Jason Smith


Earth-observing satellites have fundamental size and weight design limits since they must be launched into space. These limits serve to constrain the spatial resolutions that such imaging systems can achieve with traditional telescope design strategies. Segmented and sparse-aperture imaging system designs may offer solutions to this problem. Segmented and sparse-aperture designs can be viewed as competing technologies; both approaches offer solutions for achieving finer resolution imaging from space. Segmented-aperture systems offer greater fill factor, and therefore greater signal-to-noise ratio (SNR), for a given encircled diameter than their sparse aperture counterparts, though their larger segments often suffer from greater optical aberration than those of smaller, sparse designs. Regardless, the use of any multi-aperture imaging system comes at a price; their increased effective aperture size and improvement in spatial resolution are offset by a reduction in image quality due to signal loss (less photon-collecting area) and aberrations introduced by misalignments between individual sub-apertures as compared with monolithic collectors. Introducing multispectral considerations to a multi-aperture imaging system further starves the system of photons and reduces SNR in each spectral band. This work explores multispectral design considerations inherent in 9-element tri-arm sparse aperture, hexagonal-element segmented aperture, and monolithic aperture imaging systems. The primary thrust of this work is to develop an objective target detection-based metric that can be used to compare the achieved image utility of these competing multi-aperture telescope designs over a designated design parameter trade space. Characterizing complex multi-aperture system designs in this way may lead to improved assessment of programmatic risk and reward in the development of higher-resolution imaging capabilities. This method assumes that the stringent requirements for limiting the wavefront error (WFE) associated with multi-aperture imaging systems when producing imagery for visual assessment, can be relaxed when employing target detection-based metrics for evaluating system utility. Simple target detection algorithms were used to determine Receiver Operating Characteristic (ROC) curves for the various simulated multi-aperture system designs that could be used in an objective assessment of each system's ability to support target detection activities. Also, a set of regressed equations was developed that allow one to predict multi-aperture system target detection performance within the bounds of the designated trade space. Suitable metrics for comparing the shapes of two individual ROC curves, such as the total area under the curve (AUC) and the sample Pearson correlation coefficient, were found to be useful tools in validating the predicted results of the trade space regression models. And lastly, some simple "rules of thumb" relating to multi-aperture system design were identified from the inspection of various points of equivalency between competing system designs, as determined from the comparison metrics employed. The goal of this work, the development of a process for simulating multi-aperture imaging systems and comparing them in terms of target detection tasks, was successfully accomplished. The process presented here could be tailored to the needs of any specific multi-aperture development effort and used as a tool for system design engineers.

Library of Congress Subject Headings

Space telescopes--Design and construction; Remote sensing--Data processing; Multispectral photography; Space telescope--Design and construction; Remote sensing--Data processing; Multispectral photography

Publication Date


Document Type


Student Type


Department, Program, or Center

Chester F. Carlson Center for Imaging Science (COS)


Schott, John


Note: imported from RIT’s Digital Media Library running on DSpace to RIT Scholar Works. Physical copy available through RIT's The Wallace Library at: TA1637 .S657 2012||TA1637 .S657 2012


RIT – Main Campus