Mapping of the areas of soybean crop based on the spectral dynamics of the culture

Abstract

The objective of the present work was to evaluate, identify and map the area under soybean cultivation in the northern region of the State of Rio Grande do Sul. The study was developed based on multispectral data from the TM/Landsat-5 sensor and reference spectra of the various phases of phenological development of culture. The algorithm of supervised classification Spectral Angle Mapper (SAM) was applied successfully in one pre-processed TM/Landsat-5 sensor image. The procedure showed efficient capacity to identify in one period areas pertaining to one class, even under differentiated conditions of development. The classification process showed that approximately 42.66% of the area is under soybean cultivation and the SAM algorithm presents great potential to estimating the area under cultivation and the productivity of the crop.