Suppose a city has installed hundreds of surveillance cameras at strategic locations in
busy streets. All the captured video (each stream being 640x480px, 24-
bits (RGB), 25fps) is streamed to a central observation post in real time.
Now an efficient system needs to be designed to allow for real time data mining and support continuous querying these video streams for abnormal events, such as explosions, fires, etc..
1) What would such a system look like?
2) What are the computational costs and memory requirements for such a system?