Centre de documentation
scientifique

Filling the voids in the SRTM elevation model: a tin-based delta surface approach / E. Luedeling in ISPRS Journal of photogrammetry and remote sensing, vol 62 n° 4 (September 2007)  

Public [article]  inISPRS Journal of photogrammetry and remote sensing > vol 62 n° 4 (September 2007) . - pp 283 - 294 Titre : Filling the voids in the SRTM elevation model: a tin-based delta surface approach Type de document : Article/Communication Auteurs : E. Luedeling, Auteur ; S. Siebert, Auteur ; A. Buerkert, Auteur Année de publication : 2007 Article en page(s) : pp 283 - 294 Note générale : Bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image [Termes IGN] altitude [Termes IGN] analyse comparative [Termes IGN] erreur systématique [Termes IGN] fusion de données [Termes IGN] interpolation spatiale [Termes IGN] MNS SRTM [Termes IGN] modèle numérique de surface [Termes IGN] montagne [Termes IGN] Oman [Termes IGN] positionnement par GPS [Termes IGN] Triangulated Irregular Network Résumé : (Auteur) The Digital Elevation Model (DEM) derived from NASA's Shuttle Radar Topography Mission is the most accurate near-global elevation model that is publicly available. However, it contains many data voids, mostly in mountainous terrain. This problem is particularly severe in the rugged Oman Mountains. This study presents a method to fill these voids using a fill surface derived from Russian military maps. For this we developed a new method, which is based on Triangular Irregular Networks (TINs). For each void, we extracted points around the edge of the void from the SRTM DEM and the fill surface. TINs were calculated from these points and converted to a base surface for each dataset. The fill base surface was subtracted from the fill surface, and the result added to the SRTM base surface. The fill surface could then seamlessly be merged with the SRTM DEM. For validation, we compared the resulting DEM to the original SRTM surface, to the fill DEM and to a surface calculated by the International Center for Tropical Agriculture (CIAT) from the SRTM data. We calculated the differences between measured GPS positions and the respective surfaces for 187,500 points throughout the mountain range (?GPS). Comparison of the means and standard deviations of these values showed that for the void areas, the fill surface was most accurate, with a standard deviation of the ?GPS from the mean ?GPS of 69 m, and only little accuracy was lost by merging it to the SRTM surface (standard deviation of 76 m). The CIAT model was much less accurate in these areas (standard deviation of 128 m). The results show that our method is capable of transferring the relative vertical accuracy of a fill surface to the void areas in the SRTM model, without introducing uncertainties about the absolute elevation of the fill surface. It is well suited for datasets with varying altitude biases, which is a common problem of older topographic information. Copyright ISPRS Numéro de notice : A2007-429 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1016/j.isprsjprs.2007.05.004 En ligne : https://doi.org/10.1016/j.isprsjprs.2007.05.004 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=28792 [article]

Exemplaires(1)

Code-barres	Cote	Support	Localisation	Section	Disponibilité
081-07061	SL	Revue	Centre de documentation	Revues en salle	Disponible