半干旱草场的多角度多波段反射率遥感模型

A Spectral-directional Reflectance Remote Sensing Model of the Semiarid Landscape

从几何光学(GO)模型出发,针对半干旱的草场植被,引入了多光谱冠层反射模型MSRM、土壤反射光谱Price模型以及土壤方向性反射Walthall模型,提出一种适用于区域尺度的多角度多波段反射率模型,并采用MISR和MODIS卫星数据进行验证,验证结果表明模型有较好的适用性。这为多角度、多波段遥感数据用于半干旱草场植被的定量研究提供了新的方法。

It’s a widely studied field to get the structural parameters of the vegetation with remote sensing data. In this paper, multi-spectral information is used in the multi-angular remote sensing research. A spectral-directional reflectance remote sensing model of regional-scale is presented for the semiarid grassland. Starting with the Geometric-Optical modeling and considering the difference between the tree and the grass, 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) for calculating the canopy reflectance, the soil spectral reflectance Price model and the soil BRDF Walthall model for calculating the soil reflectance. Poisson model is used to calculate the percentage of each component of remote sensing pixel. Validation has been done using MISR and MODIS data for the model. The result shows that RMS between modeled multi-angular reflectance and MISR data, and RMS between modeled multi-spectral reflectance and MODIS data are 5.31%, 6.37%, 4.69% and 5.14%, 4.68%, 5.71% for no degraded grassland, slightly degraded grassland and badly degraded grassland respectively. The precision of our model can meet requirement of research work.The sub model of soil reflectance is experimental. The study has shown that it can be replaced by the mechanism model, for the percentage of soil is dominated especially in the badly degraded grassland. In conclusion, it provides a new approach of application of remote sensing apllication to quantitative study and quick evaluation of grassland degradation in the semiarid region.