On Wednesday, October 10, 2018, Hoang Viet Tuan NGUYEN, PhD student in Information and Communication Science and Technology (ICST) Informatics at the Laboratory of Informatics, Systems, Information and Knowledge Processing (LISTIC), will support his thesis "Taking into account data quality when extracting and selecting changes in time series of displacement fields in satellite imagery".
The defense will be held at 11am, in the amphitheatre 052, at the IAE Savoie Mont Blanc, 4 chemin de Bellevue - Annecy-le-Vieux 74944 Annecy.
Summary of the thesis
This thesis work deals with the discovery of knowledge from Time Series of Displacement Fields (STCD) obtained from satellite imagery. Such series are nowadays central to the study and monitoring of natural phenomena such as earthquakes, volcanic eruptions or glacier displacement. Indeed, these series are rich in both spatial and temporal information and can now be produced regularly at lower cost thanks to space programmes such as the European Copernicus programme and its flagship Sentine! satellites. Our proposals are based on the extraction of Sequential Frequent Grouped Patterns (SFG). These patterns, originally defined for the extraction of knowledge from the Temporal Series of Satelite Images (TSSI), have shown their potential in early work to strip a STCD. However, they do not allow the use of the confidence indices intrinsic to STCDs and the randomized swap method used to select the most promising patterns does not take into account their spatiotemporal complementarities, each pattern being evaluated individually. Our contribution is thus twofold. A first proposal aims first of all at associating a reliability measure to each pattern using confidence indices. This measure makes it possible to select patterns carried by data that are on average sufficiently reliable. We propose a corresponding algorithm to perform extractions under reliability constraints. It is based on an efficient search of the most reliable occurrences by dynamic programming and on a pruning of the search space thanks to a partial push strategy, which allows to consider significant STCDs. This new method has been implemented on the basis of the existing prototype SITS-P2miner, developed within the LISTIC and LIRIS to extract and classify SFG patterns. A second contribution aimed at selecting the most promising patterns is also presented. This contribution, based on an informational criterion, takes into account both confidence indices and the way in which patterns complement each other spatially and temporally. For this purpose, confidence indices are interpreted as probabilities, and STCDs as probability databases with only partial distributions. The informational gain associated with a pattern is then defined in terms of the ability of its occurrences to complete/refine the distributions characterizing the data. On this basis, a heuristic is proposed in order to select informational and complementary patterns. This method allows to provide a set of patterns that are slightly redundant and therefore easier to interpret than those provided by random swap. It has been implemented in a dedicated prototype. Both proposals are evaluated both quantitatively and qualitatively using a reference STCD covering Greenland glaciers constructed from Landsat optical data. Another STCD that we have built from TerraSAR-X radar data covering the Mont-Blanc massif is also used. In addition to the fact that they are built from different data and remote sensing techniques, these series differ drastically in terms of confidence indices, the series covering the Mont-Blanc massif being at very low confidence levels. For both STCDs, the proposed methods were implemented under standard conditions in terms of resource consumption (time, space), and the experts' knowledge of the areas studied was confirmed and completed.