The Eyes Have It
Machine vision–and its ramifications for security both inside and outside the enterprise–is on the cusp of widescale implementation thanks to advancements in camera, lighting, and processor technology primarily sponsored by the Defense Advanced Research Projects Agency.
Unlike flesh-and-blood security staff, machine vision technology is free from distraction, does not forget or become exhausted, and is scalable, easily upgradeable, searchable, archivable, and network-compatible; furthermore, the technology is becoming more and more affordable.
The machine vision application with the most momentum is intelligent optical character recognition, which can recognize characters emblazoned on physical objects, and its most critical application seems to be license plate recognition.
Machine vision’s biggest potential application is centered around object and behavior recognition, which could be employed to detect signs of tampering and damage.
The U.S. Bureau of Customs and Border Protection has upgraded its security by installing people recognition systems, and bureau representative Bill Anthony reports that the number of false positives along the Mexican border has dropped to zero thanks to the upgrade.
People recognition systems can also be deployed to identify incidences of tailgating and piggybacking.
One of the ways in which machine vision is expected to revolutionize corporate security is to make CSOs more cognizant of signs of unusual activities, or flags, rather than outright violations.
The enterprise-wide perspective afforded by machine vision could also benefit departments and divisions outside of security: For instance, observations of employees working overtime could be useful to the personnel department, while the detection of an odd vibration might tip off facilities management that a key piece of equipment is about to malfunction.
Abstracted by the National Law Enforcement and Corrections Technology Center(NLECTC) from CSO Magazine (12/03); Hapgood, Fred.