Mapeamento de vegetação do estado de Minas Gerais utilizando dados MODIS / Vegetation mapping of Minas Gerais state using MODIS data

AUTOR(ES)
DATA DE PUBLICAÇÃO

2006

RESUMO

The use of moderate spatial resolution and almost temporal daily resolution satellite images is a viable option to monitor natural or cultivated vegetation and its dynamics. The State of Minas Gerais in Brazil, due to it´s diverse biomes and fisionomies was selected as the study area for dissertation. The aims of the present work were: (a) to analyze the vegetation spectral dynamics; (b) to evaluate the image restoration technique applied to MODIS data; (c) to evaluate the classification´s accuracy based on the field data; and (d) to compare the obtained results with the reference map generated by IEF/UFLA. The MODerate Imaging Spectroradiometer (MODIS) sensor on board of Terra and Aqua platforms provides the MOD13A1 product (Vegetation Index in 16 days composition with four spectral bands with 250m spatial resolution). With this MODIS product, it was organized a dataset for 2004. In the obtained results, it was observed that: vegetation seasonal behaviors were differentiated alongo the year, indicating possible ifluence of seasonality. In addition the restoration techniques apllied on MOD13 images improved the spatial resolution, modifying the variance and preserving the mean values of the data. The validation of the generated maps, based on data collected during field work, presented a kappa index of 0.73 for the map derived from MODIS images and 0.84 for the reference map. From the applied methodology, the obtained results allowed to conclude that MODIS sensor data can be used for continuous vegetation monitoring in region level, and for effective land use classing mapping in Minas Gerais State.

ASSUNTO(S)

minas gerais (estate) remote sensing mapping vegetation sensoriamento remoto restauração modis vegetação minas gerais (estado) mapeamento

ACESSO AO ARTIGO

Documentos Relacionados