Bidirectional reflectance measurement is very important in the BRDF model validation and inversion . For many reasons

there exists many errors in the measured data. As in the laboratory

there is no wind and cloud

thus the measured data is more stable than in vivo. These data have been used to validated and invert many BRDF models. A BRDF process system based on MS-Windows has been developed

which include measured bidirectional reflectance data management and preprocess

data 2D and 3D analysis

BRDF model validation

and model inversion . In this paper we chose four objects as test samples which are sparse dense cotton

cotton

soybean

and corn

to validate four empirical and semiempirical models which are Walthall

Roujean

Verstraete

Rahman model. The validated result fit measured data well. We also used SAIL model to try our inversion test. At first

the parameter sensitivity and their distribution in view positions were analyzed

the according to analysis result we determined three cases: the most sensitivity area

the least sensitivity area and principal plane. Finally

the SAIL model parameters LAI

LAD

ρ

τ were inverted in three cases . The results show that :1. The inversion of sensitive area is better than insensitive area. 2. The more sensitive model parameter is

the more precise and stable the inversion is. Therefore

whether a BRDF model model is invertable or not is not only depend on its mathematical invertibility

but also depend on what parameters you want to invert and what measured data you are using.