Análise de segmentação e validação de dados do Landsat 8 para Goiás e Distrito Federal/Brasil

Abstract

This study aimed at analyzing the most adequate methods to make the classification / segmentation between two distinct computer programs, Spring and Envi, and to analyze and validate data mapping of TerraClass Cerrado 2013 from classes of pasture, annual agriculture and perennial crops. Twenty-four scenes of Landsat 8 OLI corresponding to the surfaces of State of Goias and Federal District were selected. Several targeting tests were made with several images to investigate both feasibility and reliability of the software under study. As a result, mapping made from TerraClass Cerrado 2013 validated the data from the three classes. It has been used Visual inspection techniques using high resolution satellite images and time series. Results showed reliable, available, and free techniques for analysis and validation of data mapping.