COMMERCIAL UNDERWATER PHOTOGRAMMETRY
THE FIRST DECADE
John Leatherdale, Photogrammetric Consultant,
10 Bartrams Lane, Hadley Wood, Herts, EN4 OEH, England
and
John Turner, Managing Director,
Camera Alive Limited,
21
Robert Leonard Centre,
Kirkhill Industrial Estate, ABERDEEN, AB2 OES, Scotland
Telephone (0224) 771188
FAX (0224) 771718
ISPRS Commission V, Working Group Vj4, Kyoto, Japan, July 1988
INTRODUCTION
The first major commercial contract for underwater photogrammetry
was completed in 1978 and reported to ISP at the Hamburg Congress
by Welsh et aI,
1980 (1). This task involved the urgent measurement of the damaged leg of a new oil production platform in the
North Sea in order to prefabricate a repair plate,
and was
successfully accomplished by photogrammetry after other methods
of measurement underwater had failed.
Ten years
later it is appropriate to review the developments
which established photogrammetry as a
reliable,
accurate and
economical method of measurement underwater.
The combination of
stereoscopic underwater photography and photogrammetric analysis
is now recognised and used operationally by all the major oil
producers in the North Sea and is now being adopted in other
offshore areas where permanent structures have to be maintained.
Video cameras,
currently used to navigate remotely operated
vehicles
(ROVs)
and to control underwater operations from the
surface,
are also being used to obtain stereoscopic imagery with
the promise of real-time videogrammetry for measurement where
instant results are more important than resolution and accuracy.
This may be the major achievement of the next decade.
Following that first contract,
an association was formed in 1980
to bring together the three essential components for a successful
operational
service:
expertise in diving
and
underwater
photography provided by Camera Alive Limited of Aberdeen;
underwater camera manufacture by Underwater and Marine Equipment
Limited; and photogrammetric analysis by Hunting Surveys Limited.
The early operational experience of the association has been
described in Turner and Leatherdale 1982 (2) and Leatherdale and
Turner 1983 (3).
An alternative system, described by Baldwin and
Newton 1982 (4),
was operational for a short time, and a general
review of the status of underwater photogrammetry was given by
Newton 1984 (5) at the ISPRS Congress in Rio. Since that time the
manufacture of UMEL cameras has been taken over by Camera Alive
Marine Equipment Limited (CAMEL) and,
following the closure of
Hunting
Surveys,
the
photogrammetric analysis
has
been
transferred to Plowman Craven and Associates. In ten years photogrammetry has become well established as an economic and reliable
method of inspection and measurement underwater amongst offshore
oil and gas producers around the world and is attracting the
attention of other disciplines involved in underwater operations.
Development in all aspects of underwater technology has been
spectacular over
this period,
in response to the needs of
offshore oil production extending into ever deeper and more
difficult waters.
The role of photogrammetry in this development
has been identified by the Society of Underwater Technology and
the Offshore Energy Technology Board who chose to present
the
first
David Partridge Award for Underwater Technology to
John
Turner in 1988 in recognition of a major achievement in underwater technology.
ADVANTAGES AND EXPECTATIONS OF PHOTOGRAMMETRY
Photogrammetry is in its element underwater.
It solves the very
problem which is uppermost in the minds of offshore operators,
engineers,
marine biologists,
oceanographers and underwater
archaeologists
- a means of reducing the horrendous cost
and
uncertainty
of success in recording and measuring
things
underwater.
Saturation divers may cost up to
3000 per hour in
the North Sea,
and remotely operated vehicles
(ROVs)
between
200 and
500 per hour.
Large quantities of photographs can be taken in a very short
space of time,
for subsequent analysis by experts in the comfort
of a well equipped office.
The reductions in operating costs are
spectacular:
structures which took days days to inspect
and
measure by physical contact,
can be photographed in a few hours.
Against these
savings in underwater operations,
the cost of
photogrammetric analysis is trivial.
