I have two photos of the same scene, taken from different places. I also have pairs of matching 2D points, which correspond to a same 3D point.
I found a fundamental matrix $F$ and epipoles $\textbf{e}, \textbf{e'}$.
F = 3.248066811936e-7, -0.000001956055020665, -0.001234916325071597;
-0.000004377971254, 8.358233379722482e-7, 0.019146763431467432;
0.000165564399727, -0.017625490920276023, 1.000000000000000000
I computed camera matrices $P$ and $P'$ according to the bottom of page 256.
$P=[I \mid \textbf{0} ], \\ P' = [[\textbf{e'}]_\times F + \textbf{e'}\textbf{v}^T \mid \lambda \textbf{e'}]$
I used $\textbf{v}=(0,0,0)^T$ and $\lambda=1$.
Some 3D point is at a pixel (395,128) in a first image, and at a pixel (530,108) in a second image. How can I find 3D coordinates of such point?
I see, that there are many degrees of freedom. I see, that $\lambda$ controls the distance between cameras, and the scale of the whole world. But what does the vector $v$ correspond to?