Retrieval of soil moisture content with the use of neural network-scattering model

This research is a part of the PacRim-II AIRSAR experiment carried out over Korean peninsula in September 2000 and consists of two parts. One is the analysis of surface roughness characteristics and spatial distribution of soil moisture contents of the test site, acquired as a part of the ground truth survey in Jeju island. The other is the test of neural network-scattering model based estimator for the surface parameter retrieval from polarimetric SAR data. At the time of preparing this paper, we have not received the actual polarimetric SAR data yet and this paper discusses the research approach and research plans. It was tested that when backscatter coefficients of surface are represented as fuzzy numbers, the physical surface parameters are also estimated as fuzzy membership functions.