Centre dedocumentation
scientifique

Deep SAR-Net: learning objects from signals / Zhongling Huang in ISPRS Journal of photogrammetry and remote sensing, vol 161 (March 2020)  

Public [article]  inISPRS Journal of photogrammetry and remote sensing > vol 161 (March 2020) . - pp 179 - 193 Titre : Deep SAR-Net: learning objects from signals Type de document : Article/Communication Auteurs : Zhongling Huang, Auteur ; Mihai Datcu, Auteur ; Zongxu Pan, Auteur ; Bin Lei, Auteur Année de publication : 2020 Article en page(s) : pp 179 - 193 Note générale : Bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image radar et applications [Termes descripteurs IGN] apprentissage profond [Termes descripteurs IGN] classification par réseau neuronal convolutif [Termes descripteurs IGN] image radar moirée [Termes descripteurs IGN] image Sentinel-SAR [Termes descripteurs IGN] image Terra [Termes descripteurs IGN] matrice de covariance [Termes descripteurs IGN] micro-onde [Termes descripteurs IGN] polarisation [Termes descripteurs IGN] temps-fréquence Résumé : (Auteur) This paper introduces a novel Synthetic Aperture Radar (SAR) specific deep learning framework for complex-valued SAR images. The conventional deep convolutional neural networks based methods usually take the amplitude information of single-polarization SAR images as the input to learn hierarchical spatial features automatically, which may have difficulties in discriminating objects with similar texture but discriminative scattering patterns. Our novel deep learning framework, Deep SAR-Net, takes complex-valued SAR images into consideration to learn both spatial texture information and backscattering patterns of objects on the ground. On the one hand, we transfer the detected SAR images pre-trained layers to extract spatial features from intensity images. On the other hand, we dig into the Fourier domain to learn physical properties of the objects by joint time-frequency analysis on complex-valued SAR images. We evaluate the effectiveness of Deep SAR-Net on three complex-valued SAR datasets from Sentinel-1 and TerraSAR-X satellite and demonstrate how it works better than conventional deep CNNs, especially on man-made objects classes. The proposed datasets and the trained Deep SAR-Net model with all codes are provided. Numéro de notice : A2020-065 Thématique : IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1016/j.isprsjprs.2020.01.016 date de publication en ligne : 23/01/2020 En ligne : https://doi.org/10.1016/j.isprsjprs.2020.01.016 Format de la ressource électronique : URL Article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=94583 [article]

Efficient match pair selection for oblique UAV images based on adaptive vocabulary tree / San Jiang in ISPRS Journal of photogrammetry and remote sensing, vol 161 (March 2020)  

Public [article]  inISPRS Journal of photogrammetry and remote sensing > vol 161 (March 2020) . - pp 61 - 75 Titre : Efficient match pair selection for oblique UAV images based on adaptive vocabulary tree Type de document : Article/Communication Auteurs : San Jiang, Auteur ; Wanshou Jiang, Auteur Année de publication : 2020 Article en page(s) : pp 61 - 75 Note générale : Bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Photogrammétrie [Termes descripteurs IGN] analyse des correspondances [Termes descripteurs IGN] appariement d'images [Termes descripteurs IGN] image aérienne oblique [Termes descripteurs IGN] image de drone [Termes descripteurs IGN] photogrammétrie aérienne [Termes descripteurs IGN] seuillage [Termes descripteurs IGN] structure-from-motion Résumé : (Auteur) The primary contribution of this paper is an efficient match pair selection method for oblique unmanned aerial vehicle (UAV) images. First, high overlap degrees and spatial resolutions cause image and feature redundancies in vocabulary tree building and image indexing. To cope with this issue, an image selection strategy and a feature selection strategy are designed to decrease the total number of features. Second, by analysing the distribution of the similarity scores, an adaptive threshold selection method is implemented to determine the number of candidate match pairs for each query image, and it avoids the disadvantages of the fixed number and fixed proportion methods. Then, an algorithm, termed AVT-Expansion, is proposed for the match pair selection and simplification where the initial match pairs are first selected by using the adaptive vocabulary tree (AVT). To simplify the initial match pairs, the AVT method is integrated with our previous MST-Expansion algorithm, which is used to extract a match graph by analysing the image topological connection network. Finally, the proposed method is verified using three UAV datasets captured with different oblique multi-camera systems. Experimental results demonstrate that the efficiency of the vocabulary tree building is improved, with speed-up ratios ranging from 14 to 16, and that high image retrieval precision values are obtained for the three datasets. For match pair selection of oblique UAV images, the proposed method is an efficient solution. Numéro de notice : A2020-062 Thématique : IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1016/j.isprsjprs.2019.12.013 date de publication en ligne : 15/01/2020 En ligne : https://doi.org/10.1016/j.isprsjprs.2019.12.013 Format de la ressource électronique : URL Article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=94578 [article]

