image:
· t ¸
˜ 1
e =P
2 2 1
² The projection of point on infinity along the optical ray < t ;x > on to the
1 1
second image is given by:
x =P P −1x
2 2 1 1
² The epipolar line < e ;x > is given by the cross product e ×x .
2 2 2 2
² If [e ] is the 3 × 3 antisymmetric matrix representing cross product with e ,
2 × 2
then we have that the epipolar line is given by
[e ] P P −1x =Fx
2 × 2 1 1 1
² Any point x on this epipolar line satisfies
2
x TFx =0
2 1
² F is called the fundamental matrix. It is of rank 2 and can be computed
from 8 point correspondences.
² Clearly Fe = 0 (degenerate epipolar line) and e TF = 0. The epipoles are
1 2
obtained as the null spaces of F.
1.2 Epipolar geometry: calibrated case
² There are two camera coordinate systems related by R;T
′
X =RX+T
′
² Taking the vector product with T followed by the scalar product with X
′
X·(T×RX)=0
′ ′ ′
which expresses that vectors OX, O X and OO are coplanar.
² This can be written as
′T
X EX=0
where
E=[T] R
×
is the Essential matrix.
² Image points and rays in Euclidean 3-space are related by:
x X x′ ′ X′
′ ′
y = C Y and y =C Y′
1 Z 1 Z
² Hence, we have
′T ′−T −1
x C EC x=0
² Thus, the relation between the essential and fundamental matrix is:
′−T −1
F=C EC
no reviews yet
Please Login to review.