3.Material & METHODOLOGY
Every work has a definite method and rule. In the same way, this program follows a specific method to finish it
successfully. This work is a study of mangrove forest cover change of chokoria sundarbans by using Remote Sensing and GIS. This project represents Mangrove forest cover changes of Chokoria Sundarbans between 1997 to 2006 using satellite images of Land Sat TM. Now, the basic working rules are represented continuously in the following.
3.1 Material used:
Ø Landsat TM 1997
Ø Landsat TM 2006
3.1.2Software used :
The methodology adopted for this study involves both the Digital Image Processing and GIS based analysis. Both Digital Image Processing and GIS analysis were carried out in the computer system having the following software
configuration:
* PC Based ERDAS Imagine Image processing software (Version 8.4) available at
SPARRSO laboratory.
* ARC/INFO 4.0
* ARCVIEW 4.0
3.1 .3 Basic work of ERDAS
ERDAS imagine software generally used to get an idea about earth surface. An image of earth surface is taken by satellite and it is inserted into the software. Then, the image is
analyzed and classified by the software. The software helps about the following concerns. (LGED,2001)
§ To the study of forest area investigation
§ To indicate water bodies.
§ Other various investigations.
3.2 Visual Interpretation of the satellite imageries:
help of available maps.
Creation of the GIS data base:
covering the study area .
v The imagery interpretations were digitized.v The polygon topologies were built for all the above coverage’s.
3.2.1 Visual Interpretation of LANDSAT TM image-1997 :
For this image False colour composited (FCC) using red (3), green (2),abd blue(1) has been selected . Fine red and light red are vegetation cover. Dark blue represents water bodies and wetlands. Light blue refers to land under water. Deep green are marshy lands.(Shbhan,1993)
3.2.2 Visual Interpretation of LANDSAT TM image-2006 :
For this image FCC using red (3),green(2),abd blue(1) has been selected . Fine green and light green are vegetation cover. Dark blue represents water bodies and wetlands. Light blue refers to land under water. Deep green are marshy lands.(Shbhan,1993)
3.3 Geometric correction:
To make the images workable the data needed to be transformed to a uniform ground co-ordinate system of a chosen map projection. with the help of topo maps of 100000 a good number of Destination Ground Control Points (GCP) were selected mainly from distinctive features like crossing the river, road river crossing etc,and the corresponding points were selected in the Source of images. Then the topo-points were digitized and the image is represented in the Bangladesh Transverse Mercator (BTM) projection.
The Geometric correction of the Landsat-TM imagery has been done using a set of ground control points (GCP) generated from the orbital parameters. After the geometric correction and georeferencing.
3.4 Subsetting satellite image according to project area
:
The image has fit to the window. Then, add district area and name of the district (Cox’s bazer) of Bangladesh over the image by Vector layer and vector (view in properties) tools. Hence, the main aim is to collect the image of the study area. Therefore, at first we cut the image of the chokoria sundarbans upazila with the help of inquire box. But, it is a rectangular shape. Then, select of the area around the chokoria sundarbans upazila and cut it with the help of AOl Tools. Therefore, the image of the chakaria sundarbans upazila has achieved.
3.5 Unsupervised classification of image:
3.5.1 Unsupervised Image Classification (UC)
§ Definition: identification of natural groups or structures within multispectral data Does not use training data for individual information classes as the basis forclassification
§ Image pixels are examined and aggregated into a number of spectral classesbased on natural clustering in multi-dimensional space.
§ UC is the definition/identification/labeling and mapping of natural spectral classes.
§ To determine spectrally separable classes and then define their informational usefulness.
3.5.2 Stages to Unsupervised Classification
1. Definition of minimum and maximum number of categories to be generated by
the particular classification algorithm .
3. Algorithm then finds distances between pixels and forms initial estimates of cluster centers as permitted by user defined criteria.
4. As pixels are added to the initial estimates new class means are calculated. This is an iterative process until the mean does not change significantly from one iteration to the next (a kind of wearing down of the classification problem by repetitive application assignment and reassignment of pixels to groups)
3.6 Area measurement and change detection:
From the IMAGINE attribute table areas from all the maps were calculated. The area obtained from the management plan of 1997 to2006 of the study site was compared with the area obtained from the Landsat TM imagery of the year 1997 to2006 . It was considered as abase and with which the change detection and performed.
3.7 Data analysis :
Now, it is necessary to analysis the setallite image of the year 1997to2006. This work involves a clear observation of the amount of mangrove forest decreasing. it also includes calculation of future data by using present and previous data.
Here we represent a flow chart on the total overview of the study :(
Figure-3.1)
Figure No 3.1 : Flow chart of the Methodological steps of
the study
