Previous | Next --- Slide 78 of 112
Back to Lecture Thumbnails
nssampat

Where did the blue equations for epipoles come from? I understand all lines l' will contain e', but am not seeing the connection to the epipole equations. Also, don't the blue equations imply that the epipole and its corresponding epipolar line are orthogonal? I don't think this is true but making the substitution e' = x' into x'Ex = 0, we get e'l' = 0, and this implies they are orthogonal.

motoole2

As you stated, all epipolar lines $l'$ will contain the epipole $e'$. Mathematically this means that $e'^Tl' = e'^T E x = 0$ for all $x$. Therefore, $e'^T E = \mathbf{0}$.

What does it mean for a point $e'$ to be orthogonal to a line $l'$? I would instead put it this way. $e'$ is a homogeneous 2D coordinate, and $l'$ contains the parameters to a line equation. $e'^Tl' = 0$ simply means that the point $e'$ is a point on the line $l'$. (See this slide for more details.)