Centre de documentation
scientifique

Generation of three-dimensional deformation maps from InSAR data using spectral diversity techniques / E. Erten in ISPRS Journal of photogrammetry and remote sensing, vol 65 n° 4 (July - August 2010)  

Public [article]  inISPRS Journal of photogrammetry and remote sensing > vol 65 n° 4 (July - August 2010) . - pp 388 - 394 Titre : Generation of three-dimensional deformation maps from InSAR data using spectral diversity techniques Type de document : Article/Communication Auteurs : E. Erten, Auteur ; A. Reigber, Auteur ; O. Hellwich, Auteur Année de publication : 2010 Article en page(s) : pp 388 - 394 Note générale : Bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image radar et applications [Termes IGN] déformation de la croute terrestre [Termes IGN] données localisées 3D [Termes IGN] image Envisat-ASAR [Termes IGN] image radar moirée [Termes IGN] interferométrie différentielle [Termes IGN] interféromètrie par radar à antenne synthétique [Termes IGN] Iran [Termes IGN] séisme Résumé : (Auteur) The capability of DInSAR (Differential Interferometric SAR) for precise large-scale deformation analysis has been shown in various case studies. Generally, DInSAR possesses a high potential for monitoring deformation, but only the velocity component parallel to the line-of-sight direction can be measured. An alternative approach, capable to retrieve the deformation velocity in both range and azimuth direction, is the so-called spectral diversity technique. Spectral diversity is based on a phase comparison between different sub-aperture interferograms of the scene and can generally be regarded as a high-performance technique for estimating the mis-registration between complex SAR images. In this paper, the following questions will be discussed: how to implement the spectral diversity technique for achieving the most accurate results; how to extract the full 3D deformation vector from a combination of ascending/descending passes and how to extract a surface deformation map if the data sets are not perfectly coherent. Finally, a statistical analysis of every individual processing step and an error propagation analysis is undertaken. In order to make a quantitative analysis of the technique, ENVISAT data sets of the Bam earthquake in Iran are used. Numéro de notice : A2010-298 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1016/j.isprsjprs.2010.04.005 En ligne : https://doi.org/10.1016/j.isprsjprs.2010.04.005 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=30492 [article]

Exemplaires(1)

Code-barres	Cote	Support	Localisation	Section	Disponibilité
081-2010041	SL	Revue	Centre de documentation	Revues en salle	Disponible

Linienextraktion aus SAR-Daten mit einem Markoff-Zufallsfeld-Modell / O. Hellwich (1997)

Public Titre : Linienextraktion aus SAR-Daten mit einem Markoff-Zufallsfeld-Modell Titre original : [Extraction de lignes à partir d'images SAR avec un modèle aléatoire de Markov] Type de document : Thèse/HDR Auteurs : O. Hellwich, Auteur Editeur : Munich : Bayerische Akademie der Wissenschaften Année de publication : 1997 Collection : DGK - C Sous-collection : Dissertationen num. 487 Importance : 118 p. Format : 21 x 30 cm ISBN/ISSN/EAN : 978-3-7696-9527-4 Note générale : Bibliographie Langues : Allemand (ger) Descripteur : [Vedettes matières IGN] Traitement d'image radar et applications [Termes IGN] champ aléatoire de Markov [Termes IGN] extraction automatique [Termes IGN] extraction de traits caractéristiques [Termes IGN] image ERS-SAR [Termes IGN] image radar [Termes IGN] image TOPSAR [Termes IGN] modèle de Markov [Termes IGN] objet géographique linéaire [Termes IGN] radargrammétrie Index. décimale : 35.20 Traitement d'image Résumé : (Auteur) The topic of this thesis is automatic line extraction from Synthetic Aperture Radar (SAR) data. It is scientifically as well as practically relevant, as line features are important for object extraction and modern radar satellites provide huge amounts of data which cannot be evaluated manually. In the past, line extraction from SAR data has proved to be particularly difficult. The reason is the speckle effect typical for SAR and all other coherent imaging methods. In this thesis, a line extraction method is developed which is more successful than previous methods. This is achieved by 1- using prior knowledge about the continuity of lines, 2- evaluating intensity and phase information of SAR scenes in a coherent framework, and 3- modelling SAR statistics. The line extraction method includes the following novelties: 1- Use of a Markov random field model for line extraction from SAR scenes, 2- Application of stochastic completion fields in a Markov random field model, 3- Utilization of the object dependent phase information in SAR scenes by fusing intensity and interferometric coherence data by means of a Bayesian approach. The continuity of lines is modelled by a Markov random field using stochastic completion fields to express dependencies between neighbouring pixels. In terms of image analysis, the method developed is a SAR specific low-level approach to the extraction of lines. Ways for further processing by mid and high-level approaches are sketched. Simulated as well as real SAR data are evaluated. The real SAR data are a TOPSAR scene of an urban area, an ERS-1 scene from Siberia, and an X-SAR scene of a suburban area comprising an airfield. By using prior knowledge in form of a Markov random field model and by combining intensity as well as coherence data a moderate improvement of the quality of line extraction is achieved. Gaps which are only a few pixels wide can mostly be closed by means of the stochastic completion fields. The disadvantage of the approach is the large amount of computation generally required by Markov random field approaches. Therefore, the contribution of the thesis is not so much a more efficient line extraction method, but rather an exploration of the potential of Markov random field models and of the fusion of intensity and coherence data for line extraction. The results show that contextual knowledge is very important for line extraction from noisy data. Furthermore, the results suggest that line extraction methods combining several data sources will be much more common in future. In addition to the line extraction method, the thesis includes a detailed introduction to SAR from the surveying engineering point of view. Numéro de notice : 28006 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Thèse étrangère Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=63353

Exemplaires(2)

Code-barres	Cote	Support	Localisation	Section	Disponibilité
28006-01	35.20	Livre	Centre de documentation	Télédétection	Disponible
28006-02	35.20	Livre	Centre de documentation	Télédétection	Disponible