Spelling suggestions: "subject:"anonlinear problems"" "subject:"forlinear problems""
1 |
Non-stationary Poiseuille type solutions for the second grade fluid flow problem in cylindrical domains / Antrojo laipsnio skysčių tekėjimo uždavinio nestacionarūs Puazeilio tipo sprendiniai cilindrinėse srityseKlovienė, Neringa 24 January 2013 (has links)
In the dissertation one of the Rivlin-Erikson differential type fluids model – the second grade fluids flow problem is considered. The problem is studied in three different unbounded domains: • the two-dimensional channel, • the three-dimensional axially symmetric pipe, • the three-dimensional pipe with an arbitrary cross section. For the two-dimensional channel and the three-dimensional axially symmetric pipe we assume that the initial data and the external force have only the last component and are independent of the coordinate x_n: u_0(x,t)=(0, …, u_{n0}(x’,t)), f(x,t)=(0, …, f_n(x’,t)). We look for an unidirectional (having just the last component) solution u(x,t) =(0, …, u_n(x’,t)), which satisfies the flux condition. Such solution we call Poiseuille type solution. In the first two cases the existence of a unique unidirectional Poiseuille type solution is proved and the relation between the flux of the velocity field and the pressure drop (the gradient of the pressure) is found. The analogous results were obtained for the time periodic problem in the two-dimensional channel. It is shown that in the three-dimensional pipe with an arbitrary cross section the unidirectional solution does not exists even if data are unidirectional. However, for sufficiently small data in this case exists a unique solution having all three components u(x’,t)=( u_1, u_2, u_3). To analyze the problem we use Galerkin method with the special bases. / Disertacijoje nagrinėjamas vienas iš Rivlin-Eriksono diferencialinio tipo skysčių matematinių modelių – antrojo laipsnio skysčių tekėjimo uždavinys. Problema analizuojama su papildomai užduota srauto sąlyga trijose skirtingose srityse: • begalinėje juostoje, • begaliniame sukimosi cilindre, • begaliniame vamzdyje su bet kokiu skerspjūviu. Tariama, kad pradinio greičio ir išorės jėgų vektoriai nepriklauso nuo paskutinės koordinatės ir yra išreikšti pavidalu u_0(x,t)=(0, …, u_{n0}(x’,t)), f(x,t)=(0, …, f_n(x’,t)). Ieškoma antrojo laipsnio skysčių tekėjimo uždavinio Puazeilio tipo u(x,t) =(0, …, u_n(x’,t)) sprendinio. Begalinėje dvimatėje juostoje ir begaliniame trimačiame sukimosi cilindre įrodytas kryptinio Puazeilio tipo sprendinio egzistavimas ir rastas sąryšis tarp srauto ir slėgio gradiento. Analogiški rezultatai gauti pradiniam ir kraštiniam antrojo laipsnio skysčių tekėjimo uždaviniui periodinėje pagal laiką begalinėje dvimatėje juostoje. Darbe parodyta, kad begaliniame trimačiame vamzdyje, su bet kokiu skerspjūviu, kryptinis (priklausantis tik nuo paskutinės komponentės) Puazeilio tipo sprendinys neegzistuoja net jei pradiniai duomenys yra kryptiniai. Nagrinėjamas bendresnis atvejis, kai Puazeilio tipo sprendinys priklauso nuo visų trijų komponenčių u(x’,t)=( u_1, u_2, u_3). Disertacijoje įrodyta, kad esant mažiems pradiniams duomenims egzistuoja vienintelis uždavinio sprendinys. Sprendžiant buvo naudojamas Galiorkino aproksimacijų metodas ir specialios bazės.
|
2 |
ε-SUPERPOSITION AND TRUNCATION DIMENSIONS IN AVERAGE AND PROBABILISTIC SETTINGS FOR ∞-VARIATE LINEAR PROBLEMSDingess, Jonathan M. 01 January 2019 (has links)
This thesis is a representation of my contribution to the paper of the same name I co-author with Dr. Wasilkowski. It deals with linear problems defined on γ-weighted normed spaces of functions with infinitely many variables. In particular, I describe methods and discuss results for ε-truncation and ε-superposition methods. I show through these results that the ε-truncation and ε-superposition dimensions are small under modest error demand ε. These positive results are derived for product weights and the so-called anchored decomposition.
