High-speed digital three-dimensional (3-D) imagery is possible using multiple independent charge-coupled device (CCD) cameras with sequentially triggered acquisition and individual field storage capability. The system described here utilizes sixteen independent cameras, providing versatility in configuration and image acquisition. By aligning the cameras in nearly coincident lines-of-sight, a sixteen frame two-dimensional (2-D) sequence can be captured. The delays can be individually adjusted lo yield a greater number of acquired frames during the more rapid segments of the event. Additionally, individual integration periods may be adjusted to ensure adequate radiometric response while minimizing image blur. An alternative alignment and triggering scheme arranges the cameras into two angularly separated banks of eight cameras each. By simultaneously triggering correlated stereo pairs, an eight-frame sequence of stereo images may be captured. In the first alignment scheme the camera lines-of-sight cannot be made precisely coincident. Thus representation of the data as a monocular sequence introduces the issue of independent camera coordinate registration with the real scene. This issue arises more significantly using the stereo pair method to reconstruct quantitative 3-D spatial information of the event as a function of time. The principal development here will be the derivation and evaluation of a solution transform and its inverse for the digital data which will yield a 3-D spatial mapping as a function of time.

There are three fundamental mechanisms in optical systems that contribute to image degradation: aperture diffraction, geometrical aberrations caused by residual design errors, and scattering effects due to optical fabrication errors. Diffraction effects, as well as optical design errors and fabrication errors that are laterally large in nature (generally referred to as figure errors), are accurately modeled using conventional ray trace analysis codes. However, these ray-trace codes fall short of providing a complete picture of image degradation; they routinely ignore fabrication-induced errors with spatial periods that are too small to be considered figure errors. These errors are typically referred to as mid-spatial-frequency (ripple) and high-spatial-frequency (micro-roughness) surface errors. These overlooked, but relevant, fabrication-induced errors affect image quality in different ways. Mid-spatial-frequency errors produce small-angle scatter that tends to widen the diffraction-limited image core (i.e. for a system with a circular exit pupil, this is the central lobe of the Airy pattern), and in doing so, reduces the optical resolution of a system. High-spatial-frequency errors tend to scatter energy out of the image core into a wide-angle halo, causing a reduction in image contrast. Micro-roughness and ripple are inherent aspects of the less conventional, small-tool-based optical fabrication approaches. It is especially important in these cases to specify these errors accurately during the design phase of a project, and deterministically monitor and control them during the fabrication phase of a project. Surprisingly, most current approaches to this issue employ some guessing and "gut feel" based on past experience, because accurate theories and analysis tools are not readily available. This dissertation takes the first step towards solving this problem by describing a Fourier-based approach for classifying and quantifying surface errors that can be present in a fabricated optical surface. Classical scalar diffraction theories and scatter theories are reviewed and their strengths, weaknesses and misuses are discussed. Then, this dissertation focuses on the development of more accurate surface scatter theories. Modified surface scatter theories are presented that do not exhibit the small angle or smooth surface limitations that are inherent in other theories. These improvements are especially critical for surfaces considered rough with respect to the test wavelength or for systems where large scatter and/or incidence angles are present. Predictions from these modified theories are then compared to and shown to be in excellent agreement with experimental measurements.

Methods for the restoration of images corrupted by blur and noise are presented. During transmission through an optical or electrical channel, images become corrupted by blur and noise as a result of channel limitations (i.e. optical aberrations or a bandlimit). If images are treated as a matrix whose elements (pixels) assume a finite number of values then there is a large but finite set of possible images that can be transmitted. By treating this finite set as a 'signal' set, digital communications methods may be used to estimate the uncorrupted image given a blurred and noisy version. Specifically, row-by-row estimation, decision-feedback and vector-quantization are used to extend the 1D sequence estimation ability of the a-posteriori probability (APP) and Viterbi algorithm (VA) to the estimation of 2D images. Simulations show the 2D VA and APP algorithms return near-optimal estimates of binary images as well as improved estimates of greyscale images when compared with the conventional Wiener filter (WF) estimates. Unlike the WF, the VA and APP algorithms are shown to be capable of super-resolution and adaptable for use with signal-dependent Poisson noise corruption. Restorations of experimental data gathered from an optical imaging system are presented to support simulation results.

