An approach for imperfection propagation: Application to land cover change prediction

Amine Bouatay (1) , Wadii Boulila (2, 1) , Imed Riadh Farah (2, 1)

1 ENSI - École Nationale des Sciences de l'Informatique [Manouba]
2 ITI - Département Image et Traitement Information

We propose an approach that propagates imperfection throughout a model of land cover change prediction. The proposed approach is based on Polynomial Collocation method. The proposed approach estimates the imperfection in the output of the prediction model from the imperfection in its inputs. It incorporates two steps: 1. Computing membership functions for input variables for the model of land cover change prediction, and 2. Propagating mperfections of input variables throughout this model and determining the effect of these imperfection in the model. A probabilistic collocation method is used to propagate imperfection. Experimental results show the ffectiveness of the proposed approach in improving both computation time and prediction of the land cover change of the Saint-Denis region, Reunion Island.

Mots clés: Imperfection propagation; Satellite images; Change prediction; Probabilistic collocation method

Domaines: Traitement du signal et de l'image [eess.SP]

Amine Bouatay, Wadii Boulila, Imed Riadh Farah. An approach for imperfection propagation: Application to land cover change prediction. ICAISC 2014 : proceedings of the 13th International Conference on Artificial Intelligence and Soft Computing - Part I, Jun 2014, Zakopane, Poland. pp.637 - 648, ⟨10.1007/978-3-319-07173-2_54⟩. ⟨hal-01219488⟩