Showing papers on "Homography (computer vision) published in 1991"

TL;DR: 3D measures in a scene can be performed using measures on images, without any camera calibration, and relies only on an invariant by homography.