Tropical mangrove mapping using fully-polarimetric radar data

Abstract

Although mangrove is one of important ecosystems in the world, it has been abused and exploited by human for various purposes. Monitoring mangrove is therefore required to maintain a balance between economy and conservation and provides up-to-date information for rehabilitation. Optical remote sensing data have delivered such information, however ever-changing atmospheric disturbance may significantly decrease thematic content. In this research, Synthetic Aperture Radar (SAR) fully polarimetric data were evaluated to present an alternative for mangrove mapping. Assessment using three statistical trees was performed on both tonal and textural data. It was noticeable that textural data delivered fairly good improvement which reduced the error rate to around 5-6% at L-band. This suggests that insertion of textural data is more important than any information derived from decomposition algorithm.