Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

No Cloud on the Horizon: Probabilistic Gap Filling in Satellite Image Series

: Fischer, Raphael; Piatkowski, Nico; Pelletier, Charlotte; Webb, Geoffrey I.; Petitjean, François; Morik, Katharina


Webb, G. ; Institute of Electrical and Electronics Engineers -IEEE-; IEEE Computer Society:
IEEE 7th International Conference on Data Science and Advanced Analytics, DSAA 2020. Proceedings : 6-9 October 2020, Sydney, Australia
Los Alamitos, Calif.: IEEE Computer Society Conference Publishing Services (CPS), 2020
ISBN: 978-1-7281-8206-3
ISBN: 978-1-7281-8207-0
International Conference on Data Science and Advanced Analytics (DSAA) <7, 2020, Online>
Bundesministerium für Bildung und Forschung BMBF (Deutschland)
01IS18038 A/B/C; ML2R
Fraunhofer IAIS ()
probabilistic machine learning; gap filling; spatio-temporal; graphical model; remote sensing

Spatio-temporal data sets such as satellite image series are of utmost importance for understanding global developments like climate change or urbanization. However, incompleteness of data can greatly impact usability and knowledge discovery. In fact, there are many cases where not a single data point in the set is fully observed. For filling gaps, we introduce a novel approach that utilizes Markov random fields(MRFs). We extend the probabilistic framework to also consider empirical prior information, which allows to train even on highly incomplete data. Moreover, we devise a way to make discrete MRFs predict continuous values via state superposition. Experiments on real-world remote sensing imagery suffering from cloud cover show that the proposed approach outperforms state-of-the-art gap filling techniques.