Actually, determining camera position is done quite differently and more efficiently: https://en.wikipedia.org/wiki/Simultaneous_localization_and_...

(I spent about a year writing an optimiser for camera localisation from multiple views)