Avaliação comparativa da precisão vertical de modelos digitais de elevação derivados de satélites globais gratuitos

Abstract

Digital Elevation Models (DEM) make it possible to identify geomorphological results and model the dynamics of processes on the earth's surface. Identifying and analyzing the vertical accuracy of DEMs is relevant to understand the possibilities of using remote sensing products, which are periodically updated in different versions to improve quality. is research compared different DEM, such as NASADEM, TOPODATA, ASTER GDEM and ALOS AW3D30 in the Cerrado biome area, in the lower slopes. e analysis consisted of sampling 1002 homologous points from GNSS control points with high precision, and sampling in a vertical profile in 149 homologous points, towards the mouth-east direction of the Uberabinha river section. Statistical analyzes were performed using RMSE and Box-Plot and line graphs.  The results showed that NASADEM was the most accurate vertically (RMSE 3.8) than the other DEMs, with TOPODATA being the least accurate DEM (RMSR 8.81). It was found that vertical accuracy is more related to the (re)processing of the analyzed DEMs than to the type of acquisition of information by sensors, such as the InSar method and photogrammetry.