What if you might get the shape of any object from its photos?
Even more, What if it is also possible from its video? It is not a kind of magic, it can be done with Photogrammetry, which is the science of making measurements from photographs. The bases of this technology is the concept of some spatial points called “features”, which are tracked on several consecutive photographs. Features have similar visual properties kept invariant for a while during the recording, which means for several consecutive photographs. From this information it is possible to obtain the 3D coordinates of every feature computed.
In fact it is also possible to compute the position and orientation of every camera involved. In addition to this, the visual properties of this features like color can be used to add more realistic appearance to the features, and therefore to the whole object.
Photogrammetry is not a new science, but the use of Artificial Intelligence, in particular Computer Vision Technology, has been a huge improvement. Nowadays it is not necessary to calibrate the cameras, only a set of photographs taken from one or more cameras is enough to use this technology.
The applications of Photogrammetry in Archeology are becoming more and more important, as it is easy to imagine. As an example, it is possible to obtain the 3D shape of an archaeological site from a video taken with a drone, at least from a theoretical point of view. In the same way a similar application is possible in Underwater Archeology, mutatis mutandi, that is, using a ROV or simply a camera operated by a scuba diver.
In this case, it is necessary a good visibility of the recorded area. An applied example of Photogrammetry on a wreck (Ustica Island, Italy) is shown here.
If the shape of a particular joint of a wreck is known, it is possible to compute the mechanical stresses in that joint when the ship was sailing by using the Finite Elements Method. Some results are shown in the above-mentioned video (Jules Verne 7 wreck, about 300 years bC).