Descripteur
Termes IGN > mathématiques > statistique mathématique > probabilités > stochastique > modèle stochastique
modèle stochastiqueSynonyme(s)modèle probabiliste |
Documents disponibles dans cette catégorie (180)
Ajouter le résultat dans votre panier
Visionner les documents numériques
Affiner la recherche Interroger des sources externes
Etendre la recherche sur niveau(x) vers le bas
Analyse de séries temporelles d’images Sentinel et intégration de connaissances pour la classification en milieu agricole / Simon Bailly (2017)
Titre : Analyse de séries temporelles d’images Sentinel et intégration de connaissances pour la classification en milieu agricole Type de document : Mémoire Auteurs : Simon Bailly , Auteur ; Sébastien Giordano , Encadrant ; Loïc Landrieu , Encadrant ; Nesrine Chehata , Encadrant ; Olivier Michel, Encadrant Editeur : Grenoble : Institut National Polytechnique de Grenoble INPG Année de publication : 2017 Importance : 59 p. Note générale : bibliographie
Projet de Fin d’Etudes, Grenoble INP - PhelmaLangues : Français (fre) Descripteur : [Vedettes matières IGN] Traitement d'image mixte
[Termes IGN] Alpes-de-haute-provence (04)
[Termes IGN] champ aléatoire conditionnel
[Termes IGN] classification dirigée
[Termes IGN] image Sentinel-MSI
[Termes IGN] image Sentinel-SAR
[Termes IGN] occupation du sol
[Termes IGN] parcelle agricole
[Termes IGN] Registre parcellaire graphique
[Termes IGN] Seine-et-Marne (77)Résumé : (auteur) Le sujet de stage propose l’utilisation d’images fournies par les satellites Sentinel pour l’étude de l’occupation du sol en milieu agricole. Dans le cadre de la refonte de la politique agricole commune (PAC) de l’Union Européenne en 2020, chaque état membre de l’UE doit proposer une réflexion sur de nouveaux modes de gestion. Une piste de travail envisagée concerne la déclaration des types de culture au sein du Registre Parcellaire Graphique (RPG), un système d’information géographique qui regroupe l’ensemble des informations relatives aux parcelles agricoles. A l’heure actuelle, cette déclaration est faite manuellement par les agriculteurs ; l’objectif est de l’automatiser le plus possible grâce aux images Sentinel. Nous proposons pour cela un processus fondé sur la classification supervisée de séries temporelles d’images Sentinel multi-capteurs (radar et optique), en utilisant le RPG pour l’apprentissage et pour la validation. Nous réalisons une étude sur deux zones distinctes qui présentent des règles agronomiques différentes (Alpes de Haute-Provence et Seine et Marne), avec la nomenclature la plus complète possible (28 types de culture), dans l’optique d’une implantation France Entière. Dans le but d’améliorer la robustesse du modèle, nous choisissons d’intégrer l’information relative aux rotations de culture (suite de cultures échelonnées au fil des années sur une même parcelle). Il s’agit donc d’un problème de classification structurée que nous modélisons comme un champ aléatoire conditionnel (CRF). Nous obtenons des résultats intéressants dans l’optique de l’automatisation du processus de déclaration : 96,9% de bonne classification sur la zone située en Seine et Marne (11 classes) et 64,9% sur la zone située dans les Alpes de Haute-Provence (17 classes). Ces résultats sont d’autre part sensiblement améliorés avec l’intégration de la structure temporelle relative aux rotations de culture. Note de contenu : Introduction
1- Données et sites d'étude
2- Etat de l'art
3- Classification à la parcelle
4- Classification structurée
ConclusionNuméro de notice : 17323 Affiliation des auteurs : non IGN Thématique : IMAGERIE/INFORMATIQUE Nature : Mémoire ingénieur Organisme de stage : MATIS (IGN) DOI : sans Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=98347 Documents numériques
peut être téléchargé
Analyse de séries temporelles d’images Sentinel ... - pdf auteurAdobe Acrobat PDF Comparison of belief propagation and graph-cut approaches for contextual classification of 3D LIDAR point cloud data / Loïc Landrieu (2017)
Titre : Comparison of belief propagation and graph-cut approaches for contextual classification of 3D LIDAR point cloud data Type de document : Article/Communication Auteurs : Loïc Landrieu , Auteur ; Clément Mallet , Auteur ; Martin Weinmann, Auteur Editeur : New York : Institute of Electrical and Electronics Engineers IEEE Année de publication : 2017 Autre Editeur : Saint-Mandé : Institut national de l'information géographique et forestière - IGN (2012-) Projets : 1-Pas de projet / Conférence : IGARSS 2017, IEEE International Geoscience And Remote Sensing Symposium 23/07/2017 28/07/2017 Fort Worth Texas - Etats-Unis Proceedings IEEE Importance : pp 2768 - 2771 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Lasergrammétrie
