The accurate detection of static object boundaries such as contours or surfaces and dynamic tunnels of moving objects via deformable models is an ongoing research topic in computer vision. Most deformable models attempt to converge towards a desired solution by minimizing the sum of internal (prior) and external (measurement) energy terms. Such an approach is elegant, but frequently mis-converges in the presence of noise or complex boundaries and typically requires careful semi-dependent parameter tuning and initialization. Furthermore, current deformable model based approaches are computationally demanding which precludes real-time use. To address these limitations, a decoupled deformable model (DDM) is developed which optimizes the two energy terms separately. Essentially, the DDM consists of a measurement update step, employing a Hidden Markov Model (HMM) and Maximum Likelihood (ML) estimator, followed by a separate prior step, which modifies the updated deformable model based on the relative strengths of the measurement uncertainty and the non-stationary prior. The non-stationary prior is generated by using a curvature guided importance sampling method to capture high curvature regions. By separating the measurement and prior steps, the algorithm is less likely to mis-converge; furthermore, the use of a non-iterative ML estimator allows the method to converge more rapidly than energy-based iterative solvers. The full functionality of the DDM is developed in three phases. First, a DDM in 2D called the decoupled active contour (DAC) is developed to accurately identify the boundary of a 2D object in the presence of noise and background clutter. To carry out this task, the DAC employs the Viterbi algorithm as a truncated ML estimator, curvature guided importance sampling as a non-stationary prior generator, and a linear Bayesian estimator to fuse the non-stationary prior with the measurements. Experimental results clearly demonstrate that the DAC is robust to noise, can capture regions of very high curvature, and exhibits limited dependence on contour initialization or parameter settings. Compared to three other published methods and across many images, the DAC is found to be faster and to offer consistently accurate boundary identification. Second, a fast decoupled active contour (FDAC) is proposed to accelerate the convergence rate and the scalability of the DAC without sacrificing the accuracy by employing computationally efficient and scalable techniques to solve the three primary steps of DAC. The computational advantage of the FDAC is demonstrated both experimentally and analytically compared to three computationally efficient methods using illustrative examples. Finally, an extension of the FDAC from 2D to 3D called a decoupled active surface (DAS) is developed to precisely identify the surface of a volumetric 3D image and the tunnel of a moving 2D object. To achieve the objectives of the DAS, the concepts of the FDAC are extended to 3D by using a specialized 3D deformable model representation scheme and a computationally and storage efficient estimation scheme. The performance of the DAS is demonstrated using several natural and synthetic volumetric images and a sequence of moving objects.

The accurate detection of static object boundaries such as contours or surfaces and dynamic tunnels of moving objects via deformable models is an ongoing research topic in computer vision. Most deformable models attempt to converge towards a desired solution by minimizing the sum of internal (prior) and external (measurement) energy terms. Such an approach is elegant, but frequently mis-converges in the presence of noise or complex boundaries and typically requires careful semi-dependent parameter tuning and initialization. Furthermore, current deformable model based approaches are computationally demanding which precludes real-time use. To address these limitations, a decoupled deformable model (DDM) is developed which optimizes the two energy terms separately. Essentially, the DDM consists of a measurement update step, employing a Hidden Markov Model (HMM) and Maximum Likelihood (ML) estimator, followed by a separate prior step, which modifies the updated deformable model based on the relative strengths of the measurement uncertainty and the non-stationary prior. The non-stationary prior is generated by using a curvature guided importance sampling method to capture high curvature regions. By separating the measurement and prior steps, the algorithm is less likely to mis-converge; furthermore, the use of a non-iterative ML estimator allows the method to converge more rapidly than energy-based iterative solvers. The full functionality of the DDM is developed in three phases. First, a DDM in 2D called the decoupled active contour (DAC) is developed to accurately identify the boundary of a 2D object in the presence of noise and background clutter. To carry out this task, the DAC employs the Viterbi algorithm as a truncated ML estimator, curvature guided importance sampling as a non-stationary prior generator, and a linear Bayesian estimator to fuse the non-stationary prior with the measurements. Experimental results clearly demonstrate that the DAC is robust to noise, can capture regions of very high curvature, and exhibits limited dependence on contour initialization or parameter settings. Compared to three other published methods and across many images, the DAC is found to be faster and to offer consistently accurate boundary identification. Second, a fast decoupled active contour (FDAC) is proposed to accelerate the convergence rate and the scalability of the DAC without sacrificing the accuracy by employing computationally efficient and scalable techniques to solve the three primary steps of DAC. The computational advantage of the FDAC is demonstrated both experimentally and analytically compared to three computationally efficient methods using illustrative examples. Finally, an extension of the FDAC from 2D to 3D called a decoupled active surface (DAS) is developed to precisely identify the surface of a volumetric 3D image and the tunnel of a moving 2D object. To achieve the objectives of the DAS, the concepts of the FDAC are extended to 3D by using a specialized 3D deformable model representation scheme and a computationally and storage efficient estimation scheme. The performance of the DAS is demonstrated using several natural and synthetic volumetric images and a sequence of moving objects.

