Abstract:
The objective of this work was to evaluate the use of the TM/Landsat 5 sensor imagens in the differentiation of commercial plantations of Eucalyptus urograndis and Eucalyptus dunnii of different ages. Plots were marked for identification of the two Eucalyptus species, in two different periods (2009 and 2011): for 3- and 5-year-old E. dunnii, and for 2.2- and 4.2-year-old E. urograndis. Evaluations were done for six bands of the TM/Landsat 5 sensor (B1, B2, B3, B4, B5, and B7), and for six vegetation indices - simple ratio (SR); and normalized difference vegetation index (NDVI); soil-adjusted vegetation index (Savi)-0.25; Savi-0.5; green NDVI (GNDVI); and the moisture-vegetation index (MVI). Image digital processing consisted of geometric, radiometric, and atmospheric corrections. The plantations of E. dunnii and E. urograndis were distinguished by five bands of Landsat (B2, B3, B4, B5, and B7) and three vegetation indices (Savi-0.5, Savi-0.25, and GNDVI), in 2009, and by four bands of Landsat (B2, B4, B5, and B7) and six vegetation indices (NDVI, SR, Savi-0.5, Savi-0.25, MVI, and GNDVI) in 2011. The spectral data extracted from TM/Landsat 5 images are effective, both to distinguish the species of eucalyptus as well as the same species in even-aged stand.
Index terms:
Eucalyptus; principal component analysis; spectral signature; forest plantations.
Thumbnail
