Détail de l'éditeur
Bayerische Akademie der Wissenschaften
localisé à :
Munich
Collections rattachées :
|
Documents disponibles chez cet éditeur (793)
Ajouter le résultat dans votre panier
Visionner les documents numériques
Affiner la recherche Interroger des sources externes
Titre : Deep learning for feature based image matching Type de document : Thèse/HDR Auteurs : Lin Chen, Auteur ; Christian Heipke, Directeur de thèse Editeur : Munich : Bayerische Akademie der Wissenschaften Année de publication : 2021 Collection : DGK - C, ISSN 0065-5325 num. 867 Importance : 159 p. Format : 21 x 30 cm Note générale : bibliographie
Diese Arbeit ist gleichzeitig veröffentlicht in: Wissenschaftliche Arbeiten der Fachrichtung Geodäsie und Geoinformatik der Leibniz UniversitätHannoverISSN 0174-1454, Nr. 369, Hannover 2021Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Photogrammétrie numérique
[Termes IGN] appariement d'images
[Termes IGN] chaîne de traitement
[Termes IGN] classification par réseau neuronal convolutif
[Termes IGN] descripteur
[Termes IGN] image aérienne oblique
[Termes IGN] orientation d'image
[Termes IGN] orthoimageRésumé : (auteur) Feature based image matching aims at finding matched features between two or more images. It is one of the most fundamental research topics in photogrammetry and computer vision. The matching features area prerequisite for applications such as image orientation, Simultaneous Localization and Mapping (SLAM) and robot vision. A typical feature based matching algorithm is composed of five steps: feature detection, affine shape estimation, orientation, description and descriptor matching. Today, the employment of deep neural network has framed those different steps as machine learning problems and the matching performance has been improved significantly. One of the main reasons why feature based image matching may still prove difficult is the complex change between different images, including geometric and radiometric transformations. If the change between images exceeds a certain level, it will also exceed the tolerance of those aforementioned separate steps and, in turn, cause feature based image matching to fail.
This thesis focuses on improving feature based image matching against large viewpoint and viewing direction change between images. In order to improve the feature based image matching performance under these circumstances, affine shape estimation, orientation and description are solved with deep learning architectures. In particular, Convolutional Neural Networks (CNN) are used. For the affine shape and orientation learning, the main contribution of this thesis is two fold. First, instead of a Siamese CNN, only one branch is needed and the loss is built based on the geometric measures calculated from the mean gradient or second moment matrix. Therefore, for each of the input patches, a global minimum, namely the canonical feature, exists. Second, both the affine shape and orientation are solved simultaneously within one network by combining the loss used for affine shape and orientation learning. To the best of the author’s knowledge, this is the first time these two modules are reported to have been successfully trained simultaneously. For the descriptor learning part, a new weak match is defined. For any input feature patch, a slightly transformed patch that lies far from the input feature patch in descriptor space is defined as a weak match feature. A weak match finder network is proposed to actively find these weak match features. In a following step, the found weak matches are used in the standard descriptor learning framework. In this way, the intra-variance of the appearance of matched feature patch pairs is explored in depth and, accordingly, the invariance of feature descriptors against viewpoint and viewing direction change is improved. The proposed feature based image matching method is evaluated on standard benchmarks and is used to solve for the parameters of image orientation. For the image orientation task, aerial oblique images are taken into account. Through analysis of the experiments conducted for small image blocks, it is shown that deep learning feature based image matching leads to more registered images, more reconstructed 3D points and a more stable block connection.Note de contenu : 1- Introduction
2- Basics
3- Related work
4- Deep learning feature representation
5- Experiments and results
6- Discussion
