International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B3. Istanbul 2004
estimation of its value leads to the determination of flow
resistance. This is not an easy task which may need extensive
studies of different circumstances and factors that have a direct
effect on Manning's n value.
Chow, (1973) and French (1994) described factors controlling
the value of n as follows: surface roughness: this factor is
directly related to the building material of the channel bed,
whether it is gravel, sand, silt, clay, or any other material. It is
not enough to estimate the surface roughness as grain size and
shape although they affect the magnitude of the resistance force
to the flow. Chow, (1973) states that, commonly fine grain
materials provide smooth channel and low value of » while
coarse grain materials give high resistance to the flow and
relatively higher values of n. LMNO (2000) and Henderson
(1966) introduce estimations of the values of n for some
materials.
3. AERIAL PHOTOGRAPHY AND LIDAR FOR
CREATING DIGITAL SURFACE MODELS
Both aerial photography and LiDAR have been used
extensively for digital surface modelling of the landscape.
Aerial photography and photogrammetry have had a long
history of producing DSM's through analogue, analytical and
digital methods (Mikhail ef a/., 2001). The more recent digital
techniques have enabled very rapid DSM to be produced
through automated image matching techniques and the quality
of these have been assessed by a number of researchers (Smith,
1996). One of the fundamental differences between LIDAR
and photogrammetry is that LiDAR is based on a range
measurement to a point from a single airborne position.
Photogrammetry however, is based on stereo matching of
images from two airborne positions. The stereo matching
process requires the matching of a ‘patch’ of pixels covering a
small area rather than a discrete point (footprint) as with
LiDAR. In addition, often the algorithms used in the
photogrammetry solution have been designed for smooth
landscape modelling rather than the rapidly changing elevations
of buildings in an urban environment. With both technologies
there is the question of what surface is being measured? An
analysis of these technologies is given in Smith ef al. (2000)
and in Asal (2003). Before the methodology for producing
coefficients of roughness could be investigated it was
considered useful to try and visualise the DSM for different
land uses. This would help to appreciate, from the information
that is available from a DSM, the nature of the texture of the
landscape. So, an analysis to see whether the different DSM’s
show different textural characteristics was undertaken. A full
analysis is given in Asal (2003).
3.1 The Test Site
The area covered by the test site at Newark-on-Trent includes a
variety of landscapes. Primarily it is on the flood plain of the
River Trent but one side of the river rises rapidly to the old
town of Newark. Typical of many old town centres, it has
narrow winding streets where it is difficult to see on to the
ground level from the air. This is particularly difficult for
photogrammetry as it requires too be able to see the ground
from two positions for stereo analysis. Along side the old town
is an industrial area and a relatively new residential area. To
the north bank of the river and beyond some mixed
development is a rural flood plan area of mainly hedged
agricultural fields, small woodlands and a ring road on top of an
716
embankment. Aerial photography at 1:25000 scale covers a
much greater area than the 1:10000 scale but common areas’
covered by the photography and LiDAR could be found.
DSM's were created at 2m postings. Unfortunately it was not
possible to obtain the different photography and the LiDAR at
the same time.
3.2 Analysis and Comparison of Digital Surface Models
Figures 1, 2 and 3 show typical results obtained. (Red lower,
green and mauve higher)
Figure 2: DSM from 1:10,000 aerial photography in a rural
area.
: E
y
gifs
e cS
WE 4
2 ie SF e
oe Lo TA i 4 he
106 ©
Figure 3: LIDAR DSM in a rural area.
Intern
Three
5 arc
investi
throug
digital
model:
RN ä
Figure
Further
drawn
of lanc
as exp
and th
differe
are a r
related
on the
41 1|
The. cc
water 1
first is
definec
compo
surface