Centre de documentation
scientifique

Dépouillements

Fusion of support vector machines for classification of multisensor data / Björn Waske in IEEE Transactions on geoscience and remote sensing, vol 45 n° 12 Tome 1 (December 2007)  

Public [article]  inIEEE Transactions on geoscience and remote sensing > vol 45 n° 12 Tome 1 (December 2007) . - pp 3858 - 3866 Titre : Fusion of support vector machines for classification of multisensor data Type de document : Article/Communication Auteurs : Björn Waske, Auteur ; Jon Atli Benediktsson, Auteur Année de publication : 2007 Article en page(s) : pp 3858 - 3866 Note générale : Bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image mixte [Termes IGN] analyse comparative [Termes IGN] classificateur non paramétrique [Termes IGN] classificateur paramétrique [Termes IGN] classification par arbre de décision [Termes IGN] classification par maximum de vraisemblance [Termes IGN] classification par séparateurs à vaste marge [Termes IGN] fusion d'images [Termes IGN] image multicapteur [Termes IGN] image optique [Termes IGN] image radar Résumé : (Auteur) The classification of multisensor data sets, consisting of multitemporal synthetic aperture radar data and optical imagery, is addressed. The concept is based on the decision fusion of different outputs. Each data source is treated separately and classified by a support vector machine (SVM). Instead of fusing the final classification outputs (i.e., land cover classes), the original outputs of each SVM discriminant function are used in the subsequent fusion process. This fusion is performed by another SVM, which is trained on the a priori outputs. In addition, two voting schemes are applied to create the final classification results. The results are compared with well-known parametric and nonparametric classifier methods, i.e., decision trees, the maximum-likelihood classifier, and classifier ensembles. The proposed SVM-based fusion approach outperforms all other approaches and significantly improves the results of a single SVM, which is trained on the whole multisensor data set. Copyright IEEE Numéro de notice : A2007-581 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1109/TGRS.2007.898446 En ligne : https://doi.org/10.1109/TGRS.2007.898446 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=28944 [article]

Exemplaires(1)

Code-barres	Cote	Support	Localisation	Section	Disponibilité
065-07121A	RAB	Revue	Centre de documentation	En réserve L003	Disponible

Border vector detection and adaptation for classification of multispectral and hyperspectral remote sensing images / N.G. Kasapoglu in IEEE Transactions on geoscience and remote sensing, vol 45 n° 12 Tome 1 (December 2007)  

Public [article]  inIEEE Transactions on geoscience and remote sensing > vol 45 n° 12 Tome 1 (December 2007) . - pp 3880 - 3893 Titre : Border vector detection and adaptation for classification of multispectral and hyperspectral remote sensing images Type de document : Article/Communication Auteurs : N.G. Kasapoglu, Auteur ; O.K. Ersoy, Auteur Année de publication : 2007 Article en page(s) : pp 3880 - 3893 Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image optique [Termes IGN] analyse comparative [Termes IGN] apprentissage dirigé [Termes IGN] classification par arbre de décision [Termes IGN] classification par réseau neuronal [Termes IGN] classification par séparateurs à vaste marge [Termes IGN] image hyperspectrale [Termes IGN] image multibande [Termes IGN] précision de la classification Résumé : (Auteur) Effective partitioning of the feature space for high classification accuracy with due attention to rare class members is often a difficult task. In this paper, the border vector detection and adaptation (BVDA) algorithm is proposed for this purpose. The BVDA consists of two parts. In the first part of the algorithm, some specially selected training samples are assigned as initial reference vectors called border vectors. In the second part of the algorithm, the border vectors are adapted by moving them toward the decision boundaries. At the end of the adaptation process, the border vectors are finalized. The method next uses the minimum distance to border vector rule for classification. In supervised learning, the training process should be unbiased to reach more accurate results in testing. In the BVDA, decision region borders are related to the initialization of the border vectors and the input ordering of the training samples. Consensus strategy can be applied with cross validation to reduce these dependencies. The performance of the BVDA and consensual BVDA were studied in comparison to other classification algorithms including neural network with backpropagation learning, support vector machines, and some statistical classification techniques. Copyright IEEE Numéro de notice : A2007-582 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1109/TGRS.2007.900699 En ligne : https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=4378538 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=28945 [article]

Exemplaires(1)

Code-barres	Cote	Support	Localisation	Section	Disponibilité
065-07121A	RAB	Revue	Centre de documentation	En réserve L003	Disponible

