Coastline extraction form remote sensing images
Speaker: Jen-Ru Liau
Abstract
The coastline change affect extended field, so it is important to realize how it change. Remote sensing is good to monitor it. There have two methods to extraction coastline. The first method uses passive remote sensing image (optical image). It depends on the characteristics of geomorphology and oceanography of interesting area to locate coastline. Row data have to do enhancement and filter in order to get edge-enhanced images. The other uses active remote sensing image (SAR image). Two SAR images are used to get an InSAR image. Coherence of InSAR image is calculated to realize properties of image. In coherence image, invariant part is the land; variant part is the sea. So segmentation process is used to get the part of interest from various kinds of image. There have two kinds of image segmentations: supervised and non-supervised. Fuzzy processing, one of supervised image segmentations, is used to classify coherence image to get land area and sea area. No matter what kind images be use, they must to calibrate before analyzing. Although analysis results are not good. Semi-automated method is more efficient than automated one. The physical, geological, biological, and chemical aspect must be involved in the method to define the position of coastline.
Reference
E. A. Loosm, K. O. Niemann, 2002. Shoreline feature extraction from remotely-sensed imagery. IEEE Trans. Geosci. Remote Sens. 6, 3417-3419.
S. Dellepiane, R. De Laurentiis, F. Giordano, 2004. Coastline extraction from SAR images and a method for the evaluation of the coastline precision. Pattern Recognition Lett. 25, 1461-1470.
S. G. Dellepiane, F. Fontana, G. L. Vernazza, 1996. Nonlinear image labeling for multivalued segmentation. IEEE Trans. Geosci. Remote Sens. 5, 429-446