An experimental investigation of super-resolution imaging from measurements of projections onto a random
basis is presented. In particular, a laboratory imaging system was constructed following an architecture that
has become familiar from the theory of compressive sensing. The system uses a digital micromirror array
located at an intermediate image plane to introduce binary matrices that represent members of a basis set.
The system model was developed from experimentally acquired calibration data which characterizes the system
output corresponding to each individual mirror in the array. Images are reconstructed at a resolution limited
by that of the micromirror array using the split Bregman approach to total-variation regularized optimization.
System performance is evaluated qualitatively as a function of the size of the basis set, or equivalently, the
number of snapshots applied in the reconstruction.
We describe an experimental laboratory system that generates and distributes random binary sequence bit streams between two optical terminals (labeled Alice and Bob). The random binary sequence is generated through probing the optical channel of a turbulent atmosphere between the two terminals with coincident laser beams. The two laser beams experience differential phase delays while propagating through the atmospheric optical channel. The differential phase delays are detected and sampled at each terminal to yield raw random bit streams. The random bit streams are processed to remove bit errors and, through privacy amplification, to yield a bit stream known only to Alice and Bob. The same chaotic physical mechanism that provides randomness also provides confidentiality. The laboratory system yielded secret key bit rates of a few bits/second . For external optical channels over longer channel lengths with atmospheric turbulence levels, secret bit rates of 10 s of bits/second are predicted.
Traditional approaches to persistent surveillance generate prodigious amounts of data, stressing storage, communication,
and analysis systems. As such, they are well suited for compressed sensing (CS) concepts. Existing
demonstrations of compressive target tracking have utilized time-sequences of random patterns, an approach
that is sub-optimal for real world dynamic scenes. We have been investigating an alternative architecture that
we term SCOUT-the Static Computational Optical Undersampled Tracker-which uses a pair of static masks
and a defocused detector to acquire a small number of measurements in parallel. We will report on our working
prototypes that have demonstrated successful target tracking at 16x compression.
Motion tracking in persistent surveillance applications enters an interesting regime when the movers are of a size
on the order of the image resolution elements or smaller. In this case, for reasonable scenes, information about the
movers is a natively sparse signal - in an observation of a scene at two closely separated time-steps, only a small
number of locations (those associated with the movers) will have changed dramatically. Thus, this particular
application is well-suited for compressive sensing techniques that attempt to efficiently measure sparse signals.
Recently, we have been investigating two different approaches to compressive measurement for this application.
The first, differential Combinatorial Group Testing (dCGT), is a natural extension of group testing ideas to
situations where signal differences are sparse. The second methodology is an ℓ-1-minimization based recovery
approach centered on recent work in random (and designed) multiplex sensing. In this manuscript we will
discuss these methods as they apply to the motion tracking problem, discuss various performance limits, present
early simulation results, and discuss notional optical architectures for implementing a compressive measurement
scheme.
We present the methodology for designing the optimal gain profiles
for slow-light systems under the given system constraints. Optimal
system designs for the multiple Lorentzian gain lines make the gain
spectrum uniform over larger bandwidth compared to the single-line
gain system. The design procedure for the multiple-line gain systems
is modified to make the gain spectrum uniform over arbitrarily broad
bandwidth and applied to the design of the gain-only and
gain+absorption slow-light media. The optimization of the
triple-line gain system improves the delay-bandwidth product 1.7
times the delay-bandwidth product for the single-line gain system.
For the broadband slow-light system, the optimal gain + absorption
design and the optimal gain-only design improve the fractional delay
performance by factors of 1.8 and 1.4, respectively, compared to the
Gaussian noise pump broadened (GNPB) system.
We describe theoretical and experimental results for a new class of optimal features for feature-specific imaging (FSI). In this paper, we theoretically solve the reconstruction problem without noise, and find a more general solution than principle component analysis (PCA). We present a generalized framework to find FSI projection matrices. Using Stochastic Tunneling, we find an optimal solution in the presence of noise and under an energy conservation constraint. We also show that a non-negativity requirement does not significantly reduce system performance. Finally, we propose an experimental system for FSI using a polarization-based optical pipeline processor.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.