Avaliação de dados SAR polarimétricos para estimativa volumétrica de florestas tropicais / Polarimetric SAR data analysis for tropical forest volume estimation

AUTOR(ES)
DATA DE PUBLICAÇÃO

2007

RESUMO

In this study it was evaluated the potential of polarimetric synthetic aperture radar (PolSAR) data for the retrieval of timber volume over tropical forests, with sights to identify areas with potential for sustainable timber production. As a secondary focus, it was made a sensitivity analysis of the PolSAR data in detecting floristic and structural variations in a selectively logged tropical forest area. The work was carried out in the Tapajos National Forest using L-band data acquired by the SAR-R99B airborne sensor. Procedures such as antenna pattern correction, polarimetric calibration and absolute calibration were initially performed in the polarimetric data treatment. Multiple regression techniques were applied to find out the predictive efficacy of the PolSAR data on timber volume estimation. In such a way, ground measurements from conventional forest inventories were used to adjust first order models for timber volume prediction. The set of explanatory variables was comprised of polarimetric attributes based on radar power returns, and on phase information. The sensitivity analysis of the PolSAR data in detecting floristic and structural variations was accomplished based on target decomposition and co-polarized responses. The results suggested the most important attributes for the timber volume modeling were the HV backscattering coefficient, the volume scattering component of the Freeman decomposition, the HH-VV phase difference, and the polarimetric coherence. A set of independent data, used for the models validation, showed the mean error in the volume estimation was below 15%. Thus, the results of this study confirm the hypothesis that L-band PolSAR data can be applied to quantify timber stocks in the Amazon forest with the same accuracy as that presented by conventional forest inventories. Concerning the sensitivity analysis, the applied techniques were not effective enough to detect the floristic and structural variability in the study area.

ASSUNTO(S)

polarimetria sensoriamento remoto synthetic aperture radar radar de abertura sintética estrutural florestal polarimetry forest inventory floresta nacional do tapajós inventário florestal amazon floresta tropical amazônia tropical forest tapajos national forest forest structure remote sensing

ACESSO AO ARTIGO

Documentos Relacionados