Centre de documentation
scientifique

Exploring multiscale object-based convolutional neural network (multi-OCNN) for remote sensing image classification at high spatial resolution / Vitor Martins in ISPRS Journal of photogrammetry and remote sensing, vol 168 (October 2020)  

Public [article]  inISPRS Journal of photogrammetry and remote sensing > vol 168 (October 2020) . - pp 56 - 73 Titre : Exploring multiscale object-based convolutional neural network (multi-OCNN) for remote sensing image classification at high spatial resolution Type de document : Article/Communication Auteurs : Vitor Martins, Auteur ; Amy L. Kaleita, Auteur ; Brian K. Gelder, Auteur ; et al., Auteur Année de publication : 2020 Article en page(s) : pp 56 - 73 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image optique [Termes IGN] analyse d'image orientée objet [Termes IGN] apprentissage profond [Termes IGN] classification par réseau neuronal convolutif [Termes IGN] données multiéchelles [Termes IGN] hétérogénéité environnementale [Termes IGN] image à haute résolution [Termes IGN] occupation du sol [Termes IGN] reconnaissance d'objets [Termes IGN] segmentation d'image [Termes IGN] segmentation sémantique [Termes IGN] squelettisation Résumé : (auteur) Convolutional Neural Network (CNN) has been increasingly used for land cover mapping of remotely sensed imagery. However, large-area classification using traditional CNN is computationally expensive and produces coarse maps using a sliding window approach. To address this problem, object-based CNN (OCNN) becomes an alternative solution to improve classification performance. However, previous studies were mainly focused on urban areas or small scenes, and implementation of OCNN method is still needed for large-area classification over heterogeneous landscape. Additionally, the massive labeling of segmented objects requires a practical approach for less computation, including object analysis and multiple CNNs. This study presents a new multiscale OCNN (multi-OCNN) framework for large-scale land cover classification at 1-m resolution over 145,740 km2. Our approach consists of three main steps: (i) image segmentation, (ii) object analysis with skeleton-based algorithm, and (iii) application of multiple CNNs for final classification. Also, we developed a large benchmark dataset, called IowaNet, with 1 million labeled images and 10 classes. In our approach, multiscale CNNs were trained to capture the best contextual information during the semantic labeling of objects. Meanwhile, skeletonization algorithm provided morphological representation (“medial axis”) of objects to support the selection of convolutional locations for CNN predictions. In general, proposed multi-OCNN presented better classification accuracy (overall accuracy ~87.2%) compared to traditional patch-based CNN (81.6%) and fixed-input OCNN (82%). In addition, the results showed that this framework is 8.1 and 111.5 times faster than traditional pixel-wise CNN16 or CNN256, respectively. Multiple CNNs and object analysis have proved to be essential for accurate and fast classification. While multi-OCNN produced a high-level of spatial details in the land cover product, misclassification was observed for some classes, such as road versus buildings or shadow versus lake. Despite these minor drawbacks, our results also demonstrated the benefits of IowaNet training dataset in the model performance; overfitting process reduces as the number of samples increases. The limitations of multi-OCNN are partially explained by segmentation quality and limited number of spectral bands in the aerial data. With the advance of deep learning methods, this study supports the claim of multi-OCNN benefits for operational large-scale land cover product at 1-m resolution. Numéro de notice : A2020-634 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1016/j.isprsjprs.2020.08.004 Date de publication en ligne : 13/08/2020 En ligne : https://doi.org/10.1016/j.isprsjprs.2020.08.004 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=96057 [article]

Exemplaires(3)

Code-barres	Cote	Support	Localisation	Section	Disponibilité
081-2020101	RAB	Revue	Centre de documentation	En réserve L003	Disponible
081-2020103	DEP-RECP	Revue	LASTIG	Dépôt en unité	Exclu du prêt
081-2020102	DEP-RECF	Revue	Nancy	Dépôt en unité	Exclu du prêt

Assessment of USGS DEMs for modelling pothole inundation in the prairie pothole region of Iowa / Priyadarshi Upadhyay in Geocarto international, vol 35 n° 9 ([01/07/2020])  

Public [article]  inGeocarto international > vol 35 n° 9 [01/07/2020] . - pp 1018 - 1032 Titre : Assessment of USGS DEMs for modelling pothole inundation in the prairie pothole region of Iowa Type de document : Article/Communication Auteurs : Priyadarshi Upadhyay, Auteur ; Amy L. Kaleita, Auteur ; M. L. Soupir, Auteur Année de publication : 2020 Article en page(s) : pp 1018 - 1032 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Applications photogrammétriques [Termes IGN] données lidar [Termes IGN] Global Multi-resolution Terrain Elevation Data 2010 [Termes IGN] inondation [Termes IGN] Iowa (Etats-Unis) [Termes IGN] mare [Termes IGN] modèle numérique de surface [Termes IGN] profondeur [Termes IGN] semis de points Résumé : (auteur) This study aims to compare inundation in two potholes using Annualized Agricultural Non-Point Source Pollution model (AnnAGNPS) with three Digital Elevation Models (DEMs): a 1 m DEM prepared from the LiDAR data which is readily available for the state of Iowa, USGS 1/9 arc-second DEM (∼3 m) which covers about 25% of the conterminous U.S. and USGS 1/3 arc-second DEM (∼10 m) which covers the entire USA. In this study, we found that the variations in water depth and presence/absence of ponding in the potholes of size greater than 1 ha can be predicted using USGS DEMs. The estimates of average water depths using USGS 3 m DEM was found to be 6% and 2% lower than the 1 m LiDAR DEM and the estimates of average water depths using USGS 10 m DEM was found to be 7% and 12% higher than the 1 m LiDAR DEM for the Walnut and Bunny potholes, respectively. Numéro de notice : A2020-429 Affiliation des auteurs : non IGN Thématique : GEOMATIQUE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1080/10106049.2019.1573852 Date de publication en ligne : 06/03/2019 En ligne : https://doi.org/10.1080/10106049.2019.1573852 Format de la ressource électronique : url article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=95497 [article]

Exemplaires(1)

Code-barres	Cote	Support	Localisation	Section	Disponibilité
059-2020091	RAB	Revue	Centre de documentation	En réserve L003	Disponible