A Semantic-Based Approach for Landscape Identification

Anne Toulet (1) , Emmanuel Roux (2, 3) , Anne-Elisabeth Laques (2, 3) , Eric Delaître (3, 2) , Laurent Demagistri (2, 3) , Isabelle Mougenot (2, 4)

1 SMILE - Système Multi-agent, Interaction, Langage, Evolution
2 UMR 228 Espace-Dev, Espace pour le développement
3 IRD - Institut de Recherche pour le Développement
4 UM - Université de Montpellier

Here we present an original method for the automation of landscape identification in a satellite image. There are two major challenges in this process. The first lies in the ability to take all expert knowledge into account for the full time it takes to analyze the image. The second is successfully structuring and persisting this knowledge so that it becomes interoperable and usable in the Semantic Web context. In this paper, we explain how the combination of several strategies associating image processing, the calculation of specific characteristics and inductive logic programming (ILP) can feed into the automation process, and how the integration of knowledge via the construction of dedicated ontologies can meet these challenges.

Domaines: Traitement des images [eess.IV]; Web; Intelligence artificielle [cs.AI]; Machine Learning [stat.ML]; Apprentissage [cs.LG]; Théorie de l'information [cs.IT]

Anne Toulet, Emmanuel Roux, Anne-Elisabeth Laques, Eric Delaître, Laurent Demagistri, et al.. A Semantic-Based Approach for Landscape Identification. Bruno Pinaud; Fabrice Guillet; Fabien Gandon; Christine Largeron. Advances in Knowledge Discovery and Management, 834, Springer, pp.119-136, 2019, Studies in Computational Intelligence, 978-3-030-18129-1. ⟨10.1007/978-3-030-18129-1_6⟩. ⟨hal-02164143⟩