Transactions on Machine Learning and Data Mining (ISSN: 1865-6781)

Volume 5 - Number 1 - July 2012 - Pages 23-44

Efficient Spatio-temporal Mining of Satellite Image Time Series for Agricultural Monitoring

A. Julea1, N. Méger2, C. Rigotti3, E. Trouvé2, R. Jolivet4, and P. Bolon2

1 Institute for Space Sciences, P.O. Box MG-23, Ro 077125, Bucharest-Magurele, Romania
2 Université de Savoie, Polytech Annecy-Chambéry, LISTIC Laboratory, BP 80439, F-74944 Annecy-le-Vieux Cedex, France
3 Université de Lyon, CNRS, INRIA, INSA-Lyon, LIRIS, UMR 5205, F-69621, Lyon, France
4 Université Joseph Fourier, ISTerre Laboratory, CNRS - UMR 5559. B.P. 53, F-38041 Grenoble Cedex 09, France


In this paper, we present a technique for helping experts in agricultural monitoring, by mining Satellite Image Time Series over culti- vated areas. We use frequent sequential patterns extended to this spatio- temporal context in order to extract sets of connected pixels sharing a similar temporal evolution. We show that a pixel connectivity constraint can be partially pushed to prune the search space, in conjunction with a support threshold. Together with a simple maximality constraint, the method reveals meaningful patterns in real datasets.

PDFDownload Paper (617 KB)

Back to Table of Contents