[ Steve McCollum wrote in 001002 MMD on this subject. ]

I forgot a part from my introduction, that the primary use of
photogrammetry is in civil engineering.  I have gotten used to the
soft tissue problem which involves the mapping of non uniform curves
onto the images.

I downloaded the demo of Photomodeler Pro.  It seems to be a model
matcher, which I was thinking of writing myself.  Where I am stuck is
that I have no measurements from the original, other than the photos and
a measurement of some rectangular speakers that appear in some of the
pictures.

The Tina program will do much of the same thing, with a far less
friendly interface.  The trick is to find points that match across the
images.  I have been trying to use automated extraction as I have been
working with research code.

I am surprised that I have not found this site in a web search.  Then
I may have missed it as many of these use stock models and perform
image mapping onto them.  This program actually does seem to be true
photogrammetry.  It will be interesting to try my photos with it to see
if I can get some realistic data.

The last approach I was playing with was to extract shape from the
shading data.  I was making good progress prior to becoming distracted.
This is my biggest problem.  I work on something for a few weeks, then
a dance or an event comes along and the earlier project becomes cold.

> I was thinking that a serious rebuilder could acquire, from
> overlapping photographs, accurate dimensions of an instrument or
> its parts without having to measure them in detail.  This might be
> especially valuable to be use to recreate some measurement that
> you forgot to make!

This is something that I am actively in the process of pursuing.
I would be more than happy to share problems and places to watch out
for.  It is one thing to say that something can be done, or is being done.
It is quite something else to work the processes out for oneself.

The real problem in using these techniques for mechanical music
instruments is that one tends to take the photos first.

Prior to my most recent trip to Europe I considered building a portable
laser rangefinder, using some lenses from old 8 mm projectors.  As it
began to take shape, the thought of taking such a thing on the plane
was a bit frightening.  Round objects with wires and electronics are
not what is expected to go through airport security inspections, as
anyone who has ever hand carried a birdbox or musical box knows all
too well!

Another approach would be to use structured light.  I have in mind
to use a flash unit to project a reticule of gridwork onto the object.
This might be more practical using infrared as well.  Most digital
cameras are really sensitive to infrared light.

The real problem is the processing of the data once it's acquired.
Perhaps I should put the MMD away for an evening and run some data
through its paces again.

Julie Porter