7- Conclusion and outlookNuméro de notice : 17673 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Thèse étrangère Note de thèse : PhD dissertation : Fachrichtung Geodäsie und Geoinformatik : Hanovre : 2021 En ligne : https://dgk.badw.de/fileadmin/user_upload/Files/DGK/docs/c-867.pdf Format de la ressource électronique : URL Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=97999 Improving smartphone-based GNSS positioning using state space augmentation techniques / Francesco Darugna (2021)
Titre : Improving smartphone-based GNSS positioning using state space augmentation techniques Type de document : Thèse/HDR Auteurs : Francesco Darugna, Auteur ; Steffen Schön, Directeur de thèse Editeur : Munich : Bayerische Akademie der Wissenschaften Année de publication : 2021 Collection : DGK - C, ISSN 0065-5325 num. 864 Importance : 189 p. Note générale : bibliographie
Diese Arbeit ist gleichzeitig veröffentlicht in:Wissenschaftliche Arbeiten der Fachrichtung Geodäsie und Geoinformatik der Universität Hannover - ISSN 0174-1454, Nr. 368, Hannover 2021Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Navigation et positionnement
[Termes IGN] antenne GNSS
[Termes IGN] étalonnage d'instrument
[Termes IGN] positionnement par GNSS
[Termes IGN] retard troposphérique zénithal
[Termes IGN] téléphone intelligentRésumé : (auteur) Low-cost receivers providing Global Navigation Satellite System (GNSS) pseudorange and carrier phase raw measurements for multiple frequencies and multiple GNSS constellations have become available on the market in the last years. This significantly has increased the number of devices equipped with the necessary sensors to perform precise GNSS positioning. GNSS pseudorange and carrier phase are used to compute user positions. While both observations are affected by different error sources, e.g. the passage through the atmosphere, only the carrier-phase has an ambiguous nature. The resolution of this ambiguity is a crucial factor to reach fast and highly precise GNSS-based positioning. Currently, several smartphones are equipped with a dual-frequency, multi-constellation receiver. The access to Android-based GNSS raw measurements has become a strong motivation to investigate the feasibility of smartphone-based high-accuracy positioning. The quality of smartphone GNSS measurements has been analyzed, suggesting that they often suffer from low signal-to-noise, inhomogeneous antenna gain and high levels of multipath. This workshows how to tackle several of the currently present obstacles and demonstrates centimeter-level positioning with a low-cost GNSS antenna and a low-cost GNSS receiver built into an off-the-shelf smartphone. Since the beginning of the research in smartphone-based positioning, the device’s GNSS antenna has been recognized as one of the main limitations. Besides Multipath (MP), the antenna radiation pattern is the main site-dependent error source of GNSS observations. An absolute antenna calibration has been performed for the dual-frequency smartphone HuaweiMate20X. Antenna Phase Center Offset (PCO), and Variations (PCV ) have been estimated to correct for the antenna impact on the L1 and L5 phase observations. Accordingly, the relevance of considering the individual PCO and PCV for the two frequencies is shown. The PCV patterns indicate absolute values up to 2 cm and 4 cm for L1 and L5, respectively. The impactof antenna corrections has been assessed in different multipath environments using a high-accuracy positioning algorithm employing an uncombined observation model and applying Ambiguity Resolution (AR). Experiments both in zero-baseline and short-baseline configurations have been performed. Instantaneous AR in the zero-baseline setup has been demonstrated, showing the potential for cm-level positioning with low-cost sensors available inside smartphones. In short-baselines configurations, no reliable AR is achieved without antenna corrections. However, after correcting for PCV, successful AR is demonstrated for a smartphone placed in a low multipath environment on the ground of a soccer field. For a rooftop open-skytest case with large multipath, AR was successful in 19 out of 35 data-sets. Overall, the antenna calibration is demonstrated being an asset for smartphone-based positioning with AR,showing cm-level 2D Root Mean Square Error (RMSE). In GNSS-based positioning, a user within a region covered by a network of reference stations can take advantage of the network-estimated augmentation parameters. Among the GNSS error sources, atmospheric delays have a strong impact