The high resolution multispectral imagery needs to be segmented into regions that can be easily interpreted and which correspond roughly to the ''ground truth''. In this paper, we segment multispectral images MSG2, provided by meteorological satellite ''Meteosat Second Generation 2'', by using an approach based on support vector Markov model witch takes into account both the spectral and the spatial information. A multi-variable Gaussian distribution is used in image processing and the Gibbs energy is used to describe the process of labeling. There are several forms of Gibbs energy. We test the best known of them and evaluate the different results using the Borsotti function which is known to be more appropriate with our visual perception.

In this paper, we propose a new technique for merging the label maps obtained by the marginal segmentation of a multi-component image. In the marginal segmentation, each component of the multi-component image is independently segmented by labeling the pixels of the same class with the same label. Therefore the number of label maps corresponds to the number of components in the image. It is then necessary to merge them in order to have a single label map, i.e. a single segmented image. In the most merging techniques, the compatibility links between these maps are performed a priori by making the correspondences between their labels. However the various components are segmented and labeled independently, label maps are considered as independent sources. It is then difficult to establish the relationship compatibilities between labels. The method we propose does not a priori assume any compatibility links. The label maps are combined by superposition. Unfortunately, an over-segmentation is produced. To cope with this problem, the insignificant regions and classes are eliminated. Finally, classes are grouped by using hierarchical agglomerative clustering algorithm. Tests performed on color and satellite images show the effectiveness of this method and its superiority compared to the vector segmentation. The self-organizing map is used during the segmentation process in both marginal and vector segmentations.

No abstract