----------------------------Original message----------------------------
Virginia Hetrick writes:
>
>I had a doctoral student working with some unrectified air photos from
>the Argentine Andes come by and ask about software to join these
>digitally so she could have a photomosaic of her study area. The
>problem with running the "stitcher" programs that most handscanners come
>with is, of course, that the air photos fall apart at the edges due to
>camera angles and differences in atmospherics.
>
Actually, there is more than just camera angles and refraction at play here
if the images fall in the Andes. The relief distortion in the photographs
has far more to do with the image-mosaicking problems than either of the
other two factors mentioned, and simple corrections which act globally on
the entire image aren't likely to help much unless it is very(!)
small-scale photography.
In order to obtain _seamless_ matches for mosaic purposes in mountainlous
terrain, your student will probably need to have digital orthophotos
created of the images. These require (at minimum) information concerning
ground control for scaling purposes and a digital elevation model covering
the area of interest. For a *rough* orthophoto (which may suit your
student's purposes), you *might* be able to scale the requisite ground
control from an existing topographic map. (It's not the most accurate
approach, but it may suffice for simple image-mosaicking purposes.)
However, you will also need digital elevation model information to drive
the orthophoto creation process, which may be a bit harder to find. (U.S.
Defence sources or governmment folks in Argentina may have such
information, or -- if you have overlapping stereo coverage -- you could
arrange to have it done by a mapping firm.) I'm not sure if there is any
public domain software available for digital orthophoto creation or not.
Many commercial systems have such capabilities but they are expensive.
Bottom line -- If you're willing to settle for less than seamless joins
across the photo edges, than you may be able to accomplish it with the
tools at hand. However, it will be expensive to accomplish *in this
particular type of terrain* unless you have the photogammetric skills,
production tools, source material and elevation data handy. Technology is
becoming simpler to use and more affordable, but I don't THINK such tools
are readily available in the public domain yet. (I'd love to be proven
wrong, though, and hope you'll let me know if she does get it accomplished
to her satisfaction with the tools available.)
Best regards,
-----
Dr. David Coleman <[log in to unmask]>
Geographical Engineering Group Tel.1-506-453-4698
Department of Geodesy and Geomatics Engineering Fax 1-506-453-4943
University of New Brunswick,
P.O. Box 4400, Fredericton, New Brunswick, CANADA E3B 5A3
|