If the orientation in space of one image is known, every image generates a bundle of rays that can be used to either texture models (standalone mode) or measure objects in space (if more than one image is available and intersect the bundle of rays coming from different pictures).