Centre de documentation
scientifique

Products and services of the Urban THEIA Scientific Expertise Centre / Anne Puissant (2022)    

Public Titre : Products and services of the Urban THEIA Scientific Expertise Centre Type de document : Article/Communication Auteurs : Anne Puissant, Auteur ; Thibault Catry, Auteur ; Rémi Cresson, Auteur ; Nadine Dessay, Auteur ; Laurent Demagistri, Auteur ; Sébastien Gadal, Auteur ; Arnaud Le Bris , Auteur ; Kenji Ose, Auteur ; Benjamin Pillot, Auteur Editeur : Strasbourg : Université de Strasbourg Année de publication : 2022 Conférence : LPS 2022, ESA Living Planet Symposium 22/05/2022 27/05/2022 Bonn Allemagne programme sans actes Note générale : projet AIMCEE (Apport de l’Imagerie satellitaire Multi-Capteurs pour répondre aux Enjeux Environnementaux et sociétaux des socio-systèmes urbains) Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Applications de télédétection [Termes IGN] données spatiotemporelles [Termes IGN] image satellite Résumé : (auteur) The THEIA data and services centre (www.theia-land.fr) is a consortium of 12 French public institutions involved in Earth observation and environmental sciences (CEA, CEREMA, CIRAD, CNES, IGN, INRA, CNRS, IRD, Irstea, Météo France, AgroParisTech, and ONERA). THEIA was initiated in 2012 with the objective of increasing the use of space data by the scientific community and the public actors. The first years allowed structuring the national science and user communities, pooling resources to facilitate access to data and processing capacities, federating various previously unrelated initiatives, and disseminating the French achievements nationally and internationally. The THEIA Land Data and Services Centre (www.theia-land.fr) is a consortium of 12 French public institutions involved in Earth observation and environmental sciences (CEA, CEREMA, CIRAD, CNES, IGN, INRAE, CNRS, IRD, Irstea, Météo France, AgroParisTech, and ONERA). THEIA has been initiated with the objective of increasing the use of space data by the scientific community and the public actors. The Scientific Expertise Centers (SEC) cluster research groups on various thematic domains. The "Urban” SEC gathers experts in multi-sensor urban remote sensing. Researchers of this group have structured their works around the development of algorithms useful for urban remote sensing using optical and SAR sensors to propose “urban products” at three different spatial scales: (1) the urban footprint, (2) the urban fabrics and (3) the urban objects. The objective of this poster is to present recent (>2019) advances of the URBAN SEC at these three scales. For the first two, the proposed methods are adapted to the geographic context of urban cities (West Cities, South Cities first and North Cities). For each spatial scale, the objective is to propose validated scientific products already available or in the near-term through the THEIA Land Service and Data Infrastructure. At the macro-scale (urban footprint), an unsupervised automated approach is currently under development at Espace-DEV - Montpellier, and funded by a CNES project (TOSCA DELICIOSA). This method is derived from the FOTO algorithm originally developed to differentiate vegetation textures in HR and VHR satellite images (Couteron et al. 2006, Lang et al., 2019). It has been optimized and packaged into the FOTOTEX Python Open-Source library. The method is very well suited for areas with no or few urban settlement data or with quickly growing informal settlements. No training dataset is required, and the urban footprint can be identified from only one satellite image as long as it is not covered by clouds. For Western Cities where training datasets are available, the Urba-Opt processing chain based on an automatic and object-oriented approach has been deployed on HPC infrastructure and produce annually (since 2018) an urban settlement product which is available through the A2S dissemination infrastructure and on the Urban SEC of Theia land data and service Infrastructure. An ongoing research between LIVE and Espace Dev Labs focused on the interest to use the FOTOTEX result as training data in the Urba-Opt processing chain to propose an updated product of urban settlement for South cities. At the scales of urban fabrics, products are under research activities The LIVE lab. In the context of an ongoing PhD thesis (ANR TIMES) and Tosca project (CNES 2019-2022) Sentinel-2 single-date images are used to assess two semantic segmentation networks (U-Net) that we combined using feature fusion between a from scratch network and a pre-trained network on ImageNet. Three spectral or textural indices have been added to the both networks in order to improve the classification results. The results showed a performance gain for the fusion methods. The research activities are ongoing in order to test the S1 imagery and temporal series for training in a deep architecture. The IGN-LaSTIG - Univ. Paris Est has focused on the use of Sentinel-2 and VHR mono-temporal SPOT products to retrieve land cover information related to urban density. First, images undergo a U-net based semantic segmentation at urban object level to retrieve ‘topographic’ classes (buildings, roads, vegetation, …). Generalized information about urban fabrics is then derived out of these land cover maps thanks to another CNN architecture. Both a building density measure and a simplified Urban Atlas like land cover map are calculated. The UMR ESPACE has focused on the machine learning modeling of the evolution of urban territories of Arctic (Yakutsk) and North-Eastern Europe (Baltic States and Kaliningrad) cities since the post-Soviet period at two scales: those of the built-up area with high spatial resolution SPOT 6/7 images, and of the urban structures based on the use of Landsat 5 TM, Landsat 8 OLI, and Sentinel 2 MSI images. Environmental (urban vegetation), economic (agricultural transformation), and morphometric indexes have been developed to characterize the processes of urban restructuring (densification, renovation) and expansion of post-Soviet cities. A comparative analysis of the machine learning algorithms used was done on the South-East Baltic cities to evaluate their performance. At the scale of urban object (3), a map of building with their functions is proposed by the TETIS laboratory. The study targets the retrieval of buildings footprint using deep convolutional neural networks for semantic segmentation, from Spot-6/7 images (1,5m spacing), on the entire France mainland. A single model has been trained and validated from 1.2k Spot-6/7 scenes and 20M images patches. The LIVE Lab has focused on the detection of urban changes from tri-stereoscopic Pléiades imagery through 2017 to 2020. A processing chain based on a Random Forest classifiers (ImCLASS) has been tested and the impact of the height attribute to detect changes has been evaluated to characterize changes into three thematic classes of changes. Numéro de notice : C2022-016 Affiliation des auteurs : UGE-LASTIG+Ext (2020- ) Thématique : IMAGERIE/INFORMATIQUE Nature : Poster nature-HAL : Poster-avec-CL DOI : sans En ligne : https://express.converia.de/frontend/index.php?page_id=22745&additions_conferenc [...] Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=100842

