Détail d'une collection
Tagungsbände
ISSN :
0942-2870
|
Documents disponibles dans la collection (2)
Ajouter le résultat dans votre panier
Visionner les documents numériques
Affiner la recherche Interroger des sources externesInvolving different neighborhood types for the analysis of low-level geometric 2D and 3D features and their relevance for point cloud classification / Martin Weinmann (2017)
Titre : Involving different neighborhood types for the analysis of low-level geometric 2D and 3D features and their relevance for point cloud classification Type de document : Article/Communication Auteurs : Martin Weinmann, Auteur ; Clément Mallet Editeur : Deutsche Gesellschaft für Photogrammetrie, Fernerkundung und Geoinformation Année de publication : 2017 Collection : Tagungsbände, ISSN 0942-2870 num. 26 Conférence : DGPF 2017, 37. Wissenschaftlich-Technische Jahrestagung der DGPF, Kulturelles Erbe erfassen und bewahren - Von der Dokumentation zum virtuellen Rundgang 08/03/2017 10/03/2017 Wurtzbourg Allemagne open access proceedings Importance : pp 179 - 191 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Termes IGN] semis de points Résumé : (auteur) In this paper, we address the automatic interpretation of 3D point cloud data in terms of associating a (semantic) class label to each 3D point. In particular, we aim at analyzing the behavior of standard handcrafted low-level geometric 2D and 3D features for different neighborhood types. For this purpose, we present a framework that considers four neighborhood definitions as the basis for calculating a set of 18 low-level geometric 2D and 3D features which, in turn, are provided as input to three classifiers relying on different learning principles. We demonstrate the performance of our framework on a benchmark dataset for which a labeling with respect to three structural classes (linear, planar and volumetric structures) as well as a labeling with respect to five semantic classes (wire, pole/trunk, façade, ground and vegetation) is available. The derived results clearly reveal that the suitability of the considered neighborhood types and thus the relevance of respectively extracted features with respect to the classification task varies significantly. Numéro de notice : C2017-041 Affiliation des auteurs : LASTIG MATIS+Ext (2012-2019) Thématique : IMAGERIE Nature : Communication nature-HAL : ComAvecCL&ActesPubliésNat DOI : sans En ligne : https://www.dgpf.de/src/tagung/jt2017/proceedings/start.html Format de la ressource électronique : URL Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=91923 Documents numériques
en open access
Involving different neighborhood types ... - pdf éditeurAdobe Acrobat PDF
Titre : Segmentation and localization of individual trees from MMS point cloud data acquired in urban areas Type de document : Article/Communication Auteurs : Martin Weinmann, Auteur ; Clément Mallet Editeur : Deutsche Gesellschaft für Photogrammetrie, Fernerkundung und Geoinformation Année de publication : 2016 Collection : Tagungsbände, ISSN 0942-2870 num. 25 Projets : IQmulus / Métral, Claudine Conférence : DGPF 2016, 36. Wissenschaftlich-Technische Jahrestagung der DGPF, Dreiländertagung der SGPF, DGPF und OVG Lösungen für eine Welt im Wandel 07/06/2016 09/06/2016 Bern Suisse OA Proceedings Importance : pp 351 - 360 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Lasergrammétrie
[Termes IGN] arbre urbain
[Termes IGN] données lidar
[Termes IGN] données localisées 3D
[Termes IGN] segmentation
[Termes IGN] semis de points
[Termes IGN] zone urbaineRésumé : (auteur) In this paper, we address tree segmentation and localization in the scope of the IQmulus Processing Contest IQPC’15. Based on the part of pre-classified 3D point cloud data which corresponds to trees, we present a novel framework which involves a downsampling of the original data, a projection of the downsampled data onto a horizontally oriented plane, a mean-shift-based segmentation of the projected points, a transfer of the segmentation results to the original data, a refinement of the segmentation results via segment-based shape analysis, and a localization of respective tree trunks. The results derived for a benchmark dataset reveal that all individual trees are correctly detected and localized with both acceptable accuracy and reasonable computational effort. Numéro de notice : C2016-061 Affiliation des auteurs : LASTIG MATIS+Ext (2012-2019) Thématique : IMAGERIE Nature : Communication nature-HAL : ComAvecCL&ActesPubliésNat DOI : sans En ligne : https://www.dgpf.de/src/tagung/jt2016/start.html Format de la ressource électronique : URL Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=91920 Documents numériques
peut être téléchargé
Segmentation and localization... - pdf auteurAdobe Acrobat PDF
