Modeling and structural parameters inversion of crop for multiangular remote sensing observations

Recently, more and more multiangular sensors have been launched into the space. These multiangular remote sensing observations can provide more directional information of land surface objectives than only nadir observation. To well model the directional observations in visible and near infrared (VNIR) band, bi-directional reflectance distribution function (BRDF) is much useful according to its characteristic. In this paper, we developed a geometric-optical and gap probability model to model the BRDF observations of crop while crop in the beginning of its life. This model validated by the field data of summer corn taking in 2000 in Luancheng, Hebei Province. The results show: this model can simulate the BRDF well, and it is simple to use for the horizontal homogeneous crop considering the change of leaf area index (LAI), leaf angle distribution (LAD) etc. The ultimate aim of the quantitative remote sensing is inversion. Although multiangular remote sensing observations can provide more directional information, the inversion is still difficult due to noise, ill-posed problems, etc. In this paper, we use the multi-stage, sample-direction dependent, target-decisions (MSDT) inversion strategy and above model to inverse, study how to use the prior knowledge, improve sign-noise ratio in inversion and how to divide the datasets and parameters during inversion process. The model and the methods we studied in this paper maybe an initial trial and need further study