Centre dedocumentation
scientifique

Meta-learning based hyperspectral target detection using siamese network / Yulei Wang in IEEE Transactions on geoscience and remote sensing, vol 60 n° 4 (April 2022)  

Public [article]  inIEEE Transactions on geoscience and remote sensing > vol 60 n° 4 (April 2022) . - n° 5527913 Titre : Meta-learning based hyperspectral target detection using siamese network Type de document : Article/Communication Auteurs : Yulei Wang, Auteur ; Xi Chen, Auteur ; Fengchao Wang, Auteur ; et al., Auteur Année de publication : 2022 Article en page(s) : n° 5527913 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] classification pixellaire [Termes IGN] détection de cible [Termes IGN] espace euclidien [Termes IGN] filtrage numérique d'image [Termes IGN] image hyperspectrale [Termes IGN] réseau neuronal siamois [Termes IGN] triplet Résumé : (auteur) When predicting data for which limited supervised information is available, hyperspectral target detection methods based on deep transfer learning expect that the network will not require considerable retraining to generalize to unfamiliar application contexts. Meta-learning is an effective and practical framework for solving this problem in deep learning. This article proposes a new meta-learning based hyperspectral target detection using Siamese network (MLSN). First, a deep residual convolution feature embedding module is designed to embed spectral vectors into the Euclidean feature space. Then, the triplet loss is used to learn the intraclass similarity and interclass dissimilarity between spectra in embedding feature space by using the known labeled source data on the designed three-channel Siamese network for meta-training. The learned meta-knowledge is updated with the prior target spectrum through a designed two-channel Siamese network to quickly adapt to the new detection task. It should be noted that the parameters and structure of the deep residual convolution embedding modules of each channel in the Siamese network are identical. Finally, the spatial information is combined, and the detection map of the two-channel Siamese network is processed by the guiding image filtering and morphological closing operation, and a final detection result is obtained. Based on the experimental analysis of six real hyperspectral image datasets, the proposed MLSN has shown its excellent comprehensive performance. Numéro de notice : A2022-381 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1109/TGRS.2022.3169970 Date de publication en ligne : 22/04/2022 En ligne : https://doi.org/10.1109/TGRS.2022.3169970 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=100649 [article]

Effective triplet mining improves training of multi-scale pooled CNN for image retrieval / Federico Vaccaro in Machine Vision and Applications, vol 33 n° 1 (January 2022)  

Public [article]  inMachine Vision and Applications > vol 33 n° 1 (January 2022) . - n° 16 Titre : Effective triplet mining improves training of multi-scale pooled CNN for image retrieval Type de document : Article/Communication Auteurs : Federico Vaccaro, Auteur ; Marco Bertini, Auteur ; Tiberio Uricchio, Auteur ; et al., Auteur Année de publication : 2022 Article en page(s) : n° 16 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image optique [Termes IGN] agrégation [Termes IGN] analyse visuelle [Termes IGN] architecture de réseau [Termes IGN] classification par réseau neuronal convolutif [Termes IGN] exploration de données [Termes IGN] extraction de traits caractéristiques [Termes IGN] recherche d'image basée sur le contenu [Termes IGN] réseau neuronal siamois [Termes IGN] triplet Résumé : (auteur) In this paper, we address the problem of content-based image retrieval (CBIR) by learning images representations based on the activations of a Convolutional Neural Network. We propose an end-to-end trainable network architecture that exploits a novel multi-scale local pooling based on the trainable aggregation layer NetVLAD (Arandjelovic et al in Proceedings of the IEEE conference on computer vision and pattern recognition CVPR, NetVLAD, 2016) and bags of local features obtained by splitting the activations, allowing to reduce the dimensionality of the descriptor and to increase the performance of retrieval. Training is performed using an improved triplet mining procedure that selects samples based on their difficulty to obtain an effective image representation, reducing the risk of overfitting and loss of generalization. Extensive experiments show that our approach, that can be effectively used with different CNN architectures, obtains state-of-the-art results on standard and challenging CBIR datasets. Numéro de notice : A2022-237 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article DOI : 10.1007/s00138-021-01260-z Date de publication en ligne : 06/01/2022 En ligne : https://doi.org/10.1007/s00138-021-01260-z Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=100153 [article]

Empirical comparison between stochastic and deterministic modifiers over the French Auvergne geoid computation test-bed / Ropesh Goyal in Survey review, vol 54 n° 382 (January 2022)  

