Spatio-Temporal Reasoning for the Classification of Satellite Image Time Series

François Petitjean (1) , Camille Kurtz (1) , Pierre Gançarski (1)

1 LSIIT - Laboratoire des Sciences de l'Image, de l'Informatique et de la Télédétection

Satellite Image Time Series (SITS) analysis is an important domain with various applications in land study. In the coming years, both high temporal and high spatial resolution SITS will be available. In the classical case, these data are studied by analysing the radiometric evolution of the pixels with time. When dealing with high spatial resolution images, object-based approaches are generally used in order to exploit the spatial relationships of the data. However, these approaches require a segmentation step to provide contextual information about the pixels. Even if the segmentation of single images is widely studied, its generalisation to a series of images remains an open-issue. This article aims at providing both temporal and spatial analysis of SITS. We propose first segmenting each image of the series, and then using these segmentations in order to characterise each pixel of the data with a spatial dimension (i.e. with contextual information). Providing spatially characterised pixels, pixel-based temporal analysis can be performed. Experiments carried out with this methodology show the relevance of this approach and the significance of the resulting extracted patterns in the context of the analysis of SITS.

Domaines: Intelligence artificielle [cs.AI]

François Petitjean, Camille Kurtz, Pierre Gançarski. Spatio-Temporal Reasoning for the Classification of Satellite Image Time Series. 2011. ⟨hal-00636814⟩