TEMPORAL ANALYSIS OF THE URBAN SPRAWL IN GUARUJÁ USING MACHINE LEARNING AND SATELLITE IMAGES

Satellite images and remote sensing techniques allow several studies, including the identification and characterization of land use and occupation and urban sprawl. Urban sprawl is usually associated with environmental degradation, and an increase in disasters has been observed. Therefore, this study analyzes the urban sprawl of Guarujá municipality using remote sensing techniques, machine learning, and Data Mining. The NDVI was performed to enhance the identification of the vegetation. A temporal analysis from 1990 to 2020 was performed, using satellite images from the Landsat series. The results indicate an increase in the urban area, and, consequently, a decrease in the vegetation. The CART algorithm correctly distinguished the urban areas, the vegetation, and the water. Moreover, the NDVI provided important information about environmental degradation and loss of biomass.