Outdoor infrared video surveillance: A novel dynamic technique for the subtraction of a changing background of IR images

For security applications, automatic detection and tracking of moving objects is an important and challenging issue especially in uncontrolled environments. Recently, due to the decreasing costs and increasing miniaturization of infrared sensors, the use of infrared imaging technology has become an interesting alternative in such applications.In this paper, a framework is proposed to detect, track and classify both pedestrians and vehicles in realistic scenarios using a stationary infrared camera. More specifically, a novel dynamic background-subtraction technique to robustly adapt detection to illumination changes in outdoor scenes is proposed. We noticed that combining results with edge detection enables to reduce considerably false alarms while this reinforces also tracking efficiency. The proposed system was implemented and tested successfully in various environmental conditions.