Mixture of multivariate Gaussian processes for classification of irregularly sampled satellite image time-series

Alexandre Constantin (1) , Mathieu Fauvel (2) , Stéphane Girard (1)

1 STATIFY - Modèles statistiques bayésiens et des valeurs extrêmes pour données structurées et de grande dimension
2 CESBIO - Centre d'études spatiales de la biosphère

The classification of irregularly sampled Satellite image time-series (SITS) is investigated in this paper. A multivariate Gaussian process mixture model is proposed to address the irregular sampling and the multivariate nature of the time-series. The spectral and temporal correlation is handled using a Kronecker structure on the covariance operator of the Gaussian process. The multivariate Gaussian process mixture model allows both for the classification of time-series and the imputation of missing values. Experimental results on simulated and real SITS data illustrate the importance of taking into account the spectral correlation to ensure a good behavior in terms of classification accuracy and reconstruction errors.

Mots clés: Multivariate Gaussian processes; Classification; Multivariate imputation of missing data; Irregular sampling; Satellite image time-series (SITS); Remote sensing

Domaines: Statistiques [math.ST]

Alexandre Constantin, Mathieu Fauvel, Stéphane Girard. Mixture of multivariate Gaussian processes for classification of irregularly sampled satellite image time-series. Statistics and Computing, 2022, 32 (5), pp.79. ⟨10.1007/s11222-022-10145-8⟩. ⟨hal-03280484v2⟩