« 2004
————
sion is
of the
high-
om the
quality
, use of
ing. In
ves and
em the
nbridge
iS local
ed their
; easily
s this
'ecialist
> Scans.
ss to a
iust be
e well-
graphs.
ial GPS
ers can
nces up
of less-
> maps,
| marks.
oints 1$
esirable
] points
1 strong
her and
DY,
between:
rior and
cessing
which
a least
995). In
rdinates,
fiducial
nywhere
exterior
natically
idicating
s can be
between
olled by
| in the
y can be
of error
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B4. Istanbul 2004
(Cooper & Cross, 1988): precision (relating to random errors),
reliability (gross errors) and accuracy (systematic errors). When
using scanned archival photography, the main factors affecting
precision include (Chandler, 1989; Scarpace & Saleh, 2000):
e geometry provided by the photographs;
e camera characteristics;
e image quality (quality of original film, geometry and
fidelity of scanner, scanning resolution);
e blurring due to image motion or use of focal plane shutter.
The applied self-calibrating bundle adjustment procedure is
capable of propagating stochastic properties through the model.
This provides a measure of precision of the output parameters.
Data reliability can only be improved by increasing the
redundancy of measurements (Hottier, 1976). Reliability can be
evaluated by inspecting the residuals of the photo- and control
coordinates. Systematic errors result from erroneous, non-
rigorous or an incomplete functional model, bad distribution of
control points, image texture, terrain characteristics, film
deformation, focal plane shutter and atmospheric effects
(Chandler, 1989). Various systematic errors can be modelled
(e.g. lens distortion), but the process is complicated as there is
often high correlation between many of the estimated
parameters (Granshaw, 1980). Accuracy can only be determined
by using independent checkpoints to compare
photogrammetrically coordinates with accepted survey-derived
values (Chandler, 1999).
3. CASE STUDY MAM TOR
3.1 Study area
Mam Tor is a 517m high hill, located two kilometres west of
Castleton (Derbyshire, UK; Ordnance Survey coordinates
SK135835), forming the head of Hope Valley. The eastern flank
is dominated by a steep scar with a large landslide beneath. The
former main road between Sheffield and Manchester (A625),
constructed across the slide, was abandoned in 1979 as a
consequence of continuous damages due to the moving ground
mass. The slope consists of predominantly sandstone sequences
(Mam Tor Beds) overlying predominantly shale units (Edale
Shales). The layers dip slightly inwards of the slope. South of
the headscar, a fault cuts through the shales, which might have
affected the slope stability (Skempton et al., 1989). Length of
the slide from scarp to toe is about 1000m, while elevation
drops from 510 to 230 m. Mean slope of the slipped mass is
12°. Maximum thickness of the mass is 30-40m (Skempton et
al., 1989). The affected area is approximate 32ha.
The initial rotational failure has been dated back to 3600BP
(Skempton et al., 1989). While advancing downslope the mass
= = 840
SS Mam Farm", M x N ‚N
"P. A \
P risen (Head over Epàte Shales) NC T
="Hillfort ipea S SX.
? ramparts: Mc
Vo.
à Ponds 4
/ /
/
(Head over
Edale Shales)
140
Figure 1. Map of the Mam Tor area; landslide is indicated by
dotted line (after Skempton et al., 1989).
broke into a complex of blocks and slices. Disintegration of the
front slices created a debris mass, sliding further down. The
unstable transition zone, overlying the steepest part of the basal
shear, is the most active part, moving on average 0.35m/a over
the last century (Rutter et al., 2003). There is evidence that the
movements are not continuous but accelerate during wet
winters, when rain-fall exceeds certain limits, i.e. more than
250mm rain in a single month. while over 750mm in the pre-
ceding six months (Waltham & Dixon, 2000).
There are several information sources available about
displacements that have taken place over the last century. Notes
about regular disturbance and repairs of the road, from 1907
until the final closure in 1979, are kept by Derbyshire County
Council. In 1977-78 an extensive survey and stability analysis
was carried out (Skempton et al., 1989). Since closure of the
road, a few (temporary) monitoring schemes were set up by
various universities: 1981-83 by Sheffield University, 1990-98
by students from Nottingham Trent University (Waltham &
Dixon, 2000) and since 1996 by Manchester University (Rutter
etal,2003) '
3.2 Acquired photographs
A search for aerial photography of Mam Tor revealed that there
are numerous images available, both oblique and vertical, from
1948 until present. Unfortunately, the oblique images usually
tend to be focussed on the archaeologically interesting site on
top of the hill (iron age ramparts) and do not cover the whole
landslide area. Vertical imagery from four different epochs were
acquired and processed. The images are of varying quality and
scales (Table 2), providing a good indication about the
potentials of ADAPT, when applying to a range of commonly
available material.
Date Source Scale Scan resolution Ground resolution Media
1053 NMR 1/12,000 42u 0.50m Scanned contact prints
1971 NMR 1/7,400 42u 0.31m Scanned contact prints
1973 CUCAP 1/4,000 15p 0.06m Scanned diapositives
1990 CUCAP 1/12,000 15u 0.18m Scanned diapositives
Table 2. Details of acquired and processed aerial photographs (NMR = National Monuments Record; CUCAP = Cambridge
University Collection of Air Photos).
477
