Abstract

Coastal vegetation is one of the important natural ecosystems that act as protection from sea wave erosion phenomena and supply oxygen to the surrounding organisms. POLSAR data with multi-band and multi-polarization is selected for these studies which are acquired on September 19, 2000, for coastal vegetation mapping. Two methods are performed to CVV, LHH and LHV bands: Polarization signature analysis and comparison of classification between two clustering techniques (Fuzzy-k-mean and iterative self organizing data analysis technique, ISODATA). The result shows that L band with HV polarization can distinguish coastal vegetation better than CVV. Furthermore, classification result of LHH and LHV show that the Fuzzy-k-mean clustering provided accuracy of 76% moreover with the kappa statistic of 70%.

Key words: Polarization synthetic aperture radar (POLSAR), coastal vegetation, polarization signature, fuzzy-k-mean, iterative self organizing data analysis technique (ISODATA).

Abbreviation

ATI, Along-track interferometric; RSO, rectified skew orthomorphic; SSI, speckle suppression index; PI, polarization indices.