三种植被冠层二向反射分布函数模型的比较

Comparisons of three models for vegetation canopy bi-directional reflectance distribution function

随着定量遥感技术的发展,描述森林冠层二向反射分布函数(BRDF)的机理模型越来越多。该研究采用3种植被冠层BRDF模型——DART模型、4SCALE模型和MGEOSAIL模型,模拟了不同郁闭度样地在红光、近红外波段各个观测角度下的场景反射率,并比较分析了不同BRDF模型的适用性和局限性。结果表明:MGEOSAIL模型只适于模拟样地郁闭度较小、林木个体较大条件下的场景反射率,且热点效果不十分明显;DART模型和4SCALE模型适于任何郁闭度条件下的场景反射率的模拟,并且精度较高;4SCALE模型模拟的场景反射率介于DART模型模拟的1次散射与5次散射之间。这3种模型在模拟近红外波段的场景反射率时,均存在"碗边"效应。

Aims With the development of quantitative remote sensing technology, multiple bi-directional reflectance distribution function （BRDF） models have been developed to describe reflectance of forest canopies. Our objective is to compare the DART, 4SCALE and MGEOSAIL models for greater understanding their proper use. Methods We used the BRDF models to simulate reflectance of the scene with different canopy densities at various observation angles in red and near-infrared bands and compared the results to analyze model applicability and limitations. Important findings The MGEOSAIL model is suitable to simulate the scene reflectance with low canopy density and large trees, and the hotspot is obvious. The DART and 4SCALE models are accurate and suitable to simulate the scene reflectance with various kinds of canopy density. Also, the scene reflectance simulated by 4SCALE is higher than DART＇s simulated first-order scattering value, but lower than DART＇s fifth-order scat- tering value. All three models show the ＂bowl＂ shape in the near-infrared band.