View-angle dependencies of vegetation isolines for higher-order standardization of spectral vegetation indices

Relationships between two data records such as spectral vegetation index (VI) from different sensors have been paid much attention for accuracy improvement in long-term data set. Such relationships vary with differences in spectral response function, spatial resolution, and viewing and illumination angles. This study focuses on the influences of observation geometries on VI relationship between two sensors. Specifically, an analytical relationship of VI between two sensors has been analyzed from both numerical and theoretical points of view. A set of numerical simulations have been conducted 1) to compare the slope of the vegetation isoline equation with the angular-based VIs, and 2) to compare the NIR-intercept with the distance-based VIs. The results clearly indicated that the slope of vegetation isoline is invariant to the variations of relative azimuthal angle with much less dependence on the view-zenith angle comparing to the BRDF effects on VIs. The theoretical justifications of those behaviors have also been shown based on a formulation of the vegetation isoline. It is concluded that those two variables (the isoline slope and intercept) can be a good candidate as the key variables for the purpose of the angular-standardization of higher-order data products.