Essential Matrix: Difference between revisions

Line 54: Line 54:


===Planar Images===
===Planar Images===
Copied from section 5 of Hartley<ref name="hartley"/>.
Here we will focus on calculating the essential matrix given 8 or more points.
It is possible to calculate the essential matrix using 7 points using a non-linear equation if
your correspondences are very accurate and are not linearly dependent.
For this, see section 5.1 of Hartley<ref name="hartley"/>.
For each correspondence <math>\mathbf{u}_i</math> and <math>\mathbf{u}'_i</math>.,
we get the equation <math>\mathbf{u}_i'^T Q \mathbf{u}_i = 0</math>.
This system of equations is linear in the entries of <math>Q</math> and can be
rewritten as <math>A\mathbf{x} = 0</math> where <math>\mathbf{x}</math> contains the entries of <math>Q</math>.<br>
Here,
<math>\mathbf{x} = \begin{pmatrix}
q_{11} \\ q_{12} \\ q_{13} \\
q_{21} \\ q_{22} \\ q_{23} \\
q_{31} \\ q_{32} \\ q_{33} \\
\end{pmatrix}</math>
and each row of <math>A</math> is
<math>\mathbf{a_i} = \begin{pmatrix}
u'_1 u_1 \\ u'_1 u_2 \\ u'_1 \\
u'_2 u_1 \\ u'_2 u_2 \\ u'_2 \\
u'_3 u_1 \\ u'_3 u_2 \\ 1 \\
\end{pmatrix}
</math>
The goal is to minimize <math>\Vert A\mathbf{x} \Vert </math> such that <math>\Vert \mathbf{x} \Vert = 1</math>