Real time tracking and detection of unusual circumstances of elderly people with RGB-d camera

In this paper, a real-time system, that monitors on home-alone elderly or sick people and produces an alarm whenever they act unusual like falling or fainting via RGB-d camera as surveillance, is designed and developed. In order to check person´s state whether s/he is standing, sitting or laying down horizontally, skeleton information that is collected by RGB-d camera is used. To determine if person was fallen or not, instant change of joints´ X and Y coordinates, that are saved in skeleton information, are evaluated.