Angular signatures of NASA/NOAA Pathfinder AVHRR land data and applications to land cover identification

The angular signatures of the NASA/NOAA Pathfinder AVHRR land data set are examined by applying several widely used empirical models and by developing a new temporal angular model. These models are compared for 10 different cover types, and results show that the temporal model performs much better than other models. The coefficients of the temporal angular model are applied to classify those cover types. The classification rate with two channel inputs is about 97% based on the quadratic discriminant analysis. If a single channel is used, the thermal channels are better than the visible and near-infrared channels.