563
26.8.3 Management Application
More advanced spatial and modeling techniques
are required to support the decisions of managers
and policymakers. This involves shifting of
emphasis from basic geographical data handling
to manipulation, analysis and modeling in order
to solve real-world problems.
26.9 Extraction of Wetland Extent
MNDWI, NDPI and NDVI images are generally
used to extract the wetland boundaries through
suitable hierarchical thresholds. Further, NDPI
and NDVI images are used to generate the vegetation and no-vegetation areas within a wetland
using a suitable threshold.
26.10 A Proposed Alternative
Methodology
For classifi cation of inland wetlands using RS
techniques, band 4 of IRS 1C LISS III image data
are to be density sliced for the separation of
waterbodies. The threshold values for water mask
are to be obtained interactively. A bit map is to be
generated for the waterbodies. This mask will be
used for further classifi cation of waterbodies into
turbidity patterns and aquatic vegetation.
Although the density slicing of band 4 provides
acceptable results in most of the cases, it may
sometimes lead to confusions with non-water
classes. A major class of confusion is the shadow
due to terrain. Such anomalies could be removed
through stratifi ed density slicing and through
contextual refi nements.
The normalised Difference Vegetation Index
(NDVI), which minimises effect of the shadow,
can also be used for separation of waterbodies, as
the wetland areas fall in a lower NDVI zone than
terrestrial vegetation. However, NDVI may also
exhibit confusing results, because of many other
non-vegetated classes like snow and barren land.
Many may exhibit NDVI values comparable to
these of a waterbody. However, an interactive
integration of band 4 and NDVI will clearly separate waterbodies.
26.10.1 Turbidity Patterns
Turbidity patterns are best refl ected by band 1of
IRS 1C, LISS III image data. The higher the DN
value in band 1, the higher is the turbidity. The
turbidity classifi cation is a subjective one, as it is
impractical to relate the quantitative values for
turbidity (which are dynamic according to the
season) with the refl ectance. Thus, determination
of the threshold for different turbidity levels
needs to be carried out by examining the major
(large-sized) waterbodies in the area.
26.10.2 Aquatic Vegetation (AV)
AV needs to be classifi ed within the waterbody
mask that is generated using band 4 of IRS LISS
III data. The NDVI (generated as: (IR−)/(IR + R)
where IR is DN value in band 3 and R is DN value
in band 2 of IRS-IC LISS III) is to be obtained for
waterbodies. The NDVI values are subjectively
divided into vegetation levels, i.e. nil, poor, moderate and high vegetation coverage.
26.11 Habitat Mapping Using
Microwave Remote Sensing
Microwave remote sensing tools have an important role to play in applications relating to wetland monitoring and mapping. In optical RS, the
visible and infrared part of the electromagnetic
spectrum is used to characterise objects of interest. However, during monsoon season, the suitable atmospheric windows for acquisition of
optical data are limited to cloud-free periods.
This is a major lacuna for wetland applications,
because wetlands are highly seasonal and
dynamic systems compared to terrestrial ecosystems. The radar imaging system overcomes
many of these limitations by providing increased
canopy penetrations and day and night acquisitions nearly independent of weather conditions.
It is, therefore, imperative to use radar data for a
better understanding of the dynamics of wetland
ecosystems as well as their assessment, monitoring and management. There are also several
advantages to using microwave data. Microwave
26.11 Habitat Mapping Using Microwave Remote Sensing
26.8.3 Management Application
More advanced spatial and modeling techniques
are required to support the decisions of managers
and policymakers. This involves shifting of
emphasis from basic geographical data handling
to manipulation, analysis and modeling in order
to solve real-world problems.
26.9 Extraction of Wetland Extent
MNDWI, NDPI and NDVI images are generally
used to extract the wetland boundaries through
suitable hierarchical thresholds. Further, NDPI
and NDVI images are used to generate the vegetation and no-vegetation areas within a wetland
using a suitable threshold.
26.10 A Proposed Alternative
Methodology
For classifi cation of inland wetlands using RS
techniques, band 4 of IRS 1C LISS III image data
are to be density sliced for the separation of
waterbodies. The threshold values for water mask
are to be obtained interactively. A bit map is to be
generated for the waterbodies. This mask will be
used for further classifi cation of waterbodies into
turbidity patterns and aquatic vegetation.
Although the density slicing of band 4 provides
acceptable results in most of the cases, it may
sometimes lead to confusions with non-water
classes. A major class of confusion is the shadow
due to terrain. Such anomalies could be removed
through stratifi ed density slicing and through
contextual refi nements.
The normalised Difference Vegetation Index
(NDVI), which minimises effect of the shadow,
can also be used for separation of waterbodies, as
the wetland areas fall in a lower NDVI zone than
terrestrial vegetation. However, NDVI may also
exhibit confusing results, because of many other
non-vegetated classes like snow and barren land.
Many may exhibit NDVI values comparable to
these of a waterbody. However, an interactive
integration of band 4 and NDVI will clearly separate waterbodies.
26.10.1 Turbidity Patterns
Turbidity patterns are best refl ected by band 1of
IRS 1C, LISS III image data. The higher the DN
value in band 1, the higher is the turbidity. The
turbidity classifi cation is a subjective one, as it is
impractical to relate the quantitative values for
turbidity (which are dynamic according to the
season) with the refl ectance. Thus, determination
of the threshold for different turbidity levels
needs to be carried out by examining the major
(large-sized) waterbodies in the area.
26.10.2 Aquatic Vegetation (AV)
AV needs to be classifi ed within the waterbody
mask that is generated using band 4 of IRS LISS
III data. The NDVI (generated as: (IR−)/(IR + R)
where IR is DN value in band 3 and R is DN value
in band 2 of IRS-IC LISS III) is to be obtained for
waterbodies. The NDVI values are subjectively
divided into vegetation levels, i.e. nil, poor, moderate and high vegetation coverage.
26.11 Habitat Mapping Using
Microwave Remote Sensing
Microwave remote sensing tools have an important role to play in applications relating to wetland monitoring and mapping. In optical RS, the
visible and infrared part of the electromagnetic
spectrum is used to characterise objects of interest. However, during monsoon season, the suitable atmospheric windows for acquisition of
optical data are limited to cloud-free periods.
This is a major lacuna for wetland applications,
because wetlands are highly seasonal and
dynamic systems compared to terrestrial ecosystems. The radar imaging system overcomes
many of these limitations by providing increased
canopy penetrations and day and night acquisitions nearly independent of weather conditions.
It is, therefore, imperative to use radar data for a
better understanding of the dynamics of wetland
ecosystems as well as their assessment, monitoring and management. There are also several
advantages to using microwave data. Microwave
26.11 Habitat Mapping Using Microwave Remote Sensing