A total of 52 field controlled aerial photographic paper
prints, and corresponding film transparencies at 1:80,000-
scale were selected, covering the image area of the four
RBV subscenes. Inspection of these photographs showed that
a total of 47 points with position descriptions were suit
able for controlling the mosaic. Eight of these 47 points
were common between adjoining subscenes. Five additional
points were also selected as tie-points between the
subscenes.
Transfer and Marking of Control
Horizontal control points were directly transferred and
marked from the field controlled film transparencies to the
appropriate RBV subscene, by using a Kern PMG-2 Zoom Point-
Transfer instrument. The zoom optics of the PMG-2 permit a
scale difference of up to five times. As the scale ratio
of RBV subscene to the aerial photographs was 6.33, the
aerial photographs were reduced to a scale of 1:160,000.
This reduction permitted the relative orientation of the
control film transparencies to the RBV film transparencies
for stereoscopic viewing in the PMG-2. The corresponding
field control photograph was used to identify on the
control transparency the exact location of each point.
Each point was then marked, using the PMG-2's high-speed
precision drill, with an aperture size of 40 micrometers.
RBV Subscenes Measurements
After the four RBV subscenes were marked, they were
measured on a Kern MK2 Monocomparator. The marked points
were easily identified through the monocular microscope
which has 28 times magnification. An Altek AC 189 data
acquisition system and Univac 1170 card punch recorded the
measurements on cards for computer processing.
Analysis of Subscenes
The measured image points on each subscene were adjusted
to the corresponding State plane rectangular coordinate
values by means of a plane-to-plane similarity transforma
tion program. This transformation produces a least square
adjustment of the measured points to the control points.
Inspection of the residuals showed that there was only one
control point having excessively large residual errors.
This was most probably due to misidentification from a
field control photograph, and the point was therefore
deleted. Individual measurement of subscenes A, B, C, and
D at 1:500,000-scale resulted in residual mean square error
(RMSE) from the similarity transformation.
Subscene Horizontal Control Points RMSE
A
20
50 m
B
2
indeterminate
C
24
43 m
D
11
43 m
Common tie points were also checked for comparison of
adjusted coordinate values on adjoining subscenes. The
measurements of the four subscenes were then processed
together in a least square semianalytical block adjustment
program.