I love the concept.* Object recognition is probably the hardest
   part; sometimes "outline" in drawing programs like Paint.NET or
   fancier ones can do some of the work for you; there's algorithms
   for that sort of thing... but I think pushing images through a
   series of automatic commandline filters would be faster and
   hidden. But then there's the cataloging too.* I love the concept
   though; but it's not a flaw in you or your abilities; it's just
   one of "those areas" that computers really suck at right now;*
   you're basically turning 2D into 3D by tagging... seeing it from
   above.. and computers really suck at anything 3D just yet.
   They're not meant for it.* Even the 3D cards with their matrix
   transformations don't 'know' objects well yet.* They need a
   series of templates or better object outlining capabilities.
   It's getting closer though to where that sort of thing wouldn't
   be hard to do.* I honestly thought standard computer tech would
   be there by now, but that's from growing up with Sci-Fi Tech.