I like seeing mainstream media reports of video analytics. Well - mainstream in a non-security industry way. Engadget is running a story on Chicago's new analytics system.
Video analytics systems are undoubtedly one of the most significant advancements to security in some time. At it's root, the systems are "smart" enough to sense specific activity and execute display commands (or other outputs) based on the activity, priority, time and other factors.
To some, that's the equivalent of computers being cops, and that brings up all sorts of interesting problems. However, all that's happening is that the system's algorithm programming is detecting behaivor based on indicators that usually mean specific human activity. If there were a market for it, the systems could alert the monitors every time someone picked their nose or scratched their ass.
One of the most overlooked factors that I see in all this is the huge amount of time and resources that's needed to set up this kind of system properly. But it's worth it. I mean, imagine every possible activity that can be sensed by the system- plot them out by importance, location, time and frequency and than program the system to alert in a way that depends on all the factors. That's huge, but at the end of the day:
"There's no way that human beings can effectively watch all those feeds, so enter video analytics. By programming algorithms, you give the camera intelligence. " (boingboing.net )