|
3 |
Blow-up pour des problèmes paraboliques semi linéaires avec un terme source localisé / Complete blow-up for a semilinear parabolic problem with a localized non linear termSawangtong, Panumart 13 December 2010 (has links)
On étudie l'existence de blow-up et l'ensemble des points de blow-up pour une équation de type chaleur dégénérée ou non avec un terme source uniforme fonction nonlinéaire de la température instantanée en un point fixé du domaine. L'étude est conduite par les méthodes d'analyse classique (fonctions de Green, développements en fonctions propres, principe du maximum) ou fonctionnelle (semi-groupes d'opérateurs linéaires). / We study existence of blow-up and blow-up sets for a (degenerate or not) heat-like equation with a uniform source term non linear function of the instantaneous temperature at a given point of the domain. The techniques are relevant from either classical analysis (Green functions, eigenfunction expansions, maximum principle) or functional analysis (semi-groups of linear operators).
|
4 |
Tauextrapolation - theoretische Grundlagen, numerische Experimente und Anwendungen auf die Navier-Stokes-GleichungenBernert, K. 30 October 1998 (has links)
The paper deals with tau-extrapolation - a modification of the
multigrid method, which leads to solutions with an improved con-
vergence order. The number of numerical operations depends
linearly on the problem size and is not much higher than for a
multigrid method without this modification. The paper starts
with a short mathematical foundation of the tau-extrapolation.
Then follows a careful tuning of some multigrid components
necessary for a successful application of tau-extrapolation. The
next part of the paper presents numerical illustrations to the
theoretical investigations for one- dimensional test problems.
Finally some experience with the use of tau-extrapolation for the
Navier-Stokes equations is given.
|
5 |
Métodos iterativos fraccionarios para la resolución de ecuaciones y sistemas no lineales: diseño, análisis y estabilidadCandelario Villalona, Giro Guillermo 16 June 2023 (has links)
[ES] El cálculo fraccionario es una extensión del cálculo clásico, donde el orden de las derivadas o integrales es un número real. Hoy en día, el cálculo fraccionario tiene numerosas aplicaciones en ciencias e ingeniería. La principal razón es el mayor grado de libertad de las herramientas del cálculo fraccionario en comparación con las herramientas del cálculo clásico. Muchos problemas reales se modelan por medio de ecuaciones diferenciales fraccionarias no lineales cuyo sistema de ecuaciones es no lineal, y por tanto, es conveniente que se adapten procedimientos iterativos para resolver problemas no lineales con el uso de derivadas fraccionarias, y observar cuál es la consecuencia en la convergencia de dicho método.
En esta Tesis Doctoral diseñamos nuevos procedimientos iterativos con derivadas fraccionarias (o su aproximación) que al menos igualen a los métodos clásicos en términos de orden de convergencia, mediante la introducción de las derivadas fraccionarias de Riemann-Liouville, de Caputo y conformable (o sus aproximaciones). También, proponemos estudiar la estabilidad de estos esquemas con el uso de planos de convergencia, y planos dinámicos en algunos casos. Finalmente, pretendemos diseñar una técnica que nos permita obtener la versión fraccionaria conformable (o versión con derivada conformable o su aproximación) de cualquier procedimiento iterativo clásico para problemas no lineales.
En el Capítulo 2 se exponen los conceptos previos que serán necesarios para el desarrollo de los siguientes capítulos: Se presentan los conceptos básicos relacionados con métodos de punto fijo, se muestran los esquemas clásicos que trataremos en esta memoria, y finalmente se introducen las herramientas del cálculo fraccionario que serán necesarias para el diseño de procedimientos iterativos fraccionarios.
En el Capítulo 3 se diseñan métodos fraccionarios (o esquemas con derivadas fraccionarias) de tipo Newton-Raphson escalares con las derivadas de Caputo, de Riemann Liouville y la conformable. También diseñamos esquemas fraccionarios de Newton-Raphson escalares de mayor orden. Finalmente, realizamos el análisis de convergencia de dichos procedimientos y estudiamos su estabilidad.