on the positioning performance and the ability to resolve ambiguities. State Space Representation (SSR) atmospheric corrections, i.e. tropospheric and ionospheric delays, are commonly estimated for the approximate user position by interpolation from values calculated for the reference stations. Widely used interpolation techniques are Inverse Distance Weighted (IDW), Ordinary Kriging (OK)and Weighted Least Squares (WLS). The interpolation quality of such techniques during severe weather events and Traveling Ionospheric Disturbances (TIDs) is analyzed. To improve the interpolation performance during such events, modified WLS methods taking advantage of the physical atmospheric behavior are proposed. To support this interpolation approach, external information from Numerical Weather Models (NWM) for tropospheric interpolation and from TID modeling for ionospheric interpolation is introduced to the algorithms. The interpolation is assessed using simulated data (considering artificial and real network geometries), and real SSR parameters generated by network computation of GNSS measurements. As examples, two severe weather events in northern Europe in 2017 and one TID eventover Japan in 2019 have been analyzed. The interpolation of SSR Zenith Tropospheric Delay(ZTD) and ionospheric parameters is evaluated. Considering the reference station positions as rover locations, the modified WLS approach marks a lower RMSE in up to 80% of the cases during sharp weather fluctuations. Also, the average error can be decreased in 64% of the cases during the TID event investigated. Improvements up to factors larger than two are observed. Furthermore, specific cases are isolated, showing particular ZTD variations where significant errors (e.g. larger than 1 cm) can be reduced by up to 20% of the total amount. As a final product of the analysis, tropospheric and ionospheric messages are proposed. The messages contain the information needed to implement the suggested interpolation. Along with the need for accurate atmospheric models, the concept of consistency in the SSR corrections is crucial. A format that can transport all the SSR corrections estimated by a network is the Geo++ SSR format (SSRZ). Exploiting the features of the SSRZ format, the impact of an error in the transported ionospheric parameters is investigated. It is shown that the position estimation strongly depends on the ionospheric modeling and mismodeling can result in cm level errors, especially in the height component. Numéro de notice : 17182 Affiliation des auteurs : non IGN Thématique : POSITIONNEMENT Nature : Thèse étrangère Note de thèse : Thesis : Geodäsie und Geoinformatik : Hanovre : 2021 En ligne : https://dgk.badw.de/fileadmin/user_upload/Files/DGK/docs/c-864.pdf Format de la ressource électronique : URL Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=98000
Titre : Robust and fast global image orientation Type de document : Thèse/HDR Auteurs : Xin Wang, Auteur ; Christian Heipke, Directeur de thèse Editeur : Munich : Bayerische Akademie der Wissenschaften Année de publication : 2021 Collection : DGK - C, ISSN 0065-5325 num. 871 Importance : 141 p. Note générale : bibliographie
Diese Arbeit ist gleichzeitig veröffentlicht in: Wissenschaftliche Arbeiten der Fachrichtung Geodäsie und Geoinformatik der Leibniz Universität Hannover ISSN 0174-1454, Nr. 373, Hannover 2021Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Photogrammétrie numérique
[Termes IGN] appariement d'images
[Termes IGN] appariement dense
[Termes IGN] chaîne de traitement
[Termes IGN] estimation de pose
[Termes IGN] méthode robuste
[Termes IGN] orientation d'image
[Termes IGN] orientation relative
[Termes IGN] rotation
[Termes IGN] structure-from-motion
[Termes IGN] translation
[Termes IGN] valeur aberranteRésumé : (auteur) The estimation of image orientation (also called pose) has always played a crucial role in the field of photogrammetry since it is a fundamental prerequisite for the subsequent works of multi-view dense matching, generating DEM and DSM, etc. In the community of computer vision, the task is also well known as Structure-from-Motion (SfM), which reveals that image pose, while positions of object points are determined interdependently. Despite a lot of efforts over the last decades, it has recently gained the photogrammetrists’ interests again due to the fast-growing number of different resources of images. New challenges are posed for accurately and efficiently orienting various image datasets (e.g., unordered datasets with a large number of images, or images compromised of critical stereo pairs). In this thesis, the relevant ambition is to develop a new fast and robust method for the estimation of image orientation which is capable of coping with different types of datasets. To achieve this goal, the two most time-consuming steps of image orientation are in particular taken care of: (a) image matching and (b) the estimation process. To accelerate the image matching process, a new method employing a random k-d forest is proposed to quickly obtain pairs of overlapping images from an unordered image set. After that, image matching and the estimation of relative orientation parameters are performed only for pairs found to be very likely overlapping. On the other hand, to estimate the image poses in a time efficient manner, a global image orientation strategy is advocated. Its basic idea is to first simultaneously solve all available images’ poses, before a final bundle adjustment is carried out once for refinement. The conventional two-step global approach is pursued in this work, separating the determination of rotation matrices and translation parameters; the former is solved by an existing popular method of Chatterjee and Govindu [2013], and the latter are estimated globally using a newly developed method: translation estimation integrating both the relative translations and tie points. Tie points within triplets are adopted to firstly calculate global unified scale factors for each available pairwise relative translation. Then, analogous to rotation estimation, translations are determined by performing an averaging operation on the scaled relative translations. In order to improve the robustness of the solution, efforts in this thesis are also focused on coping with outliers in the relative orientations (ROs), which global image orientation approaches are particularly sensitive to. A general method based on triplet compatibility with respect to loop closure errors of relative rotations and translations is presented for detecting blunders in relative orientations. Although this procedure eliminated many gross errors in the input ROs, it typically cannot sort out blunders which are caused by repetitive structures and critical configurations, such as inappropriate baselines (very short baseline or baselines parallel to the viewing direction). Therefore, another new method is proposed to eliminate wrong ROs which have resulted from repetitive structures and very short baselines. Two corresponding criteria that indicate the quality of ROs are introduced. Repetitive structure is detected based on counts of conjugate points of the various image pairs, while very short baselines are found by inspecting the intersection angles of corresponding image rays. By analyzing these two criteria, incorrect ROs are detected and eliminated. As correct ROs of image pairs with a wider baseline nearly parallel to both viewing directions can be valuable, a method to identify and keep these ROs is also a part of this research. The validation and evaluation of the proposed method are thoroughly conducted on various benchmarks including ordered and unordered sets of images, images with repetitive structures and inappropriate baselines, etc. In particular, robustness is investigated by demonstrating the efficacy of the corresponding RO outlier detection methods. The performance and time efficiency of determining image orientation are evaluated and compared with several state-of-the-art global image orientation approaches. In summary, based on the experimental results, the developed methods demonstrateto be able to accomplish the image orientation taskfast and robustlyon different kinds of datasets. Numéro de notice : 17672 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Thèse étrangère Note de thèse : PhD dissertation : Fachrichtung Geodäsie und Geoinformatik : Hanovre : 2021 En ligne : https://dgk.badw.de/fileadmin/user_upload/Files/DGK/docs/c-871.pdf Format de la ressource électronique : URL Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=97997
Titre : Exact optimization algorithms for the aggregation of spatial data Type de document : Thèse/HDR Auteurs : Johannes Oehrlein, Auteur Editeur : Munich : Bayerische Akademie der Wissenschaften Année de publication : 2020 Collection : DGK - C, ISSN 0065-5325 num. 862 Importance : 184 p. Format : 21 x 30 cm Note générale : bibliographie
Dissertation zur Erlangung des GradesDoktor-Ingenieur (Dr.-Ing.)
