When a camera system in a car is mounted behind the windshield, light rays will be refracted by the windshield. The distortion can be significant, especially for wide field-of-view cameras. Traditional approaches handle the windshield distortion along with the calibration that calculates the intrinsic and extrinsic parameters. However, these approaches do not handle the windshield distortion explicitly, and to understand the image formation, it brings to understand more about the windshield distortion effect. In this thesis, data is collected from a camera system viewed with and without the windshield. The windshield distortion effect has been studied by varying the windshield’s tilt and the camera’s setup. Points are then found in both images and matched. From this, a distortion difference is calculated and analyzed. Next, a preliminary model of the windshield distortion effect is presented and evaluated. The results show that the model works well for all cases and the two windshields considered in this thesis.

Systématiser : tel est le leitmotiv des résultats de cette thèse portant sur trois domaines d'étude en vision et en géométrie algorithmique. Dans le premier, nous étendons toute la machinerie du modèle sténopé des appareils photos classiques à un ensemble d'appareils photo (deux fentes, à balayage, oblique, une fente) jusqu'à présent étudiés séparément suivant différentes approches. Dans le deuxième, nous généralisons avec peu d'efforts aux convexes de R3 l'étude des épinglages de droites ou de boules, menée différemment selon la nature des objets considérés. Dans le troisième, nous tentons de dégager une approche systématique pour élaborer des stratégies d'évaluation polynomiale de prédicats géométriques, les méthodes actuelles étant bien souvent spécifiques à chaque prédicat étudié. De tels objectifs ne peuvent être atteints sans un certain investissement mathématique dans l'étude des congruences linéaires de droites, de propriétés différentielles des ensembles de tangentes à des convexes et de la théorie des invariants algébriques, respectivement. Ces outils ou leurs utilisations reposent sur la géométrie de P3 (R), construite dans la seconde moitié du XIXe siècle mais pas complètement assimilée en géométrie algorithmique et dont nous proposons une synthèse adaptée aux besoins de la communauté. / Systematize is the leitmotiv of the results in this thesis. Three problems are studied in the field of computer vision and computational geometry. In the first one, we extend all the machinery of the pinhole model for classical cameras to a whole set of cameras (two-slit, pushbroom, oblique, pencil), which were separately studied with different approaches. In the second one, we generalize to convex bodies in R3 the work on pinning lines by or balls, which had so far been tackled by techniques intimately linked to the geometry of the objects. In the third one, we attempt to work out a systematic approach in place of problem-specific methods in order to build polynomial evaluation trees for geometric predicates. Such goals could not be reached without a mathematical investigation in the study of linear line congruences, differential properties of sets of tangent lines to a convex and classical invariant theory respectively. These tools or their uses are mostly based on line geometry in P3 (R). This geometry was designed in the second half of the 19th century but its full power hos not yet been used by the computational geometry community. This thesis therefore also serves as an extended tutorial.

<p>This thesis presents a complete and fully automatic method for estimating the volume of an airbag, through all stages of its inflation, with multiple synchronized high-speed cameras.</p><p>Using recorded contours of the inflating airbag, its visual hull is reconstructed with a novel method: The intersections of all back-projected contours are first identified with an accelerated epipolar algorithm. These intersections, together with additional points sampled from concave surface regions of the visual hull, are then Delaunay triangulated to a connected set of tetrahedra. Finally, the visual hull is extracted by carving away the tetrahedra that are classified as inconsistent with the contours, according to a voting procedure.</p><p>The volume of an airbag's visual hull is always larger than the airbag's real volume. By projecting a known synthetic model of the airbag into the cameras, this volume offset is computed, and an accurate estimate of the real airbag volume is extracted. </p><p>Even though volume estimates can be computed for all camera setups, the cameras should be specially posed to achieve optimal results. Such poses are uniquely found for different airbag models with a separate, fully automatic, simulated annealing algorithm.</p><p>Satisfying results are presented for both synthetic and real-world data.</p>

This thesis presents a complete and fully automatic method for estimating the volume of an airbag, through all stages of its inflation, with multiple synchronized high-speed cameras. Using recorded contours of the inflating airbag, its visual hull is reconstructed with a novel method: The intersections of all back-projected contours are first identified with an accelerated epipolar algorithm. These intersections, together with additional points sampled from concave surface regions of the visual hull, are then Delaunay triangulated to a connected set of tetrahedra. Finally, the visual hull is extracted by carving away the tetrahedra that are classified as inconsistent with the contours, according to a voting procedure. The volume of an airbag's visual hull is always larger than the airbag's real volume. By projecting a known synthetic model of the airbag into the cameras, this volume offset is computed, and an accurate estimate of the real airbag volume is extracted. Even though volume estimates can be computed for all camera setups, the cameras should be specially posed to achieve optimal results. Such poses are uniquely found for different airbag models with a separate, fully automatic, simulated annealing algorithm. Satisfying results are presented for both synthetic and real-world data.

