Science and Development Network (SciDev.Net)
Recently, intelligent surveillance systems based on video analytics have been used in many places. In this paper, the authors propose an intrusion detection algorithm based on spatial and temporal relationship of neighboring objects and Region Of Interest (ROI). Information of objects such as position, size, and distances to neighboring objects are updated every frame. To track intruded objects remaining within a frame, the intrusion flag of the intruding object is transferred to the corresponding object in the next frame based on the object tracking information in temporal direction. The detection accuracy and execution speed of the proposed intrusion detection algorithm were verified with real surveillance sequences.