Canopy height model generation from historical images for monitoring of regeneration in atlantic rain forest
Nenhuma Miniatura disponível
Data
2019-01-01
Autores
Berveglieri, Adilson [UNESP]
Imai, Nilton Nobuhiro [UNESP]
Tommaselli, Antonio Maria Garcia [UNESP]
Martinez, Beatriz Fabretti [UNESP]
Título da Revista
ISSN da Revista
Título de Volume
Editor
Resumo
3D reconstruction of forest canopy for retrospective studies over time is possible from historical aerial photographs using modern photogrammetric algorithms. This enables the structural change analysis in forested areas and the conservation or regeneration monitoring in degraded areas. In this context, the article presents an application of photogrammetric techniques (combined with a terrain model derived from a LASER point cloud) to generate tree height models. Two height models, one dating from 1978 and another from 2017, were produced to evaluate the regeneration of an Atlantic Rainforest forest. A segmentation based on superpixels was made in both heights models to subdivide the canopy into homogeneous regions (coincident between the two dates). A local statistical analysis was conducted based on a classification in three levels of heights, which was validated with data of field. As a result, it was found that deforested areas are regenerating; however, such areas present difficulties to grow to the highest canopy. The error in height, estimated in the historical surface model, was 0.55 m, which is acceptable for forest studies and is comparable to results obtained in boreal forests.
Descrição
Palavras-chave
3D reconstruction, Absolute height model, Historical photogrammetry
Como citar
Anuario do Instituto de Geociencias, v. 42, n. 4, p. 208-218, 2019.