Advantage of hyperspectral EO-1 Hyperion over multispectral IKONOS, GeoEye-1, WorldView-2, Landsat ETM+, and MODIS vegetation indices in crop biomass estimation

摘要: Abstract Crop biomass is increasingly being measured with surface reflectance data derived from multispectral broadband (MSBB) and hyperspectral narrowband (HNB) space-borne remotely sensed to increase the accuracy efficiency of crop yield models used in a wide array agricultural applications. However, few studies compare ability MSBBs versus HNBs capture variability. Therefore, we standard mining techniques identify set MSBB IKONOS, GeoEye-1, Landsat ETM+, MODIS, WorldView-2 sensors compared their performance HNB EO-1 Hyperion sensor explaining variability four important field crops (rice, alfalfa, cotton, maize). The analysis employed two-band (ratio) vegetation indices (TBVIs) multiband (additive) (MBVIs) Singular Value Decomposition (SVD) stepwise regression. Results demonstrated that HNB-derived TBVIs MBVIs performed better than MSBB-derived on per basis for pooled data: overall, explained 5–31% greater when various TBVIs; 3–33% MBVIs. improved mildly, by combining spectral information across multiple involving WorldView-2. A number advance modeling were determined. Based highest factor loadings first component SVD, “red-edge” range (700–740 nm) centered at 722 nm (bandwidth = 10 nm) stood out prominently, while five additional distinct portions recorded (400–2500 nm) 539 nm, 758 nm, 914 nm, 1130 nm, 1320 nm also important. best estimation involved 549 752 nm rice (R2 = 0.91); 925 1104 nm alfalfa (R2 = 0.81); 722 732 nm cotton (R2 = 0.97); 529 895 nm maize (R2 = 0.94). higher resolution users choose outweigh benefits come spatial MSBBs.