Scene Understanding and Surveillance

Course Details

Lecturer

Martin Kampel

LVA-Nr.

193.209

Typ

Vorlesung mit Übung (VU)

Links

TISS

TUWEL

Location: Seminarraum FAV 01 A (Seminarraum 183/2), in person
1st Meeting: Wednesday March 4th, 2026; 10:00 – 12:00
Further Information:  TISS und TUWEL

In this course, the analysis of image sequences (videos, depth and thermal data) is dealt with in depth: this includes the topics of sensor technology, motion detection, object tracking, scene reconstruction and performance evaluation. Issues relating to visual, computer-aided event recognition in videos are discussed and critically scrutinised.

In this semester, the course places a particular focus on drone-based data acquisition and on analysing image and video data captured by drones. For this purpose, we use the DJI Air 3S, the DJI Matrice 4T and DJI Neo 2.

Students are welcome to do their own recordings with the drones.

DJI Matrice 4T

Content

  • Sensor technology, data acquisition, mobile platforms, multimodal data and data fusion
  • Data acquisition with drones
  • Learn to fly and record data with drones
  • Motion detection (background modelling, segmentation, etc.)
  • Object tracking (point tracking, kernel tracking, silhouette tracking, 3D tracking, etc.)
  • Scene understanding (scene reconstruction, event detection, etc.)
  • Behaviour modelling
  • Performance evaluation
  • Applications (surveillance, e-health / active assisted living, etc.)
  • Social and legal aspects