Coastlines are some of the most rapidly changing environments
in the world, their accurate delineation are important for navigational
and coastal survey purposes. Synthetic Aperture Radar (SAR)
such as the RADARSAT-1 is useful data source for the detection and
delineation of coastline. Generally the extraction of coast/shoreline
features from SAR imagery involves a series of processes that
include filtering, enhancements and image segmentation. In the
study to extract coastline from RADARSAT-1 of the Kuala Terrengganu
of Malaysia, a semi-automated process was developed.
the process modified and enhanced techniques previously established
and adopted it for a single radar image. A sequence of average,
median, sobel and edge filters was applied to image to achieve
a good separation of linear and coastline features in the grey scale
level. A thresholding process was applied first for image segmentation
and later for the separation of the land and water boundary.
The results showed the process was effective and efficient in
extracting and delineating the coastline. A comparison of the
extracted coastline with the navigational map showed general
similarity in the curvature of the coastline. Distortions were also
shown in some locations which may be attributed due to the
changes in the coastline over the years due to erosion of the
coastline. The study generally showed that, the techniques and
methods presented can be used for the identification and mapping
of coastlines to update navigational chart and other maps requiring
accurate shoreline mapping