A time-efficient method for anomaly detection in hyperspectral images / O. Duran in IEEE Transactions on geoscience and remote sensing, vol 45 n° 12 Tome 1 (December 2007)  

Public [article]  inIEEE Transactions on geoscience and remote sensing > vol 45 n° 12 Tome 1 (December 2007) . - pp 3894 - 3918 Titre : A time-efficient method for anomaly detection in hyperspectral images Type de document : Article/Communication Auteurs : O. Duran, Auteur ; M. Petrou, Auteur Année de publication : 2007 Article en page(s) : pp 3894 - 3918 Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image optique [Termes IGN] carte de Kohonen [Termes IGN] classification ISODATA [Termes IGN] détection d'anomalie [Termes IGN] image hyperspectrale [Termes IGN] segmentation d'image Résumé : (Auteur) We propose a computationally efficient method for determining anomalies in hyperspectral data. In the first stage of the algorithm, the background classes, which are the dominant classes in the image, are found. The method consists of robust clustering of a randomly chosen small percentage of the image pixels. The clusters are the representatives of the background classes. By using a subset of the pixels instead of the whole image, the computation is sped up, and the probability of including outliers in the background model is reduced. Anomalous pixels are the pixels with spectra that have large relative distances from the cluster centers. Several clustering techniques are investigated, and experimental results using realistic hyperspectral data are presented. A self-organizing map clustered using the local minima of the U-matrix (unified distance matrix) is identified as the most reliable method for background class extraction. The proposed algorithm for anomaly detection is evaluated using realistic hyperspectral data, is compared with a state-of-the-art anomaly detection algorithm, and is shown to perform significantly better. Copyright IEEE Numéro de notice : A2007-583 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1109/TGRS.2007.909205 En ligne : https://doi.org/10.1109/TGRS.2007.909205 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=28946 [article]

Exemplaires(1)

Code-barres	Cote	Support	Localisation	Section	Disponibilité
065-07121A	RAB	Revue	Centre de documentation	En réserve L003	Disponible

Improving river flood extent delineation from synthetic aperture radar using airborne laser altimetry / D.C. Mason in IEEE Transactions on geoscience and remote sensing, vol 45 n° 12 Tome 1 (December 2007)  

Public [article]  inIEEE Transactions on geoscience and remote sensing > vol 45 n° 12 Tome 1 (December 2007) . - pp 3932 - 3943 Titre : Improving river flood extent delineation from synthetic aperture radar using airborne laser altimetry Type de document : Article/Communication Auteurs : D.C. Mason, Auteur ; M.S. Horritt, Auteur ; et al., Auteur Année de publication : 2007 Article en page(s) : pp 3932 - 3943 Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image radar et applications [Termes IGN] analyse comparative [Termes IGN] détection de contours [Termes IGN] données lidar [Termes IGN] image radar moirée [Termes IGN] inondation [Termes IGN] lasergrammétrie [Termes IGN] occupation du sol Résumé : (Auteur) Flood extent maps that are derived from synthetic aperture radar (SAR) images provide spatially distributed data for validating hydraulic models of river flood flow. The accuracy of such maps is reduced by a number of factors, including variation in backscatter from the different land cover types that are adjacent to the flood, changes in returns from the water surface that are caused by different meteorological conditions, and the presence of emergent vegetation. This paper describes how improved accuracy can be achieved by modifying an existing flood extent delineation algorithm to use airborne laser altimetry [light detection and ranging (lidar)] as well as SAR data. The lidar data provide an additional constraint that water line heights should vary smoothly along the flooded reach. The method was tested on a SAR image of a flood for which contemporaneous aerial photography existed, together with lidar data of the un flooded reach. The water line heights of the SAR flood extent that was conditioned on both SAR and lidar data matched the corresponding heights from the aerial photograph water line significantly more closely than those from the SAR flood extent that was conditioned only on SAR data. For water line heights in areas of low slope and vegetation, the root-mean-square error on the height differences reduced from 221.1 cm for the latter case to 55.5 cm for the former. Numéro de notice : A2007-584 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1109/TGRS.2007.901032 En ligne : https://doi.org/10.1109/TGRS.2007.901032 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=28947 [article]

Exemplaires(1)

Code-barres	Cote	Support	Localisation	Section	Disponibilité
065-07121A	RAB	Revue	Centre de documentation	En réserve L003	Disponible

A supervised artificial immune classifier for remote-sensing imagery / Y. Zhong in IEEE Transactions on geoscience and remote sensing, vol 45 n° 12 Tome 1 (December 2007)  

