Centre de documentation
scientifique

Dépouillements

A novel semisupervised active-learning algorithm for hyperspectral image classification / Zengmao Wang in IEEE Transactions on geoscience and remote sensing, vol 55 n° 6 (June 2017)  

Public [article]  inIEEE Transactions on geoscience and remote sensing > vol 55 n° 6 (June 2017) . - pp 3071 - 3083 Titre : A novel semisupervised active-learning algorithm for hyperspectral image classification Type de document : Article/Communication Auteurs : Zengmao Wang, Auteur ; Bo Du, Auteur ; Lefei Zhang, Auteur ; et al., Auteur Année de publication : 2017 Article en page(s) : pp 3071 - 3083 Note générale : Bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image optique [Termes IGN] analyse de groupement [Termes IGN] apprentissage automatique [Termes IGN] apprentissage semi-dirigé [Termes IGN] image hyperspectrale [Termes IGN] segmentation sémantique Résumé : (Auteur) Less training samples are a challenging problem in hyperspectral image classification. Active learning and semisupervised learning are two promising techniques to address the problem. Active learning solves the problem by improving the quality of the training samples, while semisupervised learning solves the problem by increasing the quantity of the training samples. However, they pay too much attention to the discriminative information in the unlabeled data, leading to information bias to train supervised models, and much more effort to label samples. Therefore, a method to discover representativeness and discriminativeness by semisupervised active learning is proposed. It takes advantages of both active learning and semisupervised learning. The representativeness and discriminativeness are discovered with a labeling process based on a supervised clustering technique and classification results. Specifically, the supervised clustering results can discover important structural information in the unlabeled data, and the classification results are also highly confidential in the active-learning process. With these clustering results and classification results, we can assign pseudolabels to the unlabeled data. Meanwhile, the unlabeled samples that cannot be assigned with pseudolabels with high confidence at each iteration are regarded as candidates in active learning. The methodology is validated on four hyperspectral data sets. Significant improvements in classification accuracy are achieved by the proposed method with respect to the state-of-the-art methods. Numéro de notice : A2017-473 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1109/TGRS.2017.2650938 En ligne : https://doi.org/10.1109/TGRS.2017.2650938 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=86398 [article]

Angular reflectance of leaves with a dual-wavelength terrestrial lidar and its implications for leaf-bark separation and leaf moisture estimation / Steven Hancock in IEEE Transactions on geoscience and remote sensing, vol 55 n° 6 (June 2017)  

Public [article]  inIEEE Transactions on geoscience and remote sensing > vol 55 n° 6 (June 2017) . - pp 3084 - 3090 Titre : Angular reflectance of leaves with a dual-wavelength terrestrial lidar and its implications for leaf-bark separation and leaf moisture estimation Type de document : Article/Communication Auteurs : Steven Hancock, Auteur ; Rachel Gaulton, Auteur ; F. Mark Danson, Auteur Année de publication : 2017 Article en page(s) : pp 3084 - 3090 Note générale : Bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Lasergrammétrie [Termes IGN] angle d'incidence [Termes IGN] données lidar [Termes IGN] écorce [Termes IGN] indice de diversité [Termes IGN] longueur d'onde [Termes IGN] réflectance de surface [Termes IGN] réflectance végétale [Termes IGN] teneur en eau de la végétation Résumé : (Auteur) A new generation of multiwavelength lidars offers the potential to measure the structure and biochemistry of vegetation simultaneously, using range resolved spectral indices to overcome the confounding effects in passive optical measurements. However, the reflectance of leaves depends on the angle of incidence, and if this dependence varies between wavelengths, the resulting spectral indices will also vary with the angle of incidence, complicating their use in separating structural and biochemical effects in vegetation canopies. The Salford Advanced Laser Canopy Analyser (SALCA) dual-wavelength terrestrial laser scanner was used to measure the angular dependence of reflectance for a range of leaves at the wavelengths used by the new generation of multiwavelength lidars, 1063 and 1545 nm, as used by SALCA, DWEL, and the Optech Titan. The influence of the angle of incidence on the normalized difference index (NDI) of these wavelengths was also assessed. The reflectance at both wavelengths depended on the angle of incidence and could be well modelled as a cosine. The change in the NDI with the leaf angle of incidence was small compared with the observed difference in the NDI between fresh and dry leaves and between leaf and bark. Therefore, it is concluded that angular effects will not significantly impact leaf moisture retrievals or prevent leaf/bark separation for the wavelengths used in the new generation of 1063- and 1545-nm multiwavelength lidars. Numéro de notice : A2017-474 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1109/TGRS.2017.2652140 En ligne : http://dx.doi.org/10.1109/TGRS.2017.2652140 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=86399 [article]

