The epipolar constraint Involves epipolar lines
Postmultiplying F by x
Results in vector l-prime
It's the epipolar line
In the other view passing through x-prime
A three component vector
Of homogeneous design
The left and right nullspaces of F
Are the epipoles e-prime and e
All of the epipolar lines
Should pass through these
Here's a linear estimation example:
Take a set of 8 point samples
Construct a matrix, take the SVD And the elements of F are in the last column of V
If you try to estimate
F with a coplanar set of points
Your sample set will be degenerate
And will not bring you joy
When doing the estimation
If you don't perform rank deprivation Your epipolar lines
And the epipoles will not coincide
But if your scene has three views
The trifocal tensor is what you'd use
Constraints from the third view act like glue
That can't be determined from just two views
Dr Peter Kovesi, whose code turned Matlab into a video effects tool,
Peter Pakulski, the AviSynth guru,
ilesoft, whose Sydney Harbour Bridge model is scarily
accurate, and
Payashim, for the Japanese translation.
Email: fmatrix at danielwedge dot com
Feel free to play this in lectures etc, you have my permission (though I'd be interested to hear from you if you do!)
Daniel Wedge
Song: 19th October, 2008.
Video: 25th March, 2009.