Diese Arbeit ist gleichzeitig als elektronische Dissertationbei der Universitäts-und Landesbibliothek Bonn veröffentlichtLangues : Anglais (eng) Descripteur : [Termes IGN] agrégation spatiale
[Termes IGN] cycliste
[Termes IGN] données localisées
[Termes IGN] espace vert
[Termes IGN] généralisation automatique de données
[Termes IGN] programmation linéaire
[Termes IGN] réseau routier
[Termes IGN] trajet (mobilité)
[Termes IGN] zone urbaine
[Vedettes matières IGN] GénéralisationRésumé : (auteur) The aggregation of spatial data is a recurring problem in geoinformation science. Aggregating data means subsuming multiple pieces of information into a less complex representation. It is pursued for various reasons, like having a less complex data structure to apply further processing algorithms or a simpler visual representation as targeted in map generalization. In this thesis, we identify aggregation problems dealing with spatial data and formalize themas optimization problems. That means we set up a function that is capable of evaluating valid solutions to the considered problem, like a cost function for minimization problems. To each problem introduced, we present an algorithm that finds a valid solution that optimizes this objective function. In general, this superiority with respect to the quality of the solution comes at the cost of computation efficiency, a reason why non-exact approaches like heuristics are widely used for optimization. Nevertheless, the higher quality of solutions yielded by exact approaches is undoubtedly important. On the one hand, “good” solutions are sometimes not sufficient. On the other hand, exact approaches yield solutions that maybe used as benchmarks for the evaluation of non-exact approaches. This kind of application is of particular interest since heuristic approaches, for example, give no guarantee on the quality of solutions found. Furthermore, algorithms that provide exact solutions to optimization problems reveal weak spots of underlying models. A result that does not satisfy the user cannot be excused with a mediocre performance of an applied heuristic. With this motivation, we developed several exact approaches for aggregation problems, which we present in this thesis. Since we deal with spatial data, for all problems considered, the aggregation is based on both geometric and semantic aspects although the focus varies. The first problem we discuss is about visualizing a road network in the context of navigation. Given a fixed location in the network, we aim for a clear representation of the surroundings. For this purpose, we introduce an equivalence relation for destinations in the network based on which we perform the aggregation. We succeed in designing an efficient algorithm that aggregates as many equivalent destinations as possible. Furthermore, we tackle a class of similar and frequently discussed problems concerning the aggregation of areal units into larger, connected regions. Since these problems are NP-complete, i.e. extraordinarily complex, we do not aim for an efficient exact algorithm (which is suspected not to exist) but present a strong improvement to existing exact approaches. In another setup, we present an efficient algorithm for the analysis of urban green-space supply. Performing a hypothetical assignment of citizens to available green spaces, it detects local shortages and patterns in the accessibility of green space within a city. Finally, we introduce and demonstrate a tool for detecting route preferences of cyclists based on a selection of given trajectories. Examining a set of criteria forming suitable candidates, we aggregate them efficiently to the best-fitting derivable criterion. Overall, we present exact approaches to various aggregation problems. In particular, the NP-complete problem we deal with firmly underscores, as expected, the need for heuristic approaches. For applications asking for an immediate solution, it may be reasonable to apply a heuristic approach. This holds in particular due to easy and generally applicable meta-heuristics being available. However, with this thesis, we argue for applying exact approaches if possible. The guaranteed superior quality of solutions speaks for itself. Besides, we give additional examples which show that exact approaches can be applied efficiently as well. Numéro de notice : 17681 Affiliation des auteurs : non IGN Thématique : GEOMATIQUE Nature : Thèse étrangère Note de thèse : PhD dissertation : : Rheinische Friedrich-Wilhelms-Universität Bonn : 2020 En ligne : https://nbn-resolving.org/urn:nbn:de:hbz:5-60713 Format de la ressource électronique : URL Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=98023
Titre : Integrated georeferencing for precise depth map generation exploiting multi-camera image sequences from mobile mapping Type de document : Thèse/HDR Auteurs : Stefan Cavegn, Auteur Editeur : Munich : Bayerische Akademie der Wissenschaften Année de publication : 2020 Collection : DGK - C, ISSN 0065-5325 num. 863 Note générale : bibliographie
Dissertation zur Erlangung des Grades Doktor-Ingenieur (Dr.-Ing.)
