Informations et ressources scientifiques
sur le développement des zones arides et semi-arides

Accueil du site → Doctorat → Italie → Physically based approaches for monitoring vegetation from space. New generation sensors and operational perspectives

Università di Napoli "Federico II" (2007)

Physically based approaches for monitoring vegetation from space. New generation sensors and operational perspectives

Vuolo, Francesco

Titre : Physically based approaches for monitoring vegetation from space. New generation sensors and operational perspectives

Auteur : Vuolo, Francesco

Université de soutenance : Università di Napoli "Federico II"

Grade : Tesi di dottorato 2007

Résumé
The use of Earth Observation (E.O.) data to retrieve biophysical variables of land surface such as the Leaf Area Index (LAI) has been proven to be useful in many operative tools to repetitively gather information at spatial and temporal resolution suitable for agricultural applications. In the last years, the diverse capabilities of airborne and satellite remote sensing imagery have been extensively exploited and several approaches have been proposed to estimate the LAI with different accuracy at scales ranging from individual plots to large areas. So far, empirical approaches based on vegetation indices (VI) and alternative approaches based on inversion of physically based radiative transfer models of vegetation have been successfully applied using both airborne and satellite data. The main objective of the work is to exploit the rich information content of CHRIS/PROBA data, both in the directional and spectral domains, to estimate Leaf Area Index. For this purpose, inversion of a radiative transfer model was performed and results compared, in terms of accuracy and operational practicability, to a more empirical approach. Results show that the directional information content improves LAI estimation for two out of three of the analyzed crops. For the best case (corn), it was achieved a LAI RMSE of 0.41 by using 5 angles and 62 spectral bands with an improvement of almost 65% respect to 1 angle and 16 bands. Finally, the accuracy of the LAI estimation for the two approaches was demonstrated to be comparable.

Présentation

Version intégrale (1,1 MB)

Page publiée le 9 avril 2011, mise à jour le 18 mars 2019