The task of localizing and tracking people (and robots) in a dynamic environment is important in many applications: global monitoring the status of the system, supervision and surveillance, analysis of people behaviors, etc.

In this project we are studying techniques for people tracking using a stereo vision system. We have implemented a prototype and made several experiments in order to measure its precision and robustness.

The system has been experimented in different scenarios:
1) surveillance applications in office-like;
2) localization of people and robots in a domestic environment (within the
RoboCare project).