OF
ques’
SAC)
vas to
heries
iulna-
d data
s. The
3). ID)
ictical
it. V)
sis of
of the
future
‘ming.
hrimp
cating
z coastal
intage is
ueving à
1aculture
'entional
' remote
location,
fisheries,
Lanteri,
imilnadu
t Bengal
d it for
ttempt is
and GIS
ig in the
racy and
Iso been
Xf. coastal
(Quader,
project is
r Shrimp
; Remote
‘AP and
ingladesh
'anization
ranization
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B7. Istanbul 2004
1. Description of the Study Area
Site 1:Cox's Bazar: 21°25'-22°00'N, 91°50'-92°15'E.Site 2:
Khulna-Sathkhira: 22? 1 5'-22?45'N, 89?00'-89?30'E.
3.1 Data Used: In the present study various types of data
have been used. It includes different satellite data (IRS,
SPOT, Landsat TM), thematic maps, field-measured data and
other relevant published information etc. The thematic maps
on soil, land-use, land capability associations and soil salinity
were used for the study. The land-use and land capability
information is updated using IRS LISS III and PAN data.
3.2 Software Used: In the present study, the following
software's were principally used:
- ERDAS Imagine V 8.3.1 digital image processing
software integrated with the additional vector
module.
- . Arc/Info GIS has also been used for the GIS
related part.
The use of Imagine and Arc/Info GIS provided an
effective tool for the present work.
4 Methodology :In this section, methodologies used for the
general operation during the present study have been
described. However, detail descriptions for more specific
operations have been provided in the respective chapters.
41 Geometric Correction and Processing and
Classification of the Digital Images
Digital data of LISS III and PAN for Cox's' Bazar were
downloaded using PC based ERDAS software available in
SPARRSO. AII the images were geometrically corrected and
were projected to LCC system. IRS LISS, II and PAN
images were re-sampled to 6-m spatial resolution in order to
merge them with reasonable accuracy.
In the present work, both supervised and unsupervised methods
of classification have been employed. The unsupervised method
of classification is based on ISODATA algorithm available in
the ERDAS Imagine software. While, the supervised
classification method based on maximum likelihood algorithm
has been used.
4.2 Preparation of Base Maps: Remote sensing application in
various aspects of aquaculture has been demonstrated by
Loubersac (1985) who used simulated SPOT data to
demonstrate the capabilities of a high-resolution (10-20 m) data
for aquaculture siting. A Geographic Information System (GIS)
approach has recently been demonstrated through integration of
ground and satellite remote sensing data to identify area
suitable for aquaculture development etc. So, coastal wetland
and landform mapping on 1:50,000 scale using satellite data for
the two study areas of the coastal zone (i) Cox’s Bazar Area (ii)
Sathkhira-Khulna has been prepared as per the package
development of the project. These maps provide information at
the reconnaissance level and used as reference map for field
survey/verification and creating GIS layers on the monitor.
43 Techniques used to obtain macro-structured land use
classes in vector form:
- Unsupervised classification of merged image
as well as LISS and TM image.
229
- . Merging the classes to the desired number of
classes.
- Elimination of non-homogeneity and noise
using 7x7 majority spatial filter.
- Elimination of very small clusters.
- Raster to vector transformation.
- . Combination of vector layers obtained from all
the images.
- On-screen editing of the vector data.
- On-screen editing of the vector layer was
needed to correct classification error, as well as
to well shape the structures of the features.
Micro-structured features were digitized on-
screen from the images, as well as from the
base maps.
4.4 GPS based field survey: Extensive GPS based field
works have been carried out over the two study sites in
support of the satellite-derived information for their correction
and validation. The infrastructure information is also been
updated by such survey. In addition, the model-derived
outputs are verified in some specific points over the study
sites.
S. Construction of GIS Based Fisheries Environmental
Database (GISFED) for Suitability Analysis
Land surface processes have become a great concern in the
context of global change and massive environmental
degradation in different parts of the world. The human
intervention to nature and earth’s natural resources largely
modifies the composition and properties of the earth's surface
and its atmosphere. Specifically, the ever-increasing human
population resulted in over exploitation of natural resources
and thereby, causes irreversible damages to such system.
These activities often provoke various environmental and
ecological problems the world over. Massive destruction of
forests, intrusion of salinity, and rejection of chemical
pollution through urban and industrial activities degraded our
environment alarmingly. The scale, intensity and persistence
of such undesirable changes are highly variable over time and
space.
To avoid further degradation of the earth's environment and to
keep it in a livable condition such activities should be limited.
Efficient planning and management effort with sustainable
schemes should be coupled with the development activities. As
such, understanding and monitoring of these processes become an
urgent need requiring up to date and regular information over a
given geographical area. In such context, creation of a spatial
database is very essential for monitoring the characteristics of the
on-going processes and changes. It is now well recognized that an
efficient database must be established for the development and
efficient management of a given region.
5.1 Generation of Database: The creation of spatial database
in accordance with the GIS execution steps designed for the
model is an important step for the implementation of the
present project. The spatial data has been synthesized from
different sources having different resolution, projections and
feature types. Due to this, multi-dimensional spatial
mismatching has been occurred during the synthesis of the
database. In order to minimize these mismatching, a common
reference frame was created based on extensive GPS survey
in each of the study areas. The reference frame was first used
