being adjusted electronically. The area was overflown and photographed more than
once to ensure complete and, as far as possible, systematic coverage of acceptable
quality. Approximately 45 m was chosen as a mean flying height for the targets to
appear in appropriate size on the negatives (0.0h0 mm).
h. Processing of the photographic data
When considering the properties of the photography, which include 35 mm
negative size and potentially large and unknown rotations, the use of analytical
processing was thought to be most appropriate. At the same time, simple procedure
should be used whenever suitable. The possibility of analogue plotting was also
considered.
As already mentioned, the use of analytical methods was considered more
suitable as the main approach to the task of processing the non-metric photography.
Thus co-ordinate measurements on the 35 mm negatives were required. The use of an
expensive comparator would, to some extent, diminish the benefits to be derived
from the application of simple and inexpensive non-metric cameras and photographic
systems.
A solution to this problem was achieved by modifying an old travelling micro-
scope and the resulting instrument was calibrated and tested as a monocomparator.
A plate carrier for 35 mm negatives was attached to the base of a travelling
microscope (originally made by Cambridge Instruments Ltd.). The carrier itself
had a micrometer movement, which was perpendicular to that of the microscope.
Provision for a light source was made under the ground glass plate in order to
illuminate the negative. The reading accuracies of the microscope and the micro-
meter scales were 0.010 mm and 0.025 mm respectively. This was, by no means, a
necessary outcome of the modification, but arose from the fact that the attached
micrometer was graduated in Imperial units (0.001 inches=0.025 mm).
The calibration of the instrument, after its modification, was carried out
according to Makarovié (1969a and 1969b) using a Multiplex grid plate. This glass
plate carried a Yık mm grid covering its whole area. Prior to its use on the
Co-ordinatometer, the grid was itself calibrated on a Hilger and Watts stereo-
comparator. Multiple measurements to the grid intersections were carried out with
the Co-ordinatometer. The precision (measure of repeatability) of the readings
turned out to be higher than expected and was 0.005 mm in both x and y directions.
A third order error model was fitted to the measurements, according to Ghosh (1979).
The residuals were 0.003 mm in x and 0.007 mm in y. It is obvious that the
calibration results proved a high standard of precision for the Co-ordinatometer
(Georgopoulos, 1980).
Six black and white negatives and two colour slides were selected for the
measurements. The criteria for this selection were (1) the amount and quality of
ground information and (2) the possibility of stereopair formation from among the
selected negatives.
From the Co-ordinatometer observations and with the help of a specially
developed computer program, the plate co-ordinates of all ground control points
appearing on each photograph were produced. They were corrected for radial
distortion and referred to the origin of symmetry (Scott, 1977). This information
was then used as input data for resecting the measured photographs. The computer
program which was developed is quite simple. In the end it assumes known Z co-
ordinates and, reversing the procedure, it calculates the horizontal ground co-
ordinates of the control points. The residuals from the observed ground information
and their root mean square errors appear in Table 3. Moreover the collinearity
equations applied with the observed values give a set of image co-ordinates. Their
residuals from the comparator observations are also shown in Table 3.
Furthermore another computer program was developed to perform space inter-
section. This procedure was thought more appropriate than the conventional approach
of relative and absolute orientation, because consideration of each photograph as a
210
