Skip to Main Content
The normalized difference vegetation index (NDVI) has been widely applied in optical remote sensing. However, it has been demonstrated that NDVI is still partially affected by atmospheric path scattering and bidirectional (illumination and viewing geometry) effects. In this paper we present the benefit of using a bidirectional NDVI, and we discuss the problems in using the maximum NDVI composite method. Based on the assumption that a clear day has a larger NDVI value and a cloudy day has a smaller NDVI value (smaller reflectance in the near-infrared band and larger reflectance in red band due to atmospheric path scattering), the ratio of squared observed NDVI values and calculated NDVI values is used as a weight in our inversion method. The calculated NDVI values are derived from previously inverted bidirectional reflectance distribution functions (BRDFs). The inversion process will loop until all weights converge. Our research on the early Terra/MODIS data using a semiempirical kernel-driven BRDF model (the RossThick-LiTransit model) shows that this new method can improve inversion results whenever some cloudy pixels are not filtered out. As cloud detection and subpixel cloudiness are always a problem, this technique should still be very useful in improving the quality of BRDF inversion.