Community Research and Development Information Service - CORDIS

Abstract

A procedure is described to estimate time profiles of the normalized difference vegetation index (NDVI) based on weighted least squares regression methods that take into account the biased noise associated with the data and the possible troughs in the profile. The final estimator obtained with this method seems to show improvement in several aspects compared with the one obtained with an upper envelope approach. A new methodology is also presented which considers the time series structure of the NDVI signal, designed to deal with abnormal features that are presented in NDVI series. Some applications to the estimation of the profile are proposed within this framework.

Additional information

Authors: ÁLVAREZ P C, Universidad de Valladolid, Facultad de Ciencias, Departamento de Estadística e IO (ES);RUEDA C, Universidad de Valladolid, Facultad de Ciencias, Departamento de Estadística e IO (ES);GALLEGO J, JRC Ispra (IT);ALVAREZ P C, Universidad de Valladolid, Facultad de Ciencias, Departamento de Estadistica e IO (ES)
Bibliographic Reference: Article: Teledetección: Vsos y aplicaciones
Follow us on: RSS Facebook Twitter YouTube Managed by the EU Publications Office Top