The Defense Advanced Research Projects Agency (DARPA) is developing a system for use in urban combat called "Combat Zones That See" (CTS) to better protect troops fighting in urban combat zones.
The project's centerpiece would be groundbreaking computer software capable of automatically identifying vehicles by size, color, shape and license tag, or drivers and passengers by face.
The Combat zones That See (CTS) Program explores concepts, develops algorithms, and delivers systems for utilizing large numbers (1000s) of cameras to provide the close-in sensing demanded for military operations in urban terrain. Automatic video understanding will reduce the manpower needed to view and manage this impossibly large collection of data and reduce the bandwidth required to exfiltrate the data to manageable levels. The ability to track vehicles across extended distances is the key to providing actionable intelligence for military operations in urban terrain. Combat zones That See will advance the state-of-the-art for multiple-camera video tracking, to the point where expected track lengths reach city-sized distances. Trajectories and appearance information resulting from these tracks are the key elements to performing higher-level inference and motion pattern analysis on video-derived information. Combat zones That See will assemble the video understanding, motion pattern analysis, and sensing strategies into coherent systems suited to Urban Combat and Force Protection.
This project really is motivated by military needs.
Military Operations in Urban Terrain are fraught with danger. Urban canyons and abundant hide-sites yield standoff sensing from airborne and space-borne platforms ineffective. Short lines-of-sight neutralize much of the standoff and situation-awareness advantages currently rendered by U.S. forces. Large civilian populations and the ever-present risk of collateral damage preclude the use of overwhelming force. As a result, combat in cities has long been viewed as something to avoid. However, modern asymmetric threats seek to capitalize on these limitations by hiding in urban areas and forcing U.S. Forces to engage in cities. We can no longer avoid the need to be prepared to fight in cities. Combat zones That See will produce video understanding algorithms embedded in surveillance systems for automatically monitoring video feeds to generate, for the first time, the reconnaissance, surveillance and targeting information needed to provide close-in, continuous, always-on support for military operations in urban terrain.
You can read DARPA's contractor FAQ as a PDF.
DARPA says this system is being developed for use in foreign urban battlefields and is not meant for domestic use. This certainly seems like an honest statement of their motivations. However, such a disclaimer tells us little about how the system will eventually be used (though it certainly will be used for military purposes). First of all, once it is working are they going to turn down requests from, say, the City of New York, to install some cameras to watch for known terrorists? Seems unlikely. Secondly, once DARPA demonstrates some capability companies not involved in the development will rush to produce equivalent systems if the demand exists among law enforcement agencies. There are lots of engineers and scientists who could assist in the development of such a system.
However, just because DARPA's project will eventually enable large scale surveillance of cities which are not war zones (okay, at least not military war zones) does not mean that the project should be opposed by those who are opposed to increased domestic surveillance by governments. Civil libertarians who may wish try to stop the growth of the surveillance society by lobbying against government funding of the development of the enabling technologies in projects such as the DARPA CTS are at best fighting a delaying action. The ability to automatically recognize specific faces or cars or to read license plates is coming sooner or later as computers become faster, sensor quality improves, and visual pattern matching algorithms improve. DARPA's efforts might speed up the development of the needed technologies but their development is inevitable.
Update: The London Underground is about to test a software system called Intelligent Pedestrian Surveillance System that does automated computer monitoring of digital cameras at tube subway stops.
If the trial due to go live in two London Underground stations this week is a success, it could accelerate the adoption of the technology around the world. The software, which analyses CCTV footage, could help spot suicide attempts, overcrowding, suspect packages and trespassers. The hope is that by automating the prediction or detection of such events security staff, who often have as many as 60 cameras to monitor simultaneously, can reach the scene in time to prevent a potential tragedy.
The software is marketed by Ipsotek (Intelligent Pedestrian Surveillance and Observation Technologies), a firm that is a spin-off of the research done by their managing directory Dr. Sergio Velastin at Kingston University.
Dr. Sergio A Velastin obtained his doctoral degree from the University of Manchester (UK) for research on vision systems for pedestrian and road-traffic analysis. Joining the Department of Electronic Engineering in Kings College London (University of London) in October 1990, he became a Senior Lecturer and founded and led the Vision and Robotics Laboratory (VRL). In October 2001, Dr. Velastin and his VRL team joined the Digital Imaging Research Centre in Kingston University, with which he is still associated, attracted by its size and growing reputation in the field.
Note how the basic research being funded by a variety of governments in vision processing leads inevitably to automated systems that can monitor and detect patterns in human behavior.
|Share |||Randall Parker, 2003 July 10 11:54 AM Surveillance Cameras|