The focus of PAMON is to advance vision-based people tracking in large surveillance camera networks to support real-time airport passenger demand management.

The extremely challenging problem of people tracking over many non-overlapping cameras in complex indoor infrastructures is tackled with:

  • a novel 3D acquisition pipeline for both camera calibration and 3D sparse reconstruction
  • semantic 3D modeling of floors and walls
  • a novel automatic technique for selecting salient and unambiguous „good people to track‟ 
Novel statistical models of short-term capacity demand predictions in airport scenarios based on PAMON are evaluated and validated on a very large image data set. This research project is undertaken in a partnership with the Austrian Institute of Technology, Vienna.