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]

HV-LSC-ex2 : velocity field interpolation using extended least-squares collocation / Rebekka Steffen in Journal of geodesy, vol 96 n° 3 (March 2022)  

Public [article]  inJournal of geodesy > vol 96 n° 3 (March 2022) . - n° 15 Titre : HV-LSC-ex2 : velocity field interpolation using extended least-squares collocation Type de document : Article/Communication Auteurs : Rebekka Steffen, Auteur ; Juliette Legrand, Auteur ; Jonas Ågren, Auteur ; Holger Steffen, Auteur ; M. Lidberg, Auteur Année de publication : 2022 Article en page(s) : n° 15 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Géodésie physique [Termes IGN] champ de vitesse [Termes IGN] collocation par moindres carrés [Termes IGN] interpolation [Termes IGN] modèle mathématique Résumé : (auteur) Least-squares collocation (LSC) is a widely used method applied in physical geodesy to separate observations into a signal and noise part but has received only little attention when interpolating velocity fields. The advantage of the LSC is the possibility to filter and interpolate as well as extrapolate the observations. Here, we will present several extensions to the traditional LSC technique, which allows the combined interpolation of both horizontal velocity components (horizontal velocity (HV)-LSC), the separation of velocity observations on different tectonic plates, and the removal of stationarity by moving variance (the latter as HV-LSC-ex(tended)2). Furthermore, the covariance analysis, which is required to find necessary input parameters for the LSC, is extended by finding a suitable variance and correlation length using both horizontal velocity components at the same time. The traditional LSC and all extensions are tested on a synthetic dataset to find the signal at known as well as newly defined points, with stations separated on four different plates with distinct plate velocities. The methodologies are evaluated by calculation of a misfit to the input data, and implementation of a leave-one-out cross-validation and a Jackknife resampling. The largest improvement in terms of reduced misfit and stability of the interpolation can be obtained when plate boundaries are considered. In addition, any small-scale changes can be filtered out using the moving-variance approach and a smoother velocity field is obtained. In comparison with interpolation using the Kriging method, the fit is better using the new HV-LSC-ex2 technique. Numéro de notice : A2022-151 Affiliation des auteurs : non IGN Thématique : MATHEMATIQUE/POSITIONNEMENT Nature : Article DOI : 10.1007/s00190-022-01601-4 Date de publication en ligne : 04/03/2022 En ligne : http://dx.doi.org/10.1007/s00190-022-01601-4 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=100111 [article]

PCEDNet: a lightweight neural network for fast and interactive edge detection in 3D point clouds / Chems-Eddine Himeur in ACM Transactions on Graphics, TOG, Vol 41 n° 1 (February 2022)  

Public [article]  inACM Transactions on Graphics, TOG > Vol 41 n° 1 (February 2022) . - n° 10 Titre : PCEDNet: a lightweight neural network for fast and interactive edge detection in 3D point clouds Type de document : Article/Communication Auteurs : Chems-Eddine Himeur, Auteur ; Thibault Lejemble, Auteur ; Thomas Pellegrini, Auteur ; Mathias Paulin, Auteur ; Loïc Barthe, Auteur ; Nicolas Mellado, Auteur Année de publication : 2022 Article en page(s) : n° 10 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Lasergrammétrie [Termes IGN] classification par réseau neuronal convolutif [Termes IGN] détection de contours [Termes IGN] données lidar [Termes IGN] données localisées 3D [Termes IGN] matrice [Termes IGN] semis de points Résumé : (auteur) In recent years, Convolutional Neural Networks (CNN) have proven to be efficient analysis tools for processing point clouds, e.g., for reconstruction, segmentation, and classification. In this article, we focus on the classification of edges in point clouds, where both edges and their surrounding are described. We propose a new parameterization adding to each point a set of differential information on its surrounding shape reconstructed at different scales. These parameters, stored in a Scale-Space Matrix (SSM), provide a well-suited information from which an adequate neural network can learn the description of edges and use it to efficiently detect them in acquired point clouds. After successfully applying a multi-scale CNN on SSMs for the efficient classification of edges and their neighborhood, we propose a new lightweight neural network architecture outperforming the CNN in learning time, processing time, and classification capabilities. Our architecture is compact, requires small learning sets, is very fast to train, and classifies millions of points in seconds. Numéro de notice : A2022-304 Affiliation des auteurs : non IGN Autre URL associée : vers ArXiv Thématique : IMAGERIE/INFORMATIQUE Nature : Article DOI : 10.1145/3481804 Date de publication en ligne : 10/11/2021 En ligne : https://doi.org/10.1145/3481804 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=100374 [article]

The method of detection and localization of configuration defects in geodetic networks by means of Tikhonov regularization / Roman Kadaj in Reports on geodesy and geoinformatics, vol 112 n° 1 (December 2021)  