A time-series approach to estimating soil moisture from vegetated surfaces using L-band radar backscatter / Jeffrey D. Ouellette in IEEE Transactions on geoscience and remote sensing, vol 55 n° 6 (June 2017)  

Public [article]  inIEEE Transactions on geoscience and remote sensing > vol 55 n° 6 (June 2017) . - pp 3186 - 3193 Titre : A time-series approach to estimating soil moisture from vegetated surfaces using L-band radar backscatter Type de document : Article/Communication Auteurs : Jeffrey D. Ouellette, Auteur ; Joel T. Johnson, Auteur ; Anna Balenzano, Auteur ; et al., Auteur Année de publication : 2017 Article en page(s) : pp 3186 - 3193 Note générale : Bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Applications de télédétection [Termes IGN] couvert végétal [Termes IGN] détection de changement [Termes IGN] humidité du sol [Termes IGN] image radar [Termes IGN] radiométrie [Termes IGN] rétrodiffusion [Termes IGN] série temporelle [Termes IGN] traitement d'image radar Résumé : (Auteur) Many previous studies have shown the sensitivity of radar backscatter to surface soil moisture content, particularly at L-band. Moreover, the estimation of soil moisture from radar for bare soil surfaces is well-documented, but estimation underneath a vegetation canopy remains unsolved. Vegetation significantly increases the complexity of modeling the electromagnetic scattering in the observed scene, and can even obstruct the contributions from the underlying soil surface. Existing approaches to estimating soil moisture under vegetation using radar typically rely on a forward model to describe the backscattered signal and often require that the vegetation characteristics of the observed scene be provided by an ancillary data source. However, such information may not be reliable or available during the radar overpass of the observed scene (e.g., due to cloud coverage if derived from an optical sensor). Thus, the approach described herein is an extension of a change-detection method for soil moisture estimation, which does not require ancillary vegetation information, nor does it make use of a complicated forward scattering model. Novel modifications to the original algorithm include extension to multiple polarizations and a new technique for bounding the radar-derived soil moisture product using radiometer-based soil moisture estimates. Soil moisture estimates are generated using data from the Soil Moisture Active/Passive (SMAP) satellite-borne radar and radiometer data, and are compared with up-scaled data from a selection of in situ networks used in SMAP validation activities. These results show that the new algorithm can consistently achieve rms errors less than 0.07 m3/m3 over a variety land cover types. Numéro de notice : A2017-475 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1109/TGRS.2017.2663768 En ligne : https://doi.org/10.1109/TGRS.2017.2663768 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=86400 [article]

An adaptive weighted tensor completion method for the recovery of remote sensing images with missing data / Michael Kwok-Po Ng in IEEE Transactions on geoscience and remote sensing, vol 55 n° 6 (June 2017)  

