Land Use and Land Cover Mapping Using Fraction Images Derived from Annual VIIRS-NPP Dataset

This article presents a method to map the extent of annual land-use and land-cover (LULC) in Mato Grosso State, located in the Brazilian Legal Amazon. The proposed method applies the Linear Spectral Mixing Model (LSMM) to VIIRS NPP dataset to derive monthly vegetation, soil and shade fraction images for regional analysis. We used 500 m monthly image mosaics for VIIRS in 2015 year. These fraction images have the advantage to reduce the volume of data to be analyzed highlighting the target characteristics. Then we generated only one mosaic for each fraction images for VIIRS dataset computing de maximum value through the year, facilitating the classification of LULC classes. The proposed method allowed to classify three LULC classes: forest, cropland and non-forest (Savannah and pasture) areas. In addition, it allowed to map burned areas occurred during the study period. The results are very important for planning and management by the government and non-governmental organizations.