En el Capítulo 4 se diseña la versión vectorial del método de Newton-Raphson conformable visto en el Capítulo 3. Antes, es necesario definir nuevos conceptos y establecer nuevos resultados que serán necesarios para el dersarrollo de este esquema. Finalmente, realizamos el análisis de convergencia y estudiamos su estabilidad.
En el Capítulo 5 se diseñan procedimientos fraccionarios de tipo Traub escalares con derivadas de Caputo y de Riemann-Liouville. También se diseña una técnica general para obtener la versión fraccionaria conformable escalar de cualquier método clásico, y se usa esta técnica para diseñar algunos esquemas conformables multipunto escalares: de tipos Traub, Chun-Kim, Ostrowski y Chun. Por último, se realiza el análisis de convergencia y se estudia la estabilidad de tales procedimientos.
En el Capítulo 6 se diseñan métodos fraccionarios libres de derivadas escalares de tipos Steffensen y Secante (el cual tiene memoria), donde es necesario la aproximación de derivadas conformables. Aquí se usa la técnica general propuesta en el Capítulo 5 para obtener la versión conformable de cada esquema. Finalmente, realizamos el análisis de convergencia y se estudia la estabilidad de dichos procedimientos. En el Capítulo 7 se presentan las conclusiones y líneas futuras de investigación. / [CA] El càlcul fraccionari és una extensió del càlcul clàssic, on l'ordre de les derivades o integrals és un nombre real. Hui dia, el càlcul fraccionari té nombroses aplicacions en ciències i enginyeria. La principal raó és el major grau de llibertat de les eines del càlcul fraccionari en comparació amb les eines del càlcul clàssic. Molts problemes reals es modelen per mitjà d'equacions diferencials fraccionàries no lineals el sistema d'equacions de les quals és no lineal, i per tant, és convenient que s'adapten procediments iteratius per a resoldre problemes no lineals amb l'ús de derivades fraccionàries, i observar quina és la conseqüència en la convergència d'aquest mètode.
En aquesta Tesi Doctoral dissenyem nous procediments iteratius amb derivades fraccionàries (o la seua aproximació) que almenys igualen als mètodes clàssics en termes d'ordre de convergència, mitjançant la introducció de les derivades fraccionàries de Riemann-Liouville, de Caputo i conformable (o les seues aproximacions). També, proposem estudiar l'estabilitat d'aquests esquemes amb l'ús de plans de convergència, i plans dinàmics en alguns casos. Finalment, pretenem dissenyar una tècnica que ens permeta obtindre la versió fraccionària conformable (o versió amb derivada conformable o la seua aproximació) de qualsevol procediment iteratiu clàssic per a problemes no lineals.
En el Capítol 2 s'exposen els conceptes previs que seran necessaris per al desenvolupament dels següents capítols: Es presenten els conceptes bàsics relacionats amb mètodes de punt fix, es mostren els esquemes clàssics que tractarem en aquesta memòria, i finalment s'introdueixen les eines del càlcul fraccionari que seran necessàries per al disseny de procediments iteratius fraccionaris.
En el Capítol 3 es dissenyen mètodes fraccionaris (o esquemes amb derivades fraccionàries) de tipus Newton-Raphson escalars amb les derivades de Caputo, de Riemann Liouville i la conformable. També dissenyem esquemes fraccionaris de Newton-Raphson escalars de major ordre. Finalment, realitzem l'anàlisi de convergència d'aquests procediments i estudiem la seua estabilitat.
En el Capítol 4 es dissenya la versió vectorial del mètode de Newton-Raphson conformable vist en el Capítol 3. Abans, és necessari definir nous conceptes i establir nous resultats que seran necessaris per al dersarrollo d'aquest esquema. Finalment, realitzem l'anàlisi de convergència i estudiem la seua estabilitat.
En el Capítol 5 es dissenyen procediments fraccionaris de tipus Traub escalars amb derivades de Caputo i de Riemann-Liouville. També es dissenya una tècnica general per a obtindre la versió fraccionària conformable escalar de qualsevol mètode clàssic, i s'usa aquesta tècnica per a dissenyar alguns esquemes conformables multipunt escalars: de tipus Traub, Chun-Kim, Ostrowski i Chun. Finalment, es realitza l'anàlisi de convergència i s'estudia l'estabilitat de tals procediments.
