Dissertations / Theses on the topic 'Adaptive multiresolution'

Flows around immersed boundaries exhibit many complex, well defined and active dynamical structures. In fact, features such as shock waves, strong vorticity concentrations in shear layers, wakes, or boundary layer regions are critical elements in representing the dynamics of a flow field. In order to capture the correct kinematic and dynamic quantities associated with the fluid flows, one must be able to efficiently refine the computational mesh around areas containing high gradients of pressure, density, velocity, or other suitable flowfield variables that characterize distinct structures. Although there are techniques which utilize simple gradient-based Local Mesh Refinement (LMR) to adapt resolution selectively to capture structures in the flow, such methods lack the ability to refine structures based on the relative strengths and scales of structures that are presented in the flow. The inability to adequately define the strength and scale of structures typically results in the mesh being over-refined in regions of little consequence to the physical definition of the problem, under-refined in certain regions resulting in the loss of important features, or even the emergence of false features due to perturbations in the flowfield caused by unnecessary mesh refinement. On the other hand, significant user judgment is required to develop a "good enough" mesh for a given flow problem, so that important structures in the flowfield can be resolved. In order to overcome this problem, multiresolution techniques based on the wavelet transform are explored for feature identification and refinement. Properties and current uses of these functional transforms in fluid flow computations will be briefly discussed. A Multiresolution Transform (MRT) scheme is chosen for identifying coherent structures because of its ability to capture the scale and relative intensity of a structure, and its easy application on non-uniform meshes. The procedure used for implementation of the MRT on an octree/quadtree LMR mesh is discussed in detail, and techniques used for the identification and capture of jump discontinuities and scale information are also presented. High speed compressible flow simulations are presented for a number of cases using the described MRT LMR scheme. MRT based mesh refinement performance is analyzed and further suggestions are made for refinement parameters based on resulting refinement. The key contribution of this thesis is the identification of methods that lead to a robust, general (i.e. not requiring user-defined parameters) methodology to identify structures in compressible flows (shocks, slip lines, vertical patterns) and to direct refinement to adequately refine these structures. The ENO-MRT LMR scheme is shown to be a robust, automatic, and relatively inexpensive way of gaining accurate refinement of the major features contained in the flowfield.

CONSELHO NACIONAL DE DESENVOLVIMENTO CIENTÍFICO E TECNOLÓGICO
Este trabalho apresenta um método para extração de malhas poligonais adaptativas em multi-resolução, a partir de objetos volumétricos. As principais aplicações da extração de malhas estão ligadas à área médica, dinâmica de fluidos, geociências, meteorologia, dentre outras. Nestas áreas os dados podem ser representados como objetos volumétricos. Nos dados volumétricos as informações estão representadas implicitamente, o que dificulta o processamento direto dos objetos que se encontram representados dentro do volume. A extração da malha visa obter uma representação explícita dos objetos, de modo a viabilizar o processamento dos mesmos. O método apresentado na tese procura extrair a malha a partir de processos de Simplicação e Refinamento. Durante a simplificação é extraída uma representação super amostrada do objeto (na mesma resolução do volume inicial), a qual é simplificada de modo a se obter uma malha base ou malha grossa, em baixa resolução, porém contendo a topologia correta do objeto. A etapa de refinamento utiliza a transformada de distâ ncia para obter uma representação da malha em multi-resolução, ou seja, a cada instante é obtida uma malha de maior resolução que vai se adaptando progressivamente à geometria do objeto. A malha final apresenta uma série de propriedades importantes, como boa razão de aspecto dos triângulos, converge para a superfície do objeto, pode ser aplicada tanto a objetos com borda quanto a objetos sem borda, pode ser aplicada tanto a superfície conexas quanto a não conexas, dentre outras.
This work presents a method for extracting multiresolution adaptive polygonal meshes, from volumetric objects. Main aplications of this work are related to medical area, fluid dynamics, geosciences, metheorology and others. In these areas data may be represented as volumetric objects. Volumetric datasets are implicit representations of objects, so it s very dificult to apply directly any process to these objects. Mesh extraction obtains an explicit representation of the objetc, such that it s easier to process directly the objects. The presented method extracts the mesh from two main processes: Simplification and Refinement. The simplification step extracts a supersampled representation of the object (in the same volume resolution), and simplifies it in such a way to obtain a base mesh (or coarse mesh), in a low resolution, but containing the correct topology of the object. Refinement step uses the distance transform to obtain a multiresolution representation of the mesh, it means that at each instant it s obtained an adaptive higher resolution mesh. The final mesh presents a set of important properties, like good triangle aspect ratio, convergency to the object surface, may be applied as to objects with boundary and as to objects with multiple connected components, among others properties.