[Termes IGN] algorithme Graph-Cut
[Termes IGN] analyse comparative
[Termes IGN] champ aléatoire conditionnel
[Termes IGN] classification par maximum de vraisemblance
[Termes IGN] données lidar
[Termes IGN] données localisées 3D
[Termes IGN] inférence
[Termes IGN] semis de points
[Termes IGN] test de performance
[Termes IGN] théorie de Dempster-ShaferRésumé : (auteur) In this paper, we focus on the classification of lidar point cloud data acquired via mobile laser scanning, whereby the classification relies on a context model based on a Conditional Random Field (CRF). We present two approximate inference algorithms based on belief propagation, as well as a graph-cut-based approach not yet applied in this context. To demonstrate the performance of our approach, we present the classification results derived for a standard benchmark dataset. These results clearly indicate that the graph-cut-based method is able to retrieve a labeling of higher likelihood in only a fraction of the time needed for the other approaches. The higher likelihood, in turn, translates into a significant gain in the accuracy of the obtained classification. Numéro de notice : C2017-026 Affiliation des auteurs : IGN+Ext (2012-2019) Autre URL associée : vers HAL Thématique : IMAGERIE Nature : Communication nature-HAL : ComAvecCL&ActesPubliésIntl DOI : 10.1109/IGARSS.2017.8127571 Date de publication en ligne : 04/12/2017 En ligne : https://doi.org/10.1109/IGARSS.2017.8127571 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=89289 Documents numériques
en open access
Comparison of belief propagation ... - postprintAdobe Acrobat PDF Computationally efficient hyperspectral data learning based on the doubly stochastic dirichlet process / Xing Sun in IEEE Transactions on geoscience and remote sensing, vol 55 n° 1 (January 2017)
[article]
Titre : Computationally efficient hyperspectral data learning based on the doubly stochastic dirichlet process Type de document : Article/Communication Auteurs : Xing Sun, Auteur ; Nelson H. C. Yung, Auteur ; Edmund Y. Lam, Auteur ; Hayden K.-H. So, Auteur Année de publication : 2017 Article en page(s) : pp 363 - 374 Note générale : Bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image optique
[Termes IGN] classification
[Termes IGN] image hyperspectrale
[Termes IGN] modèle stochastique
[Termes IGN] problème de DirichletRésumé : (Auteur) The Dirichlet process (DP) prior is effective in modeling HSIs (HSI) and identifying land-cover classes. However, modeling a continuously varying intensity of these land covers elegantly and consistently is still a challenge. We propose a doubly stochastic DP (DSDP) as an efficient model of the global topic measurement space, which imposes a weaker assumption compared with the discrete Markov assumption, resulting in a lower computational cost than other DP-prior-based models. We also present a mixture model of DSDP, which is termed the marked sigmoidal Gaussian process (SGP) DSDP mixture model. It can be thinned from a DP mixture without massive auxiliary covariates, and the marked function prior makes the number of land-cover classes consistent, whereas the SGP function prior models the HSI land-cover variation globally. The consistency of the number of land covers is maintained for various HSIs with large-scale geographical areas. Experiments show that the model is robust and consistent on HSI identification with weak or even no supervision. Numéro de notice : A2017-020 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1109/TGRS.2016.2606575 En ligne : https://doi.org/10.1109/TGRS.2016.2606575 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=83951
in IEEE Transactions on geoscience and remote sensing > vol 55 n° 1 (January 2017) . - pp 363 - 374[article]Hyperspectral image classification with canonical correlation forests / Junshi Xia in IEEE Transactions on geoscience and remote sensing, vol 55 n° 1 (January 2017)
[article]
Titre : Hyperspectral image classification with canonical correlation forests Type de document : Article/Communication Auteurs : Junshi Xia, Auteur ; Naoto Yokoya, Auteur ; Akira Iwasaki, Auteur Année de publication : 2017 Article en page(s) : pp 421 - 431 Note générale : Bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image optique
[Termes IGN] analyse canonique
[Termes IGN] analyse en composantes indépendantes
[Termes IGN] champ aléatoire de Markov
[Termes IGN] classificateur
[Termes IGN] classification
[Termes IGN] classification par forêts d'arbres décisionnels
[Termes IGN] image hyperspectrale
