Centre de documentation
scientifique

Dépouillements

Sea-land segmentation using deep learning techniques for Landsat-8 OLI imagery / Ting Yang in Marine geodesy, Vol 43 n° 2 (March 2020)  

Public [article]  inMarine geodesy > Vol 43 n° 2 (March 2020) . - pp 105 - 133 Titre : Sea-land segmentation using deep learning techniques for Landsat-8 OLI imagery Type de document : Article/Communication Auteurs : Ting Yang, Auteur ; Zhonghua Hong, Auteur ; Yun Zhang, Auteur Année de publication : 2020 Article en page(s) : pp 105 - 133 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image optique [Termes IGN] apprentissage profond [Termes IGN] classification par réseau neuronal convolutif [Termes IGN] extraction automatique [Termes IGN] extraction de traits caractéristiques [Termes IGN] image Landsat-OLI [Termes IGN] littoral [Termes IGN] segmentation d'image [Termes IGN] segmentation sémantique [Termes IGN] trait de côte Résumé : (auteur) Automated coastline extraction from optical satellites is fundamental to coastal mapping, and sea-land segmentation is the core technology of coastline extraction. Deep convolutional neural networks (DCNNs) have performed well in semantic segmentation in recent years. However, sea-land segmentation using deep learning techniques remains a challenging task, due to the lack of a benchmark dataset and the difficulty of deciding which semantic segmentation model to use. We present a comparative framework of sea-land segmentation to Landsat-8 OLI imagery via semantic segmentation in deep learning techniques. Three issues are investigated: (1) constructing a sea-land benchmark dataset using Landsat-8 Operational Land Imager (OLI) imagery consisting of 18,000 km2 of coastline around China; (2) evaluating the feasibility and performance of sea-land segmentation by comparing the accuracy assessment, time complexity, spatial complexity and stability of state-of-the-art DCNNs methods; (3) choosing the most suitable semantic segmentation model for sea-land segmentation in accordance with Akaike information criterion (AIC) and Bayesian information criterion (BIC) model selection. Results show that the average test accuracy achieves over 99% accuracy, and the mean Intersection over Unions (mean IoU) is above 92%. These findings demonstrate that the Fully Convolutional DenseNet (FC-enseNet) performs better than other state-of-the-art methods in sea-land segmentation, based on both AIC and BIC. Considering training time efficiency, DeeplabV3+ performs better for sea-land segmentation. The sea-land segmentation benchmark dataset is available at: https://pan.baidu.com/s/1BlnHiltOLbLKe4TG8lZ5xg. Numéro de notice : A2020-220 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article DOI : 10.1080/01490419.2020.1713266 Date de publication en ligne : 20/01/2020 En ligne : https://doi.org/10.1080/01490419.2020.1713266 Format de la ressource électronique : url article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=94917 [article]

Validation of marine geoid models by utilizing hydrodynamic model and shipborne GNSS profiles / Sander Varbla in Marine geodesy, Vol 43 n° 2 (March 2020)  

Public [article]  inMarine geodesy > Vol 43 n° 2 (March 2020) . - pp 134 - 162 Titre : Validation of marine geoid models by utilizing hydrodynamic model and shipborne GNSS profiles Type de document : Article/Communication Auteurs : Sander Varbla, Auteur ; Artu Ellmann, Auteur ; Nicole Delpeche-Ellmann, Auteur Année de publication : 2020 Article en page(s) : pp 134 - 162 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Géodésie physique [Termes IGN] Baltique, mer [Termes IGN] données marégraphiques [Termes IGN] force de gravitation [Termes IGN] geoïde marin [Termes IGN] instrument embarqué [Termes IGN] instrumentation GNSS [Termes IGN] levé gravimétrique [Termes IGN] navire [Termes IGN] niveau de la mer [Termes IGN] simulation hydrodynamique Résumé : (auteur) An essential role of the FAMOS international cooperation project is to obtain new marine gravity observations over the Baltic Sea for improving gravimetric geoid modelling. To achieve targeted 5 cm modelling accuracy, it is important to acquire new gravimetric data, as the existing data over some regions are inaccurate and sparse. As the accuracy of contemporary geoid models over marine areas remains unknown, it is important to evaluate geoid modelling outcome by independent data. Thus, this study presents results of a shipborne marine gravity and GNSS campaign for validation of existing geoid models conducted in the eastern section of the Baltic Sea. Challenging aspects for utilizing shipborne GNSS profiles tend to be with quantifying vessel’s attitude, processing of noise in the data and referencing to the required datum. Consequently, the novelty of this study is in the development of methodology that considers the above-mentioned challenges. In addition, tide gauge records in conjunction with an operational hydrodynamic model are used to identify offshore sea level dynamics during the marine measurements. The results show improvements in geoid modelling due to new marine gravimetric data. It is concluded that the marine GNSS profiles can potentially provide complementary constraints in problematic geoid modelling areas. Numéro de notice : A2020-051 Affiliation des auteurs : non IGN Thématique : POSITIONNEMENT Nature : Article DOI : 10.1080/01490419.2019.1701153 Date de publication en ligne : 20/01/2020 En ligne : https://doi.org/10.1080/01490419.2019.1701153 Format de la ressource électronique : url article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=94918 [article]