abstract: There is a growing interest in the creation of three-dimensional (3D) images and videos due to the growing demand for 3D visual media in commercial markets. A possible solution to produce 3D media files is to convert existing 2D images and videos to 3D. The 2D to 3D conversion methods that estimate the depth map from 2D scenes for 3D reconstruction present an efficient approach to save on the cost of the coding, transmission and storage of 3D visual media in practical applications. Various 2D to 3D conversion methods based on depth maps have been developed using existing image and video processing techniques. The depth maps can be estimated either from a single 2D view or from multiple 2D views. This thesis presents a MATLAB-based 2D to 3D conversion system from multiple views based on the computation of a sparse depth map. The 2D to 3D conversion system is able to deal with the multiple views obtained from uncalibrated hand-held cameras without knowledge of the prior camera parameters or scene geometry. The implemented system consists of techniques for image feature detection and registration, two-view geometry estimation, projective 3D scene reconstruction and metric upgrade to reconstruct the 3D structures by means of a metric transformation. The implemented 2D to 3D conversion system is tested using different multi-view image sets. The obtained experimental results of reconstructed sparse depth maps of feature points in 3D scenes provide relative depth information of the objects. Sample ground-truth depth data points are used to calculate a scale factor in order to estimate the true depth by scaling the obtained relative depth information using the estimated scale factor. It was found out that the obtained reconstructed depth map is consistent with the ground-truth depth data. / Dissertation/Thesis / M.S. Electrical Engineering 2010

Made available in DSpace on 2014-06-12T15:59:06Z (GMT). No. of bitstreams: 2 arquivo4838_1.pdf: 7602633 bytes, checksum: 600ab77ebdd41162baaff3435aa76f53 (MD5) license.txt: 1748 bytes, checksum: 8a4605be74aa9ea9d79846c1fba20a33 (MD5) Previous issue date: 2003 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / Um dos problemas abertos mais básicos da área de corte e empacotamento é encontrar o maior número de (,w)-retângulos que podem ser empacotados ortogonalmente num retângulo maior (L,W). O termo ortogonalmente quer dizer, apenas, que cada lado de um (,w)-retângulo empacotado é paralelo ou perpendicular aos lados do retângulo maior (L,W). Motivados por este problema e suas variantes mais difíceis (ex. caso tridimensional), desenvolvemos, baseado no trabalho [2], uma abordagem heurística geral de decomposições de gblocos. Os gblocos são uma generalização dos blocos. Os blocos são simplesmente retângulos em dimensão 2 e paralelepípedos em dimensão 3 (e seus análogos em dimensões maiores). Aplicando a abordagem de gblocos para o problema bidimensional aberto que mencionamos, mostramos se tratar, em termos de otimalidade, de um método superior á melhor heurística existente até o momento: a heurística de R. Morabito e S. Morales (1998). De fato ainda não é conhecido nenhum problema (,w, L,W) para o qual a nossa abordagem em gblocos não seja ótima. Esta observação empírica levanta a dúvida de estarmos diante de um método exato para o problema. Além do caso bidimensional, sugerimos também uma abordagem em gblocos para o caso tridimensional. Melhores métodos de empacotamento têm importante implicação econômica. Hoje, caminhões, trens, navios e aviões transportam contêineres e paletes com uma carga menor do que poderiam. Esta Tese é um passo na busca de melhores métodos. Ela apresenta alguns resultados originais, formaliza uma linguagem adequada para o problema abstrato e, por fim, sugere um caminho promissor para o problema concreto no setor de transporte de carga