Diese Arbeit ist gleichzeitig veröffentlicht in: Wissenschaftliche Arbeiten der Fachrichtung Geodäsie der Universität StuttgartLangues : Anglais (eng) Résumé : (auteur) Image-based mobile mapping systems featuring multi-camera configurations allow for efficient geospatial data acquisition in both outdoor and indoor environments. We aim at accurate geospatial 3D image spaces consisting of collections of georeferenced multi-view RGB-D imagery, which may serve as basis for 3D street view services. In order to obtain high-quality depth maps, dense image matching exploiting multi-view image sequences captured with high redundancy needs to be performed. Since this process is entirely dependent on accurate image orientations, we mainly focus on pose estimation of multi-camera systems within this thesis. Nonetheless, we also present methods and investigations to obtain accurate, reliable and complete 3D scene representations based on multi-stereo mobile mapping sequences. Conventional image orientation approaches such as direct georeferencing enable absolute accuracies at the centimeter level in open areas with good GNSS coverage. However, GNSS conditions of street-based mobile mapping in urban canyons are often deteriorated by multipath effects and by shading of the signals caused by vegetation and large multi-story buildings. Moreover, indoor spaces do not even allow for any GNSS signals. Hence, we propose a powerful and versatile image orientation procedure that is able to cope with these issues encountered in challenging urban environments. Our integrated georeferencing approach extends the powerful structure-from-motion pipeline COLMAP with georeferencing capabilities. It assumes initial camera poses with sub-meter accuracy, which allow for direct triangulation of the complete scene. Such a global approach is much more efficient than an incremental structure-from-motion procedure. Furthermore, an initial image orientation solution already facilitates to georeference in a geodetic reference frame. Nevertheless, accuracies at the centimeter level can only be achieved by incorporation of ground control points. In order to obtain sub-pixel accurate relative orientations, strong tie point connections for the highly redundant multi-view image sequences are required. However, hardly overlapping fields of view, strongly varying views and weakly textured surfaces aggravate image feature matching. Hence, constraining relative orientation parameters among cameras is crucial for accurate, robust and efficient image orientation. Apart from supporting fixed multi-camera rigs, our integrated georeferencing approach that uses bundle adjustment allows for self-calibration of all relative orientation parameters or just single components. We extensively evaluated our integrated georeferencing procedure using six challenging real-world datasets in order to demonstrate its accuracy, robustness, efficiency and versatility. Four datasets were captured outdoors, one by a rail-based and three by different street-based multi-stereo camera systems. A portable mobile mapping system featuring a multi-head panorama camera collected two datasets in an indoor environment. Employing relative orientation constraints and ground control points within these indoor spaces resulted in absolute 3D accuracies of ca. 2 cm, and precisions at the millimeter level for relative 3D measurements. Depending on the use case, absolute 3D accuracy values for outdoor environments are slightly larger and amount to a few centimeters. However, determining 3D reference coordinates is a costly task. Not relying on any ground control points led to horizontal accuracies of ca. 5 cm for a scenario featuring some loops, while dropping down to a few decimeters for an extended junction area. Since the height component is even more dependent on prior camera poses from direct georeferencing, these 2D accuracies significantly decreased for the 3D case. However, incorporating just one ground control point facilitates the elimination of systematic effects, which results in 3D accuracies within the sub-decimeter range. Nevertheless, at least one additional check point is recommended in order to ensure a reliable solution. Once consistent and sub-pixel accurate relative poses of spatially adjacent images are available, in-sequence dense image matching can be performed. Aiming at precise and dense depth map generation, we evaluated several image matching configurations. Standard single stereo matching led to high accuracies, which could not significantly be improved by in-sequence matching. However, the image redundancy provided by additional epochs resulted in more complete and reliable depth maps. Note de contenu : 1- Introduction
2- From imagery to 3D geometry
3- Developped methods for integrated georeferencing
4- Evaluation of integrated georeferencing
5- Evaluation of In-sequence dense image matching
6- Conclusion and outlookNuméro de notice : 17679 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Thèse étrangère Note de thèse : PhD thesis : Ingénierie aérospatiale et géodésie : Stuttgart : 2020 En ligne : https://dx.doi.org/10.18419/opus-11210 Format de la ressource électronique : URL Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=98019 
Permalink
