Score and event identification based on ideal blob-size

[lazyleopard]

It seems to me that there are occasions when interesting events might be identified because they tend to produce blobs of a certain size. If this size is small then the final score for the event can be very low. Non-events (like the sun coming out) tend to swamp the score tables because they result in large blobs. If the score for a blob was based not on its absolute size, but on how closely it matched an expected size, then interesting events should get higher scores than non-events. The catch is that the score for each blob would have to be determined separately rather than just using the total blob area at the end of analysis.

Of course, the next step along the way would be some form of target recognition (Is it a car? A person? A fox? A cat?) but that'd probably require a system that could learn what things looked like....

I'm not quite certain how much of this is already in place, but I've burrowed a bit through zm_zone.cpp, and I don't think there's anything like it yet.