Author :
Xiaoming, Feng ; Yongkang, Chen ; Yingshi, Zhao
Abstract :
The grasslands of Neimeng Province in North China are destroyed more or less for the semiarid weather condition and human factors. It is necessary to do some supervising work. Remote sensing can provide sequential data of large area. Most of the models have their own limitation. No one in existence can model it ideally. Taking the work of Geometric-Optical model as a starting point, we arrive at a large-scale spectral-directional reflectance model aiming at the semiarid landscape. As we consider the difference between the tree and the grass, and the data characteristic of moderate-scale viewing from the space, two most important scene elements are accounted for in this model: bare soil and canopy. The model is composed of the Multi-spectral Reflectance Model (MSRM)(Nilson-Kuusk, 1994), which calculate the canopy reflectance, and the soil reflectance model. We tried two kinds of spectral-directional reflectance soil models here separately. One is mechanism model, that is SOILSPECT(S.Jacquemoud, 1992) model, the other is experience model, that is Price model (Price,1990) together with the BRDF Walthall model (Nilson-Kuusk,1989). The two coupled models are validated. The result show that coupled with SOILSPECT, the model is more efficient in modeling the angular distribution of reflectance, such as "hotspot effect", For the degraded grassland, the RMS between the MISR data and the calculated multi-angular data is 0.0137 in red band, 0.0164 in near IR band. The RMS between the MODIS data and the calculated spectral data is 0.0209. Coupled with Price+Walthall, the number of the parameters is reduced and easy to get. The RMS between the MISR data and the calculated multi-angular data is 0.0139 in red band, 0.0372 in near IR band. The RMS between the MODIS data and the calculated spectral data is 0.0258. In conclusion, We think it would be a good method that coupling the SOILSPECT model with MSRM, but calculating the parameters in SOILSPECT by Price model
Keywords :
geometrical optics; soil; vegetation mapping; BRDF Walthall model; Geometric-Optical model; IR band; MISR data; MODIS data; MSRM; Multispectral Reflectance Model; Neimeng Province; North China; Price model; SOILSPECT model; bare soil; canopy reflectance model; grass; grasslands; hotspot effect; human factors; moderate-scale viewing; multiangular data; red band; reflectance angular distribution; remote sensing reflectance model; semiarid landscape; semiarid weather condition; soil reflectance model; spectral data; spectral-directional reflectance soil model; tree; Degradation; Educational institutions; Geoscience; Human factors; Large-scale systems; MODIS; Reflectivity; Remote sensing; Soil; Solid modeling;
