Window Energy Detection for Unmanned Surveillance with PIR Sensor

A PIR sensor is widely applied for motion sensing applications such as automated door control, camera triggering and automated on-off control of lamps. In particular, it is usually used to detect the existence of moving objects in an indoor environment. However, in outdoor, false alarms come about due to environmental changes and other sources. Therefore, it is difficult to provide consistent detecting performance outdoors. This paper presents an algorithm to reduce and minimize the false alarms and provide reliable quality to a surveillance system. We gathered PIR signals outdoors, analyzed collected data and extract target feature defined as the window energy. Using this feature, we made models of target and false alarms and proposed a target decision algorithm. We simulated real PIR signals to provide the performance of the proposed algorithm.