rget 
lare 
ical 
nel, 
any 
nal 
ove 
10rs 
ck. 
the 
s, R 
s of 
phy 
the 
11 at 
res. 
ata 
s of 
late 
ated 
rder 
ble. 
lice 
road 
eing 
ould 
and 
road 
utes 
up 
and 
l'wo 
fore 
very good coordination would be required in the 
helicopter so that all the required photography at each site 
could be obtained within the ten minute restriction. It was 
decided to photograph the smaller of the two sites, the east 
portal, in the morning slot to enable experience to be 
gained before the larger site at the west portal was 
photographed in the lunchtime slot. 
The first sortie was flown in July 1991. A crew of three, 
in addition to the helicopter pilot, was employed. The 
navigator sat in the left front seat beside the pilot and 
provided guidance instructions to the pilot. This 
crewmember also operated the Linhof. The camera 
operator sat in the left rear seat with the third crew member 
on his right to assist with and organise the changing of the 
plates. An immediate problem not encountered during 
pre-flight testing was that the T-Max plates were slightly 
undersize for the UMK plate holders. This, added to their 
thinness of only 1 mm, made them vulnerable to vibration 
and movement within the plateholders. As a result a 
number of these plates cracked as the dark slides were 
being replaced in the plate holders. With the ten minute 
slot constraint per site it was not possible to take these 
exposures again. 
It had also been decided to take whatever photography was 
possible of the site from the ground using the Zeiss Jena 
10/1318 camera with 100 mm lens. This was as an 
insurance against some of the aerial photography not 
being useable because of blurred targets, etc. For this 
ground photography Agfa Avipan 100 ISO 130 mm by 
180 mm (5.1 inch by 7.1 inch) glass plates would be used. 
After the first sortie it was decided not to use the Linhof 
on future sorties. Although much simpler to use than the 
UMK with glass plates the Linhof's smaller focal length 
and resulting smaller image scale resulted in some targets 
not being resolved on all of the images at the maximum 
taking distance chosen for the UMK. Asthe UMK system 
had proved workable and had yielded good results it was 
decided that further hiring of the Linhof was an 
unnecessary expense. 
A number of modifications were carried out after the first 
sortie. Firstly the stomach plate for the UMK mount was 
greatly enlarged and curved to spread the pressure on the 
operator's abdomen. Additional handles were attached to 
the sides of the plastic mount to make pointing easier. 
Switches for firing the fiducials and shutter were attached 
to these handles. An improved sighting system was also 
built (Figure 3). The most significant change however 
was in the choice of emulsion for the UMK plates. Firstly 
it was decided not to use Kodak T-Max 100 plates again 
as the breakage rate in the first sortie was unacceptable. 
Additionally a faster emulsion was desirable as the second 
and third sorties were to take place during the winter and 
there was a strong possibility that lighting conditions 
would not be as good as during the first sortie. It was 
decided to use Kodak T-Max 400 which could be pushed 
to a rating of 1600 ISO during processing. However it was 
not possible to obtain this emulsion on plates so a decision 
was taken, somewhat reluctantly, to use film glued onto 
plates. 10 inch by 8 inch (254 mm by 203 mm) sheets of 
T-Max 400 were purchased, cut to size and mounted on 
old Avipan plates using Scotch Spray Mount. 
It was realised that the resulting film unflatness would 
result in a lower precision for the second and subsequent 
Sorties. Previous tests with the 100mm UMK using cut 
film on plates had indicated that errors in image 
  
Figure 3 UMK in redesigned and improved mount 
coordinates of up to 60 microns could be expected. 
However with the longer focal length, 300 mm, and 
resultant narrow lens angle these errors would be reduced 
to the 20 micron region and, with sufficient photographs 
in the bundle, would not result in unacceptably large errors 
in the coordinates of the targets. This loss of precision was 
deemed acceptable if the faster film rating would 
guarantee the acquisition of measurable images. 
The second sortie was flown in October 1991 and a third 
sortie, much delayed by weather, was flown in April 1992. 
4. MEASUREMENT AND ANALYSIS 
All photographs were measured on an Intergraph 
Intermap Analytic (IMA) analytical stereoplotter 
operating in monocomparator mode. Each fiducial and 
target point was measured three times and the 
observations meaned. The resultant mean standard 
deviations for the plate measurements were of the order 
of +/- 3 microns. These observations were transferred to 
a Sun Sparcstation where they were run through City 
University's General Adjustment Program (GAP). This 
is a program written in C for adjusting any combination 
of geodetic and photogrammetric observations. In this 
case there were only photogrammetric observations. A 
feature of GAP is that any combination of datum 
constraints can be used (ie points can be fixed, weighted 
or free). GAP will recognise any datum deficiency and 
automatically border the matrix of normal equations 
accordingly. In this instance no target points were held 
fixed and "inner constraints" were used where the starting 
values of the coordinates of the target points were used to 
define the datum. The camera positions were not used in 
the definition of the datum as these would vary from sortie 
to sortie. 
Table 1 lists the results from GAP for various 
combinations of data for epoch 2. These are: using just 
the aerial photography; using just the ground 
photography; and using both sets together. It can be seen 
that there is a great increase in precision when both the 
aerial and the ground photography are used. It was 
noticed that when using the a priori standard deviations 
for the plate observations for epoch 2 that the variance 
factor after running GAP with the aerial photography only