|     It may be possible, in the not-so-distant-future, to record
|     people in ultra high definition from a mile away, but the
|     'technology'  can be rendered rather useless with somthing
|     like...this
|
|     http://ramitia.files.wordpress.com/2010/02/japan-face-masks.jpg
|


>At this time, it is possible to do facial recognition at 500 meters,
>iris recognition at 50 meters, and heartbeat recognition at 5 meters.
>A newspaper open on a table can be read from orbit. 

I strongly doubt the part about reading the newspaper from orbit.  I don't doubt that the pattern of text and pictures on the  front page could be identified from orbit. ('Identifying the difference between Pravda and Izvestia'.)   An approximation I once heard is that a lens or mirror of about 4.5 inch in diameter can resolve an angle of one arc-second.  A mirror of the size of the Hubble Space Telescope (which I assume approximates that of the typical spy satellite today) is about 20x larger, so the resolution should be 20x better, or 1/20 arc-second.  That's 1/(57 degrees per radian)(3600arcseconds per degree)(20) = 1/4,100,000 radian.  From an altitude of 500 kilometers, that's about 1/8 of a meter, or 120 millimeter.  Maybe that's a pixel-pair, but it's far too large to resolve the text on a newspaper.  

The best prospect to improve on this resolution would be to use a 'multiple-mirror-telescope' technology.  Light-gathering capability isn't important in this application; high resolution is.  Making a spy-telescope out of a few different mirrors, held precisely many meters apart, could conceivable achieve resolutions substantially greater than this.

        Jim Bell