This dissertation investigates superresolution applications in optical data storage systems. The performance of standard and superresolving magneto-optic data storage system are quantified by scalar diffraction modeling and experiments. Classical resolution measures are reviewed. Background on superresolution definitions and their conceptual development in scanning optical microscopes, optical data storage, and image processing is presented. Figures of merit for quantifying the performance of the systems are reviewed, such as system transfer function, two-point response, focused spot size, and signal-to-noise ratio. The description of the scalar diffraction modeling used to simulate an optical data storage system is reviewed. Operation of the magneto-optic data storage system and tradeoffs of superresolving techniques are discussed. The signal and noise spatial distribution in the pupil of an optical data storage system are shown to be different. For a particular spatial frequency bandwidth, the signal and noise are concentrated in different regions of the pupil. This understanding allows the use of optical filters that partially equalize the system transfer function and increase the signal-to-noise ratio. The main superresolution techniques investigated are those that increase the transmission of the higher spatial frequencies, or equalize the system transfer function, without changing the system cutoff frequency. The optical methods used to achieve superresolution are amplitude and phase filters placed in strategic system locations. One location influences the properties of the focused spot such as the irradiance distribution and width of the central core. Another location does not change the focused spot at all, but does change the signal and noise properties of the system. Electronic filtering techniques are also used to increase the transmission of the high spatial frequencies. The amplitude and phase filter sensitivities to aberration are also investigated. Optical properties of a new laser diode are investigated. The new laser diode has potential superresolving properties that are inherent to the device. Potential application of this device in an optical data storage device is presented. Another method of increasing the transmission of higher spatial frequencies within the system bandwidth and beyond the system cutoff frequency is to use adaptive optical systems. Adaptive systems for optical data storage are also discussed.

Spectroscopic properties of high concentration Er³⁺-doped phosphate glasses and performance of short Er³⁺-doped phosphate fiber amplifiers were studied and characterized. Systematic studies of cooperative upconversion of Er³⁺ ions in ⁴I₁₃/₂ level and energy transfer from Yb³⁺ to Er³⁺ in phosphate glasses were performed by a rate equation formalism. The cooperative upconversion coefficient for an Er³⁺ concentration of 4 x 10²⁰ ions/cm³ was found to be 1.1 x 10⁻¹⁸ cm³/s. An energy transfer coefficient of 1.1 x 10⁻¹⁶ cm³/s was found for an Yb³⁺ concentration of 6 x 10²⁰ ions/cm3 and an Er³⁺ concentration of 2 x 10²⁰ ions/cm³. Energy transfer efficiencies from ²F₅/₂ level of Yb³⁺ ions to ⁴I₁₃/₂ level of Er³⁺ ions higher than 95% were determined from our measurements under weak excitation. The performance of high concentration Er³⁺-doped phosphate fiber amplifiers were characterized in terms of gain, noise figure, and signal saturation for a series of active fiber lengths, pump powers, signal input powers, and signal wavelengths. A net gain of 21 dB were achieved in a 71 mm Er³⁺-doped phosphate fiber with a noise figure of ∼5.3 dB by a 980 nm pump power of 244 mW. In addition, a 10 dB net gain can be obtained with a pump power of 110 mW. Performance of short Er³⁺-doped phosphate fiber amplifiers demonstrates the potential for device applications.

Semiconductor microcavities have emerged to present abundant opportunities for both device applications and basic quantum optics studies. Here we investigate several aspects of the cw and ultrafast optical response of semiconductor quantum well microcavities. The interaction of a high-finesse semiconductor microcavity mode with a quantum well (QW) exciton leads to normal mode coupling (NMC), where a periodic energy exchange develops between exciton and photon states, appearing as a double peak in the cavity transmission spectrum and a beating in the time resolved signal. The nonlinear saturation of the excitonic NMC leads to a reduction of the modulation depth of the NMC oscillations and corresponding transmission peaks with little change in oscillation period or NMC splitting. This behavior arises from excitonic broadening due to carrier-carrier and polarization scattering without reduction of the oscillator strength. The nonlinear NMC microcavity luminescence exhibits three excitation regimes, from reversible normal mode coupling, through an intermediate double-peaked emission regime, to lasing. The nonlinear PL spectrum is governed by density-dependent changes in both the bare QW emission and in the microcavity transmission. The temporal evolution of the microcavity emission is analogous to the density-dependent behavior, and can be attributed to a time-dependent carrier density which results from a combination of carrier cooling and photon emission. A strong magnetic field applied perpendicular to the plane of a QW confines electrons and holes to Landau orbits in the QW plane, transforming the QW into a quantum dot (QD) whose radius shrinks with increasing magnetic field strength. This strong magnetic confinement enhances the normal mode coupling strength in the microcavity via an increase in exciton oscillator strength. The time-resolved stimulated emission of a QW microcavity which has been transformed to a QD laser by magnetic confinement reveals a fast relaxation which is uninhibited by the magnetic field, indicating the absence of a phonon bottleneck. As a novel manifestation of cavity-modified emission, we demonstrate synchronization of the stimulated emission of a microcavity laser to the electron spin precession in a magnetic field, achieved by modulating the optical gain for the circularly polarized emission via the Larmor precession. The oscillating laser emission is locked to the completely internal electron spin precession clock, and the GHz oscillation frequencies depend only on the magnetic field strength and the QW material parameters.

