Centre de documentation
scientifique

The challenge of robust trait estimates with deep learning on high resolution RGB images / Etienne David (2021)  

Public Titre : The challenge of robust trait estimates with deep learning on high resolution RGB images Type de document : Thèse/HDR Auteurs : Etienne David, Auteur ; Frédéric Baret, Directeur de thèse Editeur : Avignon : Université d'Avignon Année de publication : 2021 Importance : 145 p. Format : 21 x 30 cm Note générale : bibliographie Thèse de Doctorat de l'Université d'Avignon, spécialité Sciences Agronomiques Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Applications de télédétection [Termes IGN] apprentissage profond [Termes IGN] base de données d'images [Termes IGN] blé (céréale) [Termes IGN] céréales [Termes IGN] comptage [Termes IGN] cultures [Termes IGN] densité de la végétation [Termes IGN] données d'entrainement (apprentissage automatique) [Termes IGN] image à haute résolution [Termes IGN] image captée par drone [Termes IGN] image RVB [Termes IGN] jeu de données [Termes IGN] surveillance agricole Index. décimale : THESE Thèses et HDR Résumé : (auteur) High throughput plant phenotyping, especially in the context of open field acquisitions, relies on the interpretation of data from different sensors implemented on various vectors such as tractors, robots or drones. Initially, these data were interpreted using remote sensing algorithms that exploit the spatial resolution of the signal. Since 2015, however, progresses of ”Deep Learning”, based on the training on examples, has already obtained promising results for measuring the rate of cover, counting plants or organs. It uses learned convolution layers, can take advantage of the spatial organization of the signal. The advantage of these methods is that they are based on Red-Green-Blue (RGB) sensors, which are much less expensive than multi- or hyperspectral imagers. However, these methods are sensitive to changes in the distribution between the data used in training and the predicted data. In practice, variable prediction errors from site to site can be observed using these methods. The objective of the thesis is to understand the causes of these variations and propose solutions for reliable phenotypic trait estimates using Deep Learning. The study focuses on detecting plants and organs from high-resolution RGB images acquired in the field. Our work first focused on the constitution of diversified image databases from different locations and stages of development for plant emergence (maize, beet, sunflower) and wheat ears, which allowed the publication of two annotated databases, grouping 27 acquisition sessions for thedrone and 47 for the ear detection. The datasets demonstrate the performances difference between the published results and ours due to the change in distribution. To go beyond the limits of the usual methods, we organized two data competitions, the Global Wheat Challenges, in 2020 and 2021, which allowed us to obtain solutions trained for robustness on a different data set than the training one. The analysis of the solutions showed the importance of the training strategies for robustness beyond the architectures used. We have also shown that these solutions can be effectively deployed as a replacement for manual counting. Finally, we have demonstrated the inefficiency of training functions designed for robust training. Our work opens the prospect of a better evaluation of Deep Learning in the context of high-throughput phenotyping and thus of confidence in its use in real-life conditions. Note de contenu : 1- Introduction 2- Evaluation of the robustness of handcrafted and deep learning methods for plant density estimation 3- Design of a large and diverse dataset for training and evaluating deep learning models: application to wheat head detection 4- Competition design to train robust Deep Learn model: the example of the Global Wheat Challenges 5- GlobalWheat-Wilds: Global Wheat Head Dataset as a benchmark of in-the-wild distribution shifts 6- Conclusion and perspectives Numéro de notice : 15244 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Thèse française Note de thèse : Thèse de Doctorat : Sciences Agronomiques : Avignon : 2021 Organisme de stage : Laboratoire EMMAH DOI : sans En ligne : https://hal.inrae.fr/tel-03431192v2/document Format de la ressource électronique : URL Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=100610

The use of deep machine learning for the automated selection of remote sensing data for the determination of areas of arable land degradation processes distribution / Dimitri I. Rukhovitch in Remote sensing, vol 13 n° 1 (January-1 2021)  