Public [article]  inIEEE Transactions on geoscience and remote sensing > vol 55 n° 6 (June 2017) . - pp 3367 - 3381 Titre : An adaptive weighted tensor completion method for the recovery of remote sensing images with missing data Type de document : Article/Communication Auteurs : Michael Kwok-Po Ng, Auteur ; Qiangqiang Yuan, Auteur ; Li Yan, Auteur ; Jing Sun, Auteur Année de publication : 2017 Article en page(s) : pp 3367 - 3381 Note générale : Bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image optique [Termes IGN] bande spectrale [Termes IGN] détection de partie cachée [Termes IGN] données spatiotemporelles [Termes IGN] image Aqua-MODIS [Termes IGN] spectroradiométrie [Termes IGN] tenseur Résumé : (Auteur) Missing information, such as dead pixel values and cloud effects, is very common image quality degradation problems in remote sensing. Missing information can reduce the accuracy of the subsequent image processing, in applications such as classification, unmixing, and target detection, and even the quantitative retrieval process. The main aim of this paper is to study an adaptive weighted tensor completion (AWTC) method for the recovery of remote sensing images with missing data. Our idea is to collectively make use of the spatial, spectral, and temporal information to build a new weighted tensor low-rank regularization model for recovering the missing data. In the model, the weights are determined adaptively by considering the contribution of the spatial, spectral, and temporal information in each dimension. Experimental results based on both simulated and real data sets are presented to verify that the proposed method can recover missing data, and its performance is found to be better than the other tested methods. In the simulated experiments, the peak signal-to-noise ratio is improved by more than 3 dB, compared with the original tensor completion model. In the real data experiments, the proposed AWTC model can better recover the dead line problem in Aqua Moderate Resolution Imaging Spectroradiometer band 6 and the scan-line corrector-off problem in enhanced thematic mapper plus images, with the smallest spectral distortion. Numéro de notice : A2017-476 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1109/TGRS.2017.2670021 En ligne : http://dx.doi.org/10.1109/TGRS.2017.2670021 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=86401 [article]

Improvements in precise orbits of altimetry satellites and their impact on mean sea level monitoring / Sergei Rudenko in IEEE Transactions on geoscience and remote sensing, vol 55 n° 6 (June 2017)  

Public [article]  inIEEE Transactions on geoscience and remote sensing > vol 55 n° 6 (June 2017) . - pp 3382 - 3395 Titre : Improvements in precise orbits of altimetry satellites and their impact on mean sea level monitoring Type de document : Article/Communication Auteurs : Sergei Rudenko, Auteur ; Karl-Hans Neumayer, Auteur ; Denise Dettmering, Auteur Année de publication : 2017 Article en page(s) : pp 3382 - 3395 Note générale : Bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Géodésie physique [Termes IGN] analyse comparative [Termes IGN] évaluation des données [Termes IGN] International Terrestrial Reference Frame [Termes IGN] orbitographie [Termes IGN] satellite d'observation de la mer [Termes IGN] satellite d'observation de la Terre Résumé : (Auteur) New, precise, consistent orbits (VER11) of altimetry satellites ERS-1, ERS-2, TOPEX/Poseidon, Envisat, Jason-1, and Jason-2 have been recently derived at the GFZ German Research Centre for Geosciences in the extended ITRF2008 terrestrial reference frame using improved models and covering the time span 1991–2015. These orbits show improved quality, as compared with GFZ previous (VER6) orbits derived in 2013. Improved macromodels reduce root mean square (RMS) fits of satellite laser ranging (SLR) observations by 2.6%, 6.9%, and 7% for TOPEX/Poseidon, Jason-1, and Jason-2, respectively. Applying Vienna Mapping Functions 1 instead of Hopfield model for tropospheric correction of Doppler Orbitography and Radiopositioning Integrated by Satellite (DORIS) observations reduces RMS fits of SLR observations by 2%–2.4% and those of DORIS observations by 2.6% for Envisat and Jason satellites. Using satellite true attitude instead of models improves Jason-1 SLR RMS fits by 41% from July 2012 until July 2013. The VER11 orbits indicate the mean values of the SLR RMS fits between 1.2 and 2.1 cm for the different missions. The internal orbit consistency in the radial direction is between 0.5 and 1.9 cm. The global mean sea level trend for the period 1993–2014 from TOPEX, Jason-1, and Jason-2 is 2.8 and 3.0 mm/year using GFZ VER6 and VER11 orbits, respectively. Regionally, the decadal trends from GFZ VER11 and external orbits vary in the order of 1 mm/year. Numéro de notice : A2017-477 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1109/TGRS.2017.2670061 En ligne : https://doi.org/10.1109/TGRS.2017.2670061 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=86402 [article]