With the widespread use of deformation simulations in medical applications, the realism of the force feedback has become an important issue. In order to reach real-time performance with sufficient realism the approach of adaptivity, solution of different parts of the system with different resolutions and refresh rates, has been commonly deployed. The change in accuracy resulting from the use of adaptivity, however, has been been paid scant attention in the deformation simulation field. Presentation of error metrics is rare, while more focus is given to the real-time stability. We propose an abstract pipeline to perform error analysis for different types of deformation techniques which can consider different simulation parameters. A case study is also performed using the pipeline, and the various uses of the error estimation are discussed.

Ce travail vise au développement de nouvelles méthodes numériques adaptatives pour la simulation numérique de phénomènes physiques multi-échelles en temps et en espace. Nous nous concentrons sur les écoulements réactifs à faible nombre de Mach, caractéristiques d'un grand nombre de configurations industrielles telles que la convection naturelle, la dynamique de fronts de flamme ou encore les décharges plasmas. La raideur associée à ce type de problèmes, que ce soit via le terme source chimique qui présente un large spectre d'échelles de temps caractéristiques ou encore via la présence de forts gradients très localisés associés aux fronts de réaction, génère des difficultés numériques considérables. Il est donc nécessaire de concevoir des méthodes sur mesure pour traiter la raideur de telles applications, afin d'obtenir des résultats d'une grande précision avec un coût calcul raisonnable. Dans ce cadre général, nous introduisons de nouvelles méthodes numériques pour la résolution des équations de Navier-Stokes incompressibles, une étape importante dans la réalisation d'un solveur hydrodynamique pour les écoulements à faible nombre de Mach. Nous construisons un solveur volumes finis avec adaptation de maillage par l'analyse de multirésolution, qui permet un contrôle a priori des erreurs générées par l'adaptation de maillage. Pour ce faire, nous développons un nouveau schéma de volumes finis collocalisé, avec un traitement original des modes de pression et de vitesse parasites qui n'affecte pas la précision de la discrétisation spatiale. Cette dernière est couplée à un nouveau schéma de Runge-Kutta additif d'ordre 3 pour les écoulements incompressibles, qui présente des propriétés de stabilité adaptées à la raideur des équations différentielles algébriques semi-explicites d'index 2. L'ensemble de cette stratégie est implémentée dans le code de calcul scientifique mrpy. Ce dernier est écrit en Python, et repose sur la librairie PETSc, écrite en C, pour le traitement des opérations d'algèbre linéaire. Nous évaluons l'efficacité algorithmique de cette stratégie par la simulation numérique d'un transport de scalaire passif dans un écoulement incompressible sur maillage adaptatif. Ce travail présente donc un nouveau solveur hydrodynamique d'ordre élevé pour les écoulements incompressibles, avec adaptation de maillage par multirésolution et contrôle d'erreur, qui peut être étendu aux écoulements à faible nombre de Mach
We address the development of new numerical methods for the efficient resolution of stiff Partial Differential Equations modelling multi-scale time/space physical phenomena. We are more specifically interested in low Mach reacting flow processes, that cover various real-world applications such as flame dynamics at low gas velocity, buoyant jet flows or plasma/flow interactions. It is well-known that the numerical simulation of these problems is a highly difficult task, due to the large spectrum of spatial and time scales caused by the presence of nonlinear The adaptive spatial discretization is coupled to a new 3rd-order additive Runge-Kutta method for the incompressible Navier-Stokes equations, combining a 3rd-order, A-stable, stiffly accurate, 4-stage ESDIRK method for the algebraic linear part of these equations, and a 4th-order explicit Runge-Kutta scheme for the nonlinear convective part. This numerical strategy is implemented from scratch in the in-house numerical code mrpy. This software is written in Python, and relies on the PETSc library, written in C, for linear algebra operations. We assess the capabilities of this mechanisms taking place into dynamic fronts. In this general context, this work introduces dedicated numerical tools for the resolution of the incompressible Navier-Stokes equations, an important first step when designing an hydrodynamic solver for low Mach flows. We build a space adaptive numerical scheme to solve incompressible flows in a finite-volume context, that relies on multiresolution analysis with error control. To this end, we introduce a new collocated finite-volume method on adaptive rectangular grids, with an original treatment of the spurious pressure and velocity modes that does not alter the precision of the discretization technique. new hydrodynamic solver in terms of speed and efficiency, in the context of scalar transport on adaptive grids. Hence, this study presents a new high-order hydrodynamics solver for incompressible flows, with grid adaptation by multiresolution, that can be extended to the more general low-Mach flow configuration