Public [article]  inRemote sensing > vol 13 n° 1 (January-1 2021) . - n° 155 Titre : The use of deep machine learning for the automated selection of remote sensing data for the determination of areas of arable land degradation processes distribution Type de document : Article/Communication Auteurs : Dimitri I. Rukhovitch, Auteur ; Polina V. Koroleva, Auteur ; Danila D. Rukhovitch, Auteur ; et al., Auteur Année de publication : 2021 Article en page(s) : n° 155 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Applications de télédétection [Termes IGN] apprentissage profond [Termes IGN] classification par réseau neuronal convolutif [Termes IGN] dégradation des sols [Termes IGN] distribution spatiale [Termes IGN] érosion [Termes IGN] image Landsat [Termes IGN] image Sentinel-MSI [Termes IGN] Normalized Difference Vegetation Index [Termes IGN] Russie [Termes IGN] surface cultivée [Termes IGN] système d'information géographique Résumé : (auteur) Soil degradation processes are widespread on agricultural land. Ground-based methods for detecting degradation require a lot of labor and time. Remote methods based on the analysis of vegetation indices can significantly reduce the volume of ground surveys. Currently, machine learning methods are increasingly being used to analyze remote sensing data. In this paper, the task is set to apply deep machine learning methods and methods of vegetation indices calculation to automate the detection of areas of soil degradation development on arable land. In the course of the work, a method was developed for determining the location of degraded areas of soil cover on arable fields. The method is based on the use of multi-temporal remote sensing data. The selection of suitable remote sensing data scenes is based on deep machine learning. Deep machine learning was based on an analysis of 1028 scenes of Landsats 4, 5, 7 and 8 on 530 agricultural fields. Landsat data from 1984 to 2019 was analyzed. Dataset was created manually for each pair of “Landsat scene”/“agricultural field number”(for each agricultural field, the suitability of each Landsat scene was assessed). Areas of soil degradation were calculated based on the frequency of occurrence of low NDVI values over 35 years. Low NDVI values were calculated separately for each suitable fragment of the satellite image within the boundaries of each agricultural field. NDVI values of one-third of the field area and lower than the other two-thirds were considered low. During testing, the method gave 12.5% of type I errors (false positive) and 3.8% of type II errors (false negative). Independent verification of the method was carried out on six agricultural fields on an area of 713.3 hectares. Humus content and thickness of the humus horizon were determined in 42 ground-based points. In arable land degradation areas identified by the proposed method, the probability of detecting soil degradation by field methods was 87.5%. The probability of detecting soil degradation by ground-based methods outside the predicted regions was 3.8%. The results indicate that deep machine learning is feasible for remote sensing data selection based on a binary dataset. This eliminates the need for intermediate filtering systems in the selection of satellite imagery (determination of clouds, shadows from clouds, open soil surface, etc.). Direct selection of Landsat scenes suitable for calculations has been made. It allows automating the process of constructing soil degradation maps. Numéro de notice : A2021-074 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.3390/rs13010155 Date de publication en ligne : 05/01/2021 En ligne : https://doi.org/10.3390/rs13010155 Format de la ressource électronique : url article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=96810 [article]

UAV photogrammetry and remote sensing / Fernando Carvajal-Ramírez (2021)  

