Skip to Main content Skip to Navigation

Représentation et classification de structures spatiales -- Application à la reconnaissance de paysages agricoles

Ludmila Mangelinck 1, 2
2 ORPAILLEUR - Knowledge representation, reasonning
INRIA Lorraine, LORIA - Laboratoire Lorrain de Recherche en Informatique et ses Applications
Abstract : We are interested in the representation and classification of spatial structures in object based representation systems allowing us to recognize landscape patterns from satellite data. In our system, some classes denote patterns of landscape and others denote qualitative spatial relations. The instances of the latter are linking the former together and both of them are organized into a specialization hierarchy. The hierarchy of spatial relations is organized as a Galois lattice that allows us to represent the conjunction, disjunction, implication and also how the conditions on spatial entities are linked to the validity of spatial relations. In this work, we give an algorithm that computes the minimal set of conditions for each spatial relation and reduces therefore the number of computations needed on the images. Additionally, this lattice allow us to represent the mathematical proprieties of topological relations such as the implication, composition, reflexivity, symmetrical implication and so on. In order to define landscape patterns, we need to restrict the cardinality of the relations between spatial entities. Thus we added four new facets in the object based representation system we used, namely \yafool. The semantic of each facet is defined and included in the classification algorithm. This algorithm was used on image of Lorraine region.
Document type :
Complete list of metadatas
Contributor : Florence Le Ber <>
Submitted on : Wednesday, June 3, 2009 - 3:15:26 PM
Last modification on : Friday, October 23, 2020 - 4:34:14 PM
Long-term archiving on: : Thursday, June 10, 2010 - 10:09:08 PM



  • HAL Id : tel-00391141, version 1
  • PRODINRA : 249651



Ludmila Mangelinck. Représentation et classification de structures spatiales -- Application à la reconnaissance de paysages agricoles. Informatique [cs]. Université Henri Poincaré - Nancy I, 1998. Français. ⟨tel-00391141⟩



Record views


Files downloads