Geography Reference
In-Depth Information
> 1.5
1.0
0.5
0
Depth [m]
Figure 9.11
Example of a bathymetric map obtained from the analysis of an hyperspatial image using the red band brightness versus
water depth relationship of Figure 9.10a.
downstream distance) between objects and features in
different images. In the case of the Geosalar hyperspatial
image database, the coarser imagery with a spatial resolu-
tion of 10 cmwas manually georeferenced over the course
of the (entire) summer following image acquisition. This
process rested on ground control points collected dur-
ing an intensive field campaign. Unfortunately, this field
effort only produced sufficient field data in order to geo-
reference the coarser imagery (two ground control points
per image). It was found that manual georeferencing
of the c. 5500 3 cm images was too labour intensive to
be feasible. As a result, the early FGIS functioned with-
out explicit georeferencing information. In the FGIS, the
inter-image spatial relationships were estimated based
on the known overlap between successive images and
the known dimensions of each pixel. In essence, the
FGIS measured distances downstream by counting pixels
and accounting for image overlaps whilst this process
was functional, it likely induced some error because it
approximated that, in the short 90m span of a single
image, the channel was straight.
Recently, theFGISwas takenbeyond theprototype stage
for applications in commercial settings. A new systemwas
developed and dubbed simply the 'Fluvial Information
System' (FIS). The FIS maintained the initial FGIS objec-
tive of integrating a range of remote sensing methods.
However, the FIS benefited from recent developments in
thefieldof georeferencing.Asdiscussed inChapter 8of this
volume,Carbonneauet al. (2010)developedanautomated
georeferencing approach which was much better suited
to large hyperspatial image databases. Furthermore, the
FIS implemented a River Coordinate System as described
by Legleiter and Kyriakidis (2006) in order to produce a
unique, locally orthogonal, coordinate systemwhich gives
the position [s,n] of any point in the river as a combination
of distance downstream [s] and distance across stream
[n]. With this coordinate system in place, the FIS then
integrates a range of remote sensing methods which are
well established in the literature (e.g. Carbonneau et al.,
2004; Fonstad andMarcus, 2005; Carbonneau et al., 2006;
Legleiter and Kyriakidis, 2006). Carbonneau et al. (2012)
successfully applied the FIS in order to demonstrate that
FRS technology is now at the point where it can contribute
to the broad range of river sciences. These authors showed
that hyperspatial image data can be successfully acquired,
managed, processed and analysed in order to produce
meaningful habitat parameters at sub-metric resolutions
for an entire river in the Scottish highlands. In addition to
the data presented in Carbonneau et al. (2012), the FIS is
also capable of producing innovative habitat visualisations
as discussed below.
9.5.2 Habitatdatavisualisation
The availability of continuous maps for two key habitat
parameters (water depth and bedmaterial grain size) open
up important new avenues in terms of habitat mapping.
In line with the recommendations of Fausch et al. (2002),
it is now possible to examine the spatial distribution of
fish in a spatially explicit description of fluvial habitat
at the riverscape scale with metric resolution. However,
Search WWH ::
Custom Search