Détail de l'éditeur
Documents disponibles chez cet éditeur
Ajouter le résultat dans votre panier Affiner la recherche Interroger des sources externes
Postprocessing von SAR-Satellitenbilddaten / Francesco Holecz (1993)
Titre : Postprocessing von SAR-Satellitenbilddaten Type de document : Monographie Auteurs : Francesco Holecz, Auteur Editeur : Zurich : Remote Sensing Laboratory Année de publication : 1993 Collection : Remote sensing series num. 23 Importance : 150 p. Format : 15 x 22 cm Note générale : Bibliographie Langues : Allemand (ger) Descripteur : [Vedettes matières IGN] Traitement d'image radar et applications
[Termes descripteurs IGN] correction d'image
[Termes descripteurs IGN] distorsion d'image
[Termes descripteurs IGN] filtrage du rayonnement
[Termes descripteurs IGN] image radar moirée
[Termes descripteurs IGN] image SeaSat
[Termes descripteurs IGN] relief
[Termes descripteurs IGN] télédétection spatiale
Résumé : (Auteur) The launch of the Seasat satellite in 1978 marked the beginning of the era of spaceborne Synthetic Aperture Radar (SAR) remote sensing. SAR systems are active imaging systems, able to illuminate and record the Earth's surface using their own energy source. They achieve a spatial resolution comparable to optical systems. This work uses the example of a Seasat SAR scene in the Jura region of the Swiss Canton of Vaud to show the capabilities of, and limitations on, the recognition and extraction of information over land. For the first time a complete postprocessing of spaceborne SAR images has been applied to a relatively large area (around 840 km2) of rough terrain possessing considerable thematic variation. A five step procedure is proposed for the postprocessing of SAR images. These are filtering, geometric recti-fication, correction of relief-induced distortions, texture analysis, and classification. Three principal distortions that influence the image interpretation are corrected. They are image noise, geometric distortions and relief-indu-ced radiometric influences.
The image noise, caused by the coherent nature of the sensor, makes detection of objects very difficult. Thus an adaptive filtering that is based on a multiplicative image model has been applied (Frost et al,, 1984). The geometric distortions, clearly visible in areas with rough terrain, cause a non linear compression of the surface in the range direction. Because geometric rectification of SAR images cannot be accomplished in the SAR processing step, geometric distortions must be corrected subsequently by taking into account the imaging geome-try together with the parameters used during the SAR processing. The range-Doppler approach, developed by Meier (Meier et al., 1989) is used. The significant differences in intensity that are caused by the SAR sensor's imaging geometry depend on the slope and exposition of the relief relative to the sensor position. Sloped areas faced toward the sensor appear brig-hter while sloped areas faced away from the sensor look darker. This leads to anisotropic signatures. The goal of the relief-induced radiometric correction is the production of isotropic signatures. Six different approaches, based on heuristic and theoretical methods, are used to demonstrate that the tonal information (e.g. the local mean backscattering), for spaceborne SAR systems, constitutes without question almost all of the useful information. Textural features from the methods that are studied show some separability, but are strongly correlated with the tonal information. The reason for the separability is that the textural features are based on the estimation of the mean value. The lack of textural information is mainly due to the poor spatial resolution of the Seasat system.
Because the textural information does not improve on results obtainable using only the tonal information, a forest classification can be carried out using only contextual information, based exclusively on the tonal informa-tion.
The results of this work show clearly that, in spite of coherent image noise and strong relief-induced distor-tions, the Seasat SAR system (L band, HH polarization) can be used for forest discrimination if a suitable postpro-cessing is carried out. However, such a postprocessing requires several work-intensive steps. Because the local mean backscattering constitutes most of the useful information, both the capabilities and the success of the digital image interpretation are dependent on the system's frequency, polarization, and spatial resolution.
Numéro de notice : 15691 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Monographie Permalink :
RéservationRéserver ce document
Code-barres Cote Support Localisation Section Disponibilité 15691-01 35.22 Livre Centre de documentation Télédétection Disponible