SAR data analysis for wave spectrum, bathymetric, and coastal information

Abstract

In this study, JERS-1 SAR data were examined and analyzed for extraction of coastal information, namely: (i) wave spectra, (ii) sea bottom topography, (iii) mangrove species classification, and (iv) mangrove forest biomass. The wave spectra were extracted using Fast Fourier Transform whilst, the TNO model were examined and analysed for the relationship between the backscatter and the sea bottom topography. The classification of SAR data to mangrove species level were performed using combined inparametric-parametric approach together with image segmentation technique. The allometric relationship between tree features and total weight were used as basis for modelling the biomass from SAR backscatter. The results of this study indicates that wave spectra can be extracted from JERS-1 SAR data, but it is rather restricted to certain sea condition to enable sea bottom topography be modelled from the backscatters. SAR data were found able to map mangrove at species level with an accuracy of 62 percent, and were also found as good estimator for biomass in the range 301-400 ton/ha with an overall 50 percent accuracy, at par with in-situ survey.