Learning to diversify deep belief networks for hyperspectral image classification / Ping Zhong in IEEE Transactions on geoscience and remote sensing, vol 55 n° 6 (June 2017)  

Public [article]  inIEEE Transactions on geoscience and remote sensing > vol 55 n° 6 (June 2017) . - pp 3516 - 3530 Titre : Learning to diversify deep belief networks for hyperspectral image classification Type de document : Article/Communication Auteurs : Ping Zhong, Auteur ; Zhiqiang Gong, Auteur ; Shutao Li, Auteur ; Carola-Bibiane Schönlieb, Auteur Année de publication : 2017 Article en page(s) : pp 3516 - 3530 Note générale : Bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image optique [Termes IGN] apprentissage non-dirigé [Termes IGN] classification par réseau neuronal [Termes IGN] image hyperspectrale [Termes IGN] réseau neuronal convolutif [Termes IGN] théorie de Dempster-Shafer Résumé : (Auteur) In the literature of remote sensing, deep models with multiple layers have demonstrated their potentials in learning the abstract and invariant features for better representation and classification of hyperspectral images. The usual supervised deep models, such as convolutional neural networks, need a large number of labeled training samples to learn their model parameters. However, the real-world hyperspectral image classification task provides only a limited number of training samples. This paper adopts another popular deep model, i.e., deep belief networks (DBNs), to deal with this problem. The DBNs allow unsupervised pretraining over unlabeled samples at first and then a supervised fine-tuning over labeled samples. But the usual pretraining and fine-tuning method would make many hidden units in the learned DBNs tend to behave very similarly or perform as “dead” (never responding) or “potential over-tolerant” (always responding) latent factors. These results could negatively affect description ability and thus classification performance of DBNs. To further improve DBN’s performance, this paper develops a new diversified DBN through regularizing pretraining and fine-tuning procedures by a diversity promoting prior over latent factors. Moreover, the regularized pretraining and fine-tuning can be efficiently implemented through usual recursive greedy and back-propagation learning framework. The experiments over real-world hyperspectral images demonstrated that the diversity promoting prior in both pretraining and fine-tuning procedure lead to the learned DBNs with more diverse latent factors, which directly make the diversified DBNs obtain much better results than original DBNs and comparable or even better performances compared with other recent hyperspectral image classification methods. Numéro de notice : A2017-478 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1109/TGRS.2017.2675902 En ligne : http://dx.doi.org/10.1109/TGRS.2017.2675902 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=86403 [article]

Can a machine generate humanlike language descriptions for a remote sensing image? / Zhenwei Shi in IEEE Transactions on geoscience and remote sensing, vol 55 n° 6 (June 2017)  

Public [article]  inIEEE Transactions on geoscience and remote sensing > vol 55 n° 6 (June 2017) . - pp 3623 - 3634 Titre : Can a machine generate humanlike language descriptions for a remote sensing image? Type de document : Article/Communication Auteurs : Zhenwei Shi, Auteur ; Zhengxia Zou, Auteur Année de publication : 2017 Article en page(s) : pp 3623 - 3634 Note générale : Bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Télédétection [Termes IGN] descripteur [Termes IGN] image à haute résolution [Termes IGN] intelligence artificielle [Termes IGN] interface en langage naturel Résumé : (Auteur) This paper investigates an intriguing question in the remote sensing field: “can a machine generate humanlike language descriptions for a remote sensing image?” The automatic description of a remote sensing image (namely, remote sensing image captioning) is an important but rarely studied task for artificial intelligence. It is more challenging as the description must not only capture the ground elements of different scales, but also express their attributes as well as how these elements interact with each other. Despite the difficulties, we have proposed a remote sensing image captioning framework by leveraging the techniques of the recent fast development of deep learning and fully convolutional networks. The experimental results on a set of high-resolution optical images including Google Earth images and GaoFen-2 satellite images demonstrate that the proposed method is able to generate robust and comprehensive sentence description with desirable speed performance. Numéro de notice : A2017-479 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1109/TGRS.2017.2677464 En ligne : http://dx.doi.org/10.1109/TGRS.2017.2677464 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=86406 [article]