Uso de imagens dos sensores AWiFS e WFI: análise comparativa para a detecção da Classe Corte Raso na Amazônia no Projeto DETER – B

Abstract. In order to improve the Amazon’s near real time deforestation monitoring, it was released in 2015 the DETER-B Project, which stands out for using images generated from two different optic sensors, with spatial resolutions capable to map polygons from 3 ha. The sensors are: the Advanced Wide Field Sensor (AWiFs) of the Indian satellite Resource-Sat2 and the Wide Field Imager (WFI) of the CBERS-4 satellite, producing images of 56m and 64m of spatial resolution, respectively, both with revisit time of 5 days. However, the two sensors used by DETER-B Project do not have their bands in the same spectral interval, which causes a significant visual difference between the RGB composition images. This difference can possibly lead to incoherencies in the patterns’ interpretations of the Project’s mapped classes. Hence, this paper aims to conduct a comparative analysis of the deforestation pattern’s detection – specifically for the clear cut class – in the images of both sensors, evaluating the results of the Digital Image Processing resources applied in the images, which are used to standardize as much as possible the classes’ detection and mapping in the images from both sensors. Provided that, it was possible to extend the discussion of the methodology techniques developed by the DETER-B Project B in the use of images of sensors with different characteristics for the same feature. Keywords: Deforestation monitoring, digital image processing, RGB composition, contrast, mixture model linear model, monitoramento do desmatamento, processamento digital de imagens, composição RGB, realce, modelo linear de mistura espectral.