In this study, we develop several procedures of image interpretation, integration
of various remote sensing tools and GIS techniques in order to analyze the impact (Figure
3-7). Landsat TM and ETM+ images are obtained; first, image pre-processing presenting
image-to-map registration and image-to-image registration are steps to match two images
to be perfectly align. In this stage, topographic map is adopted as a base map. Second, a
hybrid image classification method is used this method combines both unsupervised with
supervised classification approaches. After hybrid classification maps of pre- and posttsunami
were generated, accuracy assessment is then adopted to assess the accuracy of
classification maps by referring to aerial photographs and IKONOS imagery. These steps
are carried out by using ERDAS IMAGINE 8.7 software.
32
After we achieved the classification process, the study area is divided into four
regions based on their specific geographical characteristics. The coastline of the study
area and different buffer distances are created in ArcGIS 9.0 for impact analysis
preparation. The post-classification change detection stage is performed to assess the
tsunami impacts by using IDRISI 32 software. The final results from the change
detection process will be used to analyze the impacts of the 2004 tsunami on mangroves
and nearby areas.