Application of JERS-1 SAR Data for Tropical Forest Cover Mapping

摘要

[ABSTRACT] Monitoring of tropical forest cover is important because it is changing in many parts of the world. Synthetic Aperture Radar (SAR) is a useful tool to delineate the forest, as it can penetrate cloud, which often hinders the use of optical remote sensing in the tropics. The study area is located in the tropical forest region of south-eastern Bangladesh. The research used Japanese Earth Resource Satellite 1 (JERS-1) Synthetic Aperture Radar (SAR) imageries of 1997. Landsat Enhanced Thematic Mapper Plus (ETM+) imagery of 1999 assisted the interpretation procedure. Filed verification was done in 1998-1999. SAR imageries were geo-coded using previously rectified Landsat ETM+ pan image. Speckle noises on the imageries were removed by Lee-Sigma Filter. Three SAR scenes (Path／row: 144/264, 144/265, 145/264) were mosaiced. Relative image-to-image radiometric normalization was used to reduce the anomaly due to the differences in brightness before making the mosaic. The use of both SAR-intensity and texture image could separate forest from agriculture, shrimp farms, water and bare soil. Field-verification and the use of ETM+ optical image could differentiate eight different forest types (primary and secondary forest, bamboo, shrubs, plantation with indigenous species, Teak, Rubber and Acacia) but the use of only JERS-1 SAR image this was not possible. Finally, forest cover map prepared for the study area.