Project ArchEyeAutomatic:

Automatic aerial documentation
of historic monuments in 3-D


To get this project working, it is necessary to combine research from different areas. The main areas are Computer Science and Building Archaeology, but the part of Computer Science consists itself of different sub areas which are also diverse. Additionally there are different tasks from other research areas like Geodesy.

The main research areas are:

  • Building Archaeology
  • Robotics & Optimization
  • Computer Vision

Building Archaeology

The primary task of Building Archaeology is to document the condition of a historic building and reveal it’s history. This is a complex task and often only executed when restoration or repairing is planned. The complexion makes a thorough documentation necessary, usually consisting of sketches and drawings (including deformations) in different scales, photographs, and descriptions. Usually every stone and every joint is documented.
Using a 3D model can provide a-priori data for the researcher to get a better idea of the buildings condition as well as the extend of necessary tasks. This will also lead to better cost estimations and better plannings. A 3D model can also help in the ongoing examination for visualizing details or orientation, as well as annotations or as a ground base for plans. Even after the analysis the model is useful, as display in a museum or to compare it with models created later, to detect deformations or weathering.

Robotics & Optimization

Since we are using drones for this project, a main part is Robotics. We will create an interface between the existing Pixhawk autopilot system and the on-board computer, communicating the position and optimized route.  Optimization is a crucial point because the drone has limited energy in it’s batteries and we want to document as much as possible in one flight.

Computer Vision

The recording of the 3D data belongs to Computer Vision, or to be more specific, Stereo Imaging. We will have two equal cameras attached to the drone and use them to build a 3D world model. If available, this model will build on existing data like 3D geodata (like of NASA’s SRTM mission) or from a Laserscan.

Putting it all together

The main task of the PhD is to find existing research in the areas and connect them to get a working tool. All the single elements are connected through Scientific Computing.