Heuristic sample learning for complex urban scenes: Application to urban functional-zone mapping with VHR images and POI data / Xiuyuan Zhang in ISPRS Journal of photogrammetry and remote sensing, vol 161 (March 2020)  

Public [article]  inISPRS Journal of photogrammetry and remote sensing > vol 161 (March 2020) . - pp 1 - 12 Titre : Heuristic sample learning for complex urban scenes: Application to urban functional-zone mapping with VHR images and POI data Type de document : Article/Communication Auteurs : Xiuyuan Zhang, Auteur ; Shihong Du, Auteur ; Zhijia Zheng, Auteur Année de publication : 2020 Article en page(s) : pp 1 - 12 Note générale : Bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image [Termes descripteurs IGN] analyse d'image orientée objet [Termes descripteurs IGN] apprentissage dirigé [Termes descripteurs IGN] apprentissage semi-dirigé [Termes descripteurs IGN] cartographie urbaine [Termes descripteurs IGN] Chine [Termes descripteurs IGN] échantillonnage d'image [Termes descripteurs IGN] image à très haute résolution [Termes descripteurs IGN] méthode heuristique [Termes descripteurs IGN] point d'intérêt [Termes descripteurs IGN] scène urbaine Résumé : (Auteur) Urban functional zones are basic units of urban planning and resource allocation, and contribute to a wide range of urban studies and investigations. Existing studies on functional-zone mapping with very-high-resolution (VHR) satellite images focused much on feature representations and classification techniques, but ignored zone sampling which however was fundamental to automatic zone classifications. Functional-zone sampling is much complicated and can hardly be resolved by classical sampling methods, as functional zones are complex urban scenes which consist of heterogeneous land covers and have highly abstract categories. To resolve the issue, this study presents a novel sampling paradigm, i.e., heuristic sample learning (HSL). It first proposes a sparse topic model to select representative functional zones, then uses deep forest to select confusing zones, and finally embraces Chinese restaurant process to label these selected zones. The presented method collects both representative and confusing zone samples and identifies their categories accurately, which makes the functional-zone classification process robust and the classification results accurate. Experiments conducted in Beijing indicate that HSL is effective and efficient for functional-zone sampling and classifications. Compared to traditional manual sampling, HSL reduces the time cost by 55% and improves the classification accuracy by 11.3% on average; furthermore, HSL can reduce the variation in sampling and classification results caused by different proficiency of operators. Accordingly, HSL significantly contributes to functional-zone mapping and plays an important role in urban studies. Numéro de notice : A2020-061 Thématique : IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1016/j.isprsjprs.2020.01.005 date de publication en ligne : 13/01/2020 En ligne : https://doi.org/10.1016/j.isprsjprs.2020.01.005 Format de la ressource électronique : URL Article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=94577 [article]

Object-based incremental registration of terrestrial point clouds in an urban environment / Xuming Ge in ISPRS Journal of photogrammetry and remote sensing, vol 161 (March 2020)  

