Does anyone know of an algorithm for the DLT
or similar program, which allows you to vary the number of camera
views used for determination of unknown points? We fairly regularly
have a situation where we have, say, two clear views of a marker, and
a third which is hidden. At the moment we are faced with guessing for
the third camera, which then goes into the calculation and obviously
degrades the result. I'm willing to sort my way through the DLT if
necessary, but I thought someone else may have already done it and be
willing to share it.
Thanks for any help/comments.

Bob Marshall