Evaluating Classification Models in a Burned Areas’ Detection Approach Olga Oliveira Bittencourt, Fabiano Morelli, C´ ıcero Alves dos Santos J´ unior, and Rafael Santos (B ) INPE – National Institute for Space Research, Av. dos Astronautas, 1758, S˜ ao Jos´ e dos Campos, SP 12227-010, Brazil {olga.bittencourt,rafael.santos}@inpe.br Abstract. We present a study to improve automation and accuracy on a Woody Savannah burned areas’ classification process through the use of Machine Learning (ML) classification models. The reference method for this is to extract polygons from images through segmentation and identify changes in polygons extracted from images taken from the same area but in different times through manual labeling. However, not all dif- ferences correspond to burned areas: there are also deforestation, change in crops, and clouds. Our objective is to identify the changed areas caused by fire. We propose an approach that employs polygons’ attributes for classification and evaluation in order to identify changes caused by fire. This paper presents the more relevant classifier models to the problem, highlighting Random Forest and an Ensemble model, that achieved bet- ter results. The developed approach is validated over a study area in the Brazilian Woody Savannah against reference data derived from classifi- cations manually done by experts. The results indicate enhancement of the methods used so far, and will eventually be applied to more data from different areas and biomes. Keywords: Burned areas · Classification models · Remote sensing data 1 Introduction Woody Savannah (commonly referred to as Cerrado in Brazil) is a biodiversity rich region and one of the most threatened biomes in the country. It occupies around 204 million hectares, 24% of the Brazilian territory, and it is estimated that it lost almost half of its original vegetation cover. Deforestation and vege- tation fires are the main reasons for this process. More than 30,000 vegetation fires per year have been recorded in this biome in the last 15 years. Several stud- ies cover the fire-related aspects and their economic, social and environmental impacts [3, 6, 9, 16]. To understand the effects and impacts of fire on the Woody Savannah it is important to regularly monitor fire occurrences and to have accurate data prod- ucts. The Brazilian National Institute for Space Research (INPE) maintains a c  Springer Nature Switzerland AG 2019 S. Misra et al. (Eds.): ICCSA 2019, LNCS 11622, pp. 577–591, 2019. https://doi.org/10.1007/978-3-030-24305-0_43