Public [article]  inISPRS Journal of photogrammetry and remote sensing > vol 161 (March 2020) . - pp 218 - 232 Titre : Object-based incremental registration of terrestrial point clouds in an urban environment Type de document : Article/Communication Auteurs : Xuming Ge, Auteur ; Han Hu, Auteur Année de publication : 2020 Article en page(s) : pp 218 - 232 Note générale : Bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Lasergrammétrie [Termes descripteurs IGN] compensation par moindres carrés [Termes descripteurs IGN] conception orientée objet [Termes descripteurs IGN] données laser [Termes descripteurs IGN] données localisées 3D [Termes descripteurs IGN] primitive géométrique [Termes descripteurs IGN] scène urbaine [Termes descripteurs IGN] semis de points [Termes descripteurs IGN] télémétrie laser terrestre Résumé : (Auteur) Registration of terrestrial point clouds is essential for large-scale urban applications. The robustness, accuracy, and runtime are generally given the highest priority in the design of appropriate algorithms. Most approaches that target general scenarios can only fulfill some of these factors, that is, robustness and accuracy come at the cost of increased runtime and vice versa. This paper proposes an object-based incremental registration strategy that accomplishes all of these objectives without the need for artificial targets, aiming at a specific scenario, the urban environment. The key is to decompose the degrees of freedom for the SE(3) transformation to three separate but closely related steps, considering that scanners are generally leveled in urban scenes: (1) 2D transformation with matches from line primitives, (2) vertical offset compensation by robust least-squares optimization, and (3) full SE(3) least-squares refinement using uniformly selected local patches. The robustness is prioritized in the whole pipeline, as structured first by a primitive-based registration and two least-squares optimizations with robust estimations that do not require specific keypoints. An object-based strategy for terrestrial point clouds is used to increase the reliability of the first step by the line primitives, which significantly reduces the search space without affecting the recall ratio. The least-squares optimization contributes to achieve a global optimum for the accurate registration. The three coupling steps are also more efficient than segregated coarse-to-fine registration. Experimental evaluations for point clouds acquired in both a metropolis and in old-style cities reveal that the proposed methods are superior to or on par with the state-of-the-art in robustness, accuracy, and runtime. In addition, the methods are also agnostic to the primitives adopted. Numéro de notice : A2020-066 Thématique : IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1016/j.isprsjprs.2020.01.020 date de publication en ligne : 29/01/2020 En ligne : https://doi.org/10.1016/j.isprsjprs.2020.01.020 Format de la ressource électronique : URL Article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=94584 [article]

Thermal unmixing based downscaling for fine resolution diurnal land surface temperature analysis / Jiong Wang in ISPRS Journal of photogrammetry and remote sensing, vol 161 (March 2020)  

