Design of elder alarm system based on body posture reorganization

Abstract

A monitoring alarm system for the elder was designed in this paper. We use a series of continuous human posture detecting to determine whether the elder is in a normal condition, otherwise to give an alarm to ambulance center. The system is mainly used to help elder person to give an alarm when he or her is in suddenness or fall ill unexpectedly at home alone. In this paper we use background difference algorithm to achieve the motion area extraction, and then determine whether the motion area is a person, if it is a person, then there is a need for posture recognition. As for posture, we use support vector machines to classification and recognition. Finally, according to a series of posture, we determine the status of the elderly.