The goal of this research project was to characterize the physical and electrochemical properties of a commercially available fuel cell-based breath alcohol sensor. Characteristics of the existing sensor were compared with state of the art power generating fuel cells with the goal of understanding the factors that limit performance, lifetime and cost effectiveness of the sensors. This will guide the development of the next generation of breath alcohol sensors. The average lifetime of the current sensor falls short of the industry standards. In particular, sensors operating in dry conditions experience more rapid loss of sensitivity and failure. Two primary causes of degradation were investigated in this study. Loss of proton conductivity as a result of membrane dehydration was shown to be reversible by rehydrating the membrane in humid conditions. Loss of electrochemically active surface area of Pt is irreversible and seems to be caused by a change in sensor morphology after long-term exposure to dry conditions. / UOIT

This thesis investigates a new approach for dealing with the adverse effects of large-scale deformations in the main reflector of large Cassegrain antennas. In this method, the incident aperture distribution is imaged onto a tertiary focal plane. This is accomplished by using an optical imaging system consisting of a lens mounted behind the Cassegrain focus of the antenna. The lens forms a real image of the product of the incident aperture distribution and the pupil function of the antenna. The pupil function describes the profile of the main reflector of the antenna. If the incident aperture distribution is a plane wave, a real image of the pupil function of the main reflector will be produced at the focal plane of the image lens. Any imperfections in the main reflector will be imaged onto the tertiary focal plane but over a smaller area as defined by the magnification of the system. In principle, an active correcting element placed into the tertiary focal plane could compensate for these errors, thus preserving the maximum efficiency of the antenna. Experimental verification of this principle was carried out in the lab using a dielectric lens 152.4mm in diameter. Phase perturbations were simulated by placing dielectric shims in the incident aperture plane. The phase of these shims in most cases was measured to within 10 degrees in the image plane. This degree of accuracy is found to be quite adequate for correcting large-scale errors in the main reflector of the antenna.

Ce travail s'inscrit dans le cadre de l'amélioration de la connaissance sur la mesure de la surface des aérosols ultrafins. En effet, l'essor des nanotechnologies peut être à l'origine de situations d'exposition professionnelle aux particules nanostructurées dispersées dans l'air, ce qui soulève une problématique nouvelle de prévention. Si à ce jour aucun des trois indicateurs (masse, surface, nombre) ne fait l'objet d'un consensus, il semble que le paramètre de surface des particules permet une bonne corrélation avec les effets biologiques observés lorsqu'elles sont inhalées. Un travail théorique original a donc été mené afin de positionner le paramètre de surface vis-à-vis d'autres grandeurs caractéristiques des aérosols. En vue de caractériser des méthodes de mesure de la surface des aérosols nanostructurés, le banc d'essais CAIMAN (CAractérisation des Instruments de Mesure des Aérosols Nanostructurés) a été dimensionné et réalisé. Celui-ci permet la production d'aérosols nanostructurés de propriétés variables et maîtrisées (taille, concentration, nature chimique, morphologie, état de charge), offrant une très bonne stabilité dans le temps. Les aérosols générés en laboratoire ont été utilisés en vue d'évaluer expérimentalement la réponse des instruments de mesure étudiés (NSAM & AeroTrak 9000 TSI, LQ1-DC Matter Engineering). Les fonctions de réponse expérimentales établies sur des aérosols monodispersés présentent un bon accord avec les courbes théoriques, dans une large gamme d'étude de 15 à 520 nm. Par ailleurs, des hypothèses ont été avancées en vue d'expliquer les écarts raisonnables observés lors des mesures effectuées sur des aérosols polydispersés / This work aims at improving knowledge on ultrafine aerosols surface-area measurement. Indeed, the development of