A simple physical model of vegetation reflectance for standardising optical satellite imagery

Monitoring vegetation using remote sensing is beset with difficulties. Prim ary among these are the variations of vegetation reflectance with sun zenit h angle, view zenith angle. and terrain slope angle. To provide corrections of these effects, for visible and near-infrared light, we derive two simpl e physicial models of vegetation reflectance, both of which have three para meters. The first model., WAK I. is applicable when illumination is by dire ct light only, and the second model, WAK II, is applicable when illuminatio n is by both diffuse and direct light. The models were tested by fitting th em to reflectance data of Virginia pine, irrigated wheat. and alfalfa. at l arge sun zenith angles, and then comparing predictions of reflectance at sm all sun zenith angles with measurements. The process was then reversed to d o the fitting on the small sun zenith measurements and the prediction of th e large sun zenith measurements. Using this method of assessment, the WAK m odels were found to perform better than several other published models. The WAK models were used to derive a formula for the variation of reflectance with terrain slope angle, which was tested by measuring reflectances of mic rocanopies set up on different slopes in the laboratory. The formula permit s topographic correction of vegetation reflectance without needing to know the bidirectional reflectance factor. The simplicity and robustness of the WAK models make them useful for general correction and calibration of remot ely sensed imagery. (C) 2001 Elsevier Science Inc. All rights reserved.