Photogrammetry, orthophotos, and 3D-models

Thursday 8.9.2016

Christos (I.) and I tagged new amphorae that have become visible once the amphora layer above has been removed. I took GoPro video for documentation purposes. In the afternoon I went through that video and picked up documentation photos for each amphora that we tagged.

This photogrammetry stuff is neat. We collect finds (amphorae etc.) continuously from the trench, and the landscape changes quite fast. To keep track of all of this lots of photos are taken, and processed with photogrammetry software. We can get almost real time orthophotos, which are kind of stitched-together large photographs covering the whole site. We also get 3d-models that visualize current state of the site very well. You can rotate and zoom in the model, and look at current state of the site from different angles and distances. The images are not exact, because we do not have supercomputers or days to spend on computing. But they are handy.

Photomosaic of the Mazotos wreck as it once was. The image is distorted, because the control points were not set. The bow curves to the right in the image, but not in the actual wreck site. Once control points are set, the same photographs can be used to create a more realistic orthophoto of the site. The site does not look like this anymore, because of the excavations done there after this photo was made. 
(https://www.ucy.ac.cy/marelab/documents/Publications_/Demesticha_2010.pdf)

It is just amazing how useful this stuff is as compared to hand-made drawings of the past. All this of course requires that we have people with proper equipment to take those zillions of photographs and people who can process them into useful product in just a few hours. And then that product is used to plan for tomorrow's dives and at the end of the day get a new set of photographs. Teamwork at its best.

The photogrammetry software keeps track of all currently visible amphorae locations in 3D, and that data can then be used to create a 3D-model of the whole site. That is virtual reality at its best. You can "see" something that was newer really visible. Some of the amphorae are not visible yet, because they are under other amphorae or buried under the sea bed. Some of the amphorae are not in the wreck any more, because they have been removed, maybe years ago. But this model can bring them all back at the same time in the same 3D-model. However, the model can not be complete, unless all amphorae have been located and photographed.

If this interests you more, you can check scientific articles on the topic. For example, this article. Both Dimitrios Skarlatos and Stella Demesticha are present at this excavation.

The new stuff is presented to all us at the outside auditorium set in front of the office. We even have the cafeteria chairs arranged there so that it looks like a (very small) small movie theater. It is novel to have briefings like this.