Fakultät Informatik

Capturing Real Fluid Phenomena (Fog Machine)

The goal of this thesis is to improve the physical capturing setup in our “smoke” room. We have seven Raspberry Pi’s with cameras to capture rising fog. So far, the position detection is not very accurate and we have problems having high spatial as well as high temporal resolution in the captured videos. Based on the seven 2D videos, we want to reconstruct a 3D volume based on CT (Computed Tomography). An implementation exists already but produces a reconstruction with artifacts. This could stem from the inaccurate position detection as well as from inaccuracies in the raycaster.

Since real images from materials as fog do in general not underlie a linear image formation model, we could also consider changing our current lighting setup with diffuse light sources, e.g. by lighting the fog from behind with a LED panel, to get closer to a linear image formation.


  • Interest in Raspberry Pi's, camera calibration, physical hardware
  • C++, fundamental mathematical understanding
  • Advantageous: experiences with openCV,  mantaflow
  • Please provide your CV and transcript of records





Matthias Niessner, our new Professor from Stanford University, offers a number of interesting topics for  master theses.


PhD positions on   Computational Fabrication and 3D Printing and  Photorealistic Rendering for Deep Learning and Online Reconstruction are available at the Computer Graphics & Visualization group.