En el Capítol 6 es dissenyen mètodes fraccionaris lliures de derivades escalars de tipus Steffensen i Assecant (el qual té memòria), on és necessari l'aproximació de derivades conformables. Ací s'usa la tècnica general proposta en el Capítol 5 per a obtindre la versió conformable de cada esquema. Finalment, realitzem l'anàlisi de convergència i s'estudia l'estabilitat d'aquests procediments. En el Capítol 7 es presenten les conclusions i línies futures d'investigació. / [EN] Fractional calculus is an extension of classical calculus, where the order of the derivatives or integrals is a real number. Today, fractional calculus has numerous applications in science and engineering. The main reason is the higher degree of freedom of the fractional calculus tools compared to the classical calculus tools. Many real problems are modeled by means of nonlinear fractional differential equations whose system of equations is nonlinear, and therefore it is convenient that iterative procedures are adapted to solve nonlinear problems with the use of fractional derivatives, and observe what the consequence is in the convergence of said method.
In this Doctoral Thesis we design new iterative procedures with fractional derivatives (or their approximation) that are at least equal to the classical methods in terms of convergence order, by introducing the Riemann-Liouville, Caputo and conformable fractional derivatives (or their approximations). Also, we propose to study the stability of these schemes with the use of convergence planes, and dynamic planes in some cases. Finally, we intend to design a technique that allows us to obtain the conformable fractional version (or version with conformable derivative or its approximation) of any classical iterative procedure for nonlinear problems.
In Chapter 2 the previous concepts that will be necessary for the development of the following chapters are exposed: The basic concepts related to fixed point methods are presented, the classic schemes that we will deal with in this memory are shown, and finally the tools of the fractional calculus that will be necessary for the design of fractional iterative procedures.
In Chapter 3, scalar Newton-Raphson type fractional methods (or schemes with fractional derivatives) are designed with the Caputo, Riemann Liouville and conformable derivatives. We also design higher order scalar Newton-Raphson fractional schemes. Finally, we perform the convergence analysis of these procedures and study their stability.
In Chapter 4, the vector version of the conformable Newton-Raphson method seen in Chapter 3 is designed. Before, it is necessary to define new concepts and establish new results that will be necessary for the development of this scheme. Finally, we perform the convergence analysis and study its stability.
In Chapter 5, fractional procedures of the scalar Traub type with derivatives of Caputo and Riemann-Liouville are designed. A general technique is also designed to obtain the scalar conformable fractional version of any classical method, and this technique is used to design some scalar multipoint conformable schemes: of Traub, Chun-Kim, Ostrowski and Chun types. Finally, the convergence analysis is carried out and the stability of such procedures is studied.
In Chapter 6 free fractional methods of scalar derivatives of Steffensen and Secant types (which has memory) are designed, where the conformable derivatives approximation is necessary. Here we use the general technique proposed in Chapter 5 to obtain the conformable version of each scheme. Finally, we carry out the convergence analysis and the stability of these procedures is studied. In Chapter 7 the conclusions and future lines of research are presented. / Candelario Villalona, GG. (2023). Métodos iterativos fraccionarios para la resolución de ecuaciones y sistemas no lineales: diseño, análisis y estabilidad [Tesis doctoral]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/194270
|
6 |
Nonlinear approaches for phase retrieval in the Fresnel region for hard X-ray imagingIon, Valentina 26 September 2013 (has links) (PDF)
The development of highly coherent X-ray sources offers new possibilities to image biological structures at different scales exploiting the refraction of X-rays. The coherence properties of the third-generation synchrotron radiation sources enables efficient implementations of phase contrast techniques. One of the first measurements of the intensity variations due to phase contrast has been reported in 1995 at the European Synchrotron Radiation Facility (ESRF). Phase imaging coupled to tomography acquisition allows threedimensional imaging with an increased sensitivity compared to absorption CT. This technique is particularly attractive to image samples with low absorption constituents. Phase contrast has many applications, ranging from material science, paleontology, bone research to medicine and biology. Several methods to achieve X-ray phase contrast have been proposed during the last years. In propagation based phase contrast, the measurements are made at different sample-to-detector distances. While the intensity data can be acquired and recorded, the phase information of the signal has to be "retrieved" from the modulus data only. Phase retrieval is thus an illposed nonlinear problem and regularization techniques including a priori knowledge are necessary to obtain stable solutions. Several phase recovery methods have been developed in recent years. These approaches generally formulate the phase retrieval problem as a linear one. Nonlinear treatments have not been much investigated. The main purpose of this work was to propose and evaluate new algorithms, in particularly taking into account the nonlinearity of the direct problem. In the first part of this work, we present a Landweber type nonlinear iterative scheme to solve the propagation based phase retrieval problem. This approach uses the analytic expression of the Fréchet derivative of the phase-intensity relationship and of its adjoint, which are presented in detail. We also study the effect of projection operators on the convergence properties of the method. In the second part of this thesis, we investigate the resolution of the linear inverse problem with an iterative thresholding algorithm in wavelet coordinates. In the following, the two former algorithms are combined and compared with another nonlinear approach based on sparsity regularization and a fixed point algorithm. The performance of theses algorithms are evaluated on simulated data for different noise levels. Finally the algorithms were adapted to process real data sets obtained in phase CT at the ESRF at Grenoble.