This PhD work is dedicated to investigating the fringe effect of electrical capacitance tomography (ECT) and electrical resistance tomography (ERT) sensors systematically, exploring possible solutions and developing a novel 3D imaging technique by utilising the fringe effect of electrical tomography (ET) sensors. By referring to 2D simulation results, the fringe effect is investigated for 3D ERT and ECT sensor models with or without grounded guards, and with different axial lengths of electrodes. Simulation results reveal that increasing the electrode length or adopting grounded guards can reduce the fringe effect of ECT and ERT sensors. In this work, a voltage-excitation strategy is proposed and validated for ERT sensors instead of the conventional current-injection strategy. This approach enables ECT and ERT sensors to be integrated together as a common one to simplify the sensor design and reduce the interference between the ECT and ERT dual-modality measurements. For a conventional ERT sensor with the adjacent strategy, the fringe effect is evaluated for axially non-uniform central core and off-central core distributions at different axial positions and with different axial dimensions and conductivity contrasts. A method is proposed for compensating the fringe effect with the above typical distributions and a two-object distribution. A three-plane ERT sensor scheme is suggested for reducing the fringe effect induced by objects outside the sensor plane and the over-estimation by Landweber iteration. Both simulation and experiment have proved the effectiveness of the three-plane sensor scheme and the compensation method. Using the fringe effect, a novel 3D imaging method is proposed for ET with a single-plane sensor. It is explored to image metallic objects with ECT. The axial position of the object is derived by examining the corresponding fringe effect in a single-plane ECT sensor. Along with 2D imaging of the cross-sectional distribution, 3D imaging is achieved for a cylindrical metallic rod with known size. With this method, only the 3D position of the object needs to be calculated during the reconstruction which reduces the number of unknowns greatly and can improve the accuracy and stability of reconstruction.

3D representation nowadays has attracted much more public attention than ever before. One of the most important techniques in this field is depth extraction. In this thesis, we first introduce a well-known stereo matching method using color segmentation and belief propagation, and make an implementation of this framework. The color-segmentation based stereo matching method performs well recently, since this method can keep the object boundaries accurate, which is very important to depth map. Based on the implemented framework of segmentation based stereo matching, we proposed a color segmentation based 2D-to-3D video conversion method using high quality motion information. In our proposed scheme, the original depth map is generated from motion parallax by optical flow calculation. After that we employ color segmentation and plane estimation to optimize the original depth map to get an improved depth map with sharp object boundaries. We also make some adjustments for optical flow calculation to improve its efficiency and accuracy. By using the motion vectors extracted from compressed video as initial values for optical flow calculation, the calculated motion vectors are more accurate within a shorter time compared with the same process without initial values. The experimental results shows that our proposed method indeed gives much more accurate depth maps with high quality edge information. Optical flow with initial values provides good original depth map, and color segmentation with plane estimation further improves the depth map by sharpening its boundaries.

No description available.

No description available.

En Europe de l’Ouest, de nombreuses espèces de léporidés sont endémiques, ce qui en fait de bons témoins de l’évolution des écosystèmes terrestres dans lesquelles ont évolué les sociétés humaines préhistoriques. Paradoxalement, leur variabilité morphologique est mal connue et la phylogénie établie aujourd’hui, discutable. Ce travail propose de renseigner la diversité morphométrique des lapins et des lièvres sur près de deux millions d’années d’évolution, à travers l’application d’études ostéométriques et en morphométrie géométrique. L’analyse de restes osseux et dentaires – de populations actuelles et de 73 séries fossiles provenant de régions périméditerranéennes couvrant le Pléistocène – permet de caractériser les adaptations des léporidés face aux changements environnementaux et leurs tendances évolutives. Cette étude propose une nouvelle phylogénie pour le lapin et présente les différentes phases de dispersion des taxons à l’échelle de l’Europe occidentale. Ainsi, plusieurs événements de type expansion des populations, recolonisation des territoires depuis des zones refuges et extinctions locales, sont mis en évidence en réponse aux changements climatiques globaux. Ces résultats permettent de discuter la présence de ces petits gibiers dans l’environnement et alimente le débat sur les relations entre ces espèces et les communautés humaines. Leur augmentation significative dans la diète des groupes humains à la fin du Paléolithique supérieur, ne semble pas seulement s’expliquer par des changements cognitifs, culturels ou économiques, mais coïncide davantage avec les variations biogéographiques de ces espèces. / Fossil remains of small mammals of the Leporidae family are abundant in numerous paleontological and archaeological deposits from the Quaternary. Many species are endemic to Western Europe, which makes them reliable markers of change in the ecosystems in which prehistoric human societies evolved. Paradoxically, morphological variability of leporids is still poorly understood and current phylogeny remains a subject of debate. This work focuses on the morphometric diversity of rabbits (Oryctolagus) and hares (Lepus) over nearly two million years of evolution. We applied osteometry and geometric morphometric analyzes to bone and dental remains of current populations and 73 fossil samples from perimediterranean regions (Spain, France, Italy, Portugal) during different moments of the Pleistocene. These data provide insights concerning Leporidae adaptations to environmental change as well as more general evolutionary trends. Here we propose a new phylogeny for the genus Oryctolagus and present different dispersion phases for Western Europe. Several population expansion events coupled with the recolonization of refuge areas and local extinctions are highlighted in response to global climate change. These results allow us to discuss the presence of these small game species in the environment and contribute to the debate concerning relations between leporids and human communities. Humans have regularly consumed leporids since at least the Middle Paleolithic. Their significant increase in the diet at the end of the Upper Paleolithic, however, cannot be explained solely by cognitive, cultural or economic changes but rather coincides with biogeographic variations of these species.

