Remote Sensing Observation for a Coastal Zone Management System in the
Mediterranean Sea
Isabelle Herlin
AIR Project
INRIA Rocquencourt
The presentation emphasizes the use of remote sensing observation for a
coastal zone management system in the mediterranean sea.
INRIA is involved with environmental applications in dierent research areas
and applications. Some examples are :
- Image processing for remotely sensed data : vegetation growth, meteorology,
oceanography, seismic risks,
- Data assimilation in meteorology and oceanography,
- Modelling and control of renewable ressources : sheries, aquaculture,
phytoplakton, forest,
- Seismic and hydrological processes modelling,
- Information systems and data bases.
In image processing studies, INRIA projects may use the technical support
of two
start-up companies :
- ISTAR is the rst private company specialized in 3D SPOT data production,
- NOESIS is specialized in image processing software (VISILOG).
Moreover INRIA has a collaboration with ERCIM partners on a regular basis
and is using it to come up with new research programs with mediterranean
countries on environmental problems. The four main subjects identied for
a coastal zone management system with image processing or modelling are
:
- pollution,
- shing ressources,
- hydrological ressources,
- seismic risks,
and the presentation is dedicated to remotely sensed images in these four
applications.
It can be noted that these problems may be studied by numerous sensors :
SPOT, VEGETATION (to come), SEAWIFS (to come) and NIMBUS7 (former), ERS1-2,
TOPEX- POSEDON, NOAA, ...
These devices get dierent physical measures :
- Panchromatic or multispectral radiations (SPOT, VEGETATION),
- Altimetric measures (TOPEX-POSEIDON),
- Color of the ocean, phytoplankton (SEAWIFS),
- Radar measures, temperature, winds velocity (ERS1-2),
- Temperature (NOAA).
These data raise complex image processing concerns :
- There are multisensors and that implies data merging and use of extra
data in case of missing acquisitions,
- There are composite pixels with spatial resolutions range from 10
meters to 1 kilometer,
- There are huge images and it is consequently necessary to localize
interesting zones,
- There are temporal moving data and we are studying spatio-temporal
model, motion and deformation.