One of the exceptional advantages of spaceborne remote sensors is their regular scanning of the Earth surface, resulting thus in Satellite Image Time Series (SITS), extremely useful to monitor natural or man-made phenomena on the ground. In this chapter, after providing a brief overview of the most recent methods proposed to process and/or analyze time series of remotely sensed data, we describe methods handling two issues: the unsupervised exploration of SITS and the reconstruction of multispectral images. In particular, we first present data mining methods for extracting spatiotemporal patterns in an unsupervised way and illustrate this approach on time series of displacement measurements derived from multitemporal InSAR images. Then we present two methods which aim to reconstruct multispectral images contaminated by the presence of clouds. The first one is based on a linear contextual prediction mode that reproduces the local spectro-temporal relationships characterizing a given time series of images. The second method tackles the image reconstruction problem within a compressive sensing formulation and with different implementation strategies. A rich set of illustrations on real and simulated examples is provided and discussed.

Satellite Image Time Series: Mathematical Models for Data Mining and Missing Data Restoration

Pasolli, Edoardo;
2018

Abstract

One of the exceptional advantages of spaceborne remote sensors is their regular scanning of the Earth surface, resulting thus in Satellite Image Time Series (SITS), extremely useful to monitor natural or man-made phenomena on the ground. In this chapter, after providing a brief overview of the most recent methods proposed to process and/or analyze time series of remotely sensed data, we describe methods handling two issues: the unsupervised exploration of SITS and the reconstruction of multispectral images. In particular, we first present data mining methods for extracting spatiotemporal patterns in an unsupervised way and illustrate this approach on time series of displacement measurements derived from multitemporal InSAR images. Then we present two methods which aim to reconstruct multispectral images contaminated by the presence of clouds. The first one is based on a linear contextual prediction mode that reproduces the local spectro-temporal relationships characterizing a given time series of images. The second method tackles the image reconstruction problem within a compressive sensing formulation and with different implementation strategies. A rich set of illustrations on real and simulated examples is provided and discussed.
978-3-319-66328-9
978-3-319-66330-2
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: http://hdl.handle.net/11588/739576
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 0
  • ???jsp.display-item.citation.isi??? 0
social impact