A Computational Framework for Determination and Exploitation of Social Network Models from Wide-Area Persistent Surveillance Imagery

Wide-area persistent surveillance promises to revolutionize domestic situational awareness by providing real-time visual imagery of events, individual actors, and groups of interest to the national homeland security mission. However, fully realizing the potential of this technology requires computational tools capable of extracting actionable information from many highly dense data streams. At present, analysis of persistent surveillance imagery demands significant time and manpower resources as intelligence analysts tediously examine such data to identify suspicious activity. With the granularization of homeland security threats, the current approach and its associated strain on human resources must be reexamined with an eye towards automation. In order to realize the promise underlying wide-area persistent surveillance technology, an interdisciplinary approach is necessary. This brief argues that by 2 merging analytic tools from the normally disparate fields of behavioral science and engineering systems theory, a methodology for automatic target identification and exploitation can be realized.