Synergistic use of Landsat and Hyperion imageries for ecological site classification in rangelands

The synergistic use of multispectral and hyperspectral data is evaluated for discrimination and mapping of ecological sites in the Patagonian rangelands. The method proposed encompasses two steps. First, mixture tuned matched filtering and logistic regression analyses are used for Hyperion data processing to obtain ecological site probability images in the area covered by hyperspectral imagery. Second, neural networks are used to model the relationships between the spectral response patterns of Landsat and the probability images from Hyperion, and to extrapolate them to the entire study area.