Fall Detection and Personnel Tracking System Using Infrared Array Sensors

Body falling is a major risk of the older people. This paper proposes a non-contact scheme for detecting human body state using infrared array sensors, which has the advantages of low cost, easy implementation and high detection accuracy. For achieving real-time monitoring, rather than simple motion analysis, the system identifies suspicious points of body falling by simple positioning, which attempts to achieve the real-time implementation and reduces the analysis of irrelevant data. In addition, considering the influence of the radiation temperature result of the human body, the falling characteristics are included, and the clustering algorithm is proposed to tackle the problem of inaccurate determination of the area caused by the Otsu area detection. Then, a two-layer threshold system model is proposed for excluding jogging movements that are similar to the falling feature. Finally, the features are engaged with the random forest classifier in order to improve the classification rate. Experiments have shown that the proposed fall detection scheme is highly feasible and behaves good fall detection accuracy.