Model-Based Segmentation of Cloud Structures In Satellite Image Sequences

Abstract : Meteorological image sequences acquired with remote sensing devices contain huge amounts of available data featuring the temporal evolution of highly deformable and complex structures. In this study is presented a robust and effective model-based segmentation procedure for approximating highly deformable cloud structures. The whole image sequence may be viewed as a 3D data set, but the classical techniques for performing 3D segmentation are not meaningful here, because cloud structures are not shapes in the classical sense. The model-based segmentation presented here uses level-sets controlled by particle systems. A specific energy formulation is introduced, permitting the use of robust conjugate gradient techniques. The model is validated on a real meteorological image sequence. We outline the generalizations of the this particle system to perform image analysis tasks on truly 3D images.
Type de document :
Communication dans un congrès
ICCV - Proceedings of the IEEE International Conference on Computer Vision, Jan 1998, Mumbai, India. IEEE Computer Society, pp.77-85, 1998
Liste complète des métadonnées

https://hal.inria.fr/inria-00423792
Contributeur : H. Yahia <>
Soumis le : lundi 12 octobre 2009 - 17:51:38
Dernière modification le : vendredi 25 mai 2018 - 12:02:05

Identifiants

  • HAL Id : inria-00423792, version 1

Collections

Citation

Hussein Yahia, Jean-Paul Berroir, Gilles Mazars. Model-Based Segmentation of Cloud Structures In Satellite Image Sequences. ICCV - Proceedings of the IEEE International Conference on Computer Vision, Jan 1998, Mumbai, India. IEEE Computer Society, pp.77-85, 1998. 〈inria-00423792〉

Partager

Métriques

Consultations de la notice

137