[Termes IGN] Rotation Forest classificationRésumé : (Auteur) Multiple classifier systems or ensemble learning is an effective tool for providing accurate classification results of hyperspectral remote sensing images. Two well-known ensemble learning classifiers for hyperspectral data are random forest (RF) and rotation forest (RoF). In this paper, we proposed to use a novel decision tree (DT) ensemble method, namely, canonical correlation forest (CCF). More specifically, several individual canonical correlation trees (CCTs) that are binary DTs, which use canonical correlation components for the hyperplane splitting, are used to construct the CCF. Additionally, we adopt the projection bootstrap technique in CCF, in which the full spectral bands are retained for split selection in the projected space. The techniques aforementioned allow the CCF to improve the accuracy of member classifiers and diversity within the ensemble. Furthermore, the CCF is extended to the spectral-spatial frameworks that incorporate Markov random fields, extended multiattribute profiles (EMAPs), and the ensemble of independent component analysis and rolling guidance filter (E-ICA-RGF). Experimental results on six hyperspectral data sets are used to indicate the comparative effectiveness of the proposed method, in terms of accuracy and computational complexity, compared with RF and RoF, and it turns out that CCF is a promising approach for hyperspectral image classification not only with spectral information but also in the spectral-spatial frameworks. Numéro de notice : A2017-021 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1109/TGRS.2016.2607755 En ligne : https://doi.org/10.1109/TGRS.2016.2607755 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=83953
in IEEE Transactions on geoscience and remote sensing > vol 55 n° 1 (January 2017) . - pp 421 - 431[article]
Titre : Instantaneous estimation of attitude from GNSS Type de document : Thèse/HDR Auteurs : Hendy Fitrian Suhandri, Auteur ; Alfred Kleusberg, Directeur de thèse ; Hasanuddin Zainal Abidin, Directeur de thèse Editeur : Stuttgart : University of Stuttgart Année de publication : 2017 Importance : 143 p. Format : 21 x 30 cm Note générale : Bibliographie
thesis accepted by the Faculty of Aerospace Engineering and Geodesy of the University of Stuttgart in partial fulfilment of the requirements for the degree of Doctor of Engineering Sciences (Dr.-Ing.)Langues : Anglais (eng) Descripteur : [Termes IGN] ambiguïté entière
[Termes IGN] angle d'Euler
[Termes IGN] double différence
[Termes IGN] filtre de Kalman
[Termes IGN] méthode des moindres carrés
[Termes IGN] modèle stochastique
[Termes IGN] orientation de véhicule
[Termes IGN] positionnement cinématique
[Termes IGN] positionnement par GNSS
[Termes IGN] récepteur GNSS
[Termes IGN] simple différence
[Vedettes matières IGN] Traitement de données GNSSRésumé : (auteur) The use of the Global Navigation Satellite System (GNSS) is widely spread from position determination to attitude determination of a platform in space. This system offers time invariant estimation position. Another thing that can be an advantage is that the flexibility to operate the GNSS receiver variants, from the low-cost until the high-performance GNSS receivers. In terms of attitude determination application at least three receivers are required to determine three spatial axes, where the cost-effective GNSS attitude determination systems can be constructed with today’s receiver technology. At the moment, however, algorithms are lacking which are fast and efficient enough to estimate the position angles without delay. For this reason, the present work deals with the development of algorithms for the attitude determination in space of a platform under the help of the "GNSS" Global Positioning System (GPS). The investigation through this work is classified into three sequential parts: The first part is the estimation of the optimal configuration of baseline array as well as the estimation of the integer ambiguity of carrier phase differences. The estimated integer ambiguity is then used to estimate the high precision baseline coordinates. The second part is to estimate the attitude of the platform in space by means of quaternion using batch process, and the last part is to improve the algorithm using a recursive algorithm for the kinematic application purpose. The precise attitude determination about three spatial axes is possible if at least three GNSS receivers with fixed baselines are used in particular array configurations. Assuming that the basic lengths of the baselines are known a priori, the attitude angles can be calculated via the combination of carrier phase and