Détail de l'auteur
Auteur Kaiyu Guan |
Documents disponibles écrits par cet auteur (1)
Ajouter le résultat dans votre panier Affiner la recherche Interroger des sources externes
A 3D convolutional neural network method for land cover classification using LiDAR and multi-temporal Landsat imagery / Zewei Xu in ISPRS Journal of photogrammetry and remote sensing, vol 144 (October 2018)
[article]
Titre : A 3D convolutional neural network method for land cover classification using LiDAR and multi-temporal Landsat imagery Type de document : Article/Communication Auteurs : Zewei Xu, Auteur ; Kaiyu Guan, Auteur ; Nathan Casler, Auteur ; Bin Peng, Auteur ; Shaowen Wang, Auteur Année de publication : 2018 Article en page(s) : pp 423 - 434 Note générale : Bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Applications photogrammétriques
[Termes IGN] apprentissage profond
[Termes IGN] classification par réseau neuronal
[Termes IGN] données lidar
[Termes IGN] données localisées 3D
[Termes IGN] extraction de traits caractéristiques
[Termes IGN] Illinois (Etats-Unis)
[Termes IGN] image Landsat
[Termes IGN] image multitemporelle
[Termes IGN] réseau neuronal convolutif
[Termes IGN] semis de pointsRésumé : (Auteur) Terrestrial landscape has complex three-dimensional (3D) features that are difficult to extract using traditional methods based on 2D representations. These methods often relegate such features to raster or metric-based (two-dimensional) representations based on Digital Surface Models (DSM) or Digital Elevation Models (DEM), and thus are not suitable for resolving morphological and intensity features for fine-scale land cover mapping. Small-footprint LiDAR provides an ideal way for capturing these 3D features. This research develops a novel method of integrating airborne LiDAR derived features and multi-temporal Landsat images to classify land cover types. We tested our approach in Williamson County, Illinois, which has diverse and mixed landscape features. Specifically, our method applied a 3D convolutional neural network (CNN) approach to extract features from LiDAR point clouds by (1) creating an occupancy grid, an intensity grid at 1-meter resolution, and then (2) normalizing and incorporating data into the 3D CNN. The extracted features (e.g., morphological and intensity features) from the 3D CNN were finally combined with multi-temporal spectral data to enhance the performance of land cover classification based on a Support Vector Machine classifier. Visual interpretation from both hyper-resolution photos and point clouds was used for training and preparation of testing data. The classification results show that our method outperforms a traditional method by 2.65% (from 81.52% to 84.17%) when solely using LiDAR and 2.19% (from 90.20% to 92.57%) when combining all available imageries. We demonstrate that our method can effectively extract LiDAR features and improve fine-scale land cover mapping through fusion of complementary types of remote sensing data. Numéro de notice : A2018-405 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1016/j.isprsjprs.2018.08.005 Date de publication en ligne : 22/08/2018 En ligne : https://doi.org/10.1016/j.isprsjprs.2018.08.005 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=90859
in ISPRS Journal of photogrammetry and remote sensing > vol 144 (October 2018) . - pp 423 - 434[article]Exemplaires(3)
Code-barres Cote Support Localisation Section Disponibilité 081-2018101 RAB Revue Centre de documentation En réserve L003 Disponible 081-2018103 DEP-EXM Revue LASTIG Dépôt en unité Exclu du prêt 081-2018102 DEP-EAF Revue Nancy Dépôt en unité Exclu du prêt