Low-Resolution IR Sensor-based Occupancy Detection Scheme for Smart Buildings.

Accurate occupancy ditection plays a vital role in the control system of smart buildings, such as energy distribution, comfort adjustment, etc. Traditional indoor occupancy detection schemes are mainly based on RGB camera, PIR sensor, environmental sensors, etc. There are some weaknesses with these methods, such as privacy intrusion, difficulty in deployment, poor portability. According to investigations, low-resolution infrared radiation (IR) array sensors can solve the deficiencies of the above solutions. However, the IR array sensor-based method still has a few limitations, such as, the noise inherent in detected data, difficulty in multi-person detection, the lack of software and hardware interaction system. Based on the above defects, we design a new occupancy detection scheme based on low-resolution IR sensors. This scheme uses detection equipment placed in the doorway to determine the in and out status of people to achieve the purpose of occupancy detection. Among them, the combination of three frame difference and background subtraction deal with the difficulty of background segmentation. Then, a two-pass connected area division method is used as a classifier to detect multiple people in the target area. Finally, in order to facilitate real-time user perception in the construction-related information, we design a We Chat applet with strong interactive capabilities. A series of simulation experiments and actual tests verify the effectiveness of the proposed scheme.