|
7 |
Aprendizado de máquina baseado em separabilidade linear em sistema de classificação híbrido-nebuloso aplicado a problemas multiclasseTuma, Carlos Cesar Mansur 29 June 2009 (has links)
Made available in DSpace on 2016-06-02T19:05:36Z (GMT). No. of bitstreams: 1
2598.pdf: 3349204 bytes, checksum: 01649491fd1f03aa5a11b9191727f88b (MD5)
Previous issue date: 2009-06-29 / Financiadora de Estudos e Projetos / This master thesis describes an intelligent classifier system applied to multiclass non-linearly separable problems called Slicer. The system adopts a low computacional cost supervised learning strategy (evaluated as ) based on linear separability. During the learning period the system determines a set of hyperplanes associated to oneclass regions (sub-spaces). In classification tasks the classifier system uses the hyperplanes as a set of if-then-else rules to infer the class of the input attribute vector (non classified object). Among other characteristics, the intelligent classifier system is able to: deal with missing attribute values examples; reject noise examples during learning; adjust hyperplane parameters to improve the definition of the one-class regions; and eliminate redundant rules. The fuzzy theory is considered to design a hybrid version with features such as approximate reasoning and parallel inference computation. Different classification methods and benchmarks are considered for evaluation. The classifier system Slicer reaches acceptable results in terms of accuracy, justifying future investigation effort. / Este trabalho de mestrado descreve um sistema classificador inteligente aplicado a problemas multiclasse não-linearmente separáveis chamado Slicer. O sistema adota uma estratégia de aprendizado supervisionado de baixo custo computacional (avaliado em ) baseado em separabilidade linear. Durante o período de aprendizagem o sistema determina um conjunto de hiperplanos associados a regiões de classe única (subespaços). Nas tarefas de classificação o sistema classificador usa os hiperplanos como um conjunto de regras se-entao-senao para inferir a classe do vetor de atributos dado como entrada (objeto a ser classificado). Entre outras caracteristicas, o sistema classificador é capaz de: tratar atributos faltantes; eliminar ruídos durante o aprendizado; ajustar os parâmetros dos hiperplanos para obter melhores regiões de classe única; e eliminar regras redundantes. A teoria nebulosa é considerada para desenvolver uma versão híbrida com características como raciocínio aproximado e simultaneidade no mecanismo de inferência. Diferentes métodos de classificação e domínios são considerados para avaliação. O sistema classificador Slicer alcança resultados aceitáveis em termos de acurácia, justificando investir em futuras investigações.
