- ANR OATMIL (lead)
- 2019-2022 / ANR MATS (lead) : A huge trend in recent earth observation missions is to target high temporal and spatial resolutions (e.g. SENTINEL-2 mission by ESA). Data resulting from these missions can then be used for fine grained studies in many applications. This project focus on three key environmental issues: agricultural practices and their impact, forest preservation and air quality monitoring. Based on identified key requirements for these application settings, MATS project aims to feature a complete rethinking of the literature in machine learning for time series, with a focus on large-scale methods that could operate even when little supervised information is available. In more details, MATS introduce new paradigms in large-scale time series classification, spatio-temporal modeling and weakly supervised approaches for time series. Proposed methods cover a wide range of machine learning problems including domain adaptation, clustering, metric learning and (semi-)supervised classification, for which dedicated methodology is lacking when time series data is at stake. Methods developed in the project are made available to the scientific community as well as to practitioners through an open-source toolbox in order to help dissemination to a wide range of application areas. Moreover, the application settings considered in the project are used to showcase benefits offered by methodologies developed in MATS in terms of time series analysis.
- ANR MULTISCALE (lead)
- 2017-2019 / ANR SESAME : The surveillance of the maritime traffic is a major issue for defense contexts (e.g., surveillance of specific zones, borders,…) as well as security and monitoring contexts (e.g., monitoring of the maritime traffic, of fisheries activities). Spaceborn technologies, especially satellite ship tracking from AIS messages (Automatic Identification System) and high-resolution imaging of sea surface, open new avenues to address such monitoring and surveillance objectives. SESAME initiative aims at developing new big-data-oriented approaches to deliver novel solutions for the management, analysis and visualization of multi-source satellite data streams. It involves four main scientific and technical tasks : Hardware and software platforms for the management, processing and visualization of multi-source satellite data streams for maritime traffic surveillance (Task 1), Analysis, modeling and detection of marine vessel behaviours from AIS data streams (Task 2), AIS-Sentinel data synergies for maritime traffic surveillance (Task 3), Visualization and mining of large-scale augmented marine vessel tracking databases (Task 4). A fifth task embeds the implementation of the proposed solutions for dual case-studies representa tive of the scientific and technical objectives targeted by the project. (web page)
- 2018-2020 / ANR DEEPDETECT : This project focuses on the detection and recognition of multiple small objects from remote sensing images with a variety of unknown backgrounds. The goal is to develop a deep learning architecture based on convolutional neural networks for detection and recognition purpose, and then to define relevant criteria for efficient evaluation process. The proposed framework is expected to tackle two applications : the detection and mapping of marine mammal populations from satellite images, and the detection and recognition of small vehicles in infrared images.
- ANR 6P
- ADEME SEMMACAPE (lead)
- CNES R&T Trees (lead)
- CNES R&T Vehicles
- DGA MMT Semantic segmentation
- FEAMP Game of Trawls
- GDR MADICS action MACLEAN (co-lead)
- PHC PAMOJA (with Kenya)
- Erasmus Mundus Copernicus Master in Digital Earth (host of the GeoData Science track)
- MACLEAN action from GDR MADICS
to be completed…