The purpose of this project was to develop a system capable of tracking multiple people in real time video streams and moreover it should be capable of having an integrated robotic spotlight capable of lightning a person path.

     The implemented solution has all the desired tools to track people moving in a room monitored by a camera. Furthermore the application is capable of tracking people during occlusions with good accuracy and is able to translate that person coordinates into commands so that the robotic spotlight can lighten a person path.

     The results show that the application is reliable while is tracking people, nevertheless it faces some difficulties when new people enter the camera field of view. This work also allowed to find the good variables to set up the Histogram of Oriented Gradients and Kalman Filter algorithms that are more suitable for person tracking. Another advantage of the Person Tracker is its versatility as it can track people without the need of markers, specific environment conditions etc.

     As this is a very complex area there is still a long way to go to build a versatile and flawless person tracker and many more hours of testing would be needed to find the best way to correctly detect and track a person. Still the HOG algorithm working together with the Kalman Filter appears to be a good choice to track person movement in the tested scenarios.