|
8 |
Méthode du gradient topologique pour la détection de contours et de structures fines en imagerie / Topological gradient method applied to the detection of edges and fine structures in imagingDrogoul, Audric 08 October 2014 (has links)
Cette thèse porte sur la méthode du gradient topologique appliquée au traitement d'images. Principalement, on s'intéresse à la détection d'objets assimilés, soit à des contours si l'intensité de l'image à travers la structure comporte un saut, soit à une structure fine (filaments et points en 2D) s'il n'y a pas de saut à travers la structure. On commence par généraliser la méthode du gradient topologique déjà utilisée en détection de contours pour des images dégradées par du bruit gaussien, à des modèles non linéaires adaptés à des images contaminées par un processus poissonnien ou du bruit de speckle et par différents types de flous. On présente également un modèle de restauration par diffusion anisotrope utilisant le gradient topologique pour un domaine fissuré. Un autre modèle basé sur une EDP elliptique linéaire utilisant un opérateur anisotrope préservant les contours est proposé. Ensuite, on présente et étudie un modèle de détection de structures fines utilisant la méthode du gradient topologique. Ce modèle repose sur l'étude de la sensibilité topologique d'une fonction coût utilisant les dérivées secondes d'une régularisation de l'image solution d'une EDP d'ordre 4 de type Kirchhoff. En particulier on explicite les gradients topologiques pour des domaines 2D fissurés ou perforés, et des domaines 3D fissurés. Plusieurs applications pour des images 2D et 3D, floutées et contaminées par du bruit gaussien, montrent la robustesse et la rapidité de la méthode. Enfin on généralise notre approche pour la détection de contours et de structures fines par l'étude de la sensibilité topologique d'une fonction coût utilisant les dérivées m−ième d'une régularisation de l'image dégradée, solution d'une EDP d'ordre 2m. / This thesis deals with the topological gradient method applied in imaging. Particularly, we are interested in object detection. Objects can be assimilated either to edges if the intensity across the structure has a jump, or to fine structures (filaments and points in 2D) if there is no jump of intensity across the structure. We generalize the topological gradient method already used in edge detection for images contaminated by Gaussian noise, to quasi-linear models adapted to Poissonian or speckled images possibly blurred. As a by-product, a restoration model based on an anisotropic diffusion using the topological gradient is presented. We also present a model based on an elliptical linear PDE using an anisotropic differential operator preserving edges. After that, we study a variational model based on the topological gradient to detect fine structures. It consists in the study of the topological sensitivity of a cost function involving second order derivatives of a regularized version of the image solution of a PDE of Kirchhoff type. We compute the topological gradients associated to perforated and cracked 2D domains and to cracked 3D domains. Many applications performed on 2D and 3D blurred and Gaussian noisy images, show the robustness and the fastness of the method. An anisotropic restoration model preserving filaments in 2D is also given. Finally, we generalize our approach by the study of the topological sensitivity of a cost function involving the m − th derivatives of a regularization of the image solution of a 2m order PDE.
|
9 |
Optimalizační problémy při (max,min.)-lineárních omezeních a některé související úlohy / Optimization Problems under (max; min) - Linear Constraint and Some Related TopicsGad, Mahmoud Attya Mohamed January 2015 (has links)
Title: Optimization Problems under (max, min)-Linear Constraints and Some Related Topics. Author: Mahmoud Gad Department/Institue: Department of Probability and Mathematical Statis- tics Supervisor of the doctoral thesis: 1. Prof. RNDr. Karel Zimmermann,DrSc 2. Prof. Dr. Assem Tharwat, Cairo University, Egypt Abstract: Problems on algebraic structures, in which pairs of operations such as (max, +) or (max, min) replace addition and multiplication of the classical linear algebra have appeared in the literature approximately since the sixties of the last century. The first publications on these algebraic structures ap- peared by Shimbel [37] who applied these ideas to communication networks, Cunninghame-Green [12, 13], Vorobjov [40] and Gidffer [18] applied these alge- braic structures to problems of machine-time scheduling. A systematic theory of such algebraic structures was published probable for the first time in [14]. In recently appeared book [4] the readers can find latest results concerning theory and algorithms for (max, +)-linear systems of equations and inequalities. Since operation max replacing addition in no more a group, but a semigroup oppera- tion, it is a substantial difference between solving systems with variables on one side and systems with variables occuring on both sides of the equations....
