Geography Reference
In-Depth Information
is often wide enough to penetrate to the ground through
very dense vegetation (see Heritage and Large, 2009b;
Crutchley, 2009; Danson et al., 2009 for more detailed
discussion of the vegetation removal algorithm).
It is immediately clear from Figure 14.12c that a sim-
ple subtraction of the digital terrain model from the
digital surface model is able to differentiate tree shrub
and herbaceous vegetation types on basis of vegetation
height, mapping plantation areas of differing ages, iden-
tifying mature riparian trees and differentiating managed
floodplain pasture from bare earth and arable fields. The
bare earth digital surface model data is used to identify
locations on the floodplain where the local slope change
is rapid using an edge detection algorithm. It is notable
that a number of palaeochannels are revealed across the
floodplain surface (Figure 14.12b), features which are
more difficult to detect either from the aerial photograph
or on the ground. Newer flood banks with sharper slope
breaks are particularly well detected by the technology;
however, where the surface is naturally variable (e.g. to
the north in the image shown in Figure 14.12b) feature
identification becomes more problematic.
along with the brief comparison with studies involving
airborne technologies, demonstrates the evolution of the
discipline and its application in morphologically-similar
environments. It also shows how the technology has
potential value for feeding intomonitoring and legislation
programmes at a range of scales. A number of issues are
raised with what are nowwell-establishedmethods for the
acquisition of precise and reliable 3-D geo-information,
and overcoming these will be vital if laser scanning is to
achieve its potential in the surveying and monitoring of
river systems - systems which are spatially and temporally
variable, and whose inherent variability poses particular
challenges to the accurate collection and interpretation of
geo-information.
TLS survey data remain subject to many issues that
will generate inaccurate, misleading or inappropriate
information if not considered. Error in TLS measure-
ment is spatially variable, given the variation in survey
range, laser footprint and incidence angle onto the tar-
get surface. Linear features such as river systems present
particular challenges to surveying as they require con-
sistent collection and integration of data from multiple
viewpoints (Lim et al., 2009). The combination of sep-
arate scans, either spatially or through time, has the
potential to introduce inconsistencies in the orientation,
resolution and positioning of individual surveys. Prob-
ably because of it being in its relative infancy, of all
survey techniques available for river corridors, TLS has
the least standardised control; practices and error assess-
ments (Lichti et al., 2005). Clear attention has to be
given to planning, data collection and processing to min-
imise these disadvantages. In addition, most currently
available laser scanners are not well specified regard-
ing accuracy, resolution and performance (Hetherington,
2009) and only a minority are checked by indepen-
dent institutes regarding their performance and whether
they actually comply with manufacturer specifications
(Boehler et al., 2003).
As Lim et al. (2009) emphasise, field procedures using
TLS are concerned with two priorities absolute accuracy
and the ability to locate the scanned data correctly in
either local or global space. To ensure this, there is a need
for clear and unambiguous surveying protocols for linear
fluvial systems:
(a) consider the potential errors related to the scale of
survey from Table 14.1;
(b) minimise the scan distance to ensure greater scan
point density and ranging precision and to minimise
footprint size;
14.2.8.1 Ambl eve Valley, Belgium
A similar set of techniques was applied to LiDAR data col-
lected along the Ambl eve Valley in the Belgian Ardennes
region. Figure 14.13a details the morphology of approxi-
mately 1 km 2 of the river valley showing a contemporary
sinuous single thread channel set in a wandering channel
situation as indicated by the numerous palaeo-channel
and bar surfaces present across the floodplain. A shaded
relief map generated from the LiDAR data of the valley
bottom (Figure 14.13b) reveals the local morphologi-
cal complexity. This simple technique may be rapidly
applied to the digital terrain model and manipulation of
the aspect and azimuth of the light ray permits detailed
evaluation of subtle morphologic variation across the sur-
face (see also Crutchley, 2009). While such a technique
does not allow extraction of the location and metrics of
any of the features apparent, calculating the local relative
elevation change from the LiDAR data (Figures 14.13c,
13d) succeeds in differentiating channel and bar surfaces
and allows extraction of data metrics based on the local
elevation information.
14.2.9 Towardsaprotocol forTLSsurveying
offluvial systems
The series of case study examples associated with ground-
based laser scanning of river systems outlined above,
Search WWH ::




Custom Search