A. Besse Rimba, Fusanori Miura, Martiwi Diah Setiawati, Abu Bakar Sambah, Abd Rahman As-Syakur
Flood is one of the devastating natural disasters. By applying remote sensing technology, we determined the flood inundation areas, using the Synthetic Apertures Radar (SAR) sensor. Removing the speckles due to the phase fluctuations of the electromagnetic return radar signals by filtering 2 times image by using small kernel size to prevent losing information. The kernel size was 3×3. Removing the speckles by single filter and double filters, we applied and combined14 kind of filters (Low Pass filters, Gaussian Low Pass filters, Median filters, Sobel filters, Roberts, User-defined convolution filter, Lee filters, Enhanced Lee filters, Frost filters, Enhanced frost filters, Gamma filters, Kuan filters, Local sigma filters, and Bit error filters). Using a single filter the best filter is Gamma filter and for the double filter, Local Sigma filter combined with Low Pass filter, Enhanced Frost and Enhanced Lee. The Sobel filter, Kuan filter and Robert filter could not be applied for flood inundation analysis because these filters are unable to reduce the speckle and prevent information.
Department of Environmental Science and Engineering, Yamaguchi University, Yamaguchi, Japan; Integrated Research System for Sustainability Science, University of Tokyo, Institute for Advance Study, Tokyo, Japan; Faculty of Fisheries and Marine Science, Brawijaya University, East Java, Indonesia; Marine Science Department, Faculty of Marine and Fisheries, Udayana University, Bali, Indonesia; Center for Remote Sensing and Ocean Science (CReSOS), Udayana University, Bali, Indonesia
