Target Classification Using Pyroelectric Infrared Sensors in Unattended Wild Ground Environment

Pyroelectric Infrared (PIR) sensors are widely used as a simple presence trigger for alarms and a reliable counter for people in security home and smart office. Most paper have focused on the design and application of such systems, however, few research has be done on the detecting of targets in an outdoor environment using PIR sensors. In this paper, we realize a detecting system by means of PIR nodes to monitor outdoor targets more than 20 meters away from the PIR detectors. Furthermore, because of velocity difference, by extracting time domain amplitude, signal length, maximum frequency and corresponding frequency amplitude as features, we successfully classify people, wheeled vehicle and tracked vehicle in the unattended wild ground environment. Our detecting and classifying results confirm the average accuracy is 85.67% and 82.67% when PIR detectors are deployed 20 meters and 30 meters away from the area of interest (AoI), respectively.