int. j. remote sensing, 2002, vol. 23, no. 7, 1461–1474
Mapping tropical coastal vegetation using JERS-1 and ERS-1 radar
data with a decision tree classier
M. SIMARD*†, G. DE GRANDI‡, S. SAATCHI† and P. MAYAUX‡
†Jet Propulsion Laboratory, MS 300–227, 4800 Oak Grove Drive, Pasadena,
CA 91109, USA
‡Space Application Institute of the European Commission Joint Research
Centre, Ispra (VA), Italy
Abstract. The objective of this paper is to investigate the complementarity of
JERS-1 and ERS-1 data for mapping coastal tropical regions. We use a decision
tree classier to classify a coastal region of Gabon and describe the feature
contribution using the decision tree diagram. The JERS-1 Global Rain Forest
Mapping (GRFM) and ERS-1 Central Africa Mosaic Project (CAMP) datasets
are used. The result is a land cover map of the west coast of Gabon. The analysis
explicitly shows the complementary characteristics of the L- and C-band Synthetic
Aperture Radar (SAR) instruments. We demonstrate the usefulness of combined
use of L- and C-band data for large area mapping of coastal regions, especially
in ooded areas for discrimination of high and low mangroves as well as grasses
and tree swamps. The overall classication accuracy increases by 18% over single
band classication.
1. Introduction
The objective of this paper is to investigate the complementarity of ERS-1 and
JERS-1 Synthetic Aperture Radar (SAR) data for mapping of coastal tropical regions.
We use SAR image mosaics constructed in the frame of two continental-scale radar
mapping projects: the ERS-1 TREES CAMP (Central Africa Mosaic Project)
(Malingreau and Duchossois 1995, De Grandi et al. 1999) and JERS-1 GRFM
(Global Rain Forest Mapping) (Rosenqvist 1996, De Grandi et al. 2000a). The
former is a European Commission project executed at the Joint Research Centre
(JRC). The latter was initiated by NASDA (National Space Development Agency
of Japan) and the JRC (Joint Research Center), Jet Propulsion Laboratory and
NASA Alaska SAR Facility acted as the main processing nodes. We focus our study
on a coastal area of Gabon.
In moist tropical forest regions, the quasi-permanent cloud cover does not allow
for on-demand optical data acquisition of land cover. Thus, in these regions radar
data becomes invaluable for mapping land cover (Nezry et al. 1993, Luckman et al.
1997), hydrography and oods (Ford and Casey 1988, Hess et al. 1990, 1995, Kux
et al. 1993, Kasischke et al. 1997). Even if ERS-1 (CVV, C-band vertical polarization)
and JERS-1 (LHH, L-band horizontal polarization) are expected to individually
provide poor information on tropical forest vegetation species (Kasischke et al. 1997,
*e-mail: simard@innu.jpl.nasa.gov
International Journal of Remote Sensing
ISSN 0143-1161 print/ISSN 1366-5901 online © 2002 Taylor & Francis Ltd
http://www.tandf.co.uk/journals
DOI: 10.1080 /01431160110092984