Year: 2025
Type: article
Abstract: The Amazon forest is critical in global climate change and biodiversity preservation. However, acquiring accurate terrain data for this region is still challenging due to dense vegetation and high density of cloud cover. Digital Elevation Models (DEMs) such as ANADEM, Copernicus GLO-30, FABDEM, and SRTM V3 are employed to represent such surfaces but are often impacted by vegetation bias. This study evaluates the vertical accuracy of these DEMs in Amazonian deforested and first-growth areas, using the Amazon Radiography (RAM) data as a reference for metric estimations. The statistical analysis results demonstrated FABDEM and ANADEMs superior accuracy in dense forest regions. Findings suggest that region-specific terrain variations affect DEM precision, offering insights to enhance future topographic data for the Amazon.
Authors: Silva Junior, Eduardo José da, Santos, Daniel Rodrigues dos, Moreira, Leonardo Assumpção
Institution: Military Institute of Engineering
Language: Catalan
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
SDG: Life on land
Open Access status: green
OpenAlex ID: W7016046827