|
10 |
Nonlinear approaches for phase retrieval in the Fresnel region for hard X-ray imaging / Approches non linéaire en imagerie de phase par rayons X dans le domaine de FresnelIon, Valentina 26 September 2013 (has links)
Le développement de sources cohérentes de rayons X offre de nouvelles possibilités pour visualiser les structures biologiques à différentes échelles en exploitant la réfraction des rayons X. La cohérence des sources synchrotron de troisième génération permettent des implémentations efficaces des techniques de contraste de phase. Une des premières mesures des variations d’intensité dues au contraste de phase a été réalisée en 1995 à l’Installation Européenne de Rayonnement Synchrotron (ESRF). L’imagerie de phase couplée à l’acquisition tomographique permet une imagerie tridimensionnelle avec une sensibilité accrue par rapport à la tomographie standard basée sur absorption. Cette technique est particulièrement adaptée pour les échantillons faiblement absorbante ou bien présentent des faibles différences d’absorption. Le contraste de phase a ainsi une large gamme d’applications, allant de la science des matériaux, à la paléontologie, en passant par la médecine et par la biologie. Plusieurs techniques de contraste de phase aux rayons X ont été proposées au cours des dernières années. Dans la méthode de contraste de phase basée sur le phénomène de propagation l’intensité est mesurée pour différentes distances de propagation obtenues en déplaçant le détecteur. Bien que l’intensité diffractée puisse être acquise et enregistrée, les informations de phase du signal doivent être "récupérées" à partir seulement du module des données mesurées. L’estimation de la phase est donc un problème inverse non linéaire mal posé et une connaissance a priori est nécessaire pour obtenir des solutions stables. Si la plupart de méthodes d’estimation de phase reposent sur une linéarisation du problème inverse, les traitements non linéaires ont été eux très peu étudiés. Le but de ce travail était de proposer et d’évaluer des nouveaux algorithmes, prenant en particulier en compte la non linéarité du problème direct. Dans la première partie de ce travail, nous présentons un schéma de type Landweber non linéaire itératif pour résoudre le problème de la récupération de phase. Cette approche utilise l’expression analytique de la dérivée de Fréchet de la relation phase-intensité et de son adjoint. Nous étudions aussi l’effet des opérateurs de projection sur les propriétés de convergence de la méthode. Dans la deuxième partie de cette thèse, nous étudions la résolution du problème inverse linéaire avec un algorithme en coordonnées ondelettes basé sur un seuillage itératif. Par la suite, les deux algorithmes sont combinés et comparés avec une autre approche non linéaire basée sur une régularisation parcimonieuse et un algorithme de point fixe. Les performances des algorithmes sont évaluées sur des données simulées pour différents niveaux de bruit. Enfin, les algorithmes ont été adaptés pour traiter des données réelles acquises en tomographie de phase à l’ESRF à Grenoble. / The development of highly coherent X-ray sources offers new possibilities to image biological structures at different scales exploiting the refraction of X-rays. The coherence properties of the third-generation synchrotron radiation sources enables efficient implementations of phase contrast techniques. One of the first measurements of the intensity variations due to phase contrast has been reported in 1995 at the European Synchrotron Radiation Facility (ESRF). Phase imaging coupled to tomography acquisition allows threedimensional imaging with an increased sensitivity compared to absorption CT. This technique is particularly attractive to image samples with low absorption constituents. Phase contrast has many applications, ranging from material science, paleontology, bone research to medicine and biology. Several methods to achieve X-ray phase contrast have been proposed during the last years. In propagation based phase contrast, the measurements are made at different sample-to-detector distances. While the intensity data can be acquired and recorded, the phase information of the signal has to be "retrieved" from the modulus data only. Phase retrieval is thus an illposed nonlinear problem and regularization techniques including a priori knowledge are necessary to obtain stable solutions. Several phase recovery methods have been developed in recent years. These approaches generally formulate the phase retrieval problem as a linear one. Nonlinear treatments have not been much investigated. The main purpose of this work was to propose and evaluate new algorithms, in particularly taking into account the nonlinearity of the direct problem. In the first part of this work, we present a Landweber type nonlinear iterative scheme to solve the propagation based phase retrieval problem. This approach uses the analytic expression of the Fréchet derivative of the phase-intensity relationship and of its adjoint, which are presented in detail. We also study the effect of projection operators on the convergence properties of the method. In the second part of this thesis, we investigate the resolution of the linear inverse problem with an iterative thresholding algorithm in wavelet coordinates. In the following, the two former algorithms are combined and compared with another nonlinear approach based on sparsity regularization and a fixed point algorithm. The performance of theses algorithms are evaluated on simulated data for different noise levels. Finally the algorithms were adapted to process real data sets obtained in phase CT at the ESRF at Grenoble.
|
Page generated in 0.0846 seconds