Ecole d'Automne

L'école d'Automne MDIS a pour objectif de réaliser des démonstrations de codes d'analyses (traitements satellitaires, modélisations) sous la forme de travaux pratiques. Elle aura lieu dans les locaux de l'Université de Strasbourg sur le Campus central.

Programme and details of the training courses

Monday 14 october 2019:

09h30-12h00: Class1 – The on-demand NSBAS processing chain for S1 data (C. Lasserre, M.-P. Doin, E. Pathier & F. Thollard – LGLTPE & ISTerre)

Description: The Sentinel-1 mission opens new potentialities for an intensive use of SAR interferometry to measure ground deformations. However, its default mode of acquisition (TOPSAR) and the amount of data available in near real time, require specific processing schemes, and significant computing and storage facilities not available to every researcher. ForM@Ter, the French Solid Earth data and services centre develops an on-demand service for Sentinel-1 InSAR processing using NSBAS, a French processing chain, based on ROI_Pac, starting from level-1 (SLC) Sentinel-1 data up to time-series of displacement.  The tutorial (3h) will introduce the main steps of the NSBAS processing chain, and their outputs. We will also present the ForM@Ter web user-interface, allowing users to select the data to be processed, to follow the processing, and to access intermediate and final products.

  • 13h30-15h30: Class2 – The on-demand SNAP-StaMPS PSI processing chain for S1 data (M. Foumelis – BRGM / GeoHazard Office)

Description: In this class, open source tools shall be utilized to analyze ground displacement using Copernicus Sentinel-1 data. The aim is to apply Persistent Scatterer Interferometry (PSI) in a semi-automatic processing scheme for land deformation monitoring using the ESAs SentiNel Application Platform (SNAP) and Stanford Method for Persistent Scatterers (StaMPS). For the demonstration of SNAP-StaMPS integrated PSI processing, a site with well-known ground deformation signal shall be selected. Hands-on exercise will be performed using both local and cloud-based processing resources.

  • 16h00-19h00: Class3 – SAR imagery processing with the Orfeo Toolbox / OTB (P. Durand & V. Poughon / CNES).

Description: Orfeo ToolBox (OTB) is an open-source project for remote sensing. It can process high resolution optical, multispectral and radar images. Many applications are available: from ortho-rectification, classification, SAR processing, and more. Since 2018 new modules for InSAR processing are progressively implemented ingesting Sentinel1 (SM, IW), CosmoSkymed (StripMap, SpotLight) sensors. All of OTB's algorithms are accessible from Monteverdi, QGIS, Python, the command line or C++. Monteverdi is an easy to use visualization tool with an emphasis on hardware accelerated rendering for high resolution imagery (optical and SAR). With it, end-users can visualize huge raw imagery products and access all of the applications in the toolbox. From resource limited laptops to high performance MPI clusters, OTB is available on Linux, macOS and Windows. It is community driven, extensible and heavily documented. Orfeo ToolBox is not a black box! According to the expectations of the attendees, the following topics will be presented: Introduction to SAR imagery with Monteverdi, radiometric calibration, geometric corrections, speckle filtering, polarimetry, InSAR processing with OTB (DiapOTB) and scripting OTB from Python.


Tuesday 15 october 2019:

  • 09h30-12h00: Class4 – On-demand optical processing services for quantifying Earth surface deformation: the MPIC, DSM-OPT and ALADIM processing chains (A. Déprez, F. Provost, D. Michéa & J.-P. Malet – EOST/IPGS & ESA/Esrin).

Description: Optical (e.g. multi-spectral) satellite imagery (S2, Landsat, Spot, Pléiades) are interesting sensors to quantify Earth Surface deformation in terms of horizontal motion (e.g. MPIC processing chain based on multi-temporal image correlation), in terms of vertical motion (e.g. DSM-OPT with the calculation of high resolution surface models from satellite with stereographic capacity), and in terms of change detection to locate geomorphological processes in pre/post event imagery (e.g. ALADIM processing chain). In order to allow the Earth science community to easily and quickly process multi-spectral images, fully automated processing chains have been developed, and have been implemented for distributed computation on high performances infrastructures. The objective of this class will be to present the three processing chains through practical example with case studies on a variety of geological objects (earthquake, volcano, landslide, glaciers).

  • 13h30-16h00: Class5 – The Geohazards Exploitaiton Platform in a nutshell (F. Brito & F. Provost / TerraDue & ESA/Esrin).

Description: The Geohazards Thematic Exploitation Platform (GEP) is part of the Thematic Exploitation Platforms (TEP) initiative set up by ESA to provide an environment to process EO data and support the user community concerning data exploitation through cloud-based services. The platform is in pre-operations with an Early Adopter programme, supporting approximately 100 user organisations in 35 countries whose access is sponsored by ESA. The aim of this course is to present the operational capabilities of the GEP platform through the use of APIs within Jupyter Notebooks. The examples will consist in: 1) discovering and staging-in Sentinel data; 2) discovering the GEP processing services and invoking them; 3) analysing/post-processing the results produced and, 4) publishing the analysis/post-process to GEP as community shared objects.  

  • 16h30-19h00: Class6 -  Mining displacement times series (Y. Yan, N. Méger, F. Vernier, C. Rigotti, C. Pothier, E. Trouvé & T. Nguyen / Listic & Liris).

Description: This class will present tools for the mining of the displacement field time series (DFTS) to extract relevant information from complex spatio-temporal datasets. It will start with an overview of confidence measures associated to the computation of displacement fields using optical or radar images. The practical part will address the problem of the exploratory mining of DFTS using sequential patterns.


Wednesday 16 october 2019

8h00-10h00: Class7 – DEFVOLC - Interface for the Inversion of volcanic deformation sources (V. Cayol & F. Dabaghi / LMV)

Description: Sources of deformation in volcanic context can be of multiple origins: massive magma reservoirs, sheets intrusions or faults. These sources are sometimes deforming simultaneously, which makes the analysis of surface displacements complex. To analyse the InSAR displacements at volcanoes or in a tectonic setting, we combine 3D elastostatic boundary element models which take topographies into account and a neighbourhood optimization algorithm. We simultaneously determine non-linear model parameters (source geometry and location) and linear model parameters (source stress drop), and assess mean model parameters and confidence intervals. In order to speed up inversions set ups, we developed a graphical interface. In order to accelerate the inversions, they run on clusters. In the workshop, we will present the model and inversion method used, as well a the interface. A practical example of a Piton de la Fournaise eruption will be presented.

  • 10h30-12h30: Class8 – CSI: a python module to set up fault slip inversions (R. Jolivet & Z. Duputel / ENS Paris & EOST/IPGS)

Description: Geodetic and seismological data can inform us about a wide range of slip behaviors along active faults. CSI stands for Classic Slip Inversion. It has been designed to handle data preparation and fault parameterization for slip inversions. CSI also includes tools to conduct simple slip inversions to quickly come up with slip models. In this tutorial, we present the main features of the module and how to make the different classes work together. Practical examples of slip inversions for actual earthquakes will be presented.

Personnes connectées : 1