中文摘要：

RossThick-LiSparseR模型和RossThick-LiTransitN模型是两种具有代表性的核驱动模型,可以较好地描述地表的方向反射特性.以大豆地、果园、草地、阔叶林和玉米地五种地物为研究对象,采用变分正则化方法,反演得到了红色波段及近红外波段最优的双向反射分布函数（BRDF）模型参数.对比反演结果,分析了两种模型在不同条件下的优势.结果表明,在红色波段和近红外波段,RossThick-LiSparseR,模型的数据拟合能力比较强,反演结果更接近真实值,且稳定性高;两种模型在反演黑空反照率上有相同的优势.

英文摘要：

RossThick-LiSparseR model and RossThick-LiTransitN model are typical kernel-driven models. The models can simulate surface bidirectional reflectance accurately. Data of surface targets on soybeans, orchard, lawn, hardwood and corn were measured at visible waveband and near infrared waveband. The inversion of the kernel-driven bidirectional reflectance distribution function（BRDF） model was implemented by regularization method. The ability of two models is accessed by comparing the retrieving results. It is found that RossThick-LiSparseR model l~as advantage in retrieving albedo at visible and near infrared waveband. The two models have the same ability to retrieve Black-sky albedo.