Conditional mixed-state model for structural change analysis from very high resolution optical images

Benjamin Belmudez (1) , Veronique Prinet (1) , Jian-Feng Yao (2) , Patrick Bouthemy (3) , Xavier Descombes (4)

1 LIAMA - Laboratoire Franco-Chinois d'Informatique, d'Automatique et de Mathématiques Appliquées
2 IRMAR - Institut de Recherche Mathématique de Rennes
3 VISTAS - Spatio-Temporal Vision and Learning
4 ARIANA - Inverse problems in earth monitoring

The present work concerns the analysis of dynamic scenes from earth observation images. We are interested in building a map which, on one hand locates places of change, on the other hand, reconstructs a unique visual information of the non-change areas. We show in this paper that such a problem can naturally be takled with conditional mixed-state random field modeling (mixed-state CRF), where the "mixed state" refers to the symbolic or continous nature of the unknown variable. The maximum a posteriori (MAP) estimation of the CRF is, through the Hammersley-Clifford theorem, turned into an energy minimisation problem. We tested the model on several Quickbird images and illustrate the quality of the results.

Domaines: Traitement des images [eess.IV]

Benjamin Belmudez, Veronique Prinet, Jian-Feng Yao, Patrick Bouthemy, Xavier Descombes. Conditional mixed-state model for structural change analysis from very high resolution optical images. 2009 IEEE International Geosciences and Remote Sensing Symposium, Jul 2009, Cape Town, South Africa. ⟨inria-00398062⟩