Essential Matrix: Difference between revisions

Line 64: Line 64:
A <math>3 \times 3</math> real matrix can be factored into a product of a rotation matrix <math>R</math> and a non-zero skew symmetric matrix <math>S</math> iff <math>Q</math> is two equal non-zero singular values and one zero singular value.
A <math>3 \times 3</math> real matrix can be factored into a product of a rotation matrix <math>R</math> and a non-zero skew symmetric matrix <math>S</math> iff <math>Q</math> is two equal non-zero singular values and one zero singular value.


Let the singular value decomposition of our essential matrix <math>Q</math> be <math>U D V^T</math> where <math>D = \operatorname{diag}(k, k, 0)</math>.
Let the singular value decomposition of our essential matrix <math>Q</math> be <math>U D V^T</math> where <math>D = \operatorname{diag}(k, k, 0)</math>.<br>
Let <math>E = \begin{pmatrix}
Let <math>E = \begin{pmatrix}
0 & 1 & 0\\
0 & 1 & 0\\
Line 75: Line 75:
1 & 0 & 0 \\
1 & 0 & 0 \\
0 & 0 & 0
0 & 0 & 0
\end{pmatrix}</math>
\end{pmatrix}</math><br>
Then we have the following:
Then we have the following:
* <math>S = V Z V^T</math>
* <math>S = V Z V^T</math>