The main focus of this thesis is on fast numerical methods, where adaptivity is an important mechanism to lowering the methods' complexity. The application of the methods are in the areas of wireless communication, antenna design, radar signature computation, noise prediction, medical ultrasonography, crystal growth, flame propagation, wave propagation, seismology, geometrical optics and image processing.   We first consider high frequency wave propagation problems with a variable speed function in one dimension, modeled by the Helmholtz equation. One significant difficulty of standard numerical methods for such problems is that the wave length is very short compared to the computational domain and many discretization points are needed to resolve the solution. The computational cost, thus grows algebraically with the frequency w. For scattering problems with impenetrable scatterer in homogeneous media, new methods have recently been derived with a provably lower cost in terms of w. In this thesis, we suggest and analyze a fast numerical method for the one dimensional Helmholtz equation with variable speed function (variable media) that is based on wave-splitting. The Helmholtz equation is split into two one-way wave equations which are then solved iteratively for a given tolerance. We show rigorously that the algorithm is convergent, and that the computational cost depends only weakly on the frequency for fixed accuracy.  We next consider interface tracking problems where the interface moves by a velocity field that does not depend on the interface itself. We derive fast adaptive  numerical methods for such problems. Adaptivity makes methods robust in the sense that they can handle a large class of problems, including problems with expanding interface and problems where the interface has corners. They are based on a multiresolution representation of the interface, i.e. the interface is represented hierarchically by wavelet vectors corresponding to increasingly detailed meshes. The complexity of standard numerical methods for interface tracking, where the interface is described by marker points, is O(N/dt), where N is the number of marker points on the interface and dt is the time step. The methods that we develop in this thesis have O(dt^(-1)log N) computational cost for the same order of accuracy in dt. In the adaptive version, the cost is O(tol^(-1/p)log N), where tol is some given tolerance and p is the order of the numerical method for ordinary differential equations that is used for time advection of the interface.   Finally, we consider time-dependent Hamilton-Jacobi equations with convex Hamiltonians. We suggest a numerical method that is computationally efficient and accurate. It is based on a reformulation of the equation as a front tracking problem, which is solved with the fast interface tracking methods together with a post-processing step.  The complexity of standard numerical methods for such problems is O(dt^(-(d+1))) in d dimensions, where dt is the time step. The complexity of our method is reduced to O(dt^(-d)|log dt|) or even to O(dt^(-d)).

QC 20121116

In this thesis, we introduce a new approach to wavefront phase reconstruction in Adaptive Optics (AO) from the low-resolution gradient measurements provided by a wavefront sensor, using a non-linear approach derived from the Microcanonical Multiscale Formalism (MMF). MMF comes from established concepts in statistical physics, it is naturally suited to the study of multiscale properties of complex natural signals, mainly due to the precise numerical estimate of geometrically localized critical exponents, called the singularity exponents. These exponents quantify the degree of predictability, locally, at each point of the signal domain, and they provide information on the dynamics of the associated system. We show that multiresolution analysis carried out on the singularity exponents of a high-resolution turbulent phase (obtained by model or from data) allows a propagation along the scales of the gradients in low-resolution (obtained from the wavefront sensor), to a higher resolution. We compare our results with those obtained by linear approaches, which allows us to offer an innovative approach to wavefront phase reconstruction in Adaptive Optics.