nanotechnologies may lead to occupational exposure to airborne nanostructured particles, which involves a new prevention issue. There is currently no consensus concerning what parameter (mass, surface-area, number) should be measured. However, surface-area could be a relevant metric, since it leads to a satisfying correlation with biological effects when nanostructured particles are inhaled. Hence, an original theoretical work was performed to position the parameter of surface-area in relation to other aerosol characteristics. To investigate measurement techniques of nanostructured aerosols surface-area, the experimental facility CAIMAN (ChAracterization of Instruments for the Measurement of Aerosols of Nanoparticles) was designed and built. Within CAIMAN, it is possible to produce nanostructured aerosols with varying and controlled properties (size, concentration, chemical nature, morphology, state-of-charge), stable and reproducible in time. The generated aerosols were used to experimentally characterize the response of the instruments in study (NSAM & AeroTrak 9000 TSI, LQ1-DC Matter Engineering). The response functions measured with monodisperse aerosols show a good agreement with the corresponding theoretical curves in a large size range, from 15 to 520 nm. Furthermore, hypotheses have been formulated to explain the reasonable biases observed when measuring polydisperse aerosols

The proton exchange membrane fuel cell generates electricity from hydrogen and oxygen (from air) through electrocatalytic reactions in an electrochemical cell. The Pt/C catalyst, commonly used in PEM fuel cells, is very sensitive to impurities that can interact with the active catalyst sites and limit fuel cell performance. Unfortunately, most hydrogen is currently produced from fossil sources, and inevitably contains impurities. The subject of this thesis is the effect of hydrogen impurities on the operation of a PEM fuel cell using a Pt/C anode. The impurities studied are carbon monoxide (CO), carbon dioxide (CO2), and selected hydrocarbons. Particular focus is given to the interaction between the impurities studied and the anode catalyst. The main method used in the study involved performing cyclic voltammetry and mass spectrometry, simultaneously. Other electrochemical techniques are also employed. The results show that all the impurities studied adsorb to some extent on the Pt/C catalyst surface, and require potentials comparable to that of CO oxidation, i.e., about 0.6V, or higher to be removed by oxidation to CO2. For complete oxidation of propene, and toluene, potentials of above 0.8, and 1.0V, respectively, are required. The unsaturated hydrocarbons can be desorbed to some extent by reduction, but oxidation is required for complete removal. Adsorption of ethene, propene, and CO2 is dependent on the presence of adsorbed or gaseous hydrogen. Hydrogen inhibits ethene and propene adsorption, but facilitates CO2 adsorption. Adsorption of methane and propane is very limited and high concentrations of methane cause dilution effects only. The adlayer formed on the Pt/C anode catalyst in the presence of CO2, or moderate amounts of hydrocarbons, is found to be insffuciently complete to notably interfere with the hydrogen oxidation reaction. Higher concentrations of toluene do, however, limit the reaction. / Polymerelektrolytbränslecellen genererar elektricitet fran vätgas och syrgas (fran luft) genom elektrokatalytiska reaktioner i en elektrokemisk cell. Den platina-baserade katalysator som oftast används i dessa bränsleceller är känslig mot föroreningar, då dessa kan interagera med katalysatorns aktiva yta, och därmed begränsna bränslecellens prestanda. Tyvärr produceras dagens vätgas huvudsakligen fran fossila källor och innehåller därför oundvikligen föroreningar. Denna avhandling behandlar hur olika vätgasföroreningar påverkar katalysatorns aktivitet och bränslecellens drift. De föroreningar som studeras är kolmonoxid (CO) och koldioxid (CO2), samt ett antal mindre kolväten. Störst fokus ligger på hur dessa föroreningar interagerar med anodens Pt/C katalysator. Den metod som huvudsakligen används är cyklisk voltammetri kombinerat med masspektrometri, men flera elektrokemiska metoder har använts. Resultaten visar att alla undersökta föroreningar adsorberar på Pt/C katalysatorns yta i större