pseudorange observations. Since the carrier of the GPS signal is propagated in short-wave form, the measured phase differences are ambiguous. The multiples of the GPS signal phases together with the baseline lengths are therefore estimated and improved in a first step with the aid of the a priori baseline lengths information. The multiple-baseline float solution estimation method is used. However, the approach does not provide optimal results. Therefore, an alternative algorithm for the float solution is presented, which estimates the float solution by using the socalled the gradient based iterative method of the least-squares. It shows that method is able to give convergent estimate parameter. It is also shown here that the proposed method outperforms the conventional iterative least-squares in terms of iteration number and computational time. For instantaneous applications, the Least-squares AMBiguity Decorrelation Adjustment (LAMBDA) method is not optimal for fixing the integer multiples of the carrier phase differences for several baseline lengths. In addition, this method requires a high computational effort as soon as a larger number of baseline lines enter into the calculation. An improvement in this work is utilising the partial LAMBDA method, which only uses a subset of the integer multiples to be determined. This algorithm improves the determination of integer multiples and precise calculation of the baseline lengths. The advantages of this algorithm are discussed, and it is empirically demonstrated that the ambiguities are better resolved. Furthermore, the estimation of the attitude angles with the aid of quaternions is theoretically improved and analysed. Two processing strategies are investigated: the least-squares method and the Kalman Filter (KF) method. For the static case, the least-squares is applied and tested. Simulations show that the developed gradient based iterative method of the least-squares provides better estimates than the conventional adjustment methods. It is also shown that the number of iterations required is less and the computational time is reduced. This algorithm is not useful for kinematic applications where a fast sequence of results is required. A modified Extended Kalman Filter (EKF)-Like algorithm is used for kinematic applications. Experiments show that with this algorithm more stable quaternions can be calculated with fewer outliers than when they are determined by the least-squares method. All newly developed algorithms are theoretically analysed and subjected to extensive simulations and experimental kinematic tests in the field. Note de contenu : Introduction
1 - General mathematical model of GNSS positioning
2 - Multi-baseline GNSS estimation method
3 - GNSS based attitude determination
4 - Recursive attitude determination
5 - Experimental result of static and kinematic tests
6 - Summary, conclusion and future work suggestionNuméro de notice : 21574 Affiliation des auteurs : non IGN Thématique : POSITIONNEMENT Nature : Thèse étrangère Note de thèse : Doctor thesis : Engineering sciences : Stuttgart : 2017 DOI : 10.18419/opus-9239 En ligne : http://dx.doi.org/10.18419/opus-9239 Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=90576 Modèles géographiques avec le langage Mathematica / André Dauphiné (2017)PermalinkPré-segmentation pour la classification faiblement supervisée de scènes urbaines à partir de nuages de points 3D LIDAR / Stéphane Guinard (2017)PermalinkRandom-walker-based collaborative learning for hyperspectral image classification / Bin Sun in IEEE Transactions on geoscience and remote sensing, vol 55 n° 1 (January 2017)PermalinkWeakly supervised segmentation-aided classification of urban scenes from 3D LIDAR point clouds / Stéphane Guinard (2017)PermalinkDetermination of a terrestrial reference frame via Kalman filtering of very long baseline interferometry data / Benedikt Soja in Journal of geodesy, vol 90 n° 12 (December 2016)PermalinkAn adaptive stochastic model for GPS observations and its performance in precise point positioning / J. Z. Zheng in Survey review, vol 48 n° 349 (July 2016)PermalinkStochastic modeling of triple-frequency BeiDou signals: estimation, assessment and impact analysis / Bofeng Li in Journal of geodesy, vol 90 n° 7 (July 2016)PermalinkMarkov random field-based method for super-resolution mapping of forest encroachment from remotely sensed ASTER image / L. K. Tiwari in Geocarto international, vol 31 n° 3 - 4 (March - April 2016)PermalinkPermalinkInvestigating efficacy of robust M-estimation of deformation from observation differences / Krzysztof Nowel in Survey review, vol 48 n° 346 (January 2016)Permalink