A a ht — eh vd
LU e.
3.3 Phase three
In 1995 a third procurement was made, confirming
OSI's confidence in digital photogrammetry. This
reflected a need for additional capacity to meet deadlines
on some of the programmes. The intention was to focus
the new units on data capture for the large scale rural
database, which is being totally resurveyed. The
opportunity was taken also to upgrade the system
generally. Thus the slower DSW 100 scanner was traded
in for a second DSW200. Five new stereoscopic
workstations were purchased and the monoscopic one
was upgraded to stereo. The first DSW200 was provided
with a more powerful host computer and its existing host
redeployed as the basis of one of the DPWs. The
scanners were also equipped with dual screens and full
stereoscopic viewing, in order to run the Helava
Automated Triangulation System module. They share a
60 GB RAID array for top disk performance. Total disk
capacity on the 17 workstations is around 100 GB. Many
of the software modules were issued on a floating licence
basis to reduce costs. Importantly, considerable efforts
were made to improve network performance, on which
the transfer of the 80 MB images obtained by scanning at
a resolution of 25 pum placed considerable demands.
Switching hub technology consisting of a wave switch
with 80 MB per second total transfer rate and a fibre
optic link to the DECnet was added: each scanner has a
Fast Ethernet (100-BaseT) line and each DPW a
dedicated 10-BaseT connection. This gives the scanners
much faster links, eliminates network collisions and
enables any DPW to contact the DECnet directly. The
current system is summarised in Figure 2.
3.4 Software for quality control
Towards the end of this phase, OSI were commissioned
by the Department of Agriculture to quality control
newly generated orthophotos obtained from 1:40,000
aerial photography of Ireland. The orthophotos were
required for the validation of claims made by farmers for
subsidies under the European Common Agricultural
Policy. In all there were 27,000 orthophoto tiles covering
an area of 2 km x 1.5 km each. Each tile was 3 MB in
size and delivered by the contractor on CD-ROM. This
was a daunting task and quality control software to semi-
automate the process was commissioned from Helava.
The need for quality control of orthophotos has drawn
comment from users of digital photogrammetric systems
(Armenakis, Regan and Dow, 1995; Manzer, 1995) and,
indeed, ASPRS has instituted a committee charged with
developing appropriate standards (Nale, 1995).
The software performs two groups of straightforward,
useful checks. There are batch checks on all orthophotos:
presence; existence of associated support files;
correctness of formats, for example TIFF; confirmation
that files contain images; grey scale histograms. These
are complemented by interactive checks on a sample:
rmse of measured coordinates of ground control points;
image quality; visual checks on mosaicking and edge
matching, and image completeness. At the time of
writing, this software is operational and initial
experiences have been satisfactory.
3.5 Phase four
At the time of writing, OSI is reconsidering the
specification requirements for the rural mapping
929
database. This review will include a comprehensive
redefinition of the content and accuracy of the database.
It is likely that the specification will vary for different
parts of the country depending on terrain and population
density. Without preempting the result of the review, it is
reasonable to say that an alternative to stereoplotting is
being considered. Head-up digitising on a monoscopic
screen from digital orthophotos or monoscopic plotting
from raw imagery with on-line DTM is being examined.
OSI already has an accurate DTM generated by the
Helava systems. New approaches are being assessed so
that progress with the resurvey and the creation of the
rural database can be accelerated. Moreover, cost factors
plus the shortage of operators noted below make
stereoplotting less attractive. The review should be
complete by the end of June 1996, when it is expected
that a decision on the fourth procurement will be made.
4.0 PERSONNEL ISSUES
At first there were some doubts about digital
photogrammetry on the part of the instrument operators,
many of whom had considerable experience in
triangulation or compilation, primarily on analytical
plotters. Moreover, it proved impossible in Dublin to
recruit additional staff to operate the new equipment
being procured in the public sector, owing to an embargo
on recruitment. This problem was addressed by
decentralisation: the Kern and Leica analytical plotters
were relocated to the regional offices, as shown in Figure
2. Moreover, the change from revision to resurvey caused
the 100 revision staff to be reduced to eight, but much of
this was achieved by relocation to regional offices.
Dissatisfaction with the daily work decreased as the early
tendency to subject DTMs to excessive editing was
curbed, workflows suited to digital rather than analytical
photogrammetry emerged and the software became more
user friendly, for example the triangulation module and
ATLAS. KLT have implemented the latter on the DPWs
such that it runs as smoothly as on the analogue and
analytical workstations. As staff become “Unix literate”,
lead operators acquire skills in systems administration,
which reduce any residual fear of the technology. Today,
operators prefer digital to analytical workstations.
Figure 3. One of the two digital photogrammetry rooms,
showing a number of the DPW770s with SPARCstation
20s and Nu Vision stereoscopic viewing systems
Tremendous esprit de corps has been engendered by the
sheer size of the new installation. The digital systems
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B4. Vienna 1996
