In the present research, the authors examined the effects of the rice density and background to the results of four component decomposition (Yamaguchi et al., 2005), which is one of the most effective approaches for the polarimetric scattering analysis. Rice plants under different densities and background conditions were used, and the X-band polarimetric scattering of rice were measured in an anechoic radio wave chamber in Niigata Univ. in September 2, 2008. The measured data were decomposed into different scattering components by four component decomposition. It was found that X-band scattering had little effects of the background differences under the normal density, while the scattering under the less density was more sensitive to the background. In addition, the saturated condition of soil water without rice showed much stronger surface backscattering than the non-inundated condition without rice. As rice fields are usually inundated in most of the planting season, it can be concluded that the accurate analysis of microwave scattering from rice field requires the rice height and the density.
The possibility to estimate the vegetation cover rate of rice plants using the laser scanner measurement over the rice plants was examined. The vegetation cover rate and 3D point cloud data of rice plants were acquired during the growing stage in the test field. In order to examine a new method without the detection process of the ground, the rice plant's crown was determined by the 3D point cloud data. The observation point density corresponding to the laser pulse incidence depth below the rice plant's crown was calculated from the 3D point cloud data. The vegetation cover rate was compared with the laser pulse penetration depth. As a result, the laser pulse incidence depth was a good parameter to estimate the vegetation cover rate.In this study, the estimation error was 3% using the laser pulse incidence depth at 90% of the accumulated observation point density below the rice plant's crown.
We propose a new method to select data points on the surface through three dimensional point data obtained from the laser scanner loaded on the airplane.We prove through experiments that this method extends the conventional one (Masaharu et al., 2002) from the viewpoint of entropy, which makes it possible to select the landform including slopes and asperities such as tilted ridges and valleys, and select the slope from the building and slope of the similar difference of the height. Finally we comment on the respects in which the proposed method is improved and on future prospects.