There are other advantages.
Photographs provide an unbiased and
unselective archive of information at known times, independent of
human memory and foresight,
and can be re-interpreted in the
future for entirely different purposes. Corrosion pitting or weld
failures may only be noticed at an advanced and serious stage,
but re-examination of earlier photographs may help to reveal
their cause and rate of development.
The results of photographic
interpretation and measurement are very much more reliable,
because
they are undertaken by specialists in
relatively
unharrassed surroundings,
and can be repeated,
and subjected to
second and third opinions.
The only weakness in the system is the dependence on obtaining
satisfactory photography,
which in the case of photogrammetric
measurement means usable stereoscopic coverage.
As ROV operators
become more accustomed to taking photography for
particular
purposes their competence has improved but thorough briefing is
still essential.
There may also be obstructions in confined
spaces which prevent the acquisition of the required pattern of
photography.
The maximum stand-off distance at which acceptable
image quality can be obtained is generally about 2 metres even in
clear water which inhibits the use of the technique for extensive
seabed mapping.
V-35
Divers and ROV operators,
generally not specialist users,
can
concentrate on acquiring suitable photography and no longer have
to get involved in the interpretation and measurement.
Users can
analyse the evidence for themselves. Having removed the divers as
a
source of confusion,
the users now want to be able to
take
measurements themselves without having to specify their requirements in advance to photogrammetric specialists.
In this respect
photogrammetrists
have been conspicuously
unsuccessful
in
simplifying their expertise and equipment to make it accessible
to experts in other disciplines.
For photogrammetry to become an
every-day tool for underwater measurement and recording,
it must
be made available to the end users as black box technology.
The
photogrammetric
specialist who does not happen to be a pipeline
engineer or marine biologist is getting in the way and degrading
the extraction of information.
The challenge for
photogrammetrists is to produce a
simple and reliable system
for
interpretation and measurement which is portable and truly user
friendly.
Such a
system has been designed and described by
Baldwin 1984 (6) but is still seeking financial backing. The need
exists and users are more concerned about modest cost and ease of
operation
by non-photogrammetrists than with accuracy
and
refinement.
USER NEEDS
User needs are often remarkably simple and undemanding and it
is
a mistake for photogrammetrists to strive to extract the maximum
amount of information and the highest attainable accuracy from
the available photography
It is also counter-productive
to
specify unnecessarily complex patterns of photography which are
unlikely to be achieved.
If the system fails to produce rapid
and reliable results,
it becomes a very expensive disaster.
It
is also essential to understand users' needs and to present
the
results
as digital or graphic
information which is
both
convenient and meaningful to them.
Most measurements may be obtained from individual photomodels but
some critical dimensions or angles involve complex patterns of
photomodels.
Some
measurements
require
judgement
and
interpretation, even imagination, as well as factual observation.
The depth of corrosion pits or the wastage of annodes must
be
related to the original surface before corrosion started,
which
can only be estimated.
The observed surface of a damaged
platform member has to be related to the probable axis of
the
member before the damage occurred.
It is understandable that users would prefer
measurements and judgements for themselves.
to
make
such
FIGURE 1. CAMEL CI800 camera, 800 exposures,
9000 metres water depth.
FIGURE 2. Typical CAMEL oceanographic photographic system.
FIGURE 3. CAMEL CI7070 70mm photogrammetric camera with Ivis 48 character data display.
FIGURE 4. A pair of CI7070 cameras mounted on
a small ROV.
FIGURE 5. The CAMERA calibration system.
CAMERA DEVELOPMENTS
The performance and reliability of underwater cameras sui tab
for
photogrammetry have improved to the point where equipment
failures are very rare, except under extreme mishandling.
35 mm cameras remain popular with diving enthusiasts in shallow
water operating within modest budgets,
and for oceanographic
surveys where very large quantities of photogra
are
taken
often at great depths.
