Identifying mangrove forests using radar remote sensing data

Abstract

Mangrove is one of the ecologically significant ecosystems in coastal areas, both on environment and biological resources. Radar remote sensing demonstrates a high potential in detecting, identifying, mapping and monitoring mangrove forests. Advantages of radar remote sensing are that almost unaffected by the weather phenomena in the atmosphere, e.g. clouds so that it can acquire images at day and night times. This study considers possibilities of ALOS PALSAR (L-band) and ENVISAT ASAR APP (C-band) for identifying mangrove forests. Results show that using single-date data of ENVISAT ASAR APP including dual polarization HH&HV are difficult to classify mangrove objects; whilst single-date data of ALOS PALSAR with dual polarization HH&HV have a better classification for tree density but at species level identification (e.g. Avicenna or Rhizophora) is more difficult. Results classified according to forest cover density data with overall accuracy of 81.91.