Geography Reference
In-Depth Information
contribution of images of riparian vegetation studies is
broad; from an applied point of view the diversity of
situations encountered by managers can make it difficult
to suggest universal rules. So the question is not only
what we can know but what we want to know, and is
remote sensing the best way to go about it? Both of these
aspects are then illustrated with different applications
using different media such as aerial photographs, satellite
images or Light Detection and Ranging (LiDAR) data.
studies. Secondly, a huge development in data manage-
ment has made the imagery more easily available. Progress
has been seen most notably in computer science and GIS.
Thus, new tools have been developed for classification
such as objected-oriented procedures (Mathieu et al.,
2007; Arroyo et al., 2010) or neural networks.
Imagery has been used in riparian vegetation studies
with four main objectives: mapping vegetation types,
mapping vegetation species, mapping historical changes,
and measurement of environmental and vegetation
parameters such as tree height or Leaf Area Index.
10.2 Image analysis in riparian
vegetation studies: what
can we know?
10.2.1 Mappingvegetationtypesand landcover
The most common use of imagery in riparian vege-
tation concerns vegetation type mapping. A vegetation
type is defined as a plant community with relatively
homogeneous floristic composition and/or physiognomy,
although they may differ on other characteristics such as
stand density, biomass or percentage of shade (Muller,
Since the 1970s, riparian studies have benefited from
significant progress in remote sensing and GIS science
(Table 10.1). Firstly, the diversification of platforms and
instruments associated with an increase in spectral and
spatial resolution obviously has benefits for vegetation
Table 10.1
Chronology of remote sensing tools developed for riparian vegetation studies.
Period
Selected Tools
Uses and Advances for Riparian
Examples
Vegetation Studies
19
th
c.
Photographs
Riparian past landscape reconstitution
Grams and Schmidt, 2002
1930
Aerial photographs
Riparian past landscape reconstitution,
possible to quantify area,
composition and configuration
Miller et al., 1995, Marston et al.,
1995, Mendonca et al., 2001,
Greco and Plant, 2003
1970
Infra red sensors
Widening of the available spectra
significantly increases the amount of
information acquired
Girel, 1986, Otahel et al., 1994,
Neale, 1997
Satellite images (e.g. Landsat TM or
Spot images)
Larger scenes, wider spectral resolution
Butera, 1983, Mertes et al., 1995
1990
Active technology (e.g. Radar,
LiDAR)
New set of contextual or structural data
(e.g. flood extension or tree height)
Townsend, 2001 and 2002, Mason
et al., 2003, Gen¸ et al., 2004,
Straatsma and Middelkoop, 2007,
Antonarakis et al., 2008b
GIS development
Integration of ancillary data (network
map, DEM), for example, to preselect
area of interest
Congalton et al., 2002, Ehlers et al.,
2003, Johnson and Zelt, 2005,
Alber and Piegay, 2011
2000
Satellite: very high resolution
images (e.g. QuickBird, IKONOS
or GeoEye)
Reaching aerial photo spatial resolution
with more bands and for larger scenes
Franklin and Wulder, 2002
Object oriented classification
Combine automation and contextual
information
Johansen et al., 2007b, Arroyo et al.,
2010
2010
Light Aerial Remote Sensing (e.g;
UAV)
Centimetric resolution, high flexibility
Corbane et al., 2006, Lejot et al.,
2007, Thompson and Gregel,
2008, Dunford et al., 2009
Search WWH ::
Custom Search