作者: Lênio Soares Galvão , Marco Antônio Pizarro , José Carlos Neves Epiphanio
DOI: 10.1016/S0034-4257(00)00170-X
关键词:
摘要: Abstract The relationships between Airborne Visible/Infrared Imaging Spectrometer (AVIRIS) surface reflectance values and constituents (total iron, organic matter, TiO2, Al2O3, SiO2) of samples representative three important soil types from central Brazil [Terra Roxa Estruturada (STE), Latossolo Vermelho-Escuro (SLE), Areia Quartzosa (SAQ)] were analyzed. End member spectra for green vegetation (GV), nonphotosynthetic (NPV), water (W), the selected by inspecting scatter plots derived principal components analysis (PCA) 140 AVIRIS bands. They then used to compose a six end unmixing model characterize spectral variations associated with different scene components, spatial distribution types, effects mixing on spectral-chemical composition relationships. Finally, regression equations fitted their highly correlated bands produce maps showing chemical variability in areas dominated presence exposed soils, as indicated results model. showed very good agreement constituents. largest squared correlation obtained Fe2O3, but affected transition red near-infrared interval nonsoil residues (e.g., senescent or litter) over surfaces. In comparison light loamy sand SAQ, dark-red clay STE SLE presented higher contents consequently lower overall scene, because greater amounts opaque minerals. prediction these remote sensing data close association demonstrate importance present investigation mapping erosion studies. All Rights Reserved.