Public [article]  inSurvey review > vol 54 n° 382 (January 2022) . - pp 57 - 69 Titre : Empirical comparison between stochastic and deterministic modifiers over the French Auvergne geoid computation test-bed Type de document : Article/Communication Auteurs : Ropesh Goyal, Auteur ; Jonas Ågren, Auteur ; Will E. Featherstone, Auteur Année de publication : 2022 Article en page(s) : pp 57 - 69 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Géodésie physique [Termes IGN] analyse comparative [Termes IGN] Auvergne [Termes IGN] géoïde local [Termes IGN] harmonique sphérique [Termes IGN] méthode déterministe [Termes IGN] NGF-IGN69 [Termes IGN] nivellement par GPS [Termes IGN] processus stochastique [Termes IGN] quasi-géoïde Résumé : (auteur) Since 2006, several different groups have computed geoid and/or quasigeoid (quasi/geoid) models for the Auvergne test area in central France using various approaches. In this contribution, we compute and compare quasigeoid models for Auvergne using Curtin University of Technology’s and the Swedish Royal Institute of Technology’s approaches. These approaches differ in many ways, such as their treatment of the input data, choice of type of spherical harmonic model (combined or satellite-only), form and sequence of correction terms applied, and different modified Stokes’s kernels (deterministic or stochastic). We have also compared our results with most of the previously reported studies over Auvergne in order to seek any improvements with respect to time [exceptions are when different subsets of data have been used]. All studies considered here compare the computed quasigeoid models with the same 75 GPS-levelling heights over Auvergne. The standard deviation for almost all of the computations (without any fitting) is of the order of 30–40 mm, so there is not yet any clear indication whether any approach is necessarily better than any other nor improving over time. We also recommend more standardisation on the presentation of quasi/geoid comparisons with GPS-levelling data so that results from different approaches over the same areas can be compared more objectively. Numéro de notice : A2022-111 Affiliation des auteurs : non IGN Thématique : POSITIONNEMENT Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1080/00396265.2021.1871821 En ligne : https://doi.org/10.1080/00396265.2021.1871821 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=99628 [article]

Geoid determination through the combined least-squares adjustment of GNSS/levelling/gravity networks – a case study in Linyi, China / Dongmei Guo in Survey review, Vol 53 n° 381 (November 2021)  

Public [article]  inSurvey review > Vol 53 n° 381 (November 2021) . - pp 504 - 512 Titre : Geoid determination through the combined least-squares adjustment of GNSS/levelling/gravity networks – a case study in Linyi, China Type de document : Article/Communication Auteurs : Dongmei Guo, Auteur ; Zhixin Xue, Auteur Année de publication : 2021 Article en page(s) : pp 504 - 512 Note générale : Bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Géodésie physique [Termes IGN] analyse de variance [Termes IGN] Chine [Termes IGN] compensation par moindres carrés [Termes IGN] données GNSS [Termes IGN] géoïde altimétrique [Termes IGN] géoïde local [Termes IGN] levé gravimétrique [Termes IGN] méthode des moindres carrés [Termes IGN] modèle stochastique [Termes IGN] pondération [Termes IGN] réseau de nivellement [Termes IGN] réseau gravimétrique Résumé : (Auteur) A detailed discussion of the adjustment problems used to combine GNSS/levelling/gravity network data is provided in this paper. The two primary problems inherent to heterogeneous data networks, namely, parametric models that describe the datums and systematic distortions among the available data sets and stochastic models that describe the observational residuals, are described. For parametric models, a relationship between the transformation parameters and the effects of datums and systematic distortions inherent among different height data types is established based on a least squares criterion. For stochastic models, the stochastic errors in GNSS/levelling/gravity data are evaluated, and a Helmert variance component estimation approach is introduced to refine weighting models. Finally, the proposed model is applied to determine the hybrid geoid in Linyi, China. The numerical results validate the capability and effectiveness of the proposed combined adjustment technique for hybrid geoid computations, revealing an achievable external accuracy of ±1.22 cm compared with GNSS/levelling measurements, which can be increased by 0.44 cm compared with classic adjustments of GNSS/levelling/geoid height data. Numéro de notice : A2021-913 Affiliation des auteurs : non IGN Thématique : POSITIONNEMENT Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1080/00396265.2020.1842642 Date de publication en ligne : 16/11/2020 En ligne : https://doi.org/10.1080/00396265.2020.1842642 Format de la ressource électronique : URL Article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=99316 [article]

