CONTRIBUTIONS TO THE MAPPING AND CLASSIFICATION OF LAND USE THROUGH SEGMENTATION METHOD

Abstract

The knowledge of the dynamics of a landscape in a region is an important factor in regional and local planning. The classification and analysis of images from remote orbital sensors became an important tool in the acquisition of information for land use mapping. The objective of this research was to evaluate changes in land use in the buffer zone of State Forest Edmundo Navarro de Andrade (FEENA), located in Rio Claro / SP, taking the years 1995, 2005 and 2015 as the time frame. Classifications of different types of land uses were carried out based on the remote sensing image transfer method. It was found that FEENA's buffer zone has undergone major changes over the years, mainly due to the presence of sugarcane monoculture, in addition to the creation and expansion of neighborhoods in the municipality. Finally, the adopted methods were obtained from the error matrix and the calculation of the Kappa index, which presented the result of 80% accuracy in the classification, provided within the level established by the consulted theoretical references. The result is considered a performance with strong agreement, proving the accuracy of the maps produced.