||The objective of this thesis is to develop "Big Data" methods, especially data mining ones for analysing Satellite Image Time Series (SITS). For example, the Sentinel-1A satellite, launched by the European Space Agency on 4th April 2014, produces up to 2.4 TB of data a day and allows for an acquisition of the same area at decametric resolution for each 12 days, day and night, regardless of weather conditions. Exploration of such data is a real scientific challenge (volume, velocity, variety of information) and represents a major opportunity for the monitoring and management of territories which is of interest for the city of Lyon. The latter supports us and will participate in the project monitoring. Expected results are new methods and prototypes for SITS automatic analysis and maps expressing the evolution of Lyon's hills. Such maps will also be produced for the glaciers of Mont-Blanc massif which is a privileged research axis of the University Savoie Mont-Blanc. This thesis is supported by a grant from la Région Rhône-Alpes.