On considère des méthodes de Galerkin stochastiques pour des systèmes hyperboliques faisant intervenir des données en entrée incertaines de lois de distribution connues paramétrées par des variables aléatoires. On s'intéresse à des problèmes où un choc apparaît presque sûrement en temps fini. Dans ce cas, la solution peut développer des discontinuités dans les domaines spatial et stochastique. On utilise un schéma de Volumes Finis pour la discrétisation spatiale et une projection de Galerkin basée sur une approximation polynomiale par morceaux pour la discrétisation stochastique. On propose un solveur de type Roe avec correcteur entropique pour le système de Galerkin, utilisant une technique originale pour approcher la valeur absolue de la matrice de Roe et une adaptation du correcteur entropique de Dubois et Mehlmann. La méthode proposée reste coûteuse car une discrétisation stochastique très fine est nécessaire pour représenter la solution au voisinage des discontinuités. Il est donc nécessaire de faire appel à des stratégies adaptatives. Comme les discontinuités sont localisées en espace et évoluent en temps, on propose des représentations stochastiques dépendant de l'espace et du temps. On formule cette méthodologie dans un contexte multi-résolution basé sur le concept d'arbres binaires pour décrire la discrétisation stochastique. Les étapes d'enrichissement et d'élagage adaptatifs sont réalisées en utilisant des critères d'analyse multi-résolution. Dans le cas multidimensionnel, une anisotropie de la procédure adaptative est proposée. La méthodologie est évaluée sur le système des équations d'Euler dans un tube à choc et sur l'équation de Burgers en une et deux dimensions stochastiques

Cette thèse porte sur l'extraction de représentations géométriques à partir d'images tridimensionnelles. Ces représentations ont d'importantes applications dans les domaines médicaux (examen non invasif et simulation, détection de pathologies, chirurgie assistée par ordinateur, fabrication de prothèses, etc) et biologiques (analyse des structures microscopiques et de leur fonctionnement). Deux approches peuvent être suivies~ : - Les méthodes de reconstruction discrète fournissent rapidement une représentation géométrique de ces données, mais laissent de côté l'aspect segmentation de l'image en ses constituants. Parmi ces méthodes, les plus utilisées sont~ : le marching-cubes, qui construit une surface triangulée, et le suivi de surface, qui délimite une surface digitale. En introduisant des considérations de topologie digitale, nous montrons l'équivalence de ces deux représentations. De cette manière, leurs propriétés respectives peuvent être combinées efficacement. - Les méthodes basées sur les modèles déformables réunissent les opérations de segmentation et de reconstruction en un seul processus~ : le modèle recherche les constituants de l'image en se déformant sous l'action de contraintes externes, issues de l'image, et internes, dérivées de sa structure géométrique. Les modèles existants sont souvent limités à l'extraction de formes simples. Nous proposons un modèle déformable générique, basé sur une triangulation de surface, et capable d'adapter automatiquement la topologie de sa maille aux déformations imposées à sa géométrie. Cette capacité permet au modèle d'appréhender les formes arbitrairement complexes de l'image et de les extraire de l'esquisse aux détails par une approche multi-résolution. Enfin, nous présentons l'application de ce modèle à des données biomédicales de modalités variées. Les résultats sont comparés à ceux obtenus par reconstruction discrète, puis combinés afin de tirer parti des avantages spécifiques des deux approches.

