Visit at Marburg

Last week I visited Dr. Felix Teichner and Christoph Salzmann at the Institute for Prehistoric Archaeology of Marburg University to give a talk on my PhD project. Besides this, we did some fly-arounds with the copters, since Christoph is also working with UAVs in Archaeology and we did a lot of exchange in the last months, but only over internet. So it was just about time to meet again. He showed me his DJI Phantom and even trusted me to fly it a little bit around. It is indeed very easy to steer and is able to fly quite long with a GoPro camera: 20 minutes! Christoph also flew at castle Karlsmunt above Wetzlar in order to detect accessways to the stronghold with his Finwing Penguin. This is an airplane with 1.7 m wingspan, quite impressive how good it flew using the Pixhwak autopilot. He uses a Ricoh GR camera which has a fixed focal length, but offering a very good image quality.

We made it also in to the newspaper with this flight:,-Drohne-sucht-Spuren-_arid,511472.html

It was a nice visit and I had a lot of new impressions and ideas – and also a lot of fun! Thank you guys!

Structure-from-Motion and Flights at Allerheiligen Abbey

Last week our research group retreated for different smaller projects to Allerheiligen abbey in the Black Forrest – where we, Benjamin and me, offered a workshop on 3D modelling by Structure-From-Motion. The Abbey was a great sandbox to show the colleagues how to take the necessary photos, even at quite complex structures. We split up into small groups, each documenting different parts of the abbey.

To get a complete model of the abbey, we also planned flights with the octocopter, permission for the flights was kindly granted by the Staatliche Schlösser und Gärten Baden-Württemberg. Due to bad weather conditions, we only could fly on one day and only the eastern part of the church. Anyways we got some nice overview shots and could cover the whole area with highly overlappin nadir photos. Additionally we photographed some parts of the higher walls and details.

The photos were scaled to 2000px max and computed with four laptops in the evening to make sure the photos are sufficient for the full reconstruction. With this method we detected several missed areas which we completed on the second and third day. We also borrowed the GNSS (or DGPS) system of the Institute for Classical Archaeology (thanks to Prof. R. Stupperich!) to set ground control points to georeference the 3D model.

Overall we took 2050 photos with different cameras, providing enough data for a nice model.

Finding cropmarks

Today a colleague and me “hunted” some cropmarks with the octocopter at several spots near Heidelberg. Some results are really nice and will keep us looking for more.
We flew really high (100m above ground level) to get some good shots – both straight down and also around 45°. Some results will follow.

Flight at castle Junkernhees

Flight tracks around castle Junkernhees with the octocopter
Today the next castle on the programm. Castle Junkernhees is situated in the south east of the modern city of Kreuztal. This building is a interesting object with a long history and several building decorations and details. We flew with the octocopter several times around all sides of the building to get enough shots for a complete reconstruction using SFM.

Junkernhees_pointcloud3 Junkernhees_pointcloud4
Junkernhees_pointcloud2 Castle Junkernhees SFM Result

Ginsburg – once more

Flying at the Ginsburg castle with the octocopter
The second castle this week: the Ginsburg. It is not the most interesting stronghold and has seen many modifications and restorations after its clearance in the 1940ies. But through it’s visibility of the modern tower and the easy access it is a very good example for documenting a whole castle by Structure-From-Motion. After being here for several times since 2012, we hope to complete the 3D model with this visit.

At the ancient Site of Troizen, Peloponnes, Greece

On the 14th of April I started from Heidelberg with three colleagues to do fieldwork on the Peloponnes, Greece. The tour is separated in two parts: The first is archaeological documentation at Troizen (also known as Troezen or Trizina) while the second is to support the laserscanning fieldschool at Kalavryta organized by the LiDAR Research Group of the Geography department.

I’m involved to the Troizen Archaeology Project since it’s start in 2012 with a first survey campaign, where we tried to get as much informations about the ancient city and acropolis with the GNSS and also did a couple of 3D models of the so called “Frankish Tower” and some very well preserved roman tombs. The second visit was of very technical nature. I could manage to start a collaboration of the Institute for Classical Archaeology and the LiDAR Research Group. It was important for the project because it was possible to bring a long range laserscanner to Troizen, a Riegl VZ-6000, a unique chance to acquire the whole area of the city, the acropolis and also the surroundings in a radius of six kilometres. And this was not the only highlight – we also had a UAV from Swiss company SenseFly, the eBee, equiped with a common RGB-camera but also a multispectral camera. As tiny the bee is, as large is the area it can cover. Impressive. And last not least we had the ArchEyeAutomatic UAVs with us to complete the 3D models of the objects around but also to do some more detailed photos in several areas.

Several days, scans and flights later, we do have a lot of new data waiting for the processing and evaluation.


A hint for people visiting Poros – we were accommodated there – and like to have good and substantial food: Go to Apagio Taverna! Spyros, the host, is very nice and recommends you always excellent food.