Le portail francophone de la géomatique

Pour relancer l activite de cette liste, voici un sujet sur lequel je 
recherche un retour d experience :

Un collegue egyptien a debarqué dans mon bureau hier avec une idée 
bien precise :
Je voudrais quantifier la croissance de recifs coralliens a partir de 
l'exploitation d imagerie satellite.

Est ce que quelqu un a fait ce genre de chose par le passé ?


Voici differents liens, pour ceux que ca interesserait:
- New Worldwide Coral Reef Library : http://www.nasa.gov/vision/earth/
lookingatearth/coralreef_image.html
- Institute for Marine Remote Sensing (IMaRS)  Millennium Coral Reef 
Mapping Project  / http://imars.usf.edu/corals/
- AMaRS image Database :  http://imars.usf.edu/corals/imagedb.html


> "
>  Indeed, classification algorithms based only on spectral (color) 
> pixel-level information generally performs poorly (65-70% accuracy 
> using Landsat TM or ETM+ for a 6-8 class classification scheme) 
> (Andréfouët et al. in press, a). Such automatic processing 
> typically provides noisy products because pixel-based approaches 
> are sensitive to changes in water depth, to artifacts such as 
> breaking wave patterns, and of course they are unable to 
> discriminate classes with the same spectral properties (Bouvet et 
> al. in press). They require many corrections that can be avoided by 
> working directly at the segment-level. Therefore, classification 
> algorithms alone are not able to provide the map products 
> identified as objectives for this project.
> "
>> http://imars.usf.edu/corals/methods.html
>

- REMOTE SENSING Handbook for Tropical Coastal Management 
(extracts) : http://www.unesco.org/csi/pub/source/rs.htm


Stéphane Guéritte
Canterbury Christ Church University , Kent UK