Dans la présente étude, la conductivité thermique et le module d'élasticité de revêtementsd’YPSZ élaborés par projection plasma ont été prédits par modélisations numériques 2D et3D de type différences finies et éléments finis.L'influence de la résolution d'image, de la taille et de la valeur du seuil sur les propriétésprédites du revêtement a été étudiée. En outre, les effets de la méthode numérique et du typede condition aux limites ont été étudiés. En particulier, la quantification de l'effet Knudsen(effet de raréfaction) sur le transfert de chaleur à travers une structure poreuse a été réaliséepar modélisation numérique en combinaison avec l'analyse d'image. Les conductivitéseffectives obtenues par modélisation 3D s'avèrent plus élevées que celles obtenues en 2D, etaussi en meilleur accord avec les résultats mesurés. Une corrélation 2D/3D a été trouvéepour la modélisation de la conductivité thermique : cette corrélation permet de prédire lesvaleurs 3D à partir des valeurs calculées en 2D. / In the present study, the thermal conductivity and elastic modulus of thermal spray YPSZcoatings were predicted by 2D and 3D finite differences and finite elements numericalmodeling based on cross-sectional images.The influence of the image resolution, size and threshold on the predicted properties of thecoating was studied. Moreover, the effects of the numerical method and of the boundarycondition were investigated. In particular, the quantification of the Knudsen effect(rarefaction effect) on the heat transfer through a porous structure was realized by numericalmodeling in combination with image analysis. The predicted thermal conductivities obtainedby 3D modeling were found to be higher than those obtained by 2D modeling, and in betteragreement with the measured results. A 2D/3D correlation was sucessfully found for themodeling of thermal conductivity: this correlation allows predicting 3D computed valuesfrom 2D ones.

L’impact négatif de l’automobile sur l’environnement a entraîné une sévérisation des normes législatives pour protéger celui-ci. Les difficultés rencontrées lors de l’amélioration du rendement du moteur à allumage commandé, résultent principalement de son fonctionnement inefficace à charges partielles. La distribution variable est capable d'améliorations concernant la réduction de la consommation de carburant, surtout dans une zone d’utilisation fréquente : basses charges et bas régimes. Une alternative à l’étude expérimentale est l’approche par simulation numérique, CFD en utilisant le logiciel ANSYS-Fluent. Dans cette thèse de doctorat, le but principal a été de déterminer les vitesses d’écoulement du fluide pendant le processus d’admission pour deux lois de levée de la soupape, pour un moteur entraîné à 815 tr/min et une ouverture du papillon de gaz de 21.6°. Dans ce but, on a utilisé deux modèles de simulation numérique: un modèle bidimensionnel et un modèle tridimensionnel. L’étude réalisée par simulation numérique a permis de clarifier quelques aspects importants concernant les vitesses d’écoulement de l’air dans le cylindre et le degré de turbulence. / The negative impact of automobiles on the environment has led to increased severity in the legislation concerning environmental protection. The problems encountered in the efforts intended to improve the efficiency of the spark ignition engine are derived from its inefficient operation under partial loads. The variable intake valve lift is capable of significant changes aiming at lower fuel consumption, especially in the frequent use area: low torque, low speed. An alternative to the experimental study of fluid flow is the approach by numerical simulation, CFD, using the software ANSYS-Fluent. The main purpose of the present doctoral thesis was to determine the fluid flow velocity during the intake, for two intake valve lift laws, when the engine is running at 815 rpm, and with an opening of the throttle plate at 21.6°. To do this, we have used two numerical simulation models: one two-dimensional, and one three-dimensional. The study by numerical simulation made it possible to clarify some important issues regarding the air flow velocity into the cylinder, and the level of turbulence.

No description available.