The optical properties of Fresnel reflections from the human ocular surfaces, called Purkinje reflections, are examined. Extensive modeling of the behavior of the reflection of sources from the front of the cornea and the front of the crystalline lens with real rays in lens design software is presented. The modeling looks at the effects of various conic constant values on the ocular surfaces and rotation of the eye in particular. First and third Purkinje images were collected from 14 subjects for varying fixation positions to compare with modeling. The results showed a decrease in third Purkinje image height as the eye rotated from gazing at a point near the light sources to a point near the optical axis of the imaging camera. This matched the predictions from modeling and indicates that fixation position is an important factor in the accuracy and repeatability of comparison phakometry results. Schematic eye models were set up for each subject and the anterior lens radius of curvature and conic constant were optimized to match the collected Purkinje image height data. The mean conic constant estimate from optimization was -3.82 with a standard deviation of 1.51. The schematic eye models did not include crystalline lens tilt or individual corneal conic constant values, each of which is estimated to contribute an uncertainty of ± 0.5 in the anterior lens conic constant value. This is the first use of Purkinje images to assess anterior lens conic constant values.

This dissertation studies rear projection diffusing screens. We provide a methodology, a theoretical background, metrics, and experimental results to aid in the understanding and design of such screens. For this work, a theoretical model has been developed to predict local fluctuations of measured color that appear as image noise in projection screens. This predictive model is based on Fraunhofer diffraction along with Huygens' wavelet analysis and linear systems theory. Of importance are the figures of merit that have been defined and used to compare the theoretical predictions and experimental results. The range of validity of the model has also been determined. We set up an experiment to test the theoretical model. By experimentally varying the numerical aperture of the input illumination, color variations on screens have been measured and characterized. The results of both the experiment and the model show a strong relationship between the polychromatic image noise and the size of the illumination cone. As the size of the illumination cone was decreased from 6 to less than 0.1 millisteradians, there was more than a threefold increase in the figures of merit. Our model shows insight, validates and augments a common rule of thumb. It is often assumed that making the screen structure significantly larger than the coherence length of the source will result in a system with minimal noise. The model shows that this is correct, but it also provides predictions in the cross over region. This allows one to understand how the image noise in a projection system will change as screen designs are changed. Ultimately, this allows screen solutions to be assessed before they are reduced to practice.

Over the last few years there has been a great demand for noninvasive, dedicated, small-animal imaging systems for biomedical research applications. In this dissertation we will discuss the development and performance of two gamma-ray systems and a dual modality CT/SPECT system. Initially we introduce FASTSPECT, a stationary, scintillator-based, single-photon emission computed tomography (SPECT) system that was originally built as a brain imager. We discuss its transformation into a small-animal imaging system and validate its performance by presenting high-resolution images of phantoms and animals. Furthermore, we discuss the development of an in vivo imaging protocol for rat myocardial models using FASTSPECT. The infarct size obtained from the images is quantified and compared with the myocardial infarct size measured from histology. Semiconductor detectors can exhibit good spatial and energy resolution, and therefore offer a promising alternative to scintillation technology. We discuss the performance of a semiconductor detector system, previously developed in our group, for planar and tomographic imaging of small animals. The same gamma-ray detector is used in a dual modality system for imaging mice. The system combines an anatomical imaging modality, x-ray CT, with a functional modality, SPECT. We present the development of the CT/SPECT system and illustrate its performance by presenting high-resolution images of phantoms and mice. Finally, we introduce a procedure for evaluating estimation methods without the use of a gold standard.

In the 45 years since the dawn of the space age, there have only been two Mars lander camera designs to successfully operate on the Martian surface. Therefore information on Mars imager design and operation issues is limited. In addition, good examples of Mars lander imager calibration work are almost non-existent. This work presents instrument calibration results for a Mars lander camera originally designed to fly as an instrument onboard the 2001 Mars Surveyor lander as a robotic arm camera (RAC). Test procedures and results are described as well as techniques for improving the accuracy of the calibration data. In addition we describe camera algorithms and operations research results for optimizing imager operations on the Martian surface. Finally, the lessons learned from the 2001 RAC are applied to the preliminary design of a new Mars camera for the Artemis Mars Scout mission. The design utilizes a Bayer color mosaic filter, white light LED's and includes an optical system operating at f/13 with a maximum resolution of 0.11 mrad/pixel. It is capable of imaging in several modes including: stereo, microscopic and panoramic at a mass of 0.3 kg. It will provide planetary geologists with an unprecedented view of the Martian surface.