Figure
1
shows t
CAMEL CIBOO
oceanographic
35 mm camera which is capable of taking
1600
exposures on thin base film at depths down to 9000 metres.
Figure 2 shows a typical oceanographic system in a deployment
frame.
Neither of these applications is primarily concerned with
photogrammetric measurement.
The
70 mm format has become the offshore oil industry standard
for
inspection and photogrammetric analysis.
Cameras of
this
format are still compact enough to be mounted in pairs on ROVs or
handled by free swimming divers.
The larger image scale,
better
lenses,
and reseau plates to improve and monitor
1m flattening
significantly increase both the photographic quality and
the
accuracy of photogrammetric measurement.
Since 1980,
UMEL 70 mm
cameras have been used for inspection and photogrammetry
most
of the North Sea oil producing companies and their diving
contractors.
The basic design, with Zeiss Biogon 38 mm f4.5 lens
and domed front
port,
has not changed fundamentally but
the
current
version,
the
CAMEL CI7070 shown in
Figure
3,
incorporates many improvements. Prontor electro-magnetic shutters
provide flash synchronisation between pairs of cameras.
Daylight
loading magazines take up to 70 exposures,
or 200 with thin base
film,
and a sensor detects film jams.
An IVIS 48 character data
chamber can be fitted to record the number,
time and location of
each exposure.
The standard housing is rated for depths of 600
metres but housings for operating at depths of 6000 metres
are
available.
Over sixty of these cameras have been built and a
third of them have reseau plates for photogrammetric use and are
calibrated after each major service.
They have proved remarkably
robust and reliable and even the ones which have been lost at sea
are probably still in good working order.
LIGHTING
Artificial illumination has to be provided by powerful flash
units and is particularly critical for stereo photography. Stereo
pairs taken sequential
with a single camera are difficult
to
observe
satisfactorily if the object is illuminated
from
different directions.
Simultaneous stereo with two cameras is
preferred to achieve consistent illumination and to obtain the
required
stereoscopic overlap.
This requires both master and
slave cameras to be synchronised with a
single flash system.
Previous
flash systems were
not satisfactory and CAMEL now
manufacture a range of units of their own with power outputs of
up to 1000 joules and housings with depth ratings of 600 to 9000
metres. Another persistent cause of system failures has virtually
been eliminated.
FILM
Early work was done with black and white negative film but colour
reversal,
Kodak Ektachrome at 64 or 200 ISO,
is now used almost
exclusively
because
of
the
substantial
improvement
in
interpretation.
UNDERWATER OPERATIONS
rations are planned to minimise the time spent underwater and
simplify the procedures in every possible way.
Pairs of cameras
mounted on all kinds of ROVs are used to obtain
stereo
photography as illustrated in Figure 4.
A video camera mounted
beside the photographic cameras enables the operator at the
surface to manoeuvre the ROV into the required positions to take
photographs
Fender bars or manipulator arms may be used to
achieve the required stand-off distance and to carry scale bars
positioned within the camera field of view to provide portable
control.
Otherwise the manipulators are used to clip magnetic
scale bars onto the object itself.
Only exceptionally are divers with single cameras needed to gain
access in restricted spaces.
CALIBRATION AND CONTROL
Camera calibration and residual water refraction distortion are
determined simultaneously by photographing a three dimensional
object in a water tank as shown in Figure 5.
Cameras are recalibrated in this way after major overhauls,
but all routine
servicing can be done without disturbing the alignment of the
lens and port. Although this method has been rejected by the more
scientifically minded,
Newton 1984 (5),
Baldwin 1984 (6),
and
Welham 1984 (7), it has been shown to produce entirely acceptable
results given appropriate control, and is attractive to the users
because
prime diving time can be devoted exclusively
to
photographing the objects to be inspected and permits quite small
ROVs to be used which could not carry the cumbersome calibration
frames
required for on-site calibration.
