RE: PHOTO-3D digest 2297 -Reply - Pseudo-stereo and the like ...

[P3D Cliff Mugnier <cjmce@xxxxxxxxxxxxxxxx>]

Bob Maxie said:

>If you take a conventional stereo pair and
>switch the left and right eye views, you get
>pseudo- stereo as we all know.
>
>If you take a properly mounted pair and simply
>turn it upside down, you have essentially the
>same thing as you would have if you mounted
>the Left eye view on the right and right eye
>view on the left. When upside down, the left
>eye is looking at the right image and the right
>eye is looking at the left image. It appears
>upside down, but not in psuedo-stereo.
>
>Correct?

Yes.

>How come then, does the image appear just to
>be upside down but not in psuedo-stereo?
>

Your X-parallax displacements are still  the same; ergo, no pseudo-stereo.

On the other hand, in aerial photogrammetry, we usually employ a 
cost-saving technique called "aerotriangulation" that is designed to reduce 
the amount of surveying on the ground to provide control for the 
topographic mapping process.  As part of that process, we do stereo point 
transfer of conjugate imagery among photographs in the same flight line as 
well as to adjacent flight lines when there is an overlap that contains the 
points of interest.  This is normally done with an elaborate optical 
instrument called a "point transfer device".  The most common instrument 
around the world used for this is a WILD-Heerbrugg "PUG", where "PUG" is an 
acronym in German meaning "point transfer device".  The manufacturer is 
(was) the WILD Instrument Company in Heerbrugg, Switzerland.  The name has 
become a generic verb, "to pug" as well as an adjective "a pug point".

When one intends to pug a point on two photos, each photo is viewed by 
stereo binocular vision.  Differential zoom magnification (5X to 45X) is 
used to bring the two images in stereo at a particular area of the photos. 
 One photo is adjusted with very precise controls with respect to the other 
such that an indicator cursor (a floating dot) appears to rest exactly on 
the ground at exactly the desired location.  As one adjusts the x-parallax 
of the right-hand photo, (moving the photo ever so slightly left and 
right), the dot appears to "float in the air or dig into the ground."  We 
take extraordinary care to insure that there is absolutely no y-parallax 
between the two photos at that point, such that there is zero vertical 
separation of the floating dot.  To verify the veracity of our point 
transfer BEFORE we cause a drill to actually cut a 15-30 micrometer 
diameter hole in the emulsion to permanently mark the spot ( 0.015 mm - 
0.030 mm ), we use a pair of dove prisms in the optical train.

We turn the dove prism for each ocular so that both images rotate 180 
degrees, and we have total pseudo-stereo - hills are valleys and valleys 
are hills.  We check the dot and verify that even in pseudo-stereo, the dot 
is still exactly "on the ground".  If not, we make a slight adjustment. 
 Next, we rotate each image 90 degrees more (each in the same direction), 
and view "y-parallax" in stereo.  There appears a semblance of weird 
stereo, and we look to verify that the dot appears to be on the ground. Any 
"float or dig"  indicates a residual of Y-parallax remaining, (perhaps a 
quarter-dot). We adjust the Y-parallax knob to put the dot half the 
distance to the ground from whence it originally appears.  Next, we turn 
the dove prisms 180 degrees more in the same direction and view "pseudo 
y-parallax.  Again the view is a weird sort of stereo, and we check for any 
residual "float or dig."  Any residual parallax left is adjusted out by the 
Y-parallax knob, and the images are rotated another 90 degrees to return to 
true stereo.  Assuming we now view our floating dot to be EXACTLY on the 
ground, no float, no dig, no half or quarter dot of discernable y-parallax 
- we push the button for the electric drill and we "pug our point" on those 
two photos.  Next, we get the next photo in the flight line that overlaps 
that point and repeat the process all over again.  The same follows for any 
photos in overlapping flight lines, etc., etc.

That folks, is pseudo-stereo in spades!  Photogrammetrists do that in 
eight-hour shifts.  People ask me if I do stereo for fun?  Not very often.

Clifford J. Mugnier (cjmce@xxxxxxx)
The Topographic Engineering Laboratory
Department of Civil and Environmental Engineering
UNIVERSITY OF NEW ORLEANS
New Orleans, Louisiana  70148 -2212
Voice and Facsimile:   (504) 280-7095


-----Original Message-----
From: P3D 	P3D Bob MAXIE SMTP:BMAXEY@xxxxxxx] <>
Sent:	Thursday, September 18, 1997 1:28 PM
To:	Multiple recipients of list
Subject:	PHOTO-3D digest 2297 -Reply

Hey Gang:



------------------------------