Resumo

ORTIZ LOZANO, Nidia Esperanza  e  PEREZ GOMEZ, Uriel. ASTER IMAGES FOR DISCRIMINATION OF AGRICULTURAL USE AREAS IN COLOMBIA. Rev. Fac. Nac. Agron. Medellín [online]. 2009, vol.62, n.1, pp.4923-4935. ISSN 0304-2847.

The advance of geographical information technology has led to the placement of new sensors for earth observation. ASTER (Advanced Spaceborne Thermal Emission and Reflection Radiometer) is considered as a latest generation sensor. It has special characteristics that places it as an alternative for studies of vegetation cover on earth. This work is based on its spatial, spectral and radiometric features to discriminate agricultural areas in the irrigation district of USOCOELLO in Colombia. We used a level 1B image from 2006, geometrically corrected, resampled and its radiance values transformed into reflectance values allowing us to make two compositions: original images (VNIR-SWIR) and fusion images in IHS transformation. The thematic legend was established from the classification scheme "Corine Land Cover - Colombia", defining 10 representative coverage categories in the image. The maximum likelihood classifier was used in the allocation phase. In the process of verifying and quantifying the level of accuracy, it was used, as ground truth, the database of geographical area at the site in the same date that the image was taken, obtaining an estimating a global reliability of 75 % for VNIR - SWIR images; if the variability of phenological stages of crops (rice, maize and sorghum) in the area and an important space contrast in the fusion image are taken into account, the obtained Kappa index was 0,75, which means that there is a substantial degree of agreement.

Palavras-chave : Agricultural crops; cartography; multispectral resolution; IHS; verification; Kappa.

        · resumo em Espanhol     · texto em Espanhol     · Espanhol ( pdf )