Public Titre : UAV photogrammetry and remote sensing Type de document : Monographie Auteurs : Fernando Carvajal-Ramírez, Éditeur scientifique ; Francisco Agüera-Vega, Éditeur scientifique ; Patricio Martínez-Carricondo, Éditeur scientifique Editeur : Bâle [Suisse] : Multidisciplinary Digital Publishing Institute MDPI Année de publication : 2021 Importance : 258 p. Format : 16 x 24 cm ISBN/ISSN/EAN : 978-3-0365-1453-6 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Photogrammétrie numérique [Termes IGN] analyse d'image orientée objet [Termes IGN] apprentissage profond [Termes IGN] classification par réseau neuronal convolutif [Termes IGN] image captée par drone [Termes IGN] indice de végétation [Termes IGN] modélisation 3D du bâti BIM [Termes IGN] occupation du sol [Termes IGN] orthophotographie [Termes IGN] photogrammétrie aérienne [Termes IGN] point d'appui [Termes IGN] reconstruction 3D [Termes IGN] réseau antagoniste génératif [Termes IGN] semis de points [Termes IGN] structure-from-motion [Termes IGN] zone tampon Résumé : (éditeur) The concept of remote sensing as a way of capturing information from an object without making contact with it has, until recently, been exclusively focused on the use of Earth observation satellites.The emergence of unmanned aerial vehicles (UAV) with Global Navigation Satellite System (GNSS) controlled navigation and sensor-carrying capabilities has increased the number of publications related to new remote sensing from much closer distances. Previous knowledge about the behavior of the Earth's surface under the incidence different wavelengths of energy has been successfully applied to a large amount of data recorded from UAVs, thereby increasing the special and temporal resolution of the products obtained.More specifically, the ability of UAVs to be positioned in the air at pre-programmed coordinate points; to track flight paths; and in any case, to record the coordinates of the sensor position at the time of the shot and at the pitch, yaw, and roll angles have opened an interesting field of applications for low-altitude aerial photogrammetry, known as UAV photogrammetry. In addition, photogrammetric data processing has been improved thanks to the combination of new algorithms, e.g., structure from motion (SfM), which solves the collinearity equations without the need for any control point, producing a cloud of points referenced to an arbitrary coordinate system and a full camera calibration, and the multi-view stereopsis (MVS) algorithm, which applies an expanding procedure of sparse set of matched keypoints in order to obtain a dense point cloud. The set of technical advances described above allows for geometric modeling of terrain surfaces with high accuracy, minimizing the need for topographic campaigns for georeferencing of such products.This Special Issue aims to compile some applications realized thanks to the synergies established between new remote sensing from close distances and UAV photogrammetry. Note de contenu : 1- Using UAV-based photogrammetry to obtain correlation between the vegetation indices and chemical analysis of agricultural crops 2- Photogrammetry using UAV-mounted GNSS RTK: Georeferencing strategies without GCPs 3- Quality assessment of photogrammetric methods—A workflow for reproducible UAS orthomosaics 4- 3D reconstruction of power lines using UAV images to monitor corridor clearance 5- UAV-based terrain modeling under vegetation in the Chinese Loess Plateau: A deep learning and terrain correction ensemble frameword 6- UAV photogrammetry accuracy assessment for corridor mapping based on the number and distribution of ground control points 7- UAV + BIM: Incorporation of photogrammetric techniques in architectural projects with building information modeling versus classical work processes 8- Structure from motion of multi-angle RPAS imagery complements larger-scale airborne Lidar data for cost-effective snow monitoring in mountain forests 9- Use of UAV-photogrammetry for quasi-vertical wall surveying 10- Deep learning-based single image super-resolution: An investigation for dense scene reconstruction with UAS photogrammetry 11- Mapping heterogeneous urban landscapes from the fusion of digital surface model and unmanned aerial vehicle-based images using adaptive multiscale image segmentation and classification Numéro de notice : 28664 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Recueil / ouvrage collectif DOI : 10.3390/books978-3-0365-1453-6 En ligne : https://doi.org/10.3390/books978-3-0365-1453-6 Format de la ressource électronique : URL Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=99850

Unifying remote sensing image retrieval and classification with robust fine-tuning / Dimitri Gominski (2021)  

Public Titre : Unifying remote sensing image retrieval and classification with robust fine-tuning Type de document : Article/Communication Auteurs : Dimitri Gominski , Auteur ; Valérie Gouet-Brunet , Auteur ; Liming Chen, Auteur Editeur : Ithaca [New York - Etats-Unis] : ArXiv - Université Cornell Année de publication : 2021 Projets : Alegoria / Gouet-Brunet, Valérie Importance : 7 p. Format : 21 x 30 cm Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image optique [Termes IGN] analyse d'image numérique [Termes IGN] apprentissage profond [Termes IGN] base de données d'images [Termes IGN] classification par réseau neuronal convolutif [Termes IGN] image aérienne à axe vertical [Termes IGN] image aérienne oblique [Termes IGN] jeu de données [Termes IGN] précision de la classification [Termes IGN] recherche d'image basée sur le contenu Résumé : (auteur) Advances in high resolution remote sensing image analysisare currently hampered by the difficulty of gathering enoughannotated data for training deep learning methods, giving riseto a variety of small datasets and associated dataset-specificmethods. Moreover, typical tasks such as classification andretrieval lack a systematic evaluation on standard benchmarksand training datasets, which make it hard to identify durableand generalizable scientific contributions. We aim at uni-fying remote sensing image retrieval and classification witha new large-scale training and testing dataset, SF3001, in-cluding both vertical and oblique aerial images and madeavailable to the research community, and an associated fine-tuning method. We additionally propose a new adversarialfine-tuning method for global descriptors. We show that ourframework systematically achieves a boost of retrievalandclassification performance on nine different datasets com-pared to an ImageNet pretrained baseline, with currently noother method to compare to. Numéro de notice : P2021-003 Affiliation des auteurs : UGE-LASTIG+Ext (2020- ) Thématique : IMAGERIE Nature : Preprint nature-HAL : Préprint DOI : 10.48550/arXiv.2102.13392 En ligne : https://doi.org/10.48550/arXiv.2102.13392 Format de la ressource électronique : url article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=97283

Unsupervised vision methods based on image perceptual information / Eric Bazan (2021)  

