“so
+A +A +A
+i-1 + 1 + i+1
+B +B + B
Fig. 5. Location and notation of the scale transfer points A and B in the individual
photographs ;—1 to 24-1.
We assume the three adjacent photographs ? — 1, $, à + 1 to be
prepared. In each photograph the points A and B are marked with
fine crosses or holes. The points are located symmetrically with respect
to the principal points and with the same z'-coordinates as these points.
Of course, more transfer points can be used in order to obtain superfluous
measurements for the coordinate transformation. The accuracy of the
transformation procedure can be increased and checked in this way.
Stereoscopic measurements of the image coordinates of the transfer
points can be performed very conveniently and with high precision.
- and y'-coordinates are directly read for
,
In a stereocomparator the x
the left photograph. From the z- and y-parallaxes the z"- and y"-
coordinates then can be determined.
From the measured image coordinates the intersected coordinates are
computed according to the expressions (1) and (2). The base b is chosen
by estimation.
The triangulation is performed with the aid of a series of coordinate
transformations between the individual pairs of photographs. The
coordinate transformation can start from any pair of photographs.
Here we assume to start with the pair —1,0. The intersected coordinates
of the pair 0,1 are transformed to the system of the pair — 1,0 via the
identical transfer points 4—6 of the pair —1,0 and 3— 5 of the pair 0,1
respectively. Then the transformed coordinates of the points 4 and 6
of the pair 0,1 are used for the transformation of the image coordinates
of the pair 1,2 via the points 3 and 5 of the pair 1,2. In this way the
intersected coordinates of all individual pairs of photographs are trans-
formed into the temporary coordinate system of the pair — 1,0. In
the same system also the intersected coordinates of the control points
will become determined. We will use the notation strip coordinates for
those coordinates which are transformed into the system of one of the
individual models.
Next the strip coordinates are transformed into the coordinate system
of the ground with the aid of the control points. Since we have assumed
only two control points to be available the coordinate transformation
can be performed without adjustment.
ED