Identification and mapping of natural vegetation are major issues for biodiversity management and conservation. Remotely sensed data with very high spatial resolution are currently used to study vegetation, but most satellite sensors are limited to four spectral bands, which is insufficient to identify some natural vegetation formations. The study objectives are to discriminate natural vegetation and identify natural vegetation formations using a Worldview-2 satellite image. The classification of the Worldview-2 image and ancillary thematic data was performed using a hybrid pixel-based and object-oriented approach. A hierarchical scheme using three levels was implemented, from land cover at a field scale to vegetation formation. This method was applied on a 48 km² site located on the French Atlantic coast which includes a classified NATURA 2000 dune and marsh system. The classification accuracy was very high, the Kappa index varying between 0.90 and 0.74 at land cover and vegetation formation levels respectively. These results show that Wordlview-2 images are suitable to identify natural vegetation. Vegetation maps derived from Worldview-2 images are more detailed than existing ones. They provide a useful medium for environmental management of vulnerable areas. The approach used to map natural vegetation is reproducible for a wider application by environmental managers.

International audienceThe purpose of this study is to evaluate the potential of aerial and satellite imageries ofhigh spatial resolution for large scale vegetation mapping on Brittany, Normandy and Paysde la Loire regions at 1/25 000. Different types of images (BDORTHO® IRC, SPOT5,Worldview-2) were acquired for four sites which represent vegetation diversity in the North-West of France. Classification processes were established and their reproducibility wasassessed on another site. The classification used is the nested classification system of theNational Botanical Conservatory of Brest (CBN), that schedules a phytosociological typologyand a physiognomical typology. This nested classification system is compatible with remotesensing. An object-oriented classification was applied on images. It was combined with pixelbasedclassification when applied on the Worldview-2 image. Results were evaluated at threeclassification levels, corresponding to land cover, large vegetation types, and plant formationtypes. Best results, from more conclusive to less conclusive, were obtained with Worldview-2image, then BDORTHO® IRC and then SPOT5 images. Some results are not satisfactoryfor some classes of plant formation type level, but they could be improved in adding photointerpretationin post-processing, in using multi-date images from different sensors and inusing GIS data.L'objectif de cette étude est d'évaluer les potentialités qu'offre l'imagerie aérienne et satellitaire à haute et très haute résolution spatiale pour la cartographie des grands types de végétation des régions Bretagne, Basse-Normandie et Pays de la Loire à l’échelle du1/25 000. Différents types d’images (BDORTHO® IRC, SPOT5, Worldview-2) ont étéacquises sur quatre sites représentatifs de la diversité des végétations présentes sur ce territoire.Des procédures de classification ont été établies et leur reproductibilité à d’autres sites aété évaluée. La typologie des végétations utilisée est celle proposée par le ConservatoireBotanique National de Brest (CBN de Brest) qui articule, par une démarche « bottom-up »,la typologie phytosociologique utilisée sur le terrain avec une typologie physionomique(structurale) compatible avec des données de télédétection. Une approche de classificationorientée-objet a été privilégiée. Seules les classifications appliquées à l’image Worldview-2combinent deux approches, orientée-objet et basée sur le pixel. Les performances globales desclassifications ont été évaluées à trois niveaux typologiques correspondant à l’occupation dusol, aux grands types de végétations et aux types de formations végétales. Les résultats, des plusconcluants au moins concluants, ont été obtenus par la classification de l’image Worldview-2,puis la BDORTHO® IRC et enfin des images SPOT5. Les résultats ne sont pas satisfaisantspour certaines classes du niveau « Types de formations végétales » mais ils pourraient êtreaméliorés par une étape de photo-interprétation en post-traitement, par l’utilisation d’imagesmulti-dates de différents capteurs et d’information...

This paper presents a methodology for monitoring vegetation in the Pays de Brest using new series of Sentinel-1 satellite images combining with Sentinel-2 and SPOT-6. This work consists of establishing an interferogram method of the main types of vegetation in order to achieve the coherence of a multi-temporal Sentinel-1 radar image series, in SLC format (C band, VV and VH polarization), between 2015 and 2016. We then proceed to calculating the radar backscatter coefficient based on Sentinel 1 images in GRD format. Multi-date and multipolarized color compositions will be made to detect changes. It also shows the importance of data synergy to obtain an excellent accuracy using Random Forest classification..