Public [article]  inISPRS Journal of photogrammetry and remote sensing > vol 161 (March 2020) . - pp 76 - 89 Titre : Thermal unmixing based downscaling for fine resolution diurnal land surface temperature analysis Type de document : Article/Communication Auteurs : Jiong Wang, Auteur ; Olivier Schmitz, Auteur ; Meng Lu, Auteur ; Derek Karssenberg, Auteur Année de publication : 2020 Article en page(s) : pp 76 - 89 Note générale : Bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image optique [Termes descripteurs IGN] données spatiotemporelles [Termes descripteurs IGN] factorisation de matrice non-négative [Termes descripteurs IGN] image Landsat [Termes descripteurs IGN] image MODIS [Termes descripteurs IGN] mise à l'échelle [Termes descripteurs IGN] Pays-Bas [Termes descripteurs IGN] radiance [Termes descripteurs IGN] réduction [Termes descripteurs IGN] température de surface [Termes descripteurs IGN] variation diurne Résumé : (Auteur) Due to the limitation in the availability of airborne imagery data that are high in both spatial and temporal resolution, land surface temperature (LST) dense in both space and time can only be obtained through downscaling of frequently acquired LST with coarse resolution. Many conventional downscaling techniques are only feasible in an ideal situation, where land surface factors as LST predictors are continuously available for downscaling the LST. These techniques are also applied only at large scales ignoring sub-regional variations. Based upon unmixing based approaches, this study presents an LST downscaling workflow, where only the coarse resolution of 1 km LST image at the prediction time is required. The conceptual backbone of the study is assuming that the LST patterns are governed by thermal behaviors of a fixed set of temperature sensitive land surface components. In operation, the study focuses on central Netherlands covering an area of 90 × 90 km. The MODIS and Landsat imagery acquired simultaneously are used as a coarse-fine resolution pair to derive downscaling mechanism which is then applied to coarse imagery at a time with missing fine resolution imagery. First, an optimal number of thermal components are extracted at fine resolution through the application of the non-negative matrix factorization (NMF). These components are assumed to possess unique temperature change patterns caused by combined effects of land cover change, radiance change, or both. Given the LST change and thermal components at coarse resolution, the LST change load of each component can then be obtained at the coarse resolution by solving a system of linear equations encoding thermal component-LST relationship. Such LST change load of thermal components is further unmixed to fine resolution and linearly weighted by the component distribution at fine resolution to obtain the fine resolution LST change. During the process, the coarse LST data is used directly without any resampling practice as shown in previous studies. Thus the technique is less time consuming even with a large downscaling factor of 30. The downscaled fine resolution LST represents an R-squared of over 0.7 outperforming classic downscaling techniques. The downscaled LST differentiates temperature over major land types and captures both seasonal and diurnal LST dynamics. Numéro de notice : A2020-063 Thématique : IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1016/j.isprsjprs.2020.01.014 date de publication en ligne : 16/01/2020 En ligne : https://doi.org/10.1016/j.isprsjprs.2020.01.014 Format de la ressource électronique : URL Article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=94580 [article]

A convolutional neural network approach for counting and geolocating citrus-trees in UAV multispectral imagery / Lucas Prado Osco in ISPRS Journal of photogrammetry and remote sensing, vol 160 (February 2020)  

Permalink

Estimating wheat yields in Australia using climate records, satellite image time series and machine learning methods / Elisa Kamir in ISPRS Journal of photogrammetry and remote sensing, vol 160 (February 2020)  

Permalink

Landslide displacement mapping based on ALOS-2/PALSAR-2 data using image correlation techniques and SAR interferometry: application to the Hell-Bourg landslide (Salazie Circle, La Réunion Island) / Daniel Raucoules in Geocarto international, vol 35 n° 2 ([01/02/2020])  

Permalink

Landslide susceptibility mapping using maximum entropy and support vector machine models along the highway corridor, Garhwal Himalaya / Vijendra Kumar Pandey in Geocarto international, vol 35 n° 2 ([01/02/2020])  

Permalink

Optimising drone flight planning for measuring horticultural tree crop structure / Yu-Hsuan Tu in ISPRS Journal of photogrammetry and remote sensing, vol 160 (February 2020)  

Permalink

Plant survival monitoring with UAVs and multispectral data in difficult access afforested areas / Maria Luz Gil-Docampo in Geocarto international, vol 35 n° 2 ([01/02/2020])  

Permalink

Prediction of plant diversity in grasslands using Sentinel-1 and -2 satellite image time series / Mathieu Fauvel in Remote sensing of environment, Vol 237 (February 2020)  

Permalink

Red-edge band vegetation indices for leaf area index estimation from Sentinel-2/MSI imagery / Yuanheng Sun in IEEE Transactions on geoscience and remote sensing, vol 58 n° 2 (February 2020)  

Permalink

The "Incense Road" from Petra to Gaza: an analysis using GIS and Cost functions / Motti Zohar in International journal of geographical information science IJGIS, vol 34 n° 2 (February 2020)  

Permalink

Transferring deep learning models for cloud detection between Landsat-8 and Proba-V / Gonzalo Mateo-García in ISPRS Journal of photogrammetry and remote sensing, vol 160 (February 2020)  

Permalink