The potential impact of hydrodynamic leveling on the quality of the European vertical reference frame / Yosra Afrasteh in Journal of geodesy, vol 95 n° 8 (August 2021)  

Public [article]  inJournal of geodesy > vol 95 n° 8 (August 2021) . - n° 90 Titre : The potential impact of hydrodynamic leveling on the quality of the European vertical reference frame Type de document : Article/Communication Auteurs : Yosra Afrasteh, Auteur ; D.C. Slobbe, Auteur ; M. Verlaan, Auteur ; et al., Auteur Année de publication : 2021 Article en page(s) : n° 90 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Nivellement [Termes IGN] données marégraphiques [Termes IGN] écart type [Termes IGN] European vertical GPS reference network [Termes IGN] hydrodynamique [Termes IGN] méthode heuristique [Termes IGN] niveau moyen des mers [Termes IGN] nivellement [Termes IGN] qualité des données [Termes IGN] réseau européen unifié de nivellement Résumé : (auteur) The first objective of this paper is to assess by means of geodetic network analyses the impact of adding model-based hydrodynamic leveling data to the Unified European Leveling Network (UELN) data on the precision and reliability of the European Vertical Reference Frame (EVRF). In doing so, we used variance information from the latest UELN adjustment. The model-based hydrodynamic leveling data are assumed to be obtained from not-yet existing hydrodynamic models covering either all European seas surrounding the European mainland or parts of it that provide the required mean water level with uniform precision. A heuristic search algorithm was implemented to identify the set of hydrodynamic leveling connections that provide the lowest median of the propagated height standard deviations. In the scenario in which we only allow for connections between tide gauges located in the same sea basin, all having a precision of 3 cm, the median of the propagated height standard deviations improved by 38% compared to the spirit leveling-only solution. Except for the countries around the Black Sea, coastal countries benefit the most with a maximum improvement of 60% for Great Britain. We also found decreased redundancy numbers for the observations in the coastal areas and over the entire Great Britain. Allowing for connections between tide gauges among all European seas increased the impact to 42%. Lowering the precision of the hydrodynamic leveling data lowers the impact. The results show, however, that even in case the assumed precision is 5 cm, the overall improvement is still 29%. The second objective is to identify which tide gauges are most profitable in terms of impact. Our results show that these are the ones located in Sweden in which most height markers are located. The impact, however, hardly depends on the geographic location of the tide gauges within a country. Numéro de notice : A2021-586 Affiliation des auteurs : non IGN Thématique : POSITIONNEMENT Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1007/s00190-021-01543-3 Date de publication en ligne : 24/07/2021 En ligne : https://doi.org/10.1007/s00190-021-01543-3 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=98203 [article]

Gauss-Markov model with random parameters to adjust results of surveys of geodetic control networks / Marek Banas in Reports on geodesy and geoinformatics, vol 111 n° 1 (June 2021)  

Permalink

La nouvelle grille de conversion altimétrique RAF18b / François L'écu in XYZ, n° 167 (juin 2021)

Permalink

Le nivellement par GNSS chez SNCF Réseau / Antoine Beuvain Pacheco in XYZ, n° 166 (mars 2021)

Permalink

Bilan des jonctions entre les réseaux de nivellement français et italien / Alain Coulomb (2021)

Permalink

Descriptif technique du quasi-géoïde gravimétrique QGF16 et de la surface de conversion altimétrique RAF18b / François L'écu (2021)

Permalink

40 ans de géodésie à l'IGN (Institut Géographique National rebaptisé en 2012 Institut national de l'information géographique et forestière) : 2ème partie, la géodésie physique / Françoise Duquenne in XYZ, n° 162 (mars 2020)

Permalink

Quels plans de comparaison à Paris avant le nivellement général de la France ? / Alain Coulomb in XYZ, n° 162 (mars 2020)

Permalink

Radial interpolation of GPS and leveling data of ground deformation in a resurgent caldera: application to Campi Flegrei (Italy) / Andrea Bevilacqua in Journal of geodesy, vol 94 n°2 (February 2020)  

Permalink

Géodésie, topographie, cartographie / Bernard Lamy (2020)

Permalink

Georeferenced measurements of building objects with their simultaneous shape detection / Edward Osada in Survey review, Vol 52 n°370 (January 2020)  

Permalink