Public [article]  inReports on geodesy and geoinformatics > vol 112 n° 1 (December 2021) . - pp 19 - 25 Titre : The method of detection and localization of configuration defects in geodetic networks by means of Tikhonov regularization Type de document : Article/Communication Auteurs : Roman Kadaj, Auteur Année de publication : 2021 Article en page(s) : pp 19 - 25 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Systèmes de référence et réseaux [Termes IGN] détection d'erreur [Termes IGN] matrice inversible [Termes IGN] régularisation de Tychonoff [Termes IGN] réseau géodésique [Termes IGN] valeur aberrante Résumé : (auteur) In adjusted geodetic networks, cases of local configuration defects (defects in the geometric structure of the network due to missing data or errors in point numbering) can be encountered, which lead to the singularity of the normal equation system in the least-squares procedure. Numbering errors in observation sets cause the computer program to define the network geometry incorrectly. Another cause of a defect may be accidental omission of certain data records, causing local indeterminacy or lowering of local reliability rates in a network. Obviously, the problem of a configuration defect may be easily detectable in networks with a small number of points. However, it becomes a real problem in large networks, where manual checking of all data becomes a very expensive task. The paper presents a new strategy for the detection of configuration defects with the use of the Tikhonov regularization method. The method was implemented in 1992 in the GEONET system (www.geonet.net.pl). Numéro de notice : A2021-961 Affiliation des auteurs : non IGN Thématique : MATHEMATIQUE/POSITIONNEMENT Nature : Article DOI : 10.2478/rgg-2021-0004 Date de publication en ligne : 17/12/2021 En ligne : https://doi.org/10.2478/rgg-2021-0004 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=100113 [article]

MSegnet, a practical network for building detection from high spatial resolution images / Bo Yu in Photogrammetric Engineering & Remote Sensing, PERS, vol 87 n° 12 (December 2021)  

Public [article]  inPhotogrammetric Engineering & Remote Sensing, PERS > vol 87 n° 12 (December 2021) . - pp 901 - 906 Titre : MSegnet, a practical network for building detection from high spatial resolution images Type de document : Article/Communication Auteurs : Bo Yu, Auteur ; Fang Chen, Auteur ; Ying Dong, Auteur Année de publication : 2021 Article en page(s) : pp 901 - 906 Note générale : Bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image optique [Termes IGN] classification par réseau neuronal convolutif [Termes IGN] détection du bâti [Termes IGN] image à haute résolution [Termes IGN] matrice [Termes IGN] segmentation multi-échelle [Termes IGN] segmentation sémantique Résumé : (Auteur) Building detection in big earth data by remote sensing is crucial for urban development. However, improving its accuracy remains challenging due to complicated background objects and different viewing angles from various remotely sensed images. The hereto proposed methods predominantly focus on multi-scale feature learning, which omits features in multiple aspect ratios. Moreover, postprocessing is required to refine the segmentation performance. We propose modified semantic segmentation (MSegnet), a single-shot semantic segmentation model based on a matrix of convolution layers to extract features in multiple scales and aspect ratios. MSegnet consists of two modules: backbone feature learning and matrix convolution to conduct vertical and horizontal learning. The matrix convolution comprises a set of convolution operations with different aspect ratios. MSegnet is applied to a public building data set that is widely used for evaluation and shown to achieve satisfactory accuracy, compared with the published single-shot methods. Numéro de notice : A2021-898 Affiliation des auteurs : non IGN Thématique : IMAGERIE/URBANISME Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.14358/PERS.21-00016R2 Date de publication en ligne : 01/12/2021 En ligne : https://doi.org/10.14358/PERS.21-00016R2 Format de la ressource électronique : URL Article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=99296 [article]

Réservation

Réserver ce document

Exemplaires (1)

Code-barres	Cote	Support	Localisation	Section	Disponibilité
105-2021121	SL	Revue	Centre de documentation	Revues en salle	Disponible

A spatial model of cognitive distance in cities / Ed Manley in International journal of geographical information science IJGIS, vol 35 n° 11 (November 2021)  

Permalink

A vector-based method for drainage network analysis based on LiDAR data / Fangzheng Lyu in Computers & geosciences, vol 156 (November 2021)  

Permalink

Least squares adjustment with a rank-deficient weight matrix and Its applicability to image/Lidar data processing / Radhika Ravi in Photogrammetric Engineering & Remote Sensing, PERS, vol 87 n° 10 (October 2021)  

Permalink

A novel method based on deep learning, GIS and geomatics software for building a 3D city model from VHR satellite stereo imagery / Massimiliano Pepe in ISPRS International journal of geo-information, vol 10 n° 10 (October 2021)  

Permalink

New algorithms for spherical harmonic analysis of area mean values over blocks delineated by equiangular and Gaussian grids / Rong Sun in Journal of geodesy, vol 95 n° 5 (May 2021)  

Permalink

Parallel computing for fast spatiotemporal weighted regression / Xiang Que in Computers & geosciences, vol 150 (May 2021)  

Permalink

Detecting ground deformation in the built environment using sparse satellite InSAR data with a convolutional neural network / Nantheera Anantrasirichai in IEEE Transactions on geoscience and remote sensing, vol 59 n° 4 (April 2021)  

Permalink

Machine learning and geodesy: A survey / Jemil Butt in Journal of applied geodesy, vol 15 n° 2 (April 2021)  

Permalink

Study on offshore seabed sediment classification based on particle size parameters using XGBoost algorithm / Fengfan Wang in Computers & geosciences, vol 149 (April 2021)  

Permalink

A practical method for calculating reliable integer float estimator in GNSS precise positioning / Xianwen Yu in Survey review, Vol 53 n° 377 (February 2021)  

Permalink