Please use this identifier to cite or link to this item: http://www.alice.cnptia.embrapa.br/alice/handle/doc/32122
Research center of Embrapa/Collection: Embrapa Territorial - Artigo em anais de congresso (ALICE)
Date Issued: 2009
Type of Material: Artigo em anais de congresso (ALICE)
Authors: VICTORIA, D. D. C.
OLIVEIRA, A. F. D.
GREGO, C. R.
Additional Information: DANIEL DE CASTRO VICTORIA, CNPM; ARYEVERTON FORTES DE OLIVEIRA, CNPM; CELIA REGINA GREGO, CNPM.
Title: Análise harmônica de séries temporais de imagens NDVI/MODIS para discriminação de coberturas vegetais.
Publisher: In: SIMPÓSIO BRASILEIRO DE SENSORIAMENTO REMOTO, 14., (SBSR), 2009, Natal. Anais... São José dos Campos: INPE, 2009.
Pages: p. 1589-1596.
Language: pt_BR
Keywords: MODIS
NDVI
Time series
Fourier
Series temporais
Description: The high temporal resolution information obtained with the Moderate Resolution Imaging Spectroradiometer (MODIS) is of great value when it comes to monitoring changes in Earth surface. Since several land cover types presents a distinguished temporal pattern in it?s spectral response, MODIS high temporal resolution can be used to identify such covers. This is specially true when observing the Normalized Difference Vegetation Index (NDVI) of agricultural land covers. Fourier transformations decomposes any signal represented in time to a frequency domain. Applying this transformation in a NDVI time-series results in parameters that describe how this signal behaves along several time frequencies (annual, semestral, etc). A strong annual signal indicates a land cover with a long growth cycle, such as sugar-cane (1 to 1.5 years) while stronger semestral signals are typical of other agricultural crops (soy, corn, beans). Also, observing the annual and semestral signals, it?s possible to distinguish agricultural areas with one or two crop cycles per year. A computational routine, independent of any commercial remote sensing package, has been developed in order to calculate Fourier amplitude and phase images of a NDVI time series. Applying such analysis over a diverse agricultural region in São Paulo state (Ribeirão Preto) indicates that long and short growth period crops are easily distinguished (sugar-cane and annual crops such as soy, corn, beans). Silvicultural areas are also easily distinguished due to their long growth period (5 years) however, these are confused with natural forests. A longer time series analysis could easily solve this.
Data Created: 2009-04-24
ISBN: 978-85-17-00044-7
Appears in Collections:Artigo em anais de congresso (CNPM)

Files in This Item:
File Description SizeFormat 
2391.pdf1,1 MBAdobe PDFThumbnail
View/Open

FacebookTwitterDeliciousLinkedInGoogle BookmarksMySpace