I have the capability to do this already in hand - using off-the-
shelf, obsolete past-generation technology.  So, the capability has
been around for a long time, but I have found to my utter shock that
archaeologist are violently opposed to 3-D imaging.  After eight
years of preaching the imaging analysis mantra, I've never been
asked to apply this capability to the kind of 3-D problem presented.
So I've limited my work to 3-D reconstruction problems looking
through the microscope, but both in theory and in practice the method
applies to artifact piles in rooms or distributions of celestial
galaxies.  It's a trivialmatter of scale, but with simple
calibration, it's no problem.
 
I use ImagePlus II which is the model T ford of interactive image
analysis systems.  Once the system is in hand, the processing of 3-D
data is cost-efficient, i.e. dirt cheap, no pun intended.  The
process is simple.  A TV camera generated image (from "live" objects -
i.e. in situ, or a photo or an SEM output signal, etc.) is grabbed and
digitized automatically by my computer in about 2 seconds - not
tracing on digitizing boards, direct grabber.  Images can be
processed, edited, cut up, combined, etc. Sequences of images
can be stacked into serial sections into the three dimensional image
of an object or a pile of objects stripped layer by layer. The 3-D
image can then be rotated 360 degrees on any axis.  You can look
from above, below, behind, and even inside out. It can also be
sectioned to any level - ie. computer dissections.  The computer also
measures size and shape of each object by any number of parameters -
upwards of 20 to 30 different parameters per object.  For clusters,
it also measures clusters by frequency per unit area, percent of area
covered by objects, preferred orientation, frequency of similar
neighbors and a whole host of such things.  I have aready done things
at small scale at the rate like 150 individual objects a day,
clusters of 2000 objects in a day.  With optimum prepration I can
process as many as 5,000 objects a week with upwards of a million
measurements.  Nothing is ever near optimum and I do get bleary-
eyed and have to rest occasionally, but I've generated the equilavent
of 100,000 measures per week on large collections of objects.  So
far, I've never been asked to do a pile of archaeological stuff as a
tesselated mosaic - but it's done every day.  Fact is, I'm not
talking fantasy land here, or NASA level computers.  This stuff is
done in desktop PC's and commercially they are called virtual reality
games.  Why shouldn't it apply cheaply and easily to archaeology.  It
does, of course.
 
If the photos are good quality, in optimum computer vision, it's a
piece of cake.  Actually photos are too expensive.  Direct
videotaping is much better for two reasons.  A $3 video cassette
holds as many images as $200 - $300 worth of film photography.  Video
tape image IS a TV camera image which is grabbed and digitized
directly by the computer.
 
Contact me directly for more information, references, who has such
systems available to support archaeological research and at least
one guy who is starting a consulting business devoted to doing 3-D
imaging in archaeology.
 
Irv Rovner
[log in to unmask]
Phone: 919-515-2491
FAX: 919-515-2610