Public [article]  inIEEE Transactions on geoscience and remote sensing > vol 45 n° 12 Tome 1 (December 2007) . - pp 3957 - 3966 Titre : A supervised artificial immune classifier for remote-sensing imagery Type de document : Article/Communication Auteurs : Y. Zhong, Auteur ; L. Zhang, Auteur ; J. Gong, Auteur ; P. Li, Auteur Année de publication : 2007 Article en page(s) : pp 3957 - 3966 Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image [Termes IGN] classificateur [Termes IGN] classification dirigée [Termes IGN] image [Termes IGN] réseau neuronal artificiel [Termes IGN] système immunitaire artificiel Résumé : (Auteur) The artificial immune network (AIN), which is a new computational intelligence model based on artificial immune systems inspired by the vertebrate immune system, has been widely utilized for pattern recognition and data analysis. However, due to the inherent complexity of current AIN models, their application to remote-sensing image classification has been rather limited. This paper presents a novel supervised classification algorithm based on a multiple-valued immune network, which is a novel AIN model, to perform remote-sensing image classification. The proposed method trains the immune network using the samples of regions of interest and obtains an immune network with memory to classify the remote-sensing imagery. Two experiments with different types of images are performed to evaluate the performance of the proposed algorithm in comparison with other traditional image classification algorithms: Parallelepiped, Minimum Distance, Maximum Likelihood, and Back-Propagation Neural Network. The results evince that the proposed algorithm consistently outperforms the traditional algorithms in all the experiments and, hence, provides an effective option for processing remote-sensing imagery. Copyright IEEE Numéro de notice : A2007-585 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1109/TGRS.2007.907739 En ligne : https://doi.org/10.1109/TGRS.2007.907739 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=28948 [article]

Exemplaires(1)

Code-barres	Cote	Support	Localisation	Section	Disponibilité
065-07121A	RAB	Revue	Centre de documentation	En réserve L003	Disponible

Measuring land development in urban regions using graph theoretical and conditional statistical features / C. Unsalan in IEEE Transactions on geoscience and remote sensing, vol 45 n° 12 Tome 1 (December 2007)  

Public [article]  inIEEE Transactions on geoscience and remote sensing > vol 45 n° 12 Tome 1 (December 2007) . - pp 3989 - 3999 Titre : Measuring land development in urban regions using graph theoretical and conditional statistical features Type de document : Article/Communication Auteurs : C. Unsalan, Auteur Année de publication : 2007 Article en page(s) : pp 3989 - 3999 Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image optique [Termes IGN] appariement de graphes [Termes IGN] extraction automatique [Termes IGN] graphe [Termes IGN] image à résolution métrique [Termes IGN] image Ikonos [Termes IGN] image panchromatique [Termes IGN] matrice de co-occurrence [Termes IGN] niveau de gris (image) [Termes IGN] segment de droite [Termes IGN] surveillance de l'urbanisation Résumé : (Auteur) Inferring land use from satellite images is extensively studied by the remote sensing and pattern recognition communities. In previous studies, the focus was on classifying large regions due to the resolution of available satellite images. Nowadays, very high-resolution satellite imagery (Ikonos and Quickbird) allows researchers to focus on more complex land-use problems such as monitoring development in urban regions. Solutions to these complex problems may improve the life standards of city residents. To this end, we focus on automatically monitoring construction zones using their very high-resolution panchromatic satellite images through time. To monitor land development, we obtain sequential images of a selected region. Then, we extract features from each image in the sequence. Comparing values of these features, we expect to measure the degree of land development through time. In a similar study, we introduced graph theoretical measures over Ikonos imagery to measure organization in a given satellite image. This paper is an extension of our previous work with more powerful new features. Here, we first introduce a novel method to extract straight line segments using a least squares ellipse fitting. Then, we introduce four new graph theoretical features. More importantly, we introduce a novel method to embed the spatial information in gray-level co-occurrence matrix statistical features to measure land development. Finally, we test all our existing and new features to measure land development in 19 different urban construction zones. Our test set consists of Ikonos satellite images of these regions captured in separate times. Copyright IEEE Numéro de notice : A2007-586 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1109/TGRS.2007.897446 En ligne : https://doi.org/10.1109/TGRS.2007.897446 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=28949 [article]

Exemplaires(1)

Code-barres	Cote	Support	Localisation	Section	Disponibilité
065-07121A	RAB	Revue	Centre de documentation	En réserve L003	Disponible