eller mindre utstreckning. For att avlägsna det adsoberade skiktet genom oxidation till CO2 krävs potentialer jämförbara med CO oxidation, dvs ca 0,6V, eller högre. Fullständig oxidation av propen eller toluen kräver potentialer högre än 0,8V respektive 1,0V. De omättade kolvätena kan delvis avlägsnas genom reduktion, men fullständig avlägsning kräver oxidation. Närvaron av väte, i gasform eller adsorberat pa katalysatorn, hämmar adsorptionen av eten och propen, men främjar CO2 adsorption. Metan och propan adsorberar i mycket begränsad utstreckning på Pt/C katalysatorns yta. De prestandaförluster som uppstår av höga koncentrationer av metan förklaras av utspädning av vätgasen. Det adsorberade skiktet som bildas när Pt/C katalysatorn exponeras för CO2 eller måttliga koncentrationer av studerade kolväten, är inte tillräckligt heltäckande for att märkbart påverka vätgasreduktionen. Däremot kan höga koncentrationer av toluen begränsa reaktionen. / <p>QC 20161010</p>

Les Composés Organiques Volatils (COV), émis dans l'atmosphère sous différentes formes par les activités industrielles, sont considérés comme des polluants principaux de l'air. Pour le traitement des forts débits de gaz faiblement concentrés en COV, caractéristiques des principales sources de COV industriels, il n'existe que très peu de procédés adaptés et efficaces d'où le plasma non thermique. Sa faible consommation d'énergie et sa grande compacité font du traitement par plasma non-thermique un candidat prometteur. Ma thèse consiste à élaborer et étudier un dispositif de traitement des COV basé sur l'utilisation de décharges à barrière diélectrique de surface, décharges ayant la caractéristique de produire un vent électrique dirigé. Pour ce faire, nous avons conçu et optimisé une pompe plasma chimiquement active, composée d'un assemblage de cellules à surfaces actives, capable d'aspirer et de traiter simultanément de l'air pollué en COV. Il a fallu d'abord travailler sur l'optimisation paramétrique : électrique, géométrique et matériaux, d'une surface active élémentaire. Puis concevoir un canal actif constitué de deux surfaces actives optimales placées en vis-à-vis, pour finalement aboutir à la construction d'une pompe plasma prototype ayant un débit de pompage de 10 Nm3/h. Cette pompe originale par sa capacité de traitement chimique, constitue le prototype d'étude physique et chimique de cette thèse. Elle a permis d'effectuer des essais d'élimination de 5 COV différents injectés dans l'air : acétone, méthyl-éthyle cétone, butyraldéhyde, méthyl-valérate, méthyl-butyrate et d'en évaluer les taux d'abattement respectifs, mais aussi, d'identifier à l'aide de la chromatographie en phase gazeuse couplée à la spectrométrie de masse, les principaux sous-produits de la dégradation. / The Volatile Organic Compounds (VOC) emitted in the atmosphere by various forms is considered as principal atmospheric pollutant. In order to treat a high flow of gaz with a low concentration, few efficient methods exist like the non-thermal plasma. Its low power consumption and compactness make the non-thermal plasma treatment a promising candidate. My thesis deals with the development and study of a VOC treatment device based on the surface dielectric barrier discharges which have the characteristic of producing an oriented electric wind. To do this, we have designed and optimized a chemically active plasma pump, composed of an assembly of active surfaces, capable of drawing and treating simultaneously the air polluted in VOC. At first, we had to work on the optimization of the following parameters: electrical, geometric and material of an elementary active surface. And then conceive an active channel consisting of two optimal active surfaces disposed in a mirror effect, eventually leading to the construction of a prototype plasma pump having a flow rate approximately equal to 10 Nm3/h. This original pump by its capacity of chemical treatment consists on the physical and chemical prototype of the thesis. It allowed testing the conversion of 5 different VOCs injected into air which are: ketone, methyl ethyl ketone, butyraldehyde, methyl penatanoate, methyl butyrate and evaluate the respective abatement rates, but also identifying the main by-products of degradation, using the gas chromatography coupled to the mass spectrometry.