There is also no need
for the operation to be supervised by photogrammetrists.
Experience with the more rigorous solutions has been noted and
corrections can now be applied to the calibrated focal length for
average salinity on site.
Control is supplied by scale bars or frames placed around the
object.
For simple tasks involving individual photomodels,
two
or more scale bars are attached to the object by magnets or tied
to a manipulator arm or crash bar in front of the ROV and within
the stereoscopic coverage of the photographs.
For mUltiple
photomodel tasks,
control can be provided by scale bars in every
model;
by attaching a measured frame designed to enclose the
whole object;
or by using the fixed orientation parameters of a
rigid two-camera system determined from the controlled models to
provide control for the models without control. The last solution
is the one adopted for routine inspection programmes involving
thousands of stereo-pairs of photographs,
from which few if any
stereomodels will be required for
photogrammetric analysis.
Provided some stereo-pairs in each mission contained control
bars,
any of the photomodels can be used for
photogrammetric
measurement whenever required.
Besides saving time and cost in
underwater operations,
this
system is extremely simple and
therefore likely to succeed even without specialist supervision.
PREPARATION
Both photographic inspection and photogrammetric measurement can
be non-destructive,
which is a prime consideration for marine
biology,
oceanography and sedimentology,
but engineers usually
require marine growth to be removed from structures before
photography.
This can largely be done by ROVs using scrapers,
hydraulic brushes and high pressure water jetting, sometimes with
abrasives such as a sand slurry added to the water.
While soft
and mobile marine growth make it difficult to select fixed minor
control points for transfer between photomodels, hard growth such
as mussels and tube worm can provide useful identifiable points.
Serious problems arise with new structures of uniform texture and
areas of damage which have been ground and polished
and these
may require magnetic targets and grease paint graffiti to be
added before photography to provide sufficient
features and
texture for stereoscopic measurement.
PHOTOGRAMMETRIC ANALYSIS
From the beginning,
a Zeiss Stecometer stereocomparator was used
for observation, and orientation programmes developed by K M Keir
at Hunting Surveys needed only minor modification to handle
underwater
applications.
The lack of a real time
three
dimensional model was a disadvantage only overcome by the immense
experience of H M Hall who has observed and adjusted hundreds of
thousands of photomodels for both aerial triangulation and short
range applications and was responsible for all the underwater
photogrammetry undertaken between 1977 and 1987.
Features could only be plotted from discrete points and profiles
were interpolated.
Surface modelling and isometric views were
generated on an Intergraph system whenever more complex analysis
was required.
When Hunting Surveys closed down in 1987,
Plowman Craven and
Associates took over the photogrammetric analysis using
a
Digicart analytical plotter and McDonnell Douglas GDS graphics
system.
The analytical plotter is much more effective for
plotting features
and profiles,
and perspective views can be
plotted from all angles using the GDS system as shown in Figures
6 and 7.
A complete system with CAMEL underwater cameras and calibration
software has been licensed to GHD of Melbourne,
Australia since
1985.
FIGURE 6.
Perspective drawing of dented pipe with
outline of corrosion pipes.
Scale bars are also shown.
FIGURE 7.
Perspective drawing of damaged pipe.
1
Detailed photogrammetric analysis of critical areas of damage,
corrosion or replacement will continue to be sent to specialists,
but user demand for a simple portable system for
stereoscopic
interpretation and approximate measurement which they can use
themselves cannot be ignored.
Although quite a number of the new
generation
of
PC-based photogrammetric
systems
has
been
investigated and tested,
none has so far met
the required
standard of user friendliness,
robust portability and low cost,
which would
justify our recommending them to our clients for
occasional use by engineers without photogrammetric training.
Hopefully improved systems will be presented at Kyoto.