Nous avons développé un système intégré permettant l'acquisition simultanée de la forme 3D ainsi que de la réflectance des surfaces des objets scannés. Nous appelons ce système un scanner 3D multispectral du fait qu'il combine, dans un couple stéréoscopique, une caméra multispectrale et un système projecteur de lumière structurée. Nous voyons plusieurs possibilités d'application pour un tel système mais nous mettons en avant des applications dans le domaine de l'archivage et la diffusion numériques des objets du patrimoine. Dans le manuscrit, nous présentons d'abord ce système ainsi que tous les calibrages et traitements nécessaires à sa mise en oeuvre. Ensuite, une fois que le système est fonctionnel, les données qui en sont générées sont riches d'informations, hétérogènes (maillage + réflectances, etc.) et surtout occupent beaucoup de place. Ce fait rend problématiques le stockage et la transmission, notamment pour des applications en ligne de type musée virtuel. Pour cette raison, nous étudions les différentes possibilités de représentation et de codage des données acquises par ce système pour en adopter la plus pertinente. Puis nous examinons les stratégies les plus appropriées à la compression de telles données, sans toutefois perdre la généralité sur d'autres données (type satellitaire). Nous réalisons un benchmark des stratégies de compression en proposant un cadre d'évaluation et des améliorations sur les stratégies classiques existantes. Cette première étude nous permettra de proposer une approche adaptative qui se révélera plus efficace pour la compression et notamment dans le cadre de la stratégie que nous appelons Full-3D.

碩士
國立中央大學
資訊工程研究所
93
For a large terrain model, it is generally impractical to load the whole terrain model into the main memory. We here partition a large terrain model into blocks and then dynamically load the necessary terrain blocks into the memory for rendering. To improve the visualization performance without reducing the visual quality, every terrain block has its own variant resolution and the variable resolution is automatically adapted based on the screen-space error. That is, every terrain block possesses the view-dependent multiresolution modeling function. In such a case, cracks between two adjacent blocks with different resolutions are always generated. We here propose a re-triangulated boundary method to overcome the problem. During fly-through applications, the viewpoint may move toward or away from the terrain model, then the view-scope is changed. To further improve the visualization performance, the active block number is automatically adapted according to the height of the viewpoint. In a one-processor computer, the fly through always suspends during loading necessary terrain blocks. To reduce the phenomenon, we predict the flight path to pre-load the necessary blocks in advance based on the last flying directions.

碩士
國立成功大學
電機工程學系碩博士班
94
Motion-adaptive de-interlacing algorithm selects from inter-field and intra-field interpolations according to motion. Correct determination of motion information is essential for this purpose. Fine textures, having high local pixel variation, tend to cause false detection of motion. This thesis proposed a texture detection mechanism utilizing multiresolution technique to improve the correctness of motion detection. A modification of 5-tap ELA algorithm is also proposed using the texture information to yield better spatial interpolation results, while the occlusion problem is improved by the inter-field interpolation selection. Finally, a memory-efficient recursive 3-field motion detection algorithm is developed to achieve close-to 4-field detection result but little overhead comparing with traditional 3-field detection. The proposed algorithm provides higher PSNR and better perceptual visual quality than other deinterlacing algorithms as shown by the experimental results. 　　Block-based process in hardware design allows the two main datapath cores to collaborate at 6x6 block level. The DWT core performs (inverse) discrete wavelet transform to 9x9 input blocks and produces 6x7 (one row overlap) texture indices for the corresponding block to be deinterlaced by the DEI core. Bus bandwidth and DRAM access is kept minimal by direct passing of texture indices between the two datapath cores. The hardware operates at 54 MHz and is synthesized with Artisan UMC .18 um cell library.

碩士
國立中央大學
資訊工程研究所
90
The real-time rendering of large terrain models on a low-level personal computer is a challenge for flight simulation. In this paper, a multiresolution terrain modeling based on the quadric error metrics (QEM) with dynamic loading technique is proposed to achieve the need of flight simulation. The multiresolution modeling technique can rapidly produce high-quality approximation of the original model. For a large terrain model, it is generally impractical to load the whole terrain model into memory. Thus, we partition a large terrain model into blocks and then dynamically load the necessary terrain blocks into the memory for rendering. Generally, it will cause cracks between two adjacent terrain models with different resolutions. We discuss several methods to overcome the problem. For the application of flight simulation, we build a control interface to simulate the fighter’s action; let users feel as they are aviating the fighter over the terrain.