ABSTRACT
MODIS sensor images provide data covering broad areas with high periodicity. Such features are fundamental for monitoring of strategic agricultural crops in Brazil, just as sugarcane. Mathematics techniques have been applied to time-series of remote sensing imagery to characterize vegetation phenology and to better understand its dynamics. This study aimed at assessing the dynamics of sugarcane cultivation throughout São Paulo State by means of MODIS data temporal profiles from 2004/2005 to 2011/2012 crop years. Daubechies 8 Wavelet Transformation applied to time-series MODIS EVI2 consists of a robust technique to remove noise, thereby providing a better capture of sugarcane cycle trends within the entire time-series. Smoothed EVI2 temporal profiles could be used for sugarcane cultivation monitoring and land cover change detection, accompanying seasonal variations in phenology from planting or ratoon regrowth until harvest.
remote sensing; image processing; vegetation index; time-series