Includes bibliographical references (p. 207-212).

Includes bibliographical references (p. 199-204).

Thesis (M.App.Sc.) - Carleton University, 2002. / Includes bibliographical references (p. 109-113). Also available in electronic format on the Internet.

Thesis (M.S.)--Missouri University of Science and Technology, 2009. / Vita. The entire thesis text is included in file. Title from title screen of thesis/dissertation PDF file (viewed August 14, 2009) Includes bibliographical references (p. 16-17).

Includes bibliographical references (p. 70-74).

Thesis (Ph. D.)--University of Florida, 1982. / Description based on print version record. Typescript. Vita. Includes bibliographical references (leaves 95-97).

Dans cette thèse, on s’intéresse à l’application du domaine de la vision par ordinateur à la vélocimétrie de surface des rivières. Les hydrauliciens utilisent déjà plusieurs routines de traitement d’images pour traiter des vidéos de rivières. Le but ultime est d’estimer la vitesse surfacique d’un cours d’eau par une méthode sans contact. Cela permet aux chercheurs d’éviter les risques liés au jaugeage intrusif des rivières, notamment en période de crue. Dans ce but, deux enjeux sont à prendre en compte. Tout d’abord, le mouvement apparent de la rivière dans l’espace image doit être estimé. Ensuite, ce mouvement, estimé en pixels par unité de temps, doit être transformé en une vitesse réelle exprimée en mètres par seconde par exemple. Jusqu’au présent, les méthodes de vélocimétrie par images imposent quelques contraintes sur les séquences pour qu’elles soient exploitables (notamment une caméra fixe et le besoin de la présence physique des équipes hydrauliques au site de jaugeage avant ou après l’événement). Dans cette thèse, on vise à élargir ce périmètre en incluant les vidéos prises par des amateurs (c’est à dire de paramètres inconnus, et avec un mouvement potentiel de la caméra) tout en présentant de meilleures solutions pour les enjeux précédemment mentionnés. / This thesis is an application of computer vision findings to river velocimetry research. Hydraulic research scientists already use various image processing techniques to process image sequences of rivers. The ultimate goal is to estimate free surface velocity of rivers remotely. As such, many risks related to intrusive river gauging techniques could be avoided. Towards this goal, there are two major issues need be addressed. Firstly, the motion of the river in image space need to be estimated. The second issue is related to how to transform this image velocity to real world velocity. Until recently, imagebased velocimetry methods impose many requirements on images and still need considerable amount of field work to be able to estimate rivers velocity with good accuracy. We extend the perimeter of this field by including amateur videos of rivers and we provide better solutions for the aforementioned issues. We propose a motion estimation model that is based on the so-called optical flow, which is a well developed method for rigid motion estimation in image sequences. Contrary to conventional techniques used before, optical flow formulation is flexible enough to incorporate physics equations that govern rivers motion. Our optical flow is based on the scalar transport equation and is augmented with a weighted diffusion term to compensate for small scale (non-captured) contributions. Additionally, since there is no ground truth data for such type of image sequences, we present a new evaluation method to assess the results. It is based on trajectory reconstruction of few Lagrangian particles of interest and a direct comparison against their manually-reconstructed trajectories. The new motion estimation technique outperformed traditional methods in image space. Finally, we propose a specialized geometric modeling of river sites that allows complete and accurate passage from 2D velocity to world velocity, under mild assumptions. This modeling considerably reduces the field work needed before to deploy Ground Reference Points (GRPs). We proceed to show the results of two case studies in which world velocity is estimated from raw videos.

Measurement of the dynamics of flexible structures is difficult because the motion is often complex and the structures are not well suited to the attachment of sensors. As a result, non-contact optical systems are used. However, optical systems produce large amounts of data which make their use in real-time measurement difficult. Conventional computers are not well suited to the processing requirements associated with data from optical systems. In this thesis, algorithms and architectures to reduce the data bandwidth of an optical measurement system are investigated. Simulations of the ability of the algorithms to find a target on a linear-array charge-coupled device (CCD) camera are performed. The running maximum algorithm provides the best accuracy and speed and therefore is recommended. A real-time architecture to implement the running maximum algorithm is developed. The architecture allows the optical system to operate at 9700 frames/second. Experimental results from a prototype system show very good accuracy for both static and dynamic measurements. / Master of Science

Precisely positioned optical lenses are currently required for many highly repetitive mechanics and applications. Thus the need for micron-scale repetition between opto-mechanical units is evident, especially in industrial manufacturing and medical breakthroughs. In this thesis, a novel optical metrology system is proposed, designed, and built whose purpose is to precisely locate the center of a mechanical fixture and then to assemble a plano-convex optical lens into the located position of the fixture. Center location specifications up to ±3 µm decenter and ±0.001° tilting accuracy are required. Nine precisely positioned lenses and fixtures were built with eight units passing the requirements with a repetitive standard deviation of ±0.15 µm or less. The assembled units show satisfactory results.

A particle counting instrument, the Coherent Optical Particle Spectrometer (COPS) has been developed for measuring particles in aerosol systems. It optically counts and sizes single particles one at a time as they pass through an optically defined inspection region so particle size distributions can be directly measured. COPS uses the coherent nature of light available in a laser beam to measure the phase shift in the scattered light, which is fundamentally different from previous intensity based techniques. The Van-Cittert-Zernike theorem shows that scattered light from small particles will be coherent if viewed upon at the focal point of a gathering lens. Optical homodyne detection can then be used to measure the extent of the phase shift due to the particle. Scattering mechanisms can relate the phase shift to particle diameter so particle size can be determined. An optical inspection region is given by the resolution limited blur spot diameter and depth of focus of the gathering lens. Particles scattering outside this zone will not contribute to measured phase signals. Calculations show that COPS can count in concentrations of 10('9) particles per cubic centimeter with 5% coincidence error. Mie scattering calculations, coupled with homodyne theory, predict a minimum detectable particle diameter ranging from 0.03 to 0.3 micrometers, depending on optical configuration. Theory shows that small, strongly absorbing particles impart a much larger phase shift than refractive particles so a lower detection limit is predicted for particles such as soot and silicon. Particles above one micrometer show classic resonance typical of Mie calculations. An experimental COPS system verified the predicted results from the model. Resolution of particle size ranged from 25 to 60 percent of particle diameter. Preliminary experiments showed that COPS has in situ sampling possibilities and will work for liquid systems as well. Coherent detection of scattered light shows promise for in situ measurement of submicrometer aerosols in high particle laden streams with maximum sensitivity for strongly absorbing particles.