Public Titre : Unsupervised vision methods based on image perceptual information Type de document : Thèse/HDR Auteurs : Eric Bazan, Auteur ; Petr Dokladal, Directeur de thèse ; Eva Dokladalova, Directeur de thèse Editeur : Paris : Université Paris Sciences et Lettres Année de publication : 2021 Importance : 227 p. Format : 21 x 30 cm Note générale : bibliographie Thèse de doctorat de l'Université Paris Sciences et Lettres, Préparée à MINES ParisTech, spécialité Morphologie Mathématique Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image optique [Termes IGN] apprentissage automatique [Termes IGN] apprentissage non-dirigé [Termes IGN] compréhension de l'image [Termes IGN] contour [Termes IGN] couleur (variable spectrale) [Termes IGN] décomposition spectrale [Termes IGN] filtre de Gabor [Termes IGN] image captée par drone [Termes IGN] segmentation d'image [Termes IGN] texture d'image [Termes IGN] vision Index. décimale : THESE Thèses et HDR Résumé : (auteur) This thesis work deals with extracting features and low-level primitives from perceptual image information to understand scenes. Motivated by the needs and problems in Unmanned Aerial Vehicles (UAVs) vision based navigation, we propose novel methods focusing on image understanding problems. This work explores three main pieces of information in an image: intensity, color, and texture. In the first chapter of the manuscript, we work with the intensity information through image contours. We combine this information with human perception concepts, such as the Helmholtz principle and the Gestalt laws, to propose an unsupervised framework for object detection and identification. We validate this methodology in the last stage of the drone navigation, just before the landing. In the following chapters of the manuscript, we explore the color and texture information contained in the images. First, we present an analysis of color and texture as global distributions of an image. This approach leads us to study the Optimal Transport theory and its properties as a true metric for color and texture distributions comparison. We review and compare the most popular similarity measures between distributions to show the importance of a metric with the correct properties such as non-negativity and symmetry. We validate such concepts in two image retrieval systems based on the similarity of color distribution and texture energy distribution. Finally, we build an image representation that exploits the relationship between color and texture information. The image representation results from the image’s spectral decomposition, which we obtain by the convolution with a family of Gabor filters. We present in detail the improvements to the Gabor filter and the properties of the complex color spaces. We validate our methodology with a series of segmentation and boundary detection algorithms based on the computed perceptual feature space. Numéro de notice : 15285 Affiliation des auteurs : non IGN Thématique : IMAGERIE/MATHEMATIQUE Nature : Thèse française Note de thèse : Thèse de Doctorat : Morphologie Mathématique : Paris Sciences et Lettres : 2021 Organisme de stage : Centre de Morphologie Mathématique DOI : sans En ligne : https://hal.science/tel-03690309 Format de la ressource électronique : URL Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=101418

Vectorization of historical maps using deep edge filtering and closed shape extraction / Yizi Chen (2021)  

Permalink

Vegetation stratum occupancy prediction from airborne LiDAR 3D point clouds / Ekaterina Kalinicheva (2021)  

Permalink

Automatic building footprint extraction from UAV images using neural networks / Zoran Kokeza in Geodetski vestnik, vol 64 n° 4 (December 2020 - February 2021)  

Permalink

Cartographic generalization / Monika Sester in Journal of Spatial Information Science, JoSIS, n° 21 (2020)  

Permalink

A deep learning approach to improve the retrieval of temperature and humidity profiles from a ground-based microwave radiometer / Xing Yan in IEEE Transactions on geoscience and remote sensing, Vol 58 n° 12 (December 2020)  

Permalink

Deep learning for detecting and classifying ocean objects: application of YoloV3 for iceberg–ship discrimination / Frederik Hass in ISPRS International journal of geo-information, vol 9 n° 12 (December 2020)  

Permalink

Exploring the inclusion of Sentinel-2 MSI texture metrics in above-ground biomass estimation in the community forest of Nepal / Santa Pandit in Geocarto international, vol 35 n° 16 ([01/12/2020])  

Permalink

Learning from urban form to predict building heights / Nikola Milojevic-Dupont in Plos one, vol 15 n° 12 (December 2020)  

Permalink

Mapping forest tree species in high resolution UAV-based RGB-imagery by means of convolutional neural networks / Felix Schiefer in ISPRS Journal of photogrammetry and remote sensing, vol 170 (December 2020)  

Permalink

MS-RRFSegNetMultiscale regional relation feature segmentation network for semantic segmentation of urban scene point clouds / Haifeng Luo in IEEE Transactions on geoscience and remote sensing, Vol 58 n° 12 (December 2020)  

Permalink