APPLICATIONS
Over
ten years we have been asked to photograph and sometimes to
measure the seabed, natural features in the sea, and all kinds of
man-made structures,
and CAMEL cameras have been used by our
clients to take prodigious quantities of photographs,
most of
which we never see. Most of the photogrammetric analysis has been
related to the measurement of damage,
corrosion,
wastage of
sacrificial annodes,
weld fractures, replacement and redesign of
sections of oil production platforms and pipelines.
Measuring
scour and sediment around structures has also been required.
Corrosion in sea water is a continual process which eventually
weakens steel structures despite cathodic protection and must be
monitored meticulously.
Selected areas of pitting are photographed regularly and the depth and size of individual pits are
measured to calculate the rate of corrosion.
Similarly the
sacrificial annodes are measured periodically to check that the
wastage rate conforms to expectations.
Stress causes weld failures in platform structures which must be
detected
and repaired quickly to prevent
further
damage.
Systematic
photographic inspection is used
to locate weld
fractures
and photogrammetric analysis to measure and identify
the cause of the failure.
Serious cases may require clamps to be
fitted around the platform nodes. Some of the more complex multimodel photogrammetric tasks have involved circular patterns of
photographs around each tubular member to calculate the ovality,
angle and offset of each member in order to design clamps to
strengthen fractured joints.
Damage caused by storms,
collisions and falling objects require
immediate attention and repair to prevent
serious loss of
production.
The first stage is to obtain inspection photography
to assess
the nature of the problem,
followed by specific
coverage to measure the areas of damage in order to construct
repair plates or replacement parts.
Isometric views, with or
without
scale exaggeration in selected directions,
have been
found
to be particularly valuable to engineers assessing the
seriousness of the damage and in designing repairs.
V-42
Many of the oil and gas production platforms in the North Sea are
more than ten years old and require modification to adapt them to
changing operating procedures.
The original design drawings are
often inadequate to design and prefabricate replacement parts and
photogrammetry is used for this purpose.
SUMMARY
A
commercial
service
for
underwater
photography
and
photogrammetric analysis has been developed to meet the needs of
the oil and gas production industry,
first in the North Sea and
later in other offshore fields.
Now that this service exists it
is attracting the attention of other engineers,
oceanographers,
marine biologists,
archaeologists and other scientists operating
underwater.
In the next ten years,
much more widespread use of
underwater photogrammetry can be expected if portable user
friendly systems become available and videogrammetry can be
developed into a real-time measurement system.
REFERENCES
1.
2.
3.
4.
S.
6.
7.
WELSH,
N., LEADBETTER, I.K., CHEFFINS, O.W. and HALL, H.M.,
1980.
Photogrammetric procedures for a North Sea oil
rig
leg
repair.
International
Archives
of
Photogrammetry, 23(BS): 474-483.
TURNER, D. J.
and LEATHERDALE, J. D.,
1982.
Underwater
photogrammetry for inspection and maintenance of North
Sea
oil
platforms.
International
Archives
of
Photogrammetry and Remote Sensing, 24(Vj2): S07-S1S.
LEATHERDALE,
J.D.
and TURNER,
D.J.,
1983.
Underwater
photogrammetry in the North Sea. Photogrammetric Record
11(62): lSl-167.
BALDWIN,
R.A. and NEWTON, I., 1982.
A proposed system for
underwater photogrammetry from a manned submersible
camera
platform.
International
Archives
of
Photogrammetry and Remote Sensing, 24(Vj1): 39-S2.
NEWTON, I., 1984.
The current status of underwater photogrammetry. International Archives of Photogrammetry and
Remote Sensing, 2S(AS): S80-S89.
BALDWIN, R.A., 1984.
An underwater photogrammetric measurement
system for structural inspection.
International
Archives of Photogrammetry and Remote Sensing,
2S(AS):
49-S8.
WELHAM,
L.G.,
1984.
The development of an underwater
measuring capability using photogrammetric techniques.
International Archives of Photogrammetry and Remote
Sensing, 2S(AS): 7S7-766.
V-43