Cameras generally have a field of view only large enough to capture a portion of their surroundings. The goal of immersion is to replace many of your senses with virtual ones, so that the virtual environment will feel as real as possible. Panoramic cameras are used to capture the entire 360°view, also known as panoramic images.Virtual reality makes use of these panoramic images to provide a more immersive experience compared to seeing images on a 2D screen. This thesis, which is in the field of Computer vision, focuses on establishing a multi-camera geometry to generate a cylindrical panorama image and successfully implementing it with the cheapest cameras possible. The specific goal of this project is to propose the cameras geometry which will decrease artifact problems related to parallax in the panorama image. We present a new approach of cylindrical panoramic images from multiple cameras which its setup has cameras placed evenly around a circle. Instead of looking outward, which is the traditional ”radial” configuration, we propose to make the optical axes tangent to the camera circle, a ”tangential” configuration. Beside an analysis and comparison of radial and tangential geometries, we provide an experimental setup with real panoramas obtained in realistic conditions / Les caméras ont généralement un champ de vision à peine assez grand pour capturer partie de leur environnement. L’objectif de l’immersion est de remplacer virtuellement un grand nombre de sens, de sorte que l’environnement virtuel soit perçu comme le plus réel possible. Une caméra panoramique est utilisée pour capturer l’ensemble d’une vue 360°, également connue sous le nom d’image panoramique. La réalité virtuelle fait usage de ces images panoramiques pour fournir une expérience plus immersive par rapport aux images sur un écran 2D. Cette thèse, qui est dans le domaine de la vision par ordinateur, s’intéresse à la création d’une géométrie multi-caméras pour générer une image cylindrique panoramique et vise une mise en œuvre avec les caméras moins chères possibles. L’objectif spécifique de ce projet est de proposer une géométrie de caméra qui va diminuer au maximum les problèmes d’artefacts liés au parallaxe présent dans l’image panoramique. Nous présentons une nouvelle approche de capture des images panoramiques cylindriques à partir de plusieurs caméras disposées uniformément autour d’un cercle. Au lieu de regarder vers l’extérieur, ce qui est la configuration traditionnelle ”radiale”, nous proposons de rendre les axes optiques tangents au cercle des caméras, une configuration ”tangentielle”. Outre une analyse et la comparaison des géométries radiales et tangentielles, nous fournissons un montage expérimental avec de vrais panoramas obtenus dans des conditions réalistes

[pt] Com o avanço do poder de processamento gráfico e a popularização dos jogos eletrônicos, começam a surgir novas formas de entretenimento, entre elas a de espectadores de jogos. Os requisitos para o posicionamento de câmera para os espectadores visualizarem os jogos são diferentes dos aplicados aos jogadores. Enquanto a câmera para os jogadores deve atender a requisitos de jogabilidade, a câmera para os espectadores deve explorar diferentes ângulos de visualização a fim de aumentar a imersão no ambiente dos jogos e explorar a emoção das cenas. O cinema já evoluiu toda uma linguagem cinematográfica que potencializa a interpretação de cenas. Pesquisas recentes procuram usar a cinematografia na visualização de jogos. O principal desafio na adaptação da cinematografia para os jogos reside na existência de um ambiente dinâmico, onde não se pode prever o andamento da história. Podemos identificar três módulos para a adaptação da cinematografia em jogos: roteirista, responsável por identificar o que está ocorrendo na cena; diretor/editor, responsável por definir as melhores tomadas para capturar a cena; cinegrafista, responsável por posicionar a câmera no ambiente dinâmico para melhor atender às demandas do módulo diretor/editor. Este trabalho propõe a implementação de um módulo cinegrafista. A partir de requisitos para o posicionamento da câmera oriundos de um módulo externo, propõe-se um modelo de câmera que faz o posicionamento automático da câmera. Os requisitos de posicionamento são expressos em um conjunto de restrições que devem ser atendidas pelo modelo de câmera. A câmera usa um modelo físico baseado em um sistema de partículas regido pelo método de Verlet, empregando o método de relaxação para a convergência do sistema a fim de atender às restrições impostas. Experimentos computacionais demonstram a capacidade do módulo proposto de atender a sofisticadas regras de posicionamento de câmera, baseando-se em composições de restrições simples. Princípios da cinematografia, como enquadramento, posicionamento e movimento de câmeras, e respeito à linha de ação, são facilmente respeitados pelo módulo proposto. / [en] With the evolution in graphics processing power and the popularization of electronic games, new forms of entertainment, such as being a game spectator. The requirements for positioning the camera for the spectators to view the games are different from those applied to players. The camera for the players must fulfill playability requirements, while the camera for the spectators must explore different viewing angles in order to increase the immersion in the game environment and to explore the thrill of the scenes. The cinema has evolved a whole cinematographic language that optimizes the scene interpretation. Recent research has been seeking to apply cinematography to game visualization. The main challenge in the adaptation of the cinematographic language for games lies in the existence of a dynamic environment in which the story`s progress cannot be anticipated. Three modules can be identified to adopt the cinematographic language in games: screenwriter, responsible for identifying what is happening in the scene; director/editor, responsible for defining the best takes to capture the scene; and cinematographer, responsible for positioning the camera in the dynamic environment to better comply with the director/editor`s demands. The present work proposes the implementation of a cinematographer module. Based on camera position requirements obtained from an external module, we propose a camera model that automatically positions the camera. The positioning requirements are expressed by a set of constraints that must be respected by the camera model. The camera uses a physical model based on a particle system oriented by Verlet`s method, and employs the relaxation method to obtain the system`s convergence in order to comply with imposed constraints. Computational experiments have demonstrated the capacity of the proposed module to comply with sophisticated camera positioning rules based on compositions of simple constraints. Cinematography principles such as framing, camera position and movement, and respecting the line of action are easily accomplished by the proposed module.

This thesis deals with reconstruction of golf ball position using multiple cameras. Reconstruction will be used for golf simulator project. System is using fotogrametric calibration and triangulation algorithm for obtaing point coordinates. Work also discuss options for camera selection. The result is making of prototype of the simulator.

The thesis describes theoretical fundamentals of optics and problems of ray-tracing method, ray-triangle intersection computation included. Next section describes concept of three different camera models for ray-tracing method - plain camera, pinehole camea and spherical camera (fish-eye). The thesis comparse properties and capabilities of camera models and their effect to projected image of a scene. The program for raytracing was implemented for purposes of the thesis including implementation of camera models described in the thesis.