Documents numériques

peut être téléchargé Products and services of the Urban THEIA Scientific Expertise Centre - poster Adobe Acrobat PDF

SenRVM: A multi-modal deep learning regression methodology for continuous vegetation monitoring with dense temporal NDVI time series / Anatol Garioud (2022) 

Public Titre : SenRVM: A multi-modal deep learning regression methodology for continuous vegetation monitoring with dense temporal NDVI time series Type de document : Article/Communication Auteurs : Anatol Garioud , Auteur ; Silvia Valero, Auteur ; Clément Mallet , Auteur Editeur : Saint-Mandé : Institut national de l'information géographique et forestière - IGN (2012-) Année de publication : 2022 Conférence : LPS 2022, ESA Living Planet Symposium 22/05/2022 27/05/2022 Bonn Allemagne programme sans actes Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image optique [Termes IGN] analyse d'image orientée objet [Termes IGN] dynamique de la végétation [Termes IGN] image Sentinel-MSI [Termes IGN] Normalized Difference Vegetation Index [Termes IGN] phénologie [Termes IGN] régression [Termes IGN] série temporelle [Termes IGN] surveillance de la végétation Résumé : (auteur) The Earth's biosphere and the phenology of vegetation are at the heart of climatic, economic and social concerns. Human activities have led to a significant degradation of ecosystem services (e.g. carbon sequestration, biodiversity, water quality, flood, and erosion regulation) provided by various extensive ecosystems such as forests, grasslands or crops. A key parameter for relevant climate modeling, public policy implementations or commercial applications is the temporal resolution at which vegetation is observed. As a tool providing synoptic and regular coverage of Earth’s surfaces, satellite Earth Observation has been increasingly adopted, among others, for estimating biomass, yields, modeling different fluxes or detecting changes. Optical images have been historically used for vegetation monitoring, considering their efficient discrimination of phenomena related to photosynthetic activity. To deal with missing data due to clouds, many interpolation strategies integrating one or more optical sensors have been developed. Most of these strategies are based on trend modelling that does not reflect the real evolution of the vegetation cover in many cases (sudden climatic impact, man-made effects). As a result, data that may be weeks or months apart are often interpolated on areas suffering from high cloud cover. Copernicus Sentinels provide new opportunities and unprecedented observations for the monitoring of vegetation’s dynamics. In particular, concordant optical and SAR data sets provided by the Sentinel-1 and 2 satellites open the door to new multi-sensor methodologies aiming at the reconstruction of missing information. Taking into account the still numerous non-cloudy observations provided by the Sentinel-2 satellites, a deep learning regression methodology, namely the Sentinels Regression for Vegetation Monitoring (SenRVM), has been developed. Its goal is the translation of SAR features acquired regardless of the climatic conditions into NDVI. The developed architecture integrates several deep learning architectures such as Multilayer Perceptron and Recurrent Neural Networks. The SenRVM regression strategy proposes the integration of auxiliary data such as climatic and topographic features. This allows accurate NDVI time series to be predicted by minimizing effects exogenous to the vegetation’s phenology through SAR acquisitions contextualization. Object-oriented analysis of the results is carried out on large scale areas for various vegetation types with distinct phenologies (grasslands, crops and forests). The results are analyzed by taking into account spatial and temporal aspects or with an ablation study of the Network’s inputs. The proposed approach is further compared with traditional interpolation methods exploiting monomodal (Whittaker smoothing, linear weighted interpolation) or multimodal (Random Forest, Gaussian Regression Processes, single Multilayer Perceptron) features. The potential of high-temporal NDVI time series obtained by the SenRVM method for several vegetation-related applications is subsequently illustrated. In particular, the interest of the obtained time series to observe the phenology and its associated parameters of the three main vegetation classes is presented. Numéro de notice : C2022-011 Affiliation des auteurs : UGE-LASTIG+Ext (2020- ) Thématique : IMAGERIE/INFORMATIQUE Nature : Poster nature-HAL : Poster-avec-CL DOI : sans Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=100786

Documents numériques

peut être téléchargé SenRVM - poster Adobe Acrobat PDF