Community Research and Development Information Service - CORDIS

Abstract

The Global Rain Forest Mapping 9GRFM) RADAR Mosaics provide potentially an unprecedented source of information on the vegetation cover in the humid tropical ecosystems. The passage from the radar observations to quantitative thematic estimates and assessment of their error budget is the next requirement. This study tries to give an answer to this problem for a specific thematic context: the vegetation cover in the Amazon basin. The problem is attacked at several levels: first we introduce a novel technique for image segmentation based on a wavelet signal representation. Secondly the technique is applied to a number of test cases, which fully characterise the vegetation cover of interest. Quantitative information is extracted for the specific thematic context and comparison with the results obtained using a classical Isodata algorithm is performed. Finally, landsat TM-derived by the radar mosaics. This study clearly indicates that, within the stated accuracy limits, and within the boundaries of a particular thematic context, the GRFM radar mosaics offer an alternative source of information to optical data for mapping and monitoring tropical vegetation cover.

Additional information

Authors: DE GRANDI G, JRC-Space Applications Institute (SAI), Global Vegetation Monitoring Unit, Ispra (IT);EVA H, JRC-Space Applications Institute (SAI), Global Vegetation Monitoring Unit, Ispra (IT);ACHARD F, JRC-Space Applications Institute (SAI), Global Vegetation Monitoring Unit, Ispra (IT);SGRENZAROLI M, JRC-Space Applications Institute (SAI), Global Vegetation Monitoring Unit, Ispra (IT)
Bibliographic Reference: An oral report given at: Informazione Geografica: Innovazione e Formazione, 4th Conferenza Nazionale. Held at: Centro Congressi Magazzini del Contone, Genova (IT). 3-6 October, 2000
Record Number: 200013168 / Last updated on: 2001-03-20
Category: PUBLICATION
Original language: it
Available languages: it