International audienceThis paper deals with the use of an object-oriented classification approach for the identification of land cover types in a coastal area from a 10 m SPOT 5 image. A three level land cover typology is designed. Significant image objects are delineated using a multi-resolution segmentation algorithm. Decision rules using spectral, spatial and contextual information are developed to classify these image objects. The results are exported as a shapefile in a GIS software, and some attributes are modified to correspond to the three level typology. The classification accuracy is estimated with confusion matrix using ground data and aerial orthoimages. The overall classification accuracy for the whole SPOT image is 88,5 % at l...