Identification of sugar cane fields in the state of sao paulo using a time series of AVHRR/NOAA satellite images.

AUTOR(ES)
FONTE

INTERNATIONAL WORKSHOP ON THE ANALYSIS OF MULTI-TEMPORAL REMOTE SENSING IMAGES

DATA DE PUBLICAÇÃO

2011

RESUMO

Brazil is the first world producer of sugar cane. Despite the economic and social importance of agribusiness to Brazil, it is still difficult to estimate the harvest of the its main agricultural crops with the precision and anticipation needed, justifying the study and development of new methods based on the use of remote sensing data, for example. Even considering the increasing availability of high spatial resolution images in recent years, AVHRR/NOAA satellites have some characteristics suitable for application in operational methods of agricultural monitoring specially for a crop like sugar cane. Thus, this paper had the main objective to identify sugar cane areas in the state of Sao Paulo using harmonic analysis applied to a time series of AVHRR/NOAA-17 available at CEPAGRI/UNICAMP. A decision tree was calculated to search patterns that could represent the sugar cane along the crop season 2006/2007. Application of these two methods (harmonic analysis and decision tree) on the time series used was able to identify sugar cane areas in the state of Sao Paulo (Brazil) with 92% confidence when compared to ground truth based on LANDSAT/TM images and available at CANASAT?s webpage (http://www.dsr.inpe.br/canasat). This methodology can be applied in an operational way to support the official systems of harvest forecasting.

ASSUNTO(S)

cana-de-açúcar sensoriamento remoto séries temporais de avhrr/noaa-17 Árvore de decisão análise harmônica georreferenciamento navpro processamento de imagens ndvi monitoramento agrícola agricultura remote sensing sugar cane decision tree

ACESSO AO ARTIGO

Documentos Relacionados