Geography Reference
In-Depth Information
refinement process. The approach has already been
successfully applied in other operational rapid mapping
activities, such as the 'Rapid Mapping' service at
DLR's centre for satellite based crisis information (ZKI,
http://www.zki.dlr.de). Furthermore, high resolution
topographic information may be successfully combined
with this multi-scale segmentation to enhance the map-
ping performance in areas of flooded vegetation and man-
made objects as well as to remove misclassified non-water
areas - however, this is largely restricted to rural areas (see
Section 6.4.2 for near real-time detection in urban areas).
widths
500 m) has been demonstrated, for example, by
Blyth (1997); Kussul et al. (2008) within a grid comput-
ing system, and many times by the International Disaster
Charter (http://www.disasterscharter.org/). A large num-
ber of research studies have recently demonstrated the
usefulness of coarse resolution flood images for support-
ing flood modelling and management (Table 6.2).
>
6.3.1.3 Classification accuracies
of flood parameters
Many SAR image-processing techniques exist to more or
less successfully derive a flood area or extent (e.g. Aplin
et al., 1999), including simple visual interpretation (Mac-
intosh, 1995; Oberstadler et al., 1997; Brivio et al., 2002),
image histogram thresholding (e.g. Brivio et al., 2002;
Matgen et al., 2004; Schumann et al., 2005), automatic
classification algorithms (e.g. Hess et al., 1995; Bonn
et al., 2005), image texture algorithms (Schumann et al.,
2005), multi-temporal change detection methods (e.g.
Calabresi et al., 1995; Laugier et al., 1997), of which
extensive reviews are provided in Liu et al. (2004)
and Lu et al. (2004). Complex auto-logistic regression
6.3.1.2 Large area mapping with low
resolution SAR
As noted earlier, given the strong inverse relationship
between spatial resolution and revisit time, routine mon-
itoring of floods from space in near-real time seems
currently only possible through low resolution (about
100 m pixel size) SAR imagery (Di Baldassarre et al., 2011)
or satellite constellations. The fact that coarser resolution
SAR data can be used successfully to delineate flood area
and edges on larger floodplain inundations (inundation
Table 6.2
Recent research on the use of low resolution SAR imagery to support flood studies.
Aim of Study
Flood Parameters
Extraction
Location
Source
Extracted
Method Used
Near-real time
verification of a 1D
hydrodynamic model
Flood area
Histogram threshold
method
River Po, Italy
Di Baldassarre
et al., 2009a
Near-real time
verification of a 1D
hydrodynamic
model
Flood area and flood
edge heights
Water surface profiles
Histogram threshold
method fused with
SRTM topography
River Po, Italy
Schumann et al.,
2010a
Event-specific flood risk
mapping
Flood area
Multiple methods listed
in Table 3
River Dee, NE
Wales, UK
Schumann and Di
Baldassarre,
2010
Calibration of a 1D
hydrodynamic model
Flood edge heights
Automated region
growing fused with
SRTM topography
River Severn, SW
UK
Schumann et al.,
2010b
Selecting an appropriate
hydrodynamic model
structure
Flood area
Histogram threshold
method
River Po, Italy
Prestininzi et al.,
2011
Operational automated
flood detection
Flood area
Automated histogram
threshold method
combined with region
growing algorithm
River Severn, SW
UK
Red River basin